Paper 4 Artificial Sweeteners Induce Glucose Intolerance Nature 2014

8202019 Paper 4 Artificial Sweeteners Induce Glucose Intolerance Nature 2014

httpslidepdfcomreaderfullpaper-4-artificial-sweeteners-induce-glucose-intolerance-nature-2014 117

ARTICLE doi101038nature13793

Artificial sweeteners induce glucose

intolerance by altering the gut microbiota Jotham Suez1 Tal Korem2 David Zeevi2 Gili Zilberman-Schapira1 Christoph A Thaiss1 Ori Maza1 David Israeli3Niv Zmora456 Shlomit Gilad7 Adina Weinberger2 Yael Kuperman8 Alon Harmelin8 Ilana Kolodkin-Gal9 Hagit Shapiro1Zamir Halpern56 Eran Segal2 amp Eran Elinav1

Non-caloric artificial sweeteners (NAS) are among the most widely used food additives worldwide regularly consumedby lean and obese individualsalike NASconsumption is considered safe and beneficial owing to theirlowcaloric contentyet supporting scientific dataremain sparse and controversialHerewe demonstrate thatconsumption of commonlyusedNAS formulationsdrives the development of glucose intolerancethroughinductionof compositional and functional alter-ations totheintestinal microbiota TheseNAS-mediateddeleteriousmetaboliceffectsare abrogatedby antibiotictreatment

and are fullytransferrableto germ-free miceuponfaecaltransplantation of microbiota configurationsfromNAS-consuming mice or of microbiota anaerobically incubated in the presence of NAS We identify NAS-altered microbial metabolic path-ways that are linked to host susceptibility to metabolicdisease anddemonstratesimilar NAS-induced dysbiosis and glucoseintolerance in healthy human subjects Collectively our results link NAS consumption dysbiosis and metabolic abnor-malities thereby calling for a reassessment of massive NAS usage

Non-caloric artificial sweeteners(NAS) wereintroduced over a century ago as means for providing sweet taste to foods without the associatedhigh energy content of caloric sugars NAS consumption gained muchpopularity owing to their reduced costs low caloric intake and per-ceived healthbenefitsfor weight reduction andnormalization of bloodsugar levels1 For these reasons NAS are increasingly introduced intocommonly consumed foods such as diet sodas cereals and sugar-free

desserts and are being recommended for weight loss and for indivi-duals suffering from glucose intolerance and type 2 diabetes mellitus1Some studies showedbenefits forNAS consumption2 andlittle induc-

tion of a glycaemic response3 whereas othersdemonstrated associationsbetween NAS consumption and weight gain4 andincreased type 2 dia-betesrisk 5 However interpretation is complicatedby thefact that NASaretypicallyconsumed by individuals alreadysufferingfrom metabolicsyndrome manifestations Despite these controversial data theUSFoodand Drug Administration (FDA) approved six NAS products for use inthe United States

Most NAS pass through the human gastrointestinal tract withoutbeing digested by the host67 and thus directly encounter the intestinalmicrobiota which plays central roles in regulating multiple physiolo-gical processes8 Microbiota composition9 and function10are modulated

by diet in the healthylean state as well as in obesity 1112

and diabetesmellitus13 andin turn microbiota alterations have been associated withpropensity to metabolic syndrome14 Here we study NAS-mediatedmodulation of microbiota composition and functionand the resultanteffects on host glucose metabolism

Chronic NAS consumption exacerbates glucoseintoleranceTo determine the effects of NAS on glucose homeostasis we addedcommercial formulations of saccharin sucralose or aspartame to the

drinkingwaterof lean10-week-oldC57Bl6mice (Extended DataFig 1a)Since all three commercial NAS comprise5 sweetener and95glucose we used as controls mice drinking only water or water supple-mented with either glucose or sucrose Notably at week 11 the threemouse groups thatconsumed water glucoseand sucrosefeaturedcom-parable glucose tolerancecurveswhereasallthreeNAS-consumingmousegroups developed marked glucose intolerance (P 0001 Fig 1a b)

As saccharin exerted the mostpronounced effect we further studiedits role as a prototypical artificial sweetener To corroborate the find-ings in the obesity setup we fed C57Bl6 mice a high-fat diet (HFD60 kcal from fat) while consuming either commercial saccharin orpure glucose as a control (Extended Data Fig 1b) As in the lean statemicefed HFDand commercial saccharin developed glucose intolerancecompared to the control mouse group (P 003 Fig 1c and ExtendedData Fig 2a) Toexamine theeffectsof pure saccharinon glucose intol-erance we followeda cohortof 10-week-old C57Bl6 mice fedon HFDand supplemented with 01mg ml21 of pure saccharin added to theirdrinking water (Extended Data Fig 1c) This dose corresponds to theFDAacceptable dailyintake (ADI)in humans(5 mgperkg(body weight)adjustedto mouse weights seeMethods)As withcommercialsaccharinthis lower dose of pure saccharin wasassociated with impaired glucose

tolerance (P

00002 Fig 1d and Extended Data Fig 2b) starting asearly as 5 weeks afterHFD initiationSimilarly HFD-fedoutbred SwissWebster mice supplemented with or without 01 mg ml21 of pure sac-charin (Extended Data Fig 1d) showed significant glucose intoleranceafter 5 weeks of saccharin exposure as compared to controls (P 003Extended Data Fig 2c d)

Metabolic profiling of normal-chow- or HFD-fed mice in metaboliccages including liquids and chow consumption oxygen consumptionwalking distanceand energy expenditure showed similar measuresbe-tween NAS- and control-drinking mice (Extended Data Fig 3 and 4)

These authors contributed equally to this work

1Department of Immunology WeizmannInstituteof ScienceRehovot 76100 Israel2Department of ComputerScience and AppliedMathematicsWeizmann Instituteof ScienceRehovot 76100 Israel3Day

Care Unitand theLaboratory of Imagingand Brain StimulationKfarShaul hospital Jerusalem Center forMentalHealthJerusalem91060Israel4InternalMedicine DepartmentTel AvivSourasky Medical

CenterTel Aviv64239 Israel5Research CenterforDigestiveTractandLiverDiseasesTelAvivSouraskyMedical CenterSacklerFacultyof MedicineTelAvivUniversityTel Aviv69978Israel6DigestiveCenter

TelAviv Sourasky Medical CenterTel Aviv64239 Israel 7TheNancy and Stephen Grand Israel National Center forPersonalized Medicine(INCPM) WeizmannInstituteof ScienceRehovot 76100 Israel8Departmentof Veterinary Resources WeizmannInstituteof ScienceRehovot 76100 Israel 9Department of Molecular Genetics WeizmannInstitute of ScienceRehovot76100Israel

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 1

Macmillan Publishers Limited All rights reservedcopy2014



Fasting serum insulin levels and insulin tolerance were also similar inall mouse groups consuming NAS or caloric sweeteners in both thenormal-chow andHFD settings(Extended Data Fig5) Taken togetherthese results suggest that NAS promote metabolic derangements in arange of formulations doses mouse strains anddiets parallelinghumanconditions in both the lean and the obese state

Gut microbiota mediates NAS-induced glucoseintoleranceSince diet modulates the gut microbiota15 and microbiota alterationsexert profound effects on host physiology and metabolism we testedwhether the microbiota may regulate theobservedNAS effects To thisend we treated mouse groups consuming commercial or pure NAS inthe lean andHFDstates (ExtendedDataFig1ac) with a Gram-negative-targeting broad-spectrum antibiotics regimen (designated lsquoantibioticsArsquo)of ciprofloxacin (02 g l21) andmetronidazole (1 g l21)whilemain-taining mice on their diet and sweetener supplementation regimensNotablyafter4 weeks of antibiotictreatment differences in glucose in-tolerance between NAS-drinking mice and controls were abolishedboth in the lean (Fig 1a b) and the obese (Fig 1d and Extended DataFig 2b) states Similar effects were observed with the Gram-positive-targeting antibiotic vancomycin (lsquoantibiotics Brsquo 05 g l21 Fig 1a b)These resultssuggestthat NAS-inducedglucose intoleranceis mediatedthrough alterations to the commensal microbiota with contributionsfrom diverse bacterial taxa

To test whether the microbiota role is causal we performed faecaltransplantationexperiments by transferringthemicrobiota configura-

tion from mice on normal-chow diet drinking commercial saccharin

or glucose (control) into normal-chow-consuming germ-free mice(Extended Data Fig 1e) Notably recipients of microbiota from miceconsuming commercial saccharin exhibited impaired glucosetoleranceas compared to control(glucose) microbiota recipientsdetermined6 daysfollowing transfer (P 003 Fig 1e andExtended Data Fig 2e) Trans-ferring the microbiota composition of HFD-consuming mice drinking water or pure saccharin replicated the glucose intolerance phenotype(P 0004 Fig 1f and Extended Data Fig 2f) Together these resultsestablish that the metabolic derangements induced by NAS consump-tion are mediated by the intestinal microbiota

NAS mediate distinct functional alterations to themicrobiota We next examined the faecal microbiota composition of our variousmouse groupsby sequencing their 16Sribosomal RNAgeneMicedrink-ing saccharin had a distinct microbiota composition that clustered sepa-rately fromboth their starting microbiome configurations and fromallcontrol groups at week 11 (Fig 1g) Likewise microbiota in germ-freerecipients of stoolsfromsaccharin-consumingdonor mice clustered sepa-rately fromthatof germ-free recipients of glucose-drinking donor stools(Fig 1h) Compared to all control groups the microbiota of saccharin-consuming mice displayed considerable dysbiosis with more than 40operational taxonomic units(OTUs) significantly alteredin abundance(false discoveryrate (FDR) corrected P value 005 foreach OTU Ex-tendedDataFig6 SupplementaryTable1)Manyof thetaxathatincreasedin relative abundance belonged to theBacteroidesgenus and Clostridialesorder withother membersof theClostridialesordercomprising themajor-

ity of under-representedtaxa along withLactobacillusreuteriandwere

WaterPure saccharin

ndash A Antibiotics

a b

c

B l o o d g l u c o s e ( m g d l ndash 1 )

0 12090603015

Time (min)

50

150

250

350

0 12090603015

Time (min)

150

250

450

350

AspartameSucraloseSaccharin

GlucoseSucrose

Waterndash A B

Antibiotics

f

0 12090603015

Time (min)

50

150

250

200

100

g

0 12090603015

Time (min)

50

150

250

200

100 Pure saccharin

Water

Commercial saccharin

Glucose

Donors drinking

Donors drinking


0 12090603015

Time (min)

50

150

250

450

350

SaccharinGlucose

ed


GlucoseSucrose

Waterndash A B

Antibiotics

G l y c a e m i c r e s p o n s e

( A U C

times 1 0 3 )

20

40

50

30

0

10

P C

2 - 1 9 4

PC1 - 303

020

015

010

005

000

ndash005

ndash010

ndash015

ndash020ndash02 ndash 01 00 01 02 03 04

Saccharin W11

ControlsSaccharin W0

h

P C

2 - 1 9 1

PC1 - 396

020

015

010

005

000

ndash005ndash010

ndash015

ndash020ndash02 ndash01 00 01 02 03 04

SaccharinGlucose

Donorsdrinking




Figure 1 | Artificial sweeteners induce glucose intolerance transferable togerm-free mice a b Oral glucose tolerancetest (OGTT a ) andarea under thetwo-hour blood glucose response curve (AUC b) in normal-chow-fed micedrinking commercial NAS for 11 weeks before (N 520) and after antibioticsciprofloxacin and metronidazole (lsquoantibiotics Arsquo N 5 10) or vancomycin(lsquoantibiotics Brsquo N 5 5) c OGTT in mice fed HFD and commercial saccharin(N 5 10) or glucose (N 5 9) d OGTT of HFD-fed mice drinking 01mg ml21

saccharin or water for 5 weeks (N 5 20) followed by lsquoantibiotics Arsquo (N 5 10)e f OGTT of germ-free mice 6 days following transplant of microbiota fromcommercial saccharin- (N 5 12) and glucose-fed mice (N 5 11) (e) or pure

saccharin- (N 5 16) and water-fed (N 5 16) donors (f ) Symbols (OGTT) orhorizontal lines (AUC) mean error bars sem P 005 P 001P 0001 OGTT analysis of variance (ANOVA) and Bonferroni AUCANOVA and Tukey post hoc analysis Each experiment was repeated twiceg Principal coordinates analysis (PCoA) of weighted UniFrac distances basedon 16S rRNA analysis from saccharin-consuming mice at baseline (W0 black hexagons W11 blue triangles) water controls(black circles for W11 and W0)glucose (black squares for W11 and W0) or sucrose (black triangles forW11 and W0) N 5 5 in each group h PCoA of taxa in germ-free recipientsaccording to donor identity in e

RESEARCH ARTICLE

2 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4




mirrored in germ-free recipients of stools from saccharin-consuming donors (Extended Data Fig 6 right column) Likewise dysbiosis wasobserved in miceconsumingpure saccharin and HFD (Supplementary Table 1) Together these results demonstrate that saccharin consump-tion in various formulations doses and diets induces dysbiosis with over-all similar configurations

To study the functional consequencesof NASconsumption we per-formed shotgunmetagenomic sequencing of faecalsamples from before

and after 11 weeks of commercial saccharin consumption comparedto control miceconsuming either glucose or water To compare relativespeciesabundance we mapped sequencing reads to the human micro-biome project reference genome database16 In agreementwith the 16SrRNA analysis saccharintreatment inducedthelargestchangesin micro-bial relative species abundance (Fig 2a Supplementary Table 2 F-testP value 10210) These changes are unlikely to be an artefact of hori-zontal gene transfer or poorly covered genomes because changes inrelativeabundancewere observed acrossmuchof thelength ofthebacterialgenomes as exemplified by one overrepresented(BacteroidesvulgatusExtended Data Fig 7a)and one underrepresented species( Akkermansiamuciniphila Extended Data Fig 7b)

Wenextmapped themetagenomic reads to a gutmicrobialgene cata-logue evenly dividing reads mapping to more than one gene and then

groupinggenesinto KEGG(KyotoEncyclopediaof Genes andGenomes)pathwaysExamining pathways withgene coverageabove02 (115 path-ways)changes in pathway abundancewereinverselycorrelatedbetweencommercial saccharin- and glucose-consuming mice (R52045 P 1026 Fig 2b) Since commercial saccharin consists of 95 glucosethese results suggest that saccharin greatly affects microbiota functionNotably pathways overrepresented in saccharin-consuming mice in-clude a strong increase in glycan degradation pathways (Fig 2c d) inwhich glycansare fermentedto form variouscompoundsincluding short-chain fatty acids (SCFAs)17 These pathways mark enhanced energy har-

vestand their enrichment waspreviouslyassociated withobesityin mice11

and humans18 withSCFA possiblyserving as precursors andor signal-ling molecules for de novo glucose and lipid synthesis by the host19 Toidentify the underlying bacteria we annotated every read that mapped

to glycan degradation pathways by its originating bacteria Much of the increase in these pathways is attributableto reads originating fromfive Gram-negative and -positive species of which two belong to the

Bacteroides genus (Fig 2e) This is consistent with the sharp increasein theabundance of this genus in saccharin-consuming mice observedin the 16S rRNA analysis (Extended Data Fig 6) Consequently levelsof the SCFAs propionate andacetate measured in stool were markedly higher in commercialsaccharin-consumingmice compared to controlglucose-consuming mice (Fig2f g) reflectiveof the differential effectsmediated by chronicglucoseconsumption withand withoutNAS expo-sure Butyratelevels weresimilar between the groups (data not shown)

In addition to glycan degradation and similar to previous studieson humans with type 2 diabetes1320 other pathways were enriched inmicrobiomes of saccharin-consuming mice including starch and suc-rose metabolism fructose and mannose metabolism and folate gly-cerolipid and fatty acid biosynthesis (Supplementary Tables 3 and 4)whereas glucosetransport pathwayswere underrepresentedin saccharin-consuming mice(ExtendedData Fig7c) Mice consuming HFDand puresaccharin featured several enriched pathways (Extended Data Fig 7d)including ascorbate and aldarate metabolism (previously reported tobe enriched in leptin-receptor-deficient diabetic mice21) lipopolysac-charide biosynthesis (linked to metabolic endotoxemia22) andbacterialchemotaxis (previously reported to be enriched in obese mice11)

Altogether saccharin consumption resultsin distinctdiet-dependentfunctional alterations in the microbiota including normal-chow-related

expansion in glycan degradationcontributed by severalof the increasedtaxa ultimately resulting in elevated stool SCFA levels characteristic of increased microbial energy harvest11

NAS directly modulate the microbiota to induce glucoseintoleranceTo determine whether saccharin directly affects the gut microbiota wecultured faecal matterfrom naive mice under strictanaerobicconditions(75N220CO25 H2) inthe presenceof saccharin (5mg ml21) orcontrol growth media Cultures from day 9 of incubation were admi-nistered by gavage to germ-free mice (Extended Data Fig 8a) In vitrostool culture with saccharin induced an increase of the Bacteroidetesphylum and reduction in Firmicutes (Bacteroidetes 89 versus 70Firmicutes 6 versus 22Extended DataFig 8b)Transferringthis in

vitro saccharin-treated microbiota configuration into germ-free miceresultedin significantlyhigher glucose intolerance (P 0002)comparedwithgerm-free micereceiving the control culture (Fig 3a and Extended

a b

c

Saccharin

Water

Glucose

A b u n d a n c e f o l d c h a n g e

( w e e k 1 1 w e e k 0 )

100

101

102

103

10ndash3

10ndash2

10ndash1

Bacteria

Growth conditions

G r o w t h

c o n d i t i o n s

Water

Water

Glucose

Saccharin 031 ndash045

ndash069

Pearson correlation P value ( N = 115)

1 10ndash2010ndash10

W a t e r

G l u c

o s e

S a c c h a

r i n

De1047297nition

N -acetylgalactosamine-6-sulphatase

Iduronate 2-sulphatase

Heparan-α-glucosaminide N -acetyltransferase

α-N -acetylglucosaminidase

Hyaluronoglucosaminidase

N -acetylglucosamine-6-sulphatase

β-glucuronidase

Fold change(week 11week 0)

07 41

e

f

m M a c e t a t e m g f a e c e s

00

10

15

20

G l u c

o s e

C o m m e r c i a l

s a c c h a

r i n

05

g

m M p r o p i o n a t e m g f a e c e s

0

20

30

40

10

Bacteroides vulgatus

Akkermansia muciniphila

0 11

Water Glucose Sacchar in

Week 0 11 0 11

P e r c e n t o f s a m p l e

00

05

10

15

Other Unknown

Parabacteroides distasonis

Staphylococcus aureus

Providencia rettgeri


Bacteroides fragilis

G l u c

o s e

C o m m e r c i a l

s a c c h a

r i n

β-galactosidase

Sialidase-1

β-mannosidase

α-mannosidase

Glucosylceramidase

Endoglycosidase H

α-L-fucosidase

d

Figure 2 | Functional characterization of saccharin-modulated microbiotaa Species alterations in mice consuming commercial saccharin water orglucose for 11 weeks (N 5 4) shown by shotgun sequencing b Pairwisecorrelations between changes in 115 KEGG pathways across mice receiving different treatments c d Fold change in relative abundance of

glycosaminoglycan (c) or other glycan (d) degradation pathway genes

e Higher glycan degradation attributedto fivetaxa in the commercialsaccharinsetting f g Levels of faecal acetate and propionate at W11 in mice drinking commercial saccharin or glucose (N 5 5) Horizontal lines means error barssem P 001 two-sided unpaired Student t -test SCFA measurementswere performed on two independent cohorts

ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 3




Data Fig 8c) Similar to thecomposition of thesaccharin-supplementedanaerobic culture germ-free recipients of this cultured configurationfeatured over-representation of members of the Bacteroidesgenus andunder-representation of several Clostridiales (Fig 3b and Supplemen-tary Table 5)

Shotgun metagenomic sequencing analysis revealed that in vitrosaccharin treatment induced similar functional alterations to thosefound in mice consuming commercial saccharin (Fig 4c P 1024)

with glycan degradation pathways being highly enriched in both set-tings Other pathways highly enriched in both settings included thoseinvolved in sphingolipid metabolism previously shown to be over-represented in microbiomes of non-obese diabetic mice23 and com-mon under-represented pathways included glucose transport (Fig 3cand Extended Data 7c right column)

Collectively these results demonstrate that saccharin directly mod-ulates the composition and function of the microbiome and induces

dysbiosis accounting for the downstream glucose intolerance pheno-type in the mammalian host

NAS in humans associate with impaired glucosetoleranceTo study the effect of NAS in humans we examined the relationshipbetween long-term NAS consumption (based on a validated food fre-quency questionnaire see Methods) and various clinical parametersin data collected from 381 non-diabetic individuals (44 males and56 females age 4336 132) in an ongoing clinical nutritional studyWe found significant positive correlations between NAS consump-tion and several metabolic-syndrome-relatedclinicalparameters (Sup-plementary Table 6) including increased weight and waist-to-hip ratio

(measures of central obesity)higherfastingbloodglucose glycosylated

a


b

0 12090603015

Time (min)

50

150

200

100

Saccharin-treated

Control

Transplanted culture

Anaeroplasma (G)

Bacteroides (G)

Bacteroides (G)

Bacteroides (G)

Bacteroides acidifaciens (S)

Parabacteroides (G) Lactobacillus reuteri (S)

Candidatus Arthromitus (G)

Bacteroides (G)

Clostridiales (O)

Bacteroides (G)

Bacteroides (G)

Clostridiales (O)

Clostridium (G)

4390755

1135084

260410

178064

311482

4308591

355305

276149

419601

176061

190473

184753

234912

3750380

log

(fold difference)

Lowest taxonomic

level identi1047297edOTU no

log(fold difference)(saccharin-treatedcontrol)

ndash2 21

c KEGG pathway enrichment

r = 038

P lt 10ndash4

025 1 25

Mice consuming saccharin

(week 11week 0)

12

06

C u l t u r e t r a n s p l a n t

( s a c c

h a r i n P B S )

1

Sphingolipid metabolism

Glycosaminoglycan

degradation

Other glycandegradation

Glycosphingolipidbiosynthesis

PTS

Figure 3 | Saccharin directly modulates the microbiota a OGTT of germ-free mice 6 days following transplantation with saccharin-enriched or controlfaecal cultures (N 5 10 and N 59 respectively) Symbols means error barssem P 001 P 0001 ANOVA and Bonferroni post hoc analysisExperiments were repeated twice b Taxa representation in germ-free micefrom a O order G genus S species c Comparison of KEGG pathway abundance between W11 saccharin-consuming mice (compared to W0 x axis)and germ-free mice transplanted with microbiomes anaerobically cultured

with 5mg ml21 saccharin (compared to PBS y axis)

c


0 12090603015

Time (min)

50

150

200

100

Days 5ndash7Days 1ndash4

bNR1 R1NR2NR3

R2R3R4

e

g

0 12090603015

Time (min)

50

150

300

200

100

250

Day 7

Day 1

Transplanted sample

of responder 1

P C 4 - 7 7

PC2 - 163

020

010

000

ndash010

ndash020ndash03ndash02 ndash01 00 01 02 03

Responders

Non-responders

Days 1ndash4

a

Non-NAS

consumers

High-NAS

consumers

H b A 1 C

60

65

55

45

50

f

BacteroidalesClostridialesLactobacillalesYS2

RF32ErysipelotrichalesBurkholderiales

Day 1 Day 7 Day 1 Day 7

Non-responders Responders

Responders (R1ndash4)

h

0 12090603015

Time (min)

50

150

300

200

100

250 Day 7

Day 1

Transplanted sample

of non-responder 3

d

0 12090603015Time (min)

50

150

200

100


Non-responders (NR1ndash3)

1

Day

2 3 4 5 6 7


( A U C

times 1 0 2 )

80

0

20

40

60




Figure 4 | Acute saccharin consumption impairs glycaemic control inhumans by inducing dysbiosis a HbA1Cof high NASconsumers (N 5 40) versus non-consumers (N 5236) Rank sum P value 0002 b Daily incremental area under the curve (iAUC) of OGTT in 4 responders (R) and 3non-responders (NR) c d OGTT of days 1ndash4 versus days 5ndash7 in 4 responders(c) or 3 non-responders (d) e PCoA of 16S rRNA sequences from responders(N samples5 16) versus non-responders (N 5 9) during days 1ndash4 f Order-level relative abundance of taxa samples from days 1 and 7 of responders andnon-responders g h OGTT in germ-free mice (N 56) 6 days following faecaltransplantationof D1 or D7 samples of responder1 (R1 g ) or non-responder 3(NR3 h) Symbols means error bars sem P 005 P 001 ANOVAand Bonferroni post hoc analysis

RESEARCH ARTICLE

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4




haemoglobin (HbA1C)and glucose tolerance test (GTT measures of impaired glucose tolerance) and elevated serum alanine aminotrans-ferase (ALT measure of hepatic damage that is likely to be secondaryin thiscontextto non-alcoholicfatty liver disease)Moreover thelevelsof glycosylated haemoglobin (HbA1C)indicativeof glucose concen-tration over the previous 3 months were significantly increased whencomparing a subgroupof highNAS consumers (40 individuals)to non-NASconsumers (236individuals Fig4a ranksum P 0002)This in-

crease remained significant when corrected to body mass index (BMI)levels (rank sum P 0015) In this cohort we characterized the 16SrRNA in 172 randomly selected individualsNotably we found statisti-callysignificantpositivecorrelationsbetweenmultiple taxonomicentitiesand NAS consumption including the Enterobacteriaceae family (Pear-son r 5 036 FDR corrected P 1026) the Deltaproteobacteria class(Pearson r 5 033 FDR corrected P 1025) and the Actinobacteriaphylum (Pearson r 5 027 FDR corrected P 00003 Supplementary Table 7) Importantly we did not detect statistically significant corre-lations between OTU abundances and BMI suggesting that the abovecorrelations are not due to the distinct BMI of NAS consumers

Finally as an initialassessment of whether therelationshipbetweenhuman NAS consumption and blood glucose control is causative wefollowed seven healthy volunteers (5 males and2 females aged 28ndash36)

who donotnormallyconsumeNAS or NAS-containingfoodsfor 1 weekDuring this week participants consumed on days 2ndash7 the FDArsquos max-imal acceptable daily intake (ADI) of commercial saccharin (5 mg per kg (body weight)) as three divided daily doses equivalent to 120 mgand were monitored by continuous glucose measurements and daily GTT (Extended Data Fig 9a) Notably even in this short-term 7-day exposure period most individuals (4 out of 7) developed significantly poorer glycaemic responses 5ndash7 days after NAS consumption (here-after termed lsquoNASrespondersrsquo) comparedto their individual glycaemicresponse on days 1ndash4(Fig 4b c andExtended Data Fig 9b P 0001)None of the three NAS non-responders featured improved glucose tol-erance (Fig 4b d and Extended Data Fig 9c)

The microbiome configurations of NAS responders as assessed by 16S rRNA analysis clustered differently from non-responders bothbefore andafterNAS consumption (Fig 4e andExtended Data Fig 9drespectively) Moreover microbiomes from non-responders featuredlittle changesin compositionduring thestudy week whereaspronouncedcompositional changes were observed in NAS responders (Fig 4f andExtended Data Fig 9e) To study whether this NAS-induced dysbiosishas a causal role in generating glucose intolerance stool from before(day1D1)orafter(day7D7)NASexposureweretransferredfromtwoNAS responders and two NAS non-responders into groups of normal-chow-fed germ-free mice Indeed transfer of post-NAS exposure (D7)stool from NAS responders induced significant glucose intolerance inrecipient germ-free mice comparedto theresponse noted withD1 stooltransferred from the same NAS-responding individuals (Fig 4g andExtendedData Fig9f P 0004and Extended Data 9gh P 002) Incontrast D7 stools transferred into germ-free mice from the two NASnon-responders induced normal glucose tolerance which was indistin-guishable from that of mice transferred with D1 stools from the samelsquonon-respondingrsquo individuals (Fig 4h and Extended Data Fig 9indashk)Germ-free mice transplanted with lsquorespondersrsquo microbiome replicatedsome of the donor saccharin-induced dysbiosis including 20-fold rela-tive increase of Bacteroides fragilis (order Bacteroidales) and Weissellacibaria (order Lactobacillales) and approximately tenfold decrease inCandidatus Arthromitus (order Clostridiales) (Extended Data Fig 9l)

DiscussionIn summary our results suggest that NAS consumption in both miceand humans enhances the risk of glucose intolerance and that theseadverse metabolic effectsare mediated by modulation of thecomposi-tion and function of the microbiota Notably several of the bacterialtaxa that changed following NAS consumption were previously assoc-

iated with type 2 diabetes in humans1320 includingover-representation

of Bacteroides and under-representation of Clostridiales Both Gram-positive and Gram-negativetaxa contributed to the NAS-induced phe-notype (Fig 1a b) and were enriched for glycan degradation pathways(Extended Data Fig 6) previously linked to enhanced energy harvest(Fig 2c d)1124 This suggests that elaborate inter-speciesmicrobial coop-eration may functionally orchestrate the gut ecosystem and contributeto vital community activities in diverging environmental conditions(forexamplenormal-chow versus high-fat dietaryconditions) In addi-

tionwe showthatmetagenomesofsaccharin-consumingmiceareenrichedwith multiple additional pathways previously shown to associate withdiabetes mellitus23 or obesity 11 in mice and humans including sphin-golipid metabolism and lipopolysaccharide biosynthesis25

Our results fromshort-and long-termhumanNASconsumer cohorts(Fig 4Extended Data Fig9 andSupplementary Tables6 7) suggest thathuman individuals feature a personalized response to NAS possibly stemming from differences in their microbiota composition and func-tion The changes noted in our studies may be further substantiated inmice consuming different humandiets26Similarlywebelievethatotherindividualizednutritional responsesmay be drivenby personalized func-tional differences in the microbiome As such lsquopersonalized nutritionrsquoleading to lsquopersonalized medical outcomersquo may underlie the variablenutritional effects notedin many multi-factorial diseasesand warrants

further studiesArtificial sweeteners were extensively introduced into our diets withthe intention of reducing caloric intakeand normalizing blood glucoselevels without compromising the human lsquosweet toothrsquo Together withother major shifts that occurred in human nutrition this increase inNAS consumption coincides with the dramatic increase in the obesity anddiabetesepidemicsOurfindingssuggestthatNASmayhave directly contributed to enhancingthe exactepidemicthattheythemselves wereintended to fight Moreover our results point towards the need to de-

velop new nutritional strategies tailored to the individual while integ-rating personalized differences in the composition and function of thegut microbiota

Online Content Methods along with any additional Extended Data display itemsandSourceDataare available in theonline versionof thepaper references uniqueto these sections appear only in the online paper

Received 27 March accepted 28 August 2014

Published online 17 September 2014

1 Gardner C et al Nonnutritive sweeteners current use and health perspectivesDiabetes Care 35 1798ndash1808 (2012)

2 Fitch C amp Keim K S Position of the Academy ofNutrition andDietetics use ofnutritive and nonnutritive sweeteners Journal of the Academy of Nutrition and Dietetics 112 739ndash758 (2012)

3 Tordoff M G amp Alleva A M Effect of drinking soda sweetenedwithaspartame orhigh-fructose cornsyrup on food intake and body weight Am J Clin Nutr 51963ndash969 (1990)

4 HorwitzD L McLane M amp Kobe P Response to single dose of aspartame orsaccharin by NIDDM patients Diabetes Care 11 230ndash234 (1988)

5 NettletonJ Aet al Dietsoda intake and risk of incidentmetabolic syndromeandtype2 diabetesin theMulti-Ethnic Study of Atherosclerosis (MESA) Diabetes Care32 688ndash694 (2009)

6 Roberts A Renwick A G Sims J amp Snodin D J Sucralose metabolism andpharmacokinetics in man Food Chem Toxicol 38 (Suppl 2) 31ndash41 (2000)7 ByardJ L amp Goldberg L Themetabolismof saccharinin laboratoryanimals Food

Cosmet Toxicol 11 391ndash402 (1973)8 Clemente J C Ursell L K Parfrey L W amp Knight R The impact of the gut

microbiota on human health an integrative view Cell 148 1258ndash1270 (2012)9 Claesson MJ etal Gutmicrobiota composition correlates with dietand health in

the elderly Nature 488 178ndash184 (2012)10 Muegge B D et al Diet drives convergence in gut microbiome functions across

mammalian phylogeny and within humans Science 332 970ndash974 (2011)11 Turnbaugh P J et al An obesity-associated gut microbiome with increased

capacity for energy harvest Nature 444 1027ndash1031 (2006)12 Ley R E Turnbaugh P J Klein S amp Gordon J I Microbial ecology human gut

microbes associated with obesity Nature 444 1022ndash1023 (2006)13 Qin J et al A metagenome-wide association study of gut microbiota in type 2

diabetes Nature 490 55ndash60 (2012)14 Henao-MejiaJ et al Inflammasome-mediated dysbiosis regulates progressionof

NAFLD and obesity Nature 482 179ndash185 (2012)15 David L A et al Diet rapidly and reproducibly alters the human gut microbiome

Nature 505 559ndash563 (2014)

ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 5




16 Peterson J et al The NIH human microbiome project Genome Res 192317ndash2323 (2009)

17 Koropatkin N M Cameron E A amp Martens E C How glycan metabolismshapes the human gut microbiota Nature Rev Microbiol 10 323ndash335(2012)

18 Schwiertz A et al Microbiota and SCFA in lean and overweight healthy subjectsObesity 18 190ndash195 (2010)

19 BergmanE N Energycontributions ofvolatilefattyacidsfromthe gastrointestinaltract in various species Physiol Rev 70 567ndash590 (1990)

20 Karlsson F H et al Gut metagenome in European women with normal impairedand diabetic glucose control Nature 498 99ndash103 (2013)

21 Connor SC Hansen M K Corner A Smith R F amp Ryan T E Integration ofmetabolomics and transcriptomics data to aid biomarker discovery in type2diabetes Mol Biosyst 6 909ndash921 (2010)

22 Cani P D et al Changes in gut microbiota control metabolic endotoxemia-induced inflammation in high-fat diet-induced obesity and diabetes in miceDiabetes 57 1470ndash1481 (2008)

23 MarkleJ G etalSexdifferencesin the gut microbiomedrive hormone-dependentregulation of autoimmunity Science 339 1084ndash1088 (2013)

24 Sonnenburg J L et al Glycan foraging in vivo by an intestine-adapted bacterialsymbiont Science 307 1955ndash1959 (2005)

25 Cani P D et al Metabolic endotoxemia initiates obesity and insulin resistanceDiabetes 56 1761ndash1772 (2007)

26 Smith M I et al Gut microbiomes of Malawian twin pairs discordant forkwashiorkor Science 339 548ndash554 (2013)

Supplementary Information is available in the online version of the paper

Acknowledgements We thank the members of the Elinav and Segal laboratories fordiscussionsWe acknowledgeC Bar-Nathanfor germ-freemousecaretaking We thank

the Weizmann Institute management and the Nancy and Stephen Grand Israel

National Center for Personalized Medicine (INCPM) for providing financial andinfrastructuresupport We thank G Malka N Kosowerand R Bikovskyfor coordinatingthe human clinical trials and M Pevsner-Fischer T Avnit-Sagi and M Lotan-Pompanfor assistance with microbiome sample processing CAT is the recipient of aBoehringer Ingelheim Fonds PhD Fellowship GZ-S is supported by the Morris KahnFellowshipsfor SystemsBiology This workwas supported by grants fromthe NationalInstitute of Health (NIH) and the European Research Council (ERC) to ES andsupportand grants to EE providedby Y and R Ungar the Abisch FrenkelFoundationfor the Promotion of Life Sciences the Gurwin Family Fund for Scientific ResearchLeona M and Harry B Helmsley Charitable Trust Crown Endowment Fund forImmunological Research estate of J Gitlitz estate of L Hershkovich Rising Tide

foundation Minerva Stiftung foundation and the European Research Council EE isthe incumbent of the Rina Gudinski Career Development Chair

Author Contributions JS conceived the project designed and performedexperiments interpreted the results and wrote the manuscript TK DZ and GZ-Sperformed the computational and metagenomic microbiota analysisand the analysisof theretrospective and prospectivehuman study and are listed alphabetically CATOMAW and HS helped with experiments YK helped with the metabolic cageexperiments SG designed the metagenomic library protocols and generated thelibraries IK-G performed the SCFA quantification experiments DI NZ and ZHperformed and supervised human experimentation AH supervised the germ-freemouse experiments ES and EE conceived and directed the project designedexperiments interpreted the results and wrote the manuscript

Author Information Sequencing data are deposited in the European NucleotideArchive accession PRJEB6996 Reprints and permissions information is available atwwwnaturecomreprints The authors declare no competing financial interestsReadersare welcome to commenton the online version of thepaper Correspondenceandrequests formaterials should be addressed to ES(eransegalweizmannacil) or

EE (eranelinavweizmannacil)

RESEARCH ARTICLE

6 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4




METHODSMice C57Bl6 WT adult male mice were randomly assigned (without blinding)to treatment groups and were given commercial artificial sweeteners (saccharin-sucralose- or aspartame-based) or pure saccharin (Sigma Aldrich) in drinking water and fed a high-fat (HFD D12492 60 kcal from fat Research Diets) orstandard polysaccharide normal-chow diet (Harlan-Teklad) Compared groupswere always fed from the same batch of diet For antibiotic treatment mice weregiven a combination of ciprofloxacin (02g l21) and metronidazole (1g l21) or

vancomycin (05 g l21) in their drinking waterAll antibiotics were obtained from

Sigma Aldrich Adult male outbred Swiss-Webster mice (a widely used mousestrain in germ-free experiments) served as recipients for faecal transplants andwere housedin sterile isolators (Park Bioservices) Forfaecal transplantation experi-ments 200 mg of stool (from mouse pellets or human swabs) was resuspended in5 mlof PBSunderanaerobicconditionsvortexedfor3 minandallowedto settleby gravity for 2 min Transplant into recipient mice were achieved by gavage with200ml of the supernatant andmaintained on standardnormal-chow dietand waterthroughout the experiment All animal studies were approved by the WeizmannInstitute of Science Institutional Animal Care and Usage committee (IACUC)applicationnumbers 08680114-3 and 00550113-3 all animalexperiments involving transfer ofhumanmicrobiota into mice were approvedby theWeizmannInstituteof ScienceBioethicsand Embrionic Stem CellResearch oversight (ESCRO) committee

Artificial and caloric sweeteners The following commercially available NASwere dissolved in mice drinking water to obtain a 10 solution Sucrazit (5 sac-charin 95 glucose)Sucralite (5Sucralose) Sweetrsquon LowGold (4 Aspartame)

10glucose(J T Baker)and 10sucrose(Sigma Aldrich) solutionswereusedforcontrols The administered doses of 10 commercialNAS dissolved in waterwerewell below their reported toxic dose (63 g perkg (bodyweight)2716gperkg(body weight)28 and 4 g per kg (body weight)29 for saccharin sucralose and aspartamerespectively) For experiments conducted with pure saccharin (Sigma Aldrich) a01 mgml21 solution wasused inorder to meetwithFDA definedADIfor saccharinin humans (5 mg per kg (body weight)) according to the following calculation

ADI 5mg kg 1 day 1|average mouse weight 003 kg

Average daily liquid intake 2 ml ~

0075 mg ml1 01 mg ml1

Glucose and insulin tolerance tests Mice were fasted for 6 h during the lightphase with free access to water In all groups of mice where the drinking regime

was other than water it was substituted for water for the period of the fasting andglucose or insulin tolerance test Blood from the tail vein was used to measureglucose levels using a glucometer (Bayer) immediately before and 15 30 60 90and120 minafteroral feeding with 40 mg glucose (JT Baker)or intra-peritonealinjection with 01 U per kg (body weight) Insulin (Biological Industries) Plasmafasting insulin levelswere measured in sera collected immediately beforethe startof GTT using ELISA (Ultra Sensitive Mouse Insulin ELISA Kit Crystal Chem)

Metabolic studies Foodand drinkintake and energyexpenditure were measuredusing the PhenoMaster system (TSE-Systems Bad Homburg Germany) whichconsists of a combinationof sensitive feeding sensors for automatedmeasurementand a photobeam-based activity monitoring system detects and records ambulat-ory movements includingrearingand climbingin each cageAll parametersweremeasured continuously and simultaneously Mice were trained singly-housed inidentical cages before dataacquisitionTo calculatetotalcaloricintake thefollowing

values were used Chow 3 kcal g 21 sucrose 03938 kcal ml21 glucose 04kcal ml21saccharin 038kcal ml21 sucralose 0392 kcalml21 and aspartame 038 kcal ml21

In vitro anaerobic culturing Pooledfaecalmatterfrom naive adult WT C57Bl6male mice wasresuspended in 5 ml PBSin an anaerobicchamber (Coy Laboratory Products 75 N2 20 CO2 5 H2) vortexed for 3 min and allowed to settle by gravity for 2 min 500ml of the supernatant were added to a tube containing Chopped Meat Carbohydrate Broth PR II (BD) and 500 ml of a 5 mg ml21 sac-charin solution or an equal volume of PBS Every 3 days 500 ml of culture werediluted to fresh medium containing saccharin or PBS After 9 days cultures wereused for inoculation of germ-free mice

Taxonomic microbiota analysis Frozenfaecalsamples were processedfor DNAisolation using the MoBio PowerSoil kit according to the manufacturerrsquos instruc-tions1 ngof thepurified faecalDNAwas used forPCRamplification andsequencing ofthe bacterial16S rRNA gene365bpAmplicons spanningthe variable region 2(V2)of the16S rRNAgenewere generated by using thefollowingbarcoded primersFwd 59-AGAGTTTGATCCTGGCTCAG-3 9 Rev 59-TGCTGCCTCCCGTAGGAGT-39 The reactions were subsequently pooled in an equimolar ratio purified(PCR clean kit Promega)and used forIlluminaMiSeq sequencing to a depth of at

least 18000 reads per sample (mean reads per sample 1391486 5264 (sem))

Reads were then processed using the QIIME (quantitative insights into microbialecology)analysis pipeline as described3031 version18 Paired-end joined sequenceswere grouped into operational taxonomic units (OTUs) using the UCLUST algo-rithmand the GreenGenes database32 Sequenceswith distance-based similarity of 97 or greater over median sequence length of 353 bp were assigned to the sameOTU Samples were grouped according to the treatment Analysis was performedat eachtaxonomical level (phylum genus andOTU level)separately Foreach taxonG test was performedbetween the differentgroups P valueswere FDR-correctedformultiple hypothesis testing

Shotgun pyrosequencing and sequence mapping This was performed as prev-iously described33 with the following modifications 1 mg of DNA was shearedusing the Covaris 5200 system (Covaris Inc Woburn MA USA) followed by end repair ligation to adapters an 8-cycle PCR amplification (Kappa HiFi) andsequenced using an Illumina HiSeq to a minimal depth of 11773345 reads persample(mean readsper sample202960866 637379 (sem)read length51 bp)Illumina sequence reads were mapped to the human microbiome reference gen-omedatabaseoftheHumanMicrobiomeProject( httphmpdaccorgHMREFGref 16) and to a gut microbial gene catalogue34 using GEM mapper35 with the fol-lowing parameters

-m 3 ndashs 0 ndashq offset-33 ndashgem-quality-threshold 26

Microbial species abundance was measured as the fraction of reads that mapped

to a singlespecies in the database An expectationndashmaximization(EM) algorithmadapted from Pathoscope36 was employed to determinethe correct assignment of reads that mapped to more than one species We considered only species forwhich at least 10 of the genome was covered (each coverage bin was 10000-bplong) in at least oneof thegrowthconditions (saccharin wateror glucose) Readsmapped to the gut microbial gene catalogue were assigned a KEGG ID according to the mapping available with the catalogue Genes were subsequently mapped toKEGG pathways and only pathways whose gene coverage was above 02 wereincluded To calculate the contribution of different bacteria to the overrepresen-tation ofglycandegradationpathways reads thatweremappedto genes in thegutmicrobial genecataloguethat belong to glycan degrading pathways were extractedand re-mapped the HMP reference genome database seeking germs that had thehighest contribution

Short chain fatty acid quantification To determine the level of free fatty acidsanalytic HPLC (Agilent 1260) was performed as described previously 37 In brief

standard solutions of acetate butyrate and propionate (all from Sigma-Aldrich)were prepared at various concentrations (001ndash02 M) These solutions were ana-lysed using HPLC successive with quadrupole time-of-flight mass spectrometry with a step-gradient of solvent solution from 0 to 60 of CH 3CN with 01formic acid to obtain calibration curve for each fatty acid Faecal media sampleswere dissolvedwith01 formicacid andanalysedin a similar mannerto measurethe total concentration of all three free fatty acids

Analysis of the relationship between NAS consumption and clinical para-meters in humans All human studies were approved by the Tel Aviv Sourasky Medical Center InstitutionalReview Board approvalnumber TLV-0658-12 TLV-0050-13 and TLV-0522-10 Kfar Shaul Hospital Institutional Review Board ap-proval number 0-73and theWeizmann Instituteof Science Bioethicsand EmbrionicStem Cell Research oversight (ESCRO) committee The trial was reported to httpclinicaltrialsgov identifier NCT01892956 The study did not necessitate or involverandomizationFor eachindividual in the clinicalnutritional study after signinganinformed consent multiple parameters were collected including BMI body circum-

ferences fasting glucose levels general questionnaire complete blood counts and gen-eralchemistry parametersa validated long-term foodfrequencyquestionnaire38ndash40

Long-term NAS consumption was quantified directly from answers to anexplicit question regarding artificial sweeteners that participants filled out in theirfood frequency questionnaire We then used the Spearman correlation to exam-inethe relationshipbetweenNAS consumptionand each ofthe above parametersand FDR corrected for the multiple hypotheses tests performed

Statistics Thefollowing statistical analyseswere used in GTTa two-wayANOVAandBonferronipost-hoc analysiswere usedto comparebetweengroupsin differenttime-points and one-way ANOVA and Tukey post hoc analysis or unpaired two-sided Student t -testwere used to compare between AUC of multiple or twogroupsrespectively Bartlettrsquosor F-test forequal variancewere employed andno significantdifference was observed between variances of the compared groups For compar-ison of taxonomic data a G-test was used and P values were FDR-corrected formultiplehypothesistesting In metagenomicsand clinicaland taxonomicdata fromhumans Pearson andSpearman wereusedfor correlationtestsand MannndashWhitney

U wasusedto compareclinicalparameters betweengroups P 005 was considered

ARTICLE RESEARCH




significant in all analyses ( denotes P 005 P 001 P 0001) In allrelevant panels symbols or horizontal lines represent the mean and error barsrepresent sem For mouse experiments cohort sizes match common practice of the described experiments For human experiments sample size was chosen to

validate statistical analyses No mice or data points were excluded from analysesIn the human studies all humans older than 18 years of age who enrolled wereincluded Exclusion criteria included pregnancy

27 Taylor J D Richards R K amp Wiegand R G Toxicological studies with sodiumcyclamate and saccharin Food Cosmet Toxicol 6 313ndash327 (1968)

28 GoldsmithL A Acuteand subchronictoxicity of sucralose FoodChem Toxicol 38(Suppl 2) 53ndash69 (2000)

29 Magnuson B A et al Aspartame a safety evaluation based on current use levelsregulations and toxicological and epidemiological studies Crit Rev Toxicol 37629ndash727 (2007)

30 Caporaso J G et al QIIME allows analysis of high-throughput communitysequencing data Nature Methods 7 335ndash336 (2010)

31 Elinav E etal NLRP6 inflammasomeregulatescolonicmicrobial ecology andriskfor colitis Cell 145 745ndash757 (2011)

32 DeSantisT Z etal Greengenes a chimera-checked16S rRNAgenedatabase andworkbench compatiblewith ARB Appl Environ Microbiol 72 5069ndash5072 (2006)

33 Blecher-Gonen R et al High-throughput chromatin immunoprecipitation forgenome-wide mappingof invivo proteinndashDNA interactions and epigenomicstatesNature Protocols 8 539ndash554 (2013)

34 Qin J et al A human gut microbial gene catalogue established by metagenomicsequencing Nature 464 59ndash65 (2010)

35 Marco-Sola S Sammeth M Guigo R amp Ribeca P The GEM mapper fastaccurate and versatile alignment by filtration Nature Methods 9 1185ndash1188(2012)

36 Francis O E et al Pathoscope species identification and strain attribution withunassembled sequencing data Genome Res 23 1721ndash1729 (2013)

37 Kolodkin-Gal I et al D-amino acids trigger biofilm disassembly Science 328

627ndash629 (2010)38 Shahar D Fraser D Shai I amp Vardi H Development of a food frequency

questionnaire (FFQ) foran elderlypopulationbasedon a populationsurvey J Nutr133 3625ndash3629 (2003)

39 Shahar D Shai I Vardi H Brener-Azrad A amp Fraser D Development of asemi-quantitative Food Frequency Questionnaire (FFQ) to assess dietary intakeof multiethnic populations Eur J Epidemiol 18 855ndash861 (2003)

40 Shai I et al Dietary evaluation and attenuation of relative risk multiplecomparisons between blood and urinary biomarkers food frequencyand 24-hour recall questionnaires the DEARR study J Nutr 135 573ndash579(2005)

RESEARCH ARTICLE




Extended Data Figure 1 | Experimental scheme 10-week-old C57Bl6 malemice were treated with the following dietary regimes a Drinking commercially

available non-caloric artificial sweeteners (NAS saccharin sucralose andaspartame) or glucose sucrose or water as controls and fed a normal-chow (NC) diet b Drinking commercially available saccharin or glucose as control

andfed a high-fatdiet (HFD)c Drinkingpuresaccharinor waterand fedHFDd As in c but with outbred Swiss-Webster mice Glucose tolerance tests

microbiome analysis and supplementation of drinking water with antibioticswere performed on the indicated time points e Schematic of faecal transplantexperiments

ARTICLE RESEARCH




Extended Data Figure 2 | Artificial sweeteners induce glucose intolerancea AUC of mice fed HFD and commercial saccharin ( N 5 10) or glucose(N 5 9) b AUC of HFD-fed mice drinking 01mg ml21 saccharin or water for5 weeks (N 5 20) followed by lsquoantibiotics Arsquo (N 5 10) c d OGTT and AUC of HFD-fed outbred Swiss-Webster mice (N 5 5) drinking pure saccharin orwater e f Faecal samples were transferred from donor mice (N 5 10) drinking commercially available pure saccharin glucose or water controls into

8-week-old male Swiss-Webster germ-free recipient mice AUC of germ-freemice 6 days following transplant of microbiota from commercial saccharin-(N 5 12) and glucose-fed mice (N 5 11) (e) or pure saccharin- (N 5 16) andwater-fed (N 5 16) donors (f ) Symbols (GTT) or horizontal lines (AUC)means error bars sem P 005 P 001 P 0001 ANOVA andTukey post hoc analysis (GTT) or unpaired two-sided Student t-test (AUC)Each experiment was repeated twice

RESEARCH ARTICLE




Extended Data Figure 3 | Metabolic characterization of mice consuming commercial NAS formulations 10-week-old C57Bl6 mice (N 5 4) weregiven commercially available artificial sweeteners (saccharin sucralose andaspartame) or controls (water sucrose or glucose N 5 4 ineachgroup) andfednormal-chow diet After 11 weeks metabolic parameters were characterizedusing the PhenoMaster metabolic cages system for 80 h Light and dark phases

are denoted by white and black rectangles on the x -axis respectively and grey

bars for the dark phase a Liquids intake b AUC of a c Chow consumptiond AUC of c e Total caloric intake from chow and liquid during 72 h (seemethods for calculation) f Respiratory exchange rate (RER) g AUC of f h Physical activity as distance i AUC of h j Energy expenditure k Masschange compared to original mouse weight during 15 weeks (N 5 10) l AUCof k The metabolic cages characterization and weight-gain monitoring were

repeated twice

ARTICLE RESEARCH




Extended Data Figure 4 | Metabolic characterization of mice consuming HFD and pure saccharin or water 10-week-old C57Bl6 mice (N 5 8) werefedHFDwithor withoutsupplementingdrinking water with 01mg ml21 puresaccharin After 5 weeks metabolic parameters were characterized using thePhenoMaster metabolic cages system for 70h Light and dark phases are

denoted by white and black rectangles on the x -axisrespectively and grey barsfor the dark phase a Liquids intake b AUC of a c Chow consumptiondAUCof c e Respiratory exchange rate (RER)f AUCof e g Physical activity as distance h AUC of g i Energy expenditure The metabolic cagescharacterization was repeated twice

RESEARCH ARTICLE




Extended Data Figure 5 | Glucose intolerant NAS-drinking mice display normal insulin levels and tolerance a Fasting plasma insulin measured after11 weeks of commercial NAS or controls (N 5 10) b Same as a but measured

after 5 weeksof HFD and puresaccharin or water (N 520) c Insulin tolerance

test performed after 12 weeks of commercial NAS or controls (N 5 10)Horizontal lines (a b) or symbols (c) represent means error bars sem Allmeasurements were performed on two independent cohorts

ARTICLE RESEARCH




Extended Data Figure 6 | Dysbiosis in saccharin-consuming mice andgerm-free recipients Heat map representing W11 logarithmic-scale foldtaxonomic differences between commercial saccharin and water or caloricsweetener consumers (N 5 5 in each group) Right column taxonomicaldifferences in germ-free mice following faecal transplantation fromcommercial saccharin- (recipients N 5 15) or glucose-consuming mice(N 5 13) OTU number (GreenGenes) and the lowest taxonomic levelidentified are denoted

RESEARCH ARTICLE




Extended Data Figure 7 | Functional analysis of saccharin-modulatedmicrobiota a b Changes in bacterial relativeabundanceoccur throughoutthebacterial genome Shown are changes in sequencing coverage along 10000 bpgenomic regions of Bacteroides vulgatus (a ) and Akkermansia muciniphila(b) with bins ordered byabundance in week 0 of saccharin-treatedmice cFold

change in relative abundance of modules belonging to phosphotransferase

systems (PTS) between week 11 and week 0 in mice drinking commercialsaccharin glucose or water Module diagram source KEGG databased Enriched KEGG pathways (fold change 138 as cutoff) in mice consuming HFD and pure saccharin versus water compared to the fold change in relativeabundance of the same pathways in mice consuming commercial saccharin

(week 11week 0)

ARTICLE RESEARCH




Extended Data Figure 8 | Saccharin directly modulates the microbiotaa Experimental schematic b Relative taxonomic abundance of anaerobically culturedmicrobiota c AUC of germ-freemice 6 daysfollowingtransplantation

with saccharin-enriched or control faecal cultures (N 5 10 and N 59respectively) Horizontal lines means error bars sem P 001 unpairedtwo-sided Student t-test The experiment was repeated twice

RESEARCH ARTICLE




Extended Data Figure 9 | Impaired glycaemic control associated with acutesaccharin consumption in humans is transferable to germ-free micea Experimental schematic (N 5 7) b c Daily incremental AUC of days 1ndash4 versus days 5ndash7 in four responders (b) or three non-responders(c) d Principalcoordinates analysis (PCoA) of weighted UniFrac distances of 16S rRNAsequencesdemonstrating separation on principal coordinates2 (PC2)3 (PC3)and 4 (PC4) of microbiota from responders (N samples5 12) versus non-responders (N 5 8) during days 5ndash7 e Order-level relative abundance of taxasamples from days1ndash7 of responders and non-responders f AUCin germ-freemice (N 5 6) 6 daysfollowingfaecal transplantationfrom samples of responder1 (R1) collected before and after 7 days of saccharin consumption g h OGTTand AUC in germ-free mice (N 5 5) 6 days after receiving faecal samples

collected from responder 4 (R4) before and after 7 days of saccharinconsumption i AUC in germ-free mice (N 5 5) 6 days following faecaltransplantation from samples of non-responder 3 (NR3) collected before andafter 7 days of saccharin consumption j k OGTT and AUC in germ-free mice(N 5 5) 6 days after receiving faecal samples collected from non-responder 2(NR2) before and after 7 days of saccharin consumption l Fold taxonomicalabundance changes of selected OTUs altered in germ-free recipients of D7 versus D1 microbiomes from R1 Dot colour same as in e bacterial ordersSymbols (GTT) or horizontal lines (AUC) means error bars sem P 005P 001P 0001 two-way ANOVAand Bonferronipost-hoc analysis(GTT) unpaired two-sided Student t-test (AUC)

ARTICLE RESEARCH



Fasting serum insulin levels and insulin tolerance were also similar inall mouse groups consuming NAS or caloric sweeteners in both thenormal-chow andHFD settings(Extended Data Fig5) Taken togetherthese results suggest that NAS promote metabolic derangements in arange of formulations doses mouse strains anddiets parallelinghumanconditions in both the lean and the obese state

Gut microbiota mediates NAS-induced glucoseintoleranceSince diet modulates the gut microbiota15 and microbiota alterationsexert profound effects on host physiology and metabolism we testedwhether the microbiota may regulate theobservedNAS effects To thisend we treated mouse groups consuming commercial or pure NAS inthe lean andHFDstates (ExtendedDataFig1ac) with a Gram-negative-targeting broad-spectrum antibiotics regimen (designated lsquoantibioticsArsquo)of ciprofloxacin (02 g l21) andmetronidazole (1 g l21)whilemain-taining mice on their diet and sweetener supplementation regimensNotablyafter4 weeks of antibiotictreatment differences in glucose in-tolerance between NAS-drinking mice and controls were abolishedboth in the lean (Fig 1a b) and the obese (Fig 1d and Extended DataFig 2b) states Similar effects were observed with the Gram-positive-targeting antibiotic vancomycin (lsquoantibiotics Brsquo 05 g l21 Fig 1a b)These resultssuggestthat NAS-inducedglucose intoleranceis mediatedthrough alterations to the commensal microbiota with contributionsfrom diverse bacterial taxa

To test whether the microbiota role is causal we performed faecaltransplantationexperiments by transferringthemicrobiota configura-

tion from mice on normal-chow diet drinking commercial saccharin

or glucose (control) into normal-chow-consuming germ-free mice(Extended Data Fig 1e) Notably recipients of microbiota from miceconsuming commercial saccharin exhibited impaired glucosetoleranceas compared to control(glucose) microbiota recipientsdetermined6 daysfollowing transfer (P 003 Fig 1e andExtended Data Fig 2e) Trans-ferring the microbiota composition of HFD-consuming mice drinking water or pure saccharin replicated the glucose intolerance phenotype(P 0004 Fig 1f and Extended Data Fig 2f) Together these resultsestablish that the metabolic derangements induced by NAS consump-tion are mediated by the intestinal microbiota

NAS mediate distinct functional alterations to themicrobiota We next examined the faecal microbiota composition of our variousmouse groupsby sequencing their 16Sribosomal RNAgeneMicedrink-ing saccharin had a distinct microbiota composition that clustered sepa-rately fromboth their starting microbiome configurations and fromallcontrol groups at week 11 (Fig 1g) Likewise microbiota in germ-freerecipients of stoolsfromsaccharin-consumingdonor mice clustered sepa-rately fromthatof germ-free recipients of glucose-drinking donor stools(Fig 1h) Compared to all control groups the microbiota of saccharin-consuming mice displayed considerable dysbiosis with more than 40operational taxonomic units(OTUs) significantly alteredin abundance(false discoveryrate (FDR) corrected P value 005 foreach OTU Ex-tendedDataFig6 SupplementaryTable1)Manyof thetaxathatincreasedin relative abundance belonged to theBacteroidesgenus and Clostridialesorder withother membersof theClostridialesordercomprising themajor-

ity of under-representedtaxa along withLactobacillusreuteriandwere

WaterPure saccharin

ndash A Antibiotics

a b

c


0 12090603015

Time (min)

50

150

250

350

0 12090603015

Time (min)

150

250

450

350


GlucoseSucrose

Waterndash A B

Antibiotics

f

0 12090603015

Time (min)

50

150

250

200

100

g

0 12090603015

Time (min)

50

150

250

200

100 Pure saccharin

Water

Commercial saccharin

Glucose

Donors drinking

Donors drinking


0 12090603015

Time (min)

50

150

250

450

350

SaccharinGlucose

ed


GlucoseSucrose

Waterndash A B

Antibiotics


( A U C

times 1 0 3 )

20

40

50

30

0

10

P C

2 - 1 9 4

PC1 - 303

020

015

010

005

000

ndash005

ndash010

ndash015

ndash020ndash02 ndash 01 00 01 02 03 04

Saccharin W11

ControlsSaccharin W0

h

P C

2 - 1 9 1

PC1 - 396

020

015

010

005

000

ndash005ndash010

ndash015

ndash020ndash02 ndash01 00 01 02 03 04

SaccharinGlucose

Donorsdrinking




Figure 1 | Artificial sweeteners induce glucose intolerance transferable togerm-free mice a b Oral glucose tolerancetest (OGTT a ) andarea under thetwo-hour blood glucose response curve (AUC b) in normal-chow-fed micedrinking commercial NAS for 11 weeks before (N 520) and after antibioticsciprofloxacin and metronidazole (lsquoantibiotics Arsquo N 5 10) or vancomycin(lsquoantibiotics Brsquo N 5 5) c OGTT in mice fed HFD and commercial saccharin(N 5 10) or glucose (N 5 9) d OGTT of HFD-fed mice drinking 01mg ml21

saccharin or water for 5 weeks (N 5 20) followed by lsquoantibiotics Arsquo (N 5 10)e f OGTT of germ-free mice 6 days following transplant of microbiota fromcommercial saccharin- (N 5 12) and glucose-fed mice (N 5 11) (e) or pure

saccharin- (N 5 16) and water-fed (N 5 16) donors (f ) Symbols (OGTT) orhorizontal lines (AUC) mean error bars sem P 005 P 001P 0001 OGTT analysis of variance (ANOVA) and Bonferroni AUCANOVA and Tukey post hoc analysis Each experiment was repeated twiceg Principal coordinates analysis (PCoA) of weighted UniFrac distances basedon 16S rRNA analysis from saccharin-consuming mice at baseline (W0 black hexagons W11 blue triangles) water controls(black circles for W11 and W0)glucose (black squares for W11 and W0) or sucrose (black triangles forW11 and W0) N 5 5 in each group h PCoA of taxa in germ-free recipientsaccording to donor identity in e

RESEARCH ARTICLE

2 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4


















a b

c

Saccharin

Water

Glucose


( w e e k 1 1 w e e k 0 )

100

101

102

103

10ndash3

10ndash2

10ndash1

Bacteria

Growth conditions

G r o w t h

c o n d i t i o n s

Water

Water

Glucose


ndash069



W a t e r

G l u c

o s e

S a c c h a

r i n

De1047297nition







β-glucuronidase


07 41

e

f


00

10

15

20

G l u c

o s e

C o m m e r c i a l

s a c c h a

r i n

05

g


0

20

30

40

10



0 11


Week 0 11 0 11


00

05

10

15

Other Unknown






G l u c

o s e

C o m m e r c i a l

s a c c h a

r i n

β-galactosidase

Sialidase-1

β-mannosidase

α-mannosidase

Glucosylceramidase

Endoglycosidase H

α-L-fucosidase

d




ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 3











a


b

0 12090603015

Time (min)

50

150

200

100

Saccharin-treated

Control


Anaeroplasma (G)

Bacteroides (G)

Bacteroides (G)

Bacteroides (G)




Bacteroides (G)

Clostridiales (O)

Bacteroides (G)

Bacteroides (G)

Clostridiales (O)

Clostridium (G)

4390755

1135084

260410

178064

311482

4308591

355305

276149

419601

176061

190473

184753

234912

3750380

log

(fold difference)

Lowest taxonomic



ndash2 21


r = 038

P lt 10ndash4

025 1 25


(week 11week 0)

12

06


( s a c c

h a r i n P B S )

1


Glycosaminoglycan

degradation



PTS



c


0 12090603015

Time (min)

50

150

200

100


bNR1 R1NR2NR3

R2R3R4

e

g

0 12090603015

Time (min)

50

150

300

200

100

250

Day 7

Day 1

Transplanted sample

of responder 1

P C 4 - 7 7

PC2 - 163

020

010

000

ndash010


Responders

Non-responders

Days 1ndash4

a

Non-NAS

consumers

High-NAS

consumers

H b A 1 C

60

65

55

45

50

f






h

0 12090603015

Time (min)

50

150

300

200

100

250 Day 7

Day 1

Transplanted sample

of non-responder 3

d

0 12090603015Time (min)

50

150

200

100



1

Day

2 3 4 5 6 7


( A U C

times 1 0 2 )

80

0

20

40

60





RESEARCH ARTICLE

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4

































Nature 505 559ndash563 (2014)

ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 5























RESEARCH ARTICLE

6 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4











0075 mg ml1 01 mg ml1




















ARTICLE RESEARCH





















RESEARCH ARTICLE








ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH

















a b

c

Saccharin

Water

Glucose


( w e e k 1 1 w e e k 0 )

100

101

102

103

10ndash3

10ndash2

10ndash1

Bacteria

Growth conditions

G r o w t h

c o n d i t i o n s

Water

Water

Glucose


ndash069



W a t e r

G l u c

o s e

S a c c h a

r i n

De1047297nition







β-glucuronidase


07 41

e

f


00

10

15

20

G l u c

o s e

C o m m e r c i a l

s a c c h a

r i n

05

g


0

20

30

40

10



0 11


Week 0 11 0 11


00

05

10

15

Other Unknown






G l u c

o s e

C o m m e r c i a l

s a c c h a

r i n

β-galactosidase

Sialidase-1

β-mannosidase

α-mannosidase

Glucosylceramidase

Endoglycosidase H

α-L-fucosidase

d




ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 3











a


b

0 12090603015

Time (min)

50

150

200

100

Saccharin-treated

Control


Anaeroplasma (G)

Bacteroides (G)

Bacteroides (G)

Bacteroides (G)




Bacteroides (G)

Clostridiales (O)

Bacteroides (G)

Bacteroides (G)

Clostridiales (O)

Clostridium (G)

4390755

1135084

260410

178064

311482

4308591

355305

276149

419601

176061

190473

184753

234912

3750380

log

(fold difference)

Lowest taxonomic



ndash2 21


r = 038

P lt 10ndash4

025 1 25


(week 11week 0)

12

06


( s a c c

h a r i n P B S )

1


Glycosaminoglycan

degradation



PTS



c


0 12090603015

Time (min)

50

150

200

100


bNR1 R1NR2NR3

R2R3R4

e

g

0 12090603015

Time (min)

50

150

300

200

100

250

Day 7

Day 1

Transplanted sample

of responder 1

P C 4 - 7 7

PC2 - 163

020

010

000

ndash010


Responders

Non-responders

Days 1ndash4

a

Non-NAS

consumers

High-NAS

consumers

H b A 1 C

60

65

55

45

50

f






h

0 12090603015

Time (min)

50

150

300

200

100

250 Day 7

Day 1

Transplanted sample

of non-responder 3

d

0 12090603015Time (min)

50

150

200

100



1

Day

2 3 4 5 6 7


( A U C

times 1 0 2 )

80

0

20

40

60





RESEARCH ARTICLE

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4

































Nature 505 559ndash563 (2014)

ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 5























RESEARCH ARTICLE

6 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4











0075 mg ml1 01 mg ml1




















ARTICLE RESEARCH





















RESEARCH ARTICLE








ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH










a


b

0 12090603015

Time (min)

50

150

200

100

Saccharin-treated

Control


Anaeroplasma (G)

Bacteroides (G)

Bacteroides (G)

Bacteroides (G)




Bacteroides (G)

Clostridiales (O)

Bacteroides (G)

Bacteroides (G)

Clostridiales (O)

Clostridium (G)

4390755

1135084

260410

178064

311482

4308591

355305

276149

419601

176061

190473

184753

234912

3750380

log

(fold difference)

Lowest taxonomic



ndash2 21


r = 038

P lt 10ndash4

025 1 25


(week 11week 0)

12

06


( s a c c

h a r i n P B S )

1


Glycosaminoglycan

degradation



PTS



c


0 12090603015

Time (min)

50

150

200

100


bNR1 R1NR2NR3

R2R3R4

e

g

0 12090603015

Time (min)

50

150

300

200

100

250

Day 7

Day 1

Transplanted sample

of responder 1

P C 4 - 7 7

PC2 - 163

020

010

000

ndash010


Responders

Non-responders

Days 1ndash4

a

Non-NAS

consumers

High-NAS

consumers

H b A 1 C

60

65

55

45

50

f






h

0 12090603015

Time (min)

50

150

300

200

100

250 Day 7

Day 1

Transplanted sample

of non-responder 3

d

0 12090603015Time (min)

50

150

200

100



1

Day

2 3 4 5 6 7


( A U C

times 1 0 2 )

80

0

20

40

60





RESEARCH ARTICLE

4 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4

































Nature 505 559ndash563 (2014)

ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 5























RESEARCH ARTICLE

6 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4











0075 mg ml1 01 mg ml1




















ARTICLE RESEARCH





















RESEARCH ARTICLE








ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH
































Nature 505 559ndash563 (2014)

ARTICLE RESEARCH

0 0 M O N T H 2 0 1 4 | V O L 0 0 0 | N A T U R E | 5























RESEARCH ARTICLE

6 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4











0075 mg ml1 01 mg ml1




















ARTICLE RESEARCH





















RESEARCH ARTICLE








ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH






















RESEARCH ARTICLE

6 | N A T U R E | V O L 0 0 0 | 0 0 M O N T H 2 0 1 4











0075 mg ml1 01 mg ml1




















ARTICLE RESEARCH





















RESEARCH ARTICLE








ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH










0075 mg ml1 01 mg ml1




















ARTICLE RESEARCH





















RESEARCH ARTICLE








ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH




















RESEARCH ARTICLE








ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH







ARTICLE RESEARCH






RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH





RESEARCH ARTICLE







repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH






repeated twice

ARTICLE RESEARCH






RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH





RESEARCH ARTICLE







ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH






ARTICLE RESEARCH





RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH




RESEARCH ARTICLE







(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH






(week 11week 0)

ARTICLE RESEARCH






RESEARCH ARTICLE






ARTICLE RESEARCH





RESEARCH ARTICLE






ARTICLE RESEARCH





ARTICLE RESEARCH

Paper 4 Artificial Sweeteners Induce Glucose Intolerance Nature 2014

Documents

Transcript of Paper 4 Artificial Sweeteners Induce Glucose Intolerance Nature 2014