Impact of oral vancomycin on gut microbiota, bile acid metabolism, and insulin sensitivity

Research Article

Impact of oral vancomycin on gut microbiota, bile acid metabolism, and insulin sensitivity Anne Vrieze1, Carolien Out2, Susana Fuentes3, Lisanne Jonker1, Isaie Reuling1, Ruud S. Kootte4, Els van Nood1, Frits Holleman1, Max Knaapen4, Johannes A. Romijn1, Maarten R. Soeters5, Ellen E. Blaak6, Geesje M. Dallinga-Thie4, Dorien Reijnders6, Mariëtte T. Ackermans7, Mireille J. Serlie5, Filip K. Knop8,9, Jenst J. Holst9, Claude van der Ley10, Ido P. Kema10, Erwin G. Zoetendal3, Willem M. de Vos3,11, Joost B.L. Hoekstra1, Erik S. Stroes4, Albert K. Groen2, Max Nieuwdorp1,4,⇑ 1

Department of Medicine, Academic Medical Center, Amsterdam, The Netherlands; 2Department of Pediatrics, University Medical Center Groningen, University of Groningen, Groningen, The Netherlands; 3Laboratory of Microbiology, Wageningen University, Wageningen, The Netherlands; 4Department of Vascular Medicine and Experimental Vascular Medicine, Academic Medical Center, Amsterdam, The Netherlands; 5 Department of Endocrinology and Metabolism, Academic Medical Center, Amsterdam, The Netherlands; 6Department of Human Metabolism, NUTRIM, School for Nutrition, Toxicology and Metabolism, Maastricht University, The Netherlands; 7Department of Clinical Chemistry, Laboratory of Endocrinology, Academic Medical Center, Amsterdam, The Netherlands; 8Department of Internal Medicine, Gentofte Hospital, Hellerup, Denmark; 9NNF Center for Basic Metabolic Research, Department of Biomedical Sciences, The Panum Institute, University of Copenhagen, Denmark; 10Department of Laboratory Medicine, University Medical Center Groningen, Groningen, The Netherlands; 11Department of Basic Veterinary Medicine, University of Helsinki, Helsinki, Finland

Background & Aims: Obesity has been associated with changes in the composition and function of the intestinal microbiota. Modulation of the microbiota by antibiotics also alters bile acid and glucose metabolism in mice. Hence, we hypothesized that short term administration of oral antibiotics in humans would affect fecal microbiota composition and subsequently bile acid and glucose metabolism. Methods: In this single blinded randomized controlled trial, 20 male obese subjects with metabolic syndrome were randomized to 7 days of amoxicillin 500 mg t.i.d. or 7 days of vancomycin 500 mg t.i.d. At baseline and after 1 week of therapy, fecal microbiota composition (Human Intestinal Tract Chip phylogenetic microarray), fecal and plasma bile acid concentrations as well as insulin sensitivity (hyperinsulinemic euglycemic clamp using [6,6-2H2]-glucose tracer) were measured. Results: Vancomycin reduced fecal microbial diversity with a decrease of gram-positive bacteria (mainly Firmicutes) and a compensatory increase in gram-negative bacteria (mainly Proteobacteria). Concomitantly, vancomycin decreased fecal secondary bile acids with a simultaneous postprandial increase in primary bile acids in plasma (p <0.05). Moreover, changes in fecal bile acid concentrations were predominantly associated with altered

Keywords: Antibiotics; Vancomycin; Amoxicillin; Intestinal microbiota; Bile acids; Insulin resistance; Metabolic syndrome. Received 29 May 2013; received in revised form 21 November 2013; accepted 25 November 2013; available online 6 December 2013 ⇑ Corresponding author. Address: Department of Internal Medicine, Academic Medical Center, Meibergdreef 9, Room F4-159.2, 1105 AZ Amsterdam, The Netherlands. E-mail address: [email protected] (M. Nieuwdorp).

Firmicutes. Finally, administration of vancomycin decreased peripheral insulin sensitivity (p <0.05). Amoxicillin did not affect any of these parameters. Conclusions: Oral administration of vancomycin significantly impacts host physiology by decreasing intestinal microbiota diversity, bile acid dehydroxylation and peripheral insulin sensitivity in subjects with metabolic syndrome. These data show that intestinal microbiota, particularly of the Firmicutes phylum contributes to bile acid and glucose metabolism in humans. This trial is registered at the Dutch Trial Register (NTR2566). Ó 2013 European Association for the Study of the Liver. Published by Elsevier B.V. All rights reserved.

Introduction The intestinal microbiota consists of several thousands of species, which collectively maintain gut physiology and homeostasis [1], including metabolic energy balance and immune responses [2,3]. Recent studies in both mice and humans revealed a novel concept in which, depending on genotype and lifestyle, the changes in intestinal microbiota are thought to actively contribute to the development of obesity and systemic insulin resistance [4–7]. In line, obesity has been found to be associated with significant alterations in the composition of the intestinal microbiota [7–10]. Landmark experiments by Gordon and colleagues elegantly provided evidence for a regulating function of the gut microbiota in energy homeostasis [7,9]. Germ-free mice were shown to be protected from obesity and insulin resistance when consuming a high fat diet [11,12], whereas colonization led to increased body fat content with concomitant insulin resistance [13]. The fact that colonization of germ-free mice

Journal of Hepatology 2014 vol. 60 j 824–831

JOURNAL OF HEPATOLOGY with microbiota obtained from obese mice induced an even greater increase in (visceral) adipose tissue lends further support to a causal relation between gut microbiota, systemic fat and insulin homeostasis [9,14]. Recently, we have extrapolated these experimental findings towards human pathophysiology as we demonstrated that transfer of lean donor fecal microbiota to obese subjects with the metabolic syndrome significantly altered the composition of the intestinal microbiota, which was associated with an improvement of peripheral insulin sensitivity [15]. Oral antibiotic treatment results in short- and long-term changes of the intestinal microbiota in both mice and humans [16–19]. Long-term intravenous vancomycin administration was also reported to correlate with the development of obesity in a retrospective study [20]. Recently, animal and human studies implicated that prolonged antibiotic treatment early after birth is associated with an increased risk of overweight [21,22]. In line, it has been long recognized that antibiotic treatment can induce profound changes in bile acid metabolism [23]. Primary bile acids (i.e., cholate and chenodeoxycholate) are produced in the liver from cholesterol by the enzyme cholesterol 7a-hydroxylase (CYP7A1). Prior to their secretion into the small intestine, bile acids are conjugated with either taurine or glycine. Subsequently, intestinal microbiota further modify the bile acids within the intestine [24]. Several gram-positive bacterial species, such as Lactobacilli deconjugate primary bile acids [24], which is in contrast to most gram-negative intestinal bacteria, with the exception of two strains of Bacteroides [25]. After deconjugation, additional microbial modifications occur, including oxidation and dehydroxylation, giving rise to the formation of secondary bile acids [26], which is only carried out by a minor population of gram-positive anaerobic Clostridium species [27–30]. Interestingly, bile acids have recently emerged as potential regulators of systemic energy homeostasis. For example, binding of particularly secondary bile acids (i.e., deoxycholate and lithocholate) to the G protein-coupled receptor TGR5 in the intestine strongly induces secretion of the incretin GLP-1, thereby affecting glucose homeostasis [31–33]. To date, it is poorly understood whether and to what extent intestinal bacteria are involved in the regulation of human bile acid and energy homeostasis. In view of the different role of gram-positive and -negative bacteria in intestinal bile acid metabolism, modification of either of these bacteria may have distinct effects on bile acid homeostasis. We thus hypothesize that the intestinal microbiota composition can affect bile acid composition with subsequently altered FGF-19 signaling thereby affecting glucose metabolism. In the present study, we thus set out to evaluate the impact of two different antibiotic regimens known to affect murine bile acid metabolism (amoxicillin and vancomycin orally administered) on intestinal microbiota composition, bile acid homeostasis and systemic insulin sensitivity in humans. We show that reduction of the gram-positive intestinal microbiota by vancomycin is associated with a decrease in peripheral insulin sensitivity in obese subjects and that modulation of the bile acid pool may be instrumental in mediating this effect.

Patients and methods Subjects Male Caucasian obese subjects were recruited via local advertisements and screened for characteristics of the metabolic syndrome, including waist circumference >102 cm and fasting plasma glucose >5.6 mmol/L [34]. Subjects with

a history of cholecystectomy, as well as subjects who used any medication (including probiotics and/or antibiotics in the past 3 months) were excluded. All subjects had a stable weight for 3 months prior to inclusion. Written informed consent was obtained from all subjects. The study was approved by the Institutional Review Board and conducted at the Academic Medical Center Amsterdam, The Netherlands, in accordance with the Declaration of Helsinki (updated version 2008). The study was registered at the Dutch Trial Register (NTR 2566). All authors had access to the study data, reviewed and approved the final manuscript. Study design A randomized, controlled single (physician) blinded intervention study was performed. Following assessment of eligibility and baseline measurements, participants were randomized to oral intake of 500 mg three times a day (t.i.d) of either amoxicillin or vancomycin, for a period of 7 days with daily dosages based on clinical used treatment regiments for systemic infection and enterocolitis, respectively. Two days before the antibiotic treatment and 2 days after cessation (to ensure proper wash-out of antibiotics), weight and height were measured, a standardized mixed meal test (MMT) was performed (day 2 and day 9) and glucose kinetics were measured in the basal state and during a two-step hyperinsulinemic euglycemic clamp using [6,6-2H2]-glucose as an isotopic tracer (day 1 and day 10). All measurements were conducted following a 12 h overnight fast (Supplementary Fig. 1). Participants were allowed to continue their usual diet, but were asked to complete an online nutritional diary (www.dieetinzicht.nl) to monitor caloric intake including the amount of dietary carbohydrates, fat, protein, and fibers during the study. Compliance was tested by counted amount of pills returned after one week treatment. The complete methods are included in the Supplementary data. Statistical analyses Statistical analyses were performed using SPSS 19.0 (SPSS Inc., Chicago, USA). Depending on the distribution of the data, data are expressed as mean ± SEM or median with range. The primary endpoint was change in intestinal microbiota composition, whereas secondary endpoints were changes in insulin sensitivity and bile acid metabolism after antibiotic treatment. For significant changes between and within treatment groups, clinical parameters were tested with (paired) Student’s t test or Mann Whitney test (differences between groups) and Wilcoxon Signed (differences within treatment group) depending on distribution of the data. Also, correlations were calculated using Pearson (parametric) or Spearman (nonparametric) coefficients. With respect to intestinal microbiota analyses, the microarray data were normalized and further analyzed using a set of R-based scripts (http:// www.r-project.org/) in combination with a custom designed relational database, which runs under the MySQL database management system (http://www.mysql.com). Hierarchical clustering of probe profiles was carried out using Pearson correlation-based distance and complete linkage method. False discovery rate q values were calculated to account for multiple testing. Significance level are expressed as p <0.05 or q value <0.05. Gut microbial diversity was determined by Simpson reciprocal index as previously described [35].

Results Baseline characteristics Twenty male subjects fulfilling the criteria of the metabolic syndrome were randomized to either amoxicillin (500 mg t.i.d.) or vancomycin (500 mg t.i.d.) (Supplementary Fig. 1). Table 1 shows the baseline characteristics for both groups. Almost half of the volunteers in each treatment group reported diarrhea during antibiotic treatment, and no difference was found in the reported stool frequency or compliance between the two treatment groups. No other side effects were noted and there was no change in body weight or daily caloric intake after either intervention. Moreover, there were no significant changes in both plasma lipopolysaccharide binding protein (LBP) levels and (postprandial) lipid profiles during the mixed meal test following antibiotic use in both groups (Table 1).

Journal of Hepatology 2014 vol. 60 j 824–831

825

Research Article Table 1. Characteristics of study subjects before and after antibiotic treatment.

Age (yr) Height (cm) Weight (kg) Body mass index (kg/m2) Fasting plasma glucose (mmol/L) Cholesterol (mmol/L) HDLc (mmol/L) LDLc (mmol/L) TG (mmol/L) Plasma FFA (mmol/L) CRP (mg/L) LBP (µg/ml) Fasting total GLP-1 (pmol/L) Fasting total GIP (pmol/L)

Amoxicillin group (n = 10) Before After

Vancomycin group (n = 10) Before After

61 (37-65) 180 (172-194) 113 (89-129) 110 (88-130) 33.0 (26.9-40.2) 32.6 (26.6-40.1)

51 (27-61) 183 (174-198) 118 (96-140) 34.0 (29.3-44.3)

5.6 (5.2-7.8) 4.9 (3.6-8.0) 1.0 (0.7-1.4) 3.8 (2.5-6.5) 2.1 (0.9-4.5) 0.5 (0.3-0.8) 2.5 (1.6-5.9) 24.6 (11.5-31.1) 12.0 (5-17)

5.6 (4.7-7.7) 5.0 (3.6-7.8) 1.0 (0.8-1.6) 3.0 (2.2-5.6) 1.8 (0.7-6.6) 0.6 (0.4-0.9) 2.4 (0.4-8.1) 21.7 (12.0-29.4) 9.5 (6-17)

10.0 (2-15)

10.0 (2-17)

p value

118 (97-141) 34.0 (29.9-44.9)

Baseline After amoxicillin 0.06 0.26 0.46 0.16 0.99 0.17

After vancomycin

0.08 0.09

5.3 (4.6-7.6) 4.2 (3.3-7.2) 0.9 (0.7-1.4) 3.2 (2.3-5.8) 1.5 (0.7-4.2) 0.5 (0.3-0.9) 2.7 (2.0-5.3) 18.1 (9.2-33.1) 12.0 (7-18)

5.5 (4.7-8.1) 4.2 (3.6-8.3) 1.0 (0.7-1.4) 2.4 (1.4-4.4) 1.8 (0.7-6.6) 0.4 (0.3-1.0) 1.8 (0.7-6.0) 17.2 (9.9-26.4) 11.5 (8-17)

0.28 0.43 0.89 0.70 0.12 0.42 0.98 0.13 0.88

0.77 0.61 0.09 0.41 0.07 0.53 0.16 0.98 0.18

0.07 0.80 0.13 0.07 0.17 0.73 0.18 0.68 1.00

9.5 (2-14)

9.5 (2-28)

0.62

0.77

0.57

Values are expressed as median (range). HDLc, High density lipoprotein cholesterol; LDLc, Low density lipoprotein cholesterol; TG, Triglycerides; FFA, Free fatty acids; CRP, C-reactive protein; LBP, Lipopolysaccharide binding protein; GLP-1, Glucagon-like peptide-1; GIP, Glucose-dependent insulinotropic polypeptide. p Value denotes differences between group at baseline and after treatment.

Effects of antibiotics on intestinal microbiota composition To determine the impact of amoxicillin and vancomycin, respectively, on fecal microbiota, the composition was compared before and after each intervention. Following vancomycin or amoxicillin, quantitative PCR analysis showed no significant changes in total numbers of fecal bacteria (vancomycin: from 10.4 ± 0.8 to 9.8 ± 0.9 log10 bacteria/gram feces, ns vs. amoxicillin: from 10.3 ± 0.7 to 10.4 ± 0.7 log10 bacteria/gram feces, ns). Moreover, vancomycin significantly decreased gut microbial diversity (Simpson’s inverse index from 138 ± 44 to 62 ± 33, p <0.01), whereas the diversity in the amoxicillin-treated group showed a trend toward increased diversity (from 161 ± 41 to 185 ± 37, p = 0.19). In line, vancomycin reduced the absolute number of gram-positive bacteria, with a compensatory increase in gramnegative bacteria. To specify, different groups of intestinal bacteria that were affected significantly by vancomycin but not by amoxicillin (Fig. 1 and Supplementary Table 1) belonged to the Firmicutes phylum (Clostridium cluster IV and XIVa, Lactobacillus plantarum and various butyrate-producing species including Faecalibacterium prausnitzii and Eubacterium hallii) as well as known pathogens from the Proteobacteria phylum (Escherichia coli, Haemophilus and Serratia). Effect of antibiotics on bile acid metabolism Following vancomycin treatment, fecal bile acid composition was markedly changed. Vancomycin considerably decreased the excretion of the secondary bile acids deoxycholic acid (DCA), lithocholic acid (LCA) and iso-lithocholic acid (iso-LCA), whereas the amount of the primary bile acids cholic acid (CA) and chenodeoxycholic acid (CDCA) was increased (Fig. 2A). In contrast, treatment with amoxicillin did not change the fecal 826

composition of primary or secondary bile acids (Fig. 2B). Moreover, vancomycin significantly decreased postprandial plasma concentrations of secondary bile salts (AUC from 5.4 to 1.5 lmol/L, p <0.05, Fig. 3B), whereas there was a trend towards postprandial increased plasma primary (predominantly CDCA) bile salts (Fig. 3A). Treatment with amoxicillin had no effect on plasma bile salt concentrations (Fig. 3C and D). It is generally assumed that intestinal bile acids regulate the expression of fibroblast growth factor 19 (FGF19), an important regulator of bile acid synthesis in the liver, whereas another established surrogate marker for bile acid synthesis is plasma 7-alphahydroxy-4-cholesten-3-one (C4) (36). Plasma levels of basal as well as postprandial FGF19 were significantly decreased following vancomycin treatment (AUC from 674 to 555 pg/ml, p <0.01 (Supplementary Fig. 2A) suggesting an increase in bile acid synthesis, whereas no effect was seen with amoxicillin (data not shown). In line, fasted C4 plasma levels increased more than 40% after vancomycin treatment with a significant decrease upon the MMT (AUC from 185 to 283 nmol/L, p <0.01, Supplementary Fig. 2B). In contrast, no changes were observed after amoxicillin treatment (AUC from 213 to 249 nmol/L, ns). Of note, both antibiotic regiments did not have an effect on either fasting or postprandial GLP1 and GIP plasma levels (Supplementary Fig. 2C-F). As previously published [24], L. plantarum content was associated with fecal primary bile acids (Pearson correlation coefficient r = 0.44, q <0.05). Interestingly, short chain fatty acid (butyrate) producing bacteria were positively correlated with fecal secondary bile acids (F. prausnitzii and E. hallii: Pearson correlation coefficients r = 0.69 and r = 0.63 respectively, q <0.05) and inversely correlated with fecal primary bile acids (F. prausnitzii and E. hallii: Pearson correlation coefficients r = 0.68 and r = 0.72 respectively, q <0.05, see also Supplementary Fig. 3). Of note, no significant correlations

Journal of Hepatology 2014 vol. 60 j 824–831

JOURNAL OF HEPATOLOGY Amoxicillin Before

Vancomycin After

Before

After Streptococcus mitis and rel. Streptococcus intermedius and rel. Phascolarctobacterium faecium and rel. Mitsuokella multiacida and rel. Megasphaera elsdenii and rel. Klebsiella pneumoniae and rel. Enterobacter aerogenes and rel. Leminorella Akkermansia Sutterella wadsworthia and rel. Bacteroides fragilis and rel. Eubacterium cylindroides and rel. Clostridium ramosum and rel. Veillonella Proteus and rel. Desulfovibrio and rel. Moraxellaceae Megamonas hypermegale and rel. Biophila and rel. Gemella Actinomycetaceae Fusobacteria Peptostreptococcus micros and rel. Helicobacter Brachyspira Methylobacterium Asteroleplasma and rel. Campylobacter Corynebacterium Uncultured Chroococcales Clostridium thermocellum and rel. Areomonas Novosphingobium Oceanospirillum Vibrio Pseudomonas Haemophilus Serratia Yersinia and rel. Lactococcus Escherichia coli and rel. Anaerobiospirillum Lactobacillus plantarum and rel. Weissella and rel Lactobacillus salivarius and rel. Eubacterium limosum and rel Eubacterium ventriosum and rel. Clostridium nexile and rel. Lachnospira pectioschiza and rel. Eubacterium hallii and rel. Lachnobacillus bovis and rel. Eubacterium rectale and rel. Faecalibacterium prausnitzii and rel. Butyrivibrio crossotus and rel. Ruminococcus bromii and rel. Clostridium stercorarium and rel. Eubacterium biforme and rel. Collinsella Oscillospira guillermondii and rel. Clostridium cellulosi and rel. Uncultured Clostridriales II Peptococcus niger and rel.

Color key

Proteobacteria, Bacilli, Clostridium cluster IX (+ others) -4

-2 0 2 Row Z-score

4

Clostridium clusters IV and XIVa

Fig. 1. Effect of antibiotics on intestinal microbiota composition. Heatmap regarding the effect of antibiotics on intestinal microbiota composition showing the bacterial groups (at genus-like level) that changed significantly (q <0.05) after vancomycin and amoxicillin administration (n = 10 per group), respectively. Relative abundances of the groups plotted are color coded from less (blue) to more abundant (red). The color value shows log10 fold changes.

between bile acids and bacterial species were found upon amoxicillin treatment. Effect of antibiotics on insulin sensitivity Details of the hyperinsulinemic euglycemic clamp are shown in Supplementary Table 2. Basal endogenous glucose production (EGP) was not different at baseline (vancomycin group: 9.9 ± 0.8 vs. amoxicillin group: 10.4 ± 1.4 lmol/kg/min, n.s.) and insulin concentrations during the clamp were similar between and within treatment groups (Supplementary Table 2). EGP was equally suppressed in both groups during the hyperinsulinemic euglycemic clamps with no significant impact of either intervention (vancomycin group: 67.5 vs. 65.7% suppression, p = 0.45 amoxicillin group: 62.8 vs. 58.9% suppression, p = 0.72) (Fig. 4B). In contrast, one week of vancomycin administration significantly decreased peripheral insulin sensitivity (rate of glucose disposal Rd, median Rd decreased from 31.9 to 30.7 lmol/kg/min, p <0.05) and a positive correlation was found between the change in fecal secondary bile acids (DCA, LCA and iso-LCA) and change

in peripheral insulin sensitivity in the vancomycin treatment group (Spearman rho = 0.74, p <0.05). No significant effects or correlations were observed in the amoxicillin-treated group (Median Rd from 26.7 to 25.8 lmol/kg/min, p = 0.84) (Fig. 4A).

Discussion We show that following 7 days of oral vancomycin administration in male subjects with metabolic syndrome, both peripheral insulin sensitivity and bile acid dehydroxylation were significantly impaired with a concomitantly marked change in intestinal microbiota composition. In contrast, treatment with amoxicillin did not have any effect on these parameters. Grampositive bacteria (mostly affected by vancomycin) are thought to be more important in maintaining intestinal microbiota stability than gram-negative/anaerobic bacteria (more affected by amoxicillin) [37]. It is tempting to hypothesize that amoxicillin resistant intestinal bacterial species can blossom and therefore compensate for initial loss of microbial diversity. These data also

Journal of Hepatology 2014 vol. 60 j 824–831

827

Research Article

0 0

1

2 3 4 Time (h)

5

15

AUC

p = 0.08

0

20

4 2

10 5

0 0

1

2 3 4 Time (h)

*

0

5

Be fo Af re te r

AUC

20

20 10

00

1

2 3 4 Time (h)

0

5

10 30

8 6

20 AUC

AUC

5

4

10

2 0 0

1

2 3 4 Time (h)

0

5

Fig. 3. Effect of antibiotics on plasma bile acid concentrations. After vancomycin treatment the postprandial rise in plasma concentrations of primary bile acids is increased (A), whereas the concentration of secondary bile acids is decreased (B). Amoxicillin treatment did not have an effect on either postprandial primary (C) or secondary (D) plasma bile acids. Data are expressed as medians (n = 10 per group), ⁄p <0.05.

B

Amoxicillin

w

0

1

T

T

T

=

=

T

D

Vancomycin

100

Hepatic insulin sensitivity (% suppression EGP)

Amoxicillin

100

= T

0

k w 1

k w 1

= T

20

=

0

40

T

25

60

0

50

80

=

75

T

Hepatic insulin sensitivity (% suppression EGP)

C

k

60 50 40 30 20 10 0

1

=

w

0

k

p <0.05

60 50 40 30 20 10 0

=

Vancomycin

Peripheral insulin sensitivity (Rd, µmol/kg/min)

A

Intestinal microbiota and bile acid metabolism It is well established that the intestinal microbiota has profound effects on bile acid metabolism by promoting deconjugation, dehydrogenation, and dehydroxylation of primary bile acids in the distal small intestine and colon. The recently published murine data by Sayin et al. corroborate with our human findings, underscoring the potential modulating role of intestinal microbiota in improved insulin sensitivity via FXR mediated upregulation of FGF19 [38]. In this respect, gram-positive bacteria belonging to the Clostridium clusters IV and XIVa of the Firmicutes phylum have been associated with bile acid homeostasis [24]. Many gram-positive bacteria are able to deconjugate, oxidize and dehydroxylate primary bile acids into secondary bile acids [24,26], whereas no deconjugating activity has been detected in gramnegative intestinal bacteria, with the exception of two strains of Bacteroides [25]. In contrast to the widely spread trait of deconjugation, dehydroxylation is only carried out by a minor population of gram-positive anaerobic Clostridium species [27–30]. As a consequence of this differential impact of bacterial species on bile acids, eradication of either gram-positive or gram-negative bacteria can be expected to have distinct effects on bile acid homeostasis, whereas bile acids themselves have endogenous bacteriostatic effects [39]. Indeed, in the present study vancomycin significantly reduced the number of gram-positive bacteria,

30

Before amoxicillin After amoxicillin

Be fo Af re te r

10

Secondary bile acids (µmol/L)

40

0

imply that vancomycin aggravates insulin resistance, which may be caused by altered bile acid metabolism due to specific changes in intestinal microbiota (see Fig. 5).

15

Be fo Af re te r

Primary bile acids (µmol/L)

C D C

Fig. 2. Effect of antibiotics on fecal bile acid excretion. (A) Fecal excretion of the secondary bile acids iso-LCA, LCA and DCA was decreased after vancomycin treatment (n = 10), whereas the excretion of the primary bile acids CA and CDCA was increased. (B) Amoxicillin treatment did not affect fecal bile acid excretion. Data are expressed as median with range (⁄p <0.05).

D

Before amoxicillin After amoxicillin

Peripheral insulin sensitivity (Rd, µmol/kg/min)

A

A C

A C D

A

is

LC

A LC

10

40

25

6

C

500

o-

20

Before vancomycin After vancomycin

8

D C

1000

0

828

60

A

A C

A C D

A is oLC Bile acids (µmol/24 h)

80

30

C

Before amoxicillin treatment After amoxicillin treatment

1500

40

Secondary bile acids (µmol/L)

* *

*

B Before vancomycin After vancomycin

Be fo Af re te r

*

500

Primary bile acids (µmol/L)

1000

A

Bile acids (µmol/24 h)

*

0

B

A

Before vancomycin treatment After vancomycin treatment

1500

LC

A

Fig. 4. Effect of antibiotics on insulin sensitivity. Peripheral insulin sensitivity was decreased after vancomycin treatment (A) (n = 10) (p <0.05), but was unchanged after amoxicillin treatment (B) (n = 10). Hepatic insulin sensitivity was unchanged after vancomycin (C) and amoxicillin (D) treatment.

especially bacteria belonging to the Firmicutes phylum (e.g., species belonging to the Clostridium clusters IV and XIVa), which

Journal of Hepatology 2014 vol. 60 j 824–831

JOURNAL OF HEPATOLOGY Vancomycin treatment Peripheral insulin resistance ↑ 1. Gut microbiota diversity ↓ - Proteobacteria ↑ - Firmicutes ↓

Plasma: C4 ↑ FGF19 ↓

2. SCFA (butyrate) producing bacteria ↓ - E. hallii ↓ - F. prausznitzii ↓ FXR ↑ TGR5 ↓

3. Bilacid deconjugation Lactobacillus plantarum ↑ 4. Fecal bile acid hydroxylation ↓ Primary bile acids ↑ Secondary bile acids ↓

Fig. 5. Effect of vancomycin treatment (eradication of intestinal gram positive bacteria) on human metabolism. Interaction between altered intestinal microbiota and impaired bile acid hydroxylation resulting in increased primary bile acid concentrations and subsequent increased peripheral insulin resistance in obese insulin resistant subjects. SCFA, short chain fatty acids. FXR, farnesoid X receptor; TGR5, G protein-coupled receptor 5; C4, 7a-hydroxy-4-cholesten-3-one; FGF19, fibroblast growth factor 19.

was associated with a significant reduction in fecal secondary bile acid levels (Supplementary Fig. 3). The decrease in secondary bile acid content probably reduced FGF19 secretion in plasma leading to an increase in bile acid synthesis as monitored via plasma C4 levels [36]. Bile acid metabolism and Insulin sensitivity Apart from their role in intestinal fat absorption, bile acids have also been shown to affect human metabolism [31,32]. By activating specific bile acid receptors and triggering downstream signaling pathways, bile acids influence energy expenditure and glucose homeostasis, e.g., via their effects on gluconeogenesis, insulin secretion and insulin sensitivity in both mouse and human studies [31–33]. Therefore, it has been suggested that manipulation of bile acid homeostasis might be an attractive treatment target for obesity induced insulin resistance [31,32]. Although various bile acid species differ in their in vitro capacity to activate specific receptors such as the farnesoid X receptor (FXR) or TGR5, it is not known which subclasses of bile acids are responsible for these effects in humans, because most studies on the relationship between bile acid and glucose metabolism have either been performed with only one subgroup of bile acids (mostly CDCA) or with bile sequestrants [31,32]. Nevertheless, secondary bile acids were found to have a much stronger activating capacity for TGR5 than primary bile acids [31,33,40]. Moreover, TGR5 activation by the secondary bile acids LCA and DCA is one of the proposed mechanisms, by which bile acids ameliorate insulin resistance [31,32,41,42]. Our finding of identical endogenous glucose production (EGP) in both treatment groups is most likely explained by the unaffected glucagon secretion (see Supplementary Table 2) [43]. However, our data on peripheral insulin sensitivity are in line with a recently published paper showing that specific FXR agonists can indeed further improve

(peripheral) hepatic insulin sensitivity as supported by increased plasma FGF 19 and decreased levels of 7a-hydroxy-4-cholesten3-one [44]. In line, we found a marked reduction in secondary bile acids after oral vancomycin treatment administration, which correlated with a significant decrease in peripheral insulin sensitivity. This might be due to different effects on butyrate producing intestinal bacteria (e.g., F. prausnitzii and E. hallii) associated with secondary bile acids [3]. It is therefore tempting to speculate about a potential bidirectional relationship between bile acids, short chain fatty acid producing bacteria and subsequent FXR mediated glucose homeostasis [15]. Study limitations Several limitations of the present study merit further consideration. Firstly, due to the potential development of bacterial resistance against these antibiotics, it was deemed ethically unacceptable to give prolonged treatment in these otherwise healthy subjects. Since we didn’t analyse intestinal microbiota composition after cessation of antibiotics, we are therefore unable to investigate and compare long-term impact of the 7 days antibiotic exposure on human metabolism [20]. Nevertheless a previous study has shown that after 3 months intestinal microbiota composition is almost back to pre-treatment composition [17], pointing out towards potential reversibility of this adverse effect on insulin resistance. Secondly, we used the HITChip, a phylogenetic microarray, for the high throughput profiling of the microbiota in feces as described previously [2]. Its advantage is that phylogenetic profiling of microbiota goes beyond the depth of canonical pyrosequencing in terms of intestinal microbiota composition [45]. However, it can only detect differences in microbiota composition and therefore assumptions regarding mechanistic effects of (i) specific species with their unique properties and (ii) their enzymatic enriched/depleted function

Journal of Hepatology 2014 vol. 60 j 824–831

829

Research Article are speculative yet informative [46]. Moreover, possible crosshybridization and detection limits for bacterial species that are present at low levels in the feces may possibly affect accuracy. Finally, the activity of the microbiota is not addressed. Hence, functional metagenomic analysis in conjunction with repetitive (daily) fecal sampling would be an appropriate avenue for future studies. Finally, bile acids can affect energy expenditure via increased mitochondrial activity in rodents [31,42,47–50]. Unfortunately, we were not able to perform resting energy expenditure and measurements of physical exercise (by accelerometers) or muscle biopsies before and after antibiotics exposure and therefore we cannot speculate on the relative contribution of bile acids on energy expenditure and subsequent glucose metabolism. Nevertheless, this study provides advances to our understanding towards human microbial-host metabolic interaction (partly) driven by bile acids in obese subjects with the metabolic syndrome. We furthermore show that vancomycin treatment reduces the amount of gram-positive bacteria, involved in human bile acid dehydroxylation. Even after short-term treatment, oral vancomycin significantly impaired peripheral insulin sensitivity and bile acid dehydroxylation, and gram-positive bacteria, especially those belonging to the Firmicutes phylum, may be responsible for these effects on glucose metabolism via their effects on intestinal bile acid metabolism. Our findings are in line with recent epidemiological data showing a relation between use of antibiotics and obesity [22] and warrant further research on this detrimental effect of (cumulative) antibiotics exposure on development of insulin resistance and subsequent type 2 diabetes mellitus.

Financial support M. Nieuwdorp: NWO-VENI Grant 2008 (016.096.044), E. van Nood: NWO-ZONMW VEMI Grant (170881001), M.R. Soeters: DFN Ruby Grant 2011, W.M. de Vos: NWO-Spinoza Grant 2008, R.S. Kootte and D. Reijnders (TIFN G003 Grant 2011).

Conflict of interest The authors who have taken part in this study declared that they do not have anything to disclose regarding funding or conflict of interest with respect to this manuscript. Authors’ contribution M.N., F.H., J.B.L.H., A.K.G., J.A.R., E.S, E.E.B., E.G.Z, E.vanN., and W.M.deV. designed the study. A.V., C.O, L.J., I.R, R.S.K, M.R.S., and E.R. performed the research. M.K., M.T.A., M.J.S., F.K.K., G.M.D.T., E.K.S. J.J.H, C.vd.L., and I.P.K. provided analytic tools. S.F. performed the statistical analysis. M.N., A.V., C.O, W.M.deV., and A.K.G. drafted the paper; all authors critically reviewed the manuscript. Acknowledgments We are grateful to Ineke Heikamp-de Jong, Philippe Puylaert and Wilma Akkermans-van Vliet (Wageningen University) for their excellent laboratory assistance.

830

Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jhep. 2013.11.034.

References [1] Tremaroli V, Backhed F. Functional interactions between the gut microbiota and host metabolism. Nature 2012;489:242–249. [2] Arumugam M, Raes J, Pelletier E, et al. Enterotypes of the human gut microbiome. Nature 2011;473:174–180. [3] Kootte RS, Vrieze A, Holleman F, et al. The therapeutic potential of manipulating gut microbiota in obesity and type 2 diabetes mellitus. Diabetes Obes Metab 2011;14:112–120. [4] Vrieze A, Holleman F, Zoetendal EG, et al. The environment within: how gut microbiota may influence metabolism and body composition. Diabetologia 2010;53:606–613. [5] Tilg H, Moschen AR, Kaser A. Obesity and the microbiota. Gastroenterology 2009;136:1476–1483. [6] DiBaise JK, Zhang H, Crowell MD, et al. Gut microbiota and its possible relationship with obesity. Mayo Clin Proc 2008;83:460–469. [7] Ley RE, Backhed F, Turnbaugh P, et al. Obesity alters gut microbial ecology. Proc Natl Acad Sci U S A 2005;102:11070–11075. [8] Murphy EF, Cotter PD, Healy S, et al. Composition and energy harvesting capacity of the gut microbiota: relationship to diet, obesity and time in mouse models. Gut 2010;59:1635–1642. [9] Turnbaugh PJ, Backhed F, Fulton L, et al. Diet-induced obesity is linked to marked but reversible alterations in the mouse distal gut microbiome. Cell Host Microbe 2008;3:213–223. [10] Turnbaugh PJ, Ridaura VK, Faith JJ, et al. The effect of diet on the human gut microbiome: a metagenomic analysis in humanized gnotobiotic mice. Sci Transl Med 2009;1:6ra14. [11] Backhed F, Manchester JK, Semenkovich CF, et al. Mechanisms underlying the resistance to diet-induced obesity in germ-free mice. Proc Natl Acad Sci U S A 2007;104:979–984. [12] Rabot S, Membrez M, Bruneau A, et al. Germ-free C57BL/6J mice are resistant to high-fat-diet-induced insulin resistance and have altered cholesterol metabolism. FASEB J 2010;24:4948–4959. [13] Backhed F, Ding H, Wang T, et al. The gut microbiota as an environmental factor that regulates fat storage. Proc Natl Acad Sci U S A 2004;101:15718–15723. [14] Turnbaugh PJ, Ley RE, Mahowald MA, et al. An obesity-associated gut microbiome with increased capacity for energy harvest. Nature 2006;444:1027–1031. [15] Vrieze A, van Nood E, Holleman F, et al. Transfer of intestinal microbiota from lean donors increases insulin sensitivity in individuals with metabolic syndrome. Gastroenterology 2012;143:913–916. [16] Dethlefsen L, Huse S, Sogin ML, et al. The pervasive effects of an antibiotic on the human gut microbiota, as revealed by deep 16S rRNA sequencing. PLoS Biol 2008;6:2383–2400. [17] Dethlefsen L, Relman DA. Incomplete recovery and individualized responses of the human distal gut microbiota to repeated antibiotic perturbation. Proc Natl Acad Sci U S A 2011;108:4554–4561. [18] Jernberg C, Lofmark S, Edlund C, et al. Long-term ecological impacts of antibiotic administration on the human intestinal microbiota. ISME J 2007;1:56–66. [19] Murphy EF, Cotter PD, Hogan A, et al. Divergent metabolic outcomes arising from targeted manipulation of the gut microbiota in diet-induced obesity. Gut 2012;62:220–226. [20] Thuny F, Richet H, Casalta JP, et al. Vancomycin treatment of infective endocarditis is linked with recently acquired obesity. PLoS One 2010;5:e9074. [21] Cho I, Yamanishi S, Cox L, et al. Antibiotics in early life alter the murine colonic microbiome and adiposity. Nature 2012;488:621–626. [22] Trasande L, Blustein J, Liu M, et al. Infant antibiotic exposures and early-life body mass. Int J Obes (Lond) 2012;37:16–23. [23] Canzi E. Effect of lincomycin treatment on intestinal microflora composition and its bile-acid-metabolizing activity. Curr Microbiol 1985;12:1–4. [24] Begley M, Hill C, Gahan CG. Bile salt hydrolase activity in probiotics. Appl Environ Microbiol 2006;72:1729–1738.

Journal of Hepatology 2014 vol. 60 j 824–831

JOURNAL OF HEPATOLOGY [25] Jones BV, Begley M, Hill C, et al. Functional and comparative metagenomic analysis of bile salt hydrolase activity in the human gut microbiome. Proc Natl Acad Sci U S A 2008;105:13580–13585. [26] Batta AK, Salen G, Arora R, et al. Side chain conjugation prevents bacterial 7dehydroxylation of bile acids. J Biol Chem 1990;265:10925–10928. [27] Kitahara M, Sakata S, Sakamoto M, et al. Comparison among fecal secondary bile acid levels, fecal microbiota and Clostridium scindens cell numbers in Japanese. Microbiol Immunol 2004;48:367–375. [28] Wells JE, Hylemon PB. Identification and characterization of a bile acid 7alpha-dehydroxylation operon in Clostridium sp. strain TO-931, a highly active 7alpha-dehydroxylating strain isolated from human feces. Appl Environ Microbiol 2000;66:1107–1113. [29] Aragozzini F, Canzi E, Ferrari A, et al. A study on the mechanism of the epimerization at C-3 of chenodeoxycholic acid by Clostridium perfringens. Biochem J 1985;230:451–455. [30] Hylemon PB, Harder J. Biotransformation of monoterpenes, bile acids, and other isoprenoids in anaerobic ecosystems. FEMS Microbiol Rev 1998;22:475–488. [31] Houten SM, Watanabe M, Auwerx J. Endocrine functions of bile acids. EMBO J 2006;25:1419–1425. [32] Prawitt J, Caron S, Staels B. Bile acid metabolism and the pathogenesis of type 2 diabetes. Curr Diab Rep 2011;11:160–166. [33] Pols TW, Noriega LG, Nomura M, et al. The bile acid membrane receptor TGR5: a valuable metabolic target. Dig Dis 2011;29:37–44. [34] Alberti KG, Eckel RH, Grundy SM, et al. Harmonizing the metabolic syndrome: a joint interim statement of the International Diabetes Federation Task Force on Epidemiology and Prevention; National Heart, Lung, and Blood Institute; American Heart Association; World Heart Federation; International Atherosclerosis Society; and International Association for the Study of Obesity. Circulation 2009;120:1640–1645. [35] Zilliox MJ, Irizarry RA. A gene expression bar code for microarray data. Nat Methods 2007;4:911–913. [36] Sauter G, Berr F, Beuers U, et al. Serum concentrations of 7alpha-hydroxy-4cholesten-3-one reflect bile acid synthesis in humans. Hepatology 1996;24:123–126. [37] Van den Abbeele P, Grootaert C, Marzorati M, et al. Microbial community development in a dynamic gut model is reproducible, colon region specific, and selective for Bacteroidetes and Clostridium cluster IX. Appl Environ Microbiol 2010;76:5237–5246.

[38] Sayin SI, Wahlstrom A, Felin J, et al. Gut microbiota regulates bile acid metabolism by reducing the levels of tauro-beta-muricholic acid, a naturally occurring FXR antagonist. Cell Metab 2013;17:225–235. [39] Sung JY, Shaffer EA, Costerton JW. Antibacterial activity of bile salts against common biliary pathogens. Effects of hydrophobicity of the molecule and in the presence of phospholipids. Dig Dis Sci 1993;38:2104–2112. [40] Wang H, Chen J, Hollister K, et al. Endogenous bile acids are ligands for the nuclear receptor FXR/BAR. Mol Cell 1999;3:543–553. [41] Katsuma S, Hirasawa A, Tsujimoto G. Bile acids promote glucagon-like peptide-1 secretion through TGR5 in a murine enteroendocrine cell line STC1. Biochem Biophys Res Commun 2005;329:386–390. [42] Thomas C, Gioiello A, Noriega L, et al. TGR5-mediated bile acid sensing controls glucose homeostasis. Cell Metab 2009;10:167–177. [43] Raju B, Cryer PE. Maintenance of the postabsorptive plasma glucose concentration: insulin or insulin plus glucagon? Am J Physiol Endocrinol Metab 2005;289:E181–E186. [44] Mudaliar S, Henry RR, Sanyal AJ, Morrow L, Marschall HU, Kipnes M, et al. Efficacy and safety of the farnesoid x receptor agonist obeticholic acid in patients with type 2 diabetes and nonalcoholic fatty liver disease. Gastroenterology 2013;145:574–582. [45] Claesson MJ, O’Toole PW. Evaluating the latest high-throughput molecular techniques for the exploration of microbial gut communities. Gut Microbes 2010;1:277–278. [46] Le Chatelier E, Nielsen T, Qin J, et al. Richness of human gut microbiome correlates with metabolic markers. Nature 2013;500:541–546. [47] Lowell BB, Shulman GI. Mitochondrial dysfunction and type 2 diabetes. Science 2005;307:384–387. [48] Ockenga J, Valentini L, Schuetz T, et al. Plasma bile acids are associated with energy expenditure and thyroid function in humans. J Clin Endocrinol Metab 2012;97:535–542. [49] Tomlinson E, Fu L, John L, et al. Transgenic mice expressing human fibroblast growth factor-19 display increased metabolic rate and decreased adiposity. Endocrinology 2002;143:1741–1747. [50] Watanabe M, Houten SM, Mataki C, et al. Bile acids induce energy expenditure by promoting intracellular thyroid hormone activation. Nature 2006;439:484–489.

Journal of Hepatology 2014 vol. 60 j 824–831

831