SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

The aim of this study was to investigate the gut microbiota in preschool children with and without overweight and obesity. Twenty overweight or obese children and twenty children with BMI within the normal range (age: 4–5 years) were recruited from the south of Sweden. The gut microbiota was accessed by quantitative PCR (qPCR) and terminal restriction fragment length polymorphism and calprotectin was measured in feces. Liver enzymes were quantified in obese/overweight children. The concentration of the gram-negative family Enterobacteriaceae was significantly higher in the obese/overweight children (P = 0.036), whereas levels of Desulfovibrio and Akkermansia muciniphila-like bacteria were significantly lower in the obese/overweight children (P = 0.027 and P = 0.030, respectively). No significant differences were found in content of Lactobacillus, Bifidobacterium or the Bacteroides fragilis group. The diversity of the dominating bacterial community tended to be less diverse in the obese/overweight group, but the difference was not statistically significant. Concentration of Bifidobacterium was inversely correlated to alanine aminotransferase (ALT) in obese/overweight children. The fecal levels of calprotectin did not differ between the study groups. These findings indicate that the gut microbiota differed among preschool children with obesity/overweight compared with children with BMI within the normal range.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

Obesity is one major public health problem, and the increasing incidence of childhood obesity is a source of concern because it is a risk factor for overweight and obesity later in life (1). Childhood obesity has repercussions in adulthood by increasing, for example, the risk of type 2 diabetes (2).

The intestinal microbiota in relation to pediatric metabolic disorders have been poorly studied, although in recent years the gut microbial ecosystem has emerged as a significant factor involved in obesity, even if no causal relationship has been established (3,4,5). Obesity is associated with a chronic subclinical inflammation reflected in changes in biochemical markers (6). Lipopolysaccharides (LPS) in the cell wall of gram-negative bacteria have inflammatory capacity, highlighting the possible role of gram-negative bacteria in obesity (7,8). Furthermore, obese people have been reported to have lower bacterial diversity in the gastrointestinal tract compared to lean subjects (5).

The objective of the present study was to evaluate the gut microbiota of preschool children with excessive body weight in comparison to counterparts with normal body weight. Quantitative PCR (qPCR) was used to quantify bacterial groups and terminal restriction fragment length polymorphism (T-RFLP) was used to study the diversity of the dominating intestinal microbiota. Fecal calprotectin was measured as an indicator of intestinal inflammation. In obese and overweight subjects the concentrations of liver enzymes were determined.

Methods and Procedures

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

Subjects and sample collection

Twenty overweight or obese children with BMI ranging 17.6–25.8 kg/m2 (overweight/obese (OO) group) were randomly selected before start of the intervention programme “Overweight and Obesity in Preschool Children, Prevalence and Prevention—Family-Based Health Interventions for Child Health” in the south of Sweden. The 1-year intervention aims at encouraging parents of overweight and obese preschool children to adopt healthy lifestyles that can be continued throughout childhood and that have a positive long-term effect on weight development. Twenty children with BMI within the normal range (13.6–17.2 kg/m2) served as control group (C group). Overweight, obesity, and normal weight were defined according to International Task Force of Obesity (9), taking child gender and age into consideration. To obtain a standardization of the BMI-measurements, BMI-standard deviation score was calculated using age- and gender-specific reference values (10). The ages of children in this study ranged between 4 years and 5 years 3 months. Characteristics of the study population are shown in Table 1. Children in the control group were recruited from two municipal nursery schools in the south of Sweden. The study was conducted according to the Helsinki Declaration, participation was voluntary and all parents gave written informed consent. The study was approved by the Regional Ethical Review Board in the south of Sweden.

Table 1.  Characteristics of the cohort
inline image

Fecal samples from both OO and C groups were collected by the parents at home, stored at −20 °C, transported to the hospital or nursery school, respectively the next day where it was stored at −20 °C for up to 4 days before transfer to the laboratory, where samples were stored at −80 °C until processing. Blood samples from the OO group were collected on the day of inclusion to the intervention program and fecal samples from the OO group were collected within the first weeks of inclusion.

DNA extraction

DNA from fecal samples was isolated and purified in BioRobot EZ1 (EZ1 DNA Tissue Kit and tissue card; Qiagen, Hilden, Germany). In brief, stool samples were thawed on ice and 50 mg was diluted and homogenized in 500 µl sterile phosphate-buffered saline. After 10 min at room temperature, samples were vortexed and then centrifuged at 3,000g for 30 s; 200 µl of the supernatant was used for extraction in the BioRobot EZ1.

Microbial analyses

Quantitative real-time PCR. The amount of 16S ribosomal RNA (rRNA) genes of bacteria belonging to Lactobacillus, Bifidobacterium, Enterobacteriaceae, Akkermansia muciniphila-like bacteria, Desulfovibrio, and the Bacteroides fragilis group were estimated using separate qPCR assays. For all assays, each reaction contained 10 µl 2× Rotor-Gene SYBR Green PCR Master Mix (Qiagen), 0.5 µmol/l of each primer (Table 2 (11,12,13,14,15,16,17)), 2 µl of template DNA and RNAse-free water to the final volume of 20 µl. Samples, standards, and non-template controls were run in triplicate. The thermal cycling was carried out in Rotor-Gene Q (Qiagen) with a programme of 95 °C for 5 min, followed by 40 cycles with denaturation at 95 °C for 5 s, annealing and elongation at 60 °C for 10–30 s (Table 2). The fluorescent products were detected at the last step of each cycle. Melting curve analysis was carried out to ensure specific amplification. Absolute abundance of 16S rRNA genes was calculated based on standard curves using Rotor-Gene Q Series Software 1.7 (Qiagen), R2 > 0.998. Detection limit was 101 genes/reaction for the A. muciniphila-like bacteria, 102 genes/reaction for Lactobacillus, Bifidobacterium, Enterobacteriaceae, and Desulfovibrio while the B. fragilis group assay detected 103 genes/reaction. As standard curves, cloned PCR products from Lactobacillus plantarum DSM9843, Bifidobacterium infantis DSM15159, Escherichia coli CCUG29300T, B. fragilis CCUG4856T, Desulfovibrio desulfuricans subsp. desulfuricans CCUG34226T, and A. muciniphila (a clone confirmed by sequencing) were used. Tenfold dilution series of the target DNA were made in TE buffer (10 mmol/l Tris, 1 mmol/l EDTA, pH 8.0) supplemented with 0.1 µg/µl herring sperm DNA (VWR International, West Chester, PA). Number of bacteria was expressed as numbers of 16S rRNA genes/g wet weight of feces.

Table 2.  Oligonucleotide primers used in this study
inline image

T-RFLP analysis. The 16S rRNA genes were amplified with universal primers as described previously (8); 2 µl of template was used in the reaction with a final volume of 25 µl. Amplification was carried out for 30 cycles and PCR products were verified and purified as described by Karlsson et al. (18)., with the modification that 2.5 U/µl FastStart Taq polymerase (Roche Diagnostics, Mannheim, Germany) and initial enzyme activation for 3 min at 95 °C was used. DNA concentration was measured by NanoDrop ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE) using 1 µl purified PCR product. T-RFLP was performed and analyzed as previously described (8). Thresholds for internal standard and T-RFs were set to 10 and 20 fluorescence units, respectively.

Biochemical analyses

Blood samples were taken from children in the OO group, and serum alanine aminotransferase (ALT) and aspartate aminotransferase (AST) were analyzed at the Department of Clinical Chemistry and Pharmacology at Skåne University Hospital (Lund, Sweden) using standard enzyme activity methods. To ensure no ongoing intestinal inflammation, fecal calprotectin was analyzed with the Quantum Blue system according to the manufacturer's instruction (Bühlman, Basel, Switzerland). Detection limit was 30 µg/g wet weight feces.

Calculations

The relative abundance of each T-RF within a given T-RFLP pattern was calculated as the peak area of the respective T-RF divided by the total peak area of all T-RFs, in the given T-RFLP pattern, detected within a fragment length of 20–600 base pairs (bp). Relative abundance was used for calculations of diversity index. Shannon (Há) and Simpson (D) indexes were calculated for all individuals by using the equations: Há = −σ pi ln pi and 1 − D where D = σ pi2, where pi is the relative abundance of ith peak in the community (19). Results are given as medians (interquartile range).

Statistical evaluation was performed in SigmaStat 3.1 (Systat Software, Point Richmond, CA) using Student's t-test for parametric data while Mann—Whitney rank sum was used for nonparametric data. Pearson product moment correlation was used for calculations of bacterial concentrations vs. ALT and AST serum concentrations for subjects in the OO group. P < 0.05 was considered statistically significant. Multivariate data analysis with principal component analysis (PCA) (SIMCA-P+12.0.1; Umetrics, Umeå, Sweden) on the diversity indexes and qPCR data was performed to reveal differences in the microbial composition between the two groups.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

Quantification of bacterial groups in the intestinal microbiota

The OO group was found to have significantly more Enterobacteriaceae compared to the C group (P = 0.036) while Desulfovibrio and A. muciphilia-like bacteria were significantly lower in the OO group compared to the C group, (P = 0.027 and P = 0.030, respectively). No significant differences were found in number of 16S rRNA gene copies of Lactobacillus, Bifidobacterium or the B. fragilis group (Table 3).

Table 3.  Concentrations of specific bacterial groups in feces of overweight/obese and normal weight children, detected by qPCR
inline image

Diversity of the intestinal microbiota

The diversity of the dominating intestinal microbiota, accessed by calculation of diversity indexes from T-RFLP profiles, showed a tendency to be lower for the OO group compared to the C group. Median Shannon index was 2.93 (2.62–3.09) and 3.03 (2.88–3.17) for the OO and C group respectively (P = 0.091) when digested with MspI. Similar trends were obtained when Simpson index was calculated and when the restriction endonuclease AluI was used for digestion. Multivariate analysis of all diversity indexes together with the qPCR data indicated different overall composition of the intestinal microbiota in children with excessive body weight compared to the children with BMI in the normal range (Figure 1).

image

Figure 1. Principal component analysis (PCA) of the intestinal microbiota in children with normal and excessive body weight. The microbial composition is clearly different between the two groups, shown here when all diversity indexes and qPCR data were included, PC1; R2X = 0.395. C, control; OO, obese/overweight; qPCR, quantitative PCR; s.d., standard deviation.

Download figure to PowerPoint

Biochemical markers

The liver enzyme ALT was negatively correlated to fecal Bifidobacterium concentration in the OO group (P = 0.03, r = −0.479). No correlations between concentrations of other bacterial groups and ALT or AST were found. The mean ALT concentration was 0.348 (s.d 0.13) µkat/l and the mean AST concentration was 0.580 (s.d. 0.12) µkat/l. Fecal calprotectin was below detection limit in about 50% of the subjects and there was no significant difference between the groups.

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

In recent years, perturbations in the gut microbial ecosystem in relation to obesity prevalence and development has been acknowledged, although no consensus has been reached about the causative components of the microbiota. However, Cani and co-workers (7,20) suggested that increased gut permeability and LPS were triggering components for obesity development, which is mediated by an inflammation-driven process. The present study demonstrates differences in specific gut bacterial groups among preschool children with normal and excessive body weight. To the best of our knowledge, this is the first time Enterobacteriaceae has been found at significantly higher concentrations in obese/overweight humans compared to individuals with body weight within the normal range. Enterobacteriaceae is a family of gram-negative commensals that is found at various levels in the gut, and has been shown to dominate in a 1-week-old baby, represented by the species E. coli (21). E. coli has also been found at higher numbers in children with Crohn's disease and proceeding atopic eczema, compared to the pediatric control groups (22,23). In addition, it was recently observed that neonates born large for gestational age had significantly more Proteobacteria, preferentially E. coli, than those born appropriate for gestational age (24). Recently, administration of E. coli to rats from fetal life until adulthood was shown to significantly impact the physiology as well as the intestinal microbiota (8).

LPS in the outer membrane of gram-negatives like Enterobacteriaceae are potent stimulators of inflammation. Similarly, B. fragilis group, Desulfovibrio, and Akkermansia are gram-negatives. However, structural differences among LPS make them induce various stimuli (25). Akkermansia has not been correlated to any disease or sign of pathogenicity so far (26). Instead, the A. muciniphila is known to degrade mucin and is commonly found in the human gastrointestinal tract (27,28). In the present study, A. muciniphila-like bacteria was significantly higher in group C than in group OO. Recently, Akkermansia has been found to be inversely related to the severity of appendicitis (29) and also present at lower numbers in patients with inflammatory bowel disease (30). Studies in pregnant women have shown A. muciniphila-like bacteria to be positively correlated with normal weight gain over pregnancy (31), whereas no significant difference was found between normal weight and overweight pregnant women (32). The genome of A. muciniphila was recently sequenced, and it suggests high ability to metabolize a variety of complex carbohydrates as well as synthesis of all canonical amino acids and important cofactors and vitamins (28). Furthermore, bacterial fermentation of mucin is an important feature for selectively favoring growth of other microbes (26). For instance, sulphate is released upon mucin fermentation by A. muciniphila-like bacteria, thereby generating a favorable environment for growth of Desulfovibrio.

Desulfovibrio reduce sulphate to sulphide, which is toxic to epithelial cells (33) and Desulfovibrio has been suggested to be part of the aetiology of ulcerative colitis (34). However, the relationship is controversial since not all studies found difference between diseased and healthy subjects (17), and in the present study the C group had higher abundance than the OO group.

The number of 16S rRNA genes of Lactobacillus, Bifidobacterium, and B. fragilis group did not differ in relation to pediatric body weight, which was in accordance with other studies (24). Nevertheless, an inverse correlation between Bifidobacterium and ALT was observed, suggesting the Bifidobacterium to be an indication of improved liver health. The gut and the liver are closely connected, and the intestinal mucosa functions as the barrier that helps to prevent the systemic spread of bacteria and endotoxins, i.e., mostly LPS from the cell walls of gram-negative bacteria. There is evidence that portal vein endotoxemia of gut origin in minute amounts is a normal physiological phenomenon (35), and that the leakage is rapidly cleared by the liver. However, excessive LPS stimuli impact the hepatic function (7). It can be speculated that certain bacterial groups of the gut microbiota (e.g., Enterobacteriaceae) can enhance the stress on the mucosa and the liver and other bacteria such as Bifidobacterium can suppress such a stress. The liver enzymes may function as markers for the gut-derived stress on the liver. However, further investigations are needed to understand the relationship between liver status and gut microbiota. For ethical reasons blood samples were not taken from children in the control group but only from those included in the ongoing intervention study, so determination of liver enzymes was only performed among obese/overweight children. Therefore, the relationship between Bifidobacterium and ALT in nonobese children remains unknown.

Although obesity is characterized by inflammation, and calprotectin is a measure of intestinal inflammation often used as a diagnostic tool for inflammatory bowel diseases (36), no significant difference was found between the OO group and the C group. Fecal calprotectin was analyzed in all children and notably, about half of the subjects, irrespective of group, had concentrations below detection limit, suggesting no intestinal inflammation among the participants.

Although a battery of different bacterial groups was quantified, the gut microbiota is highly complex (37). Evaluation of the diversity of the dominating bacterial flora is a way to get an overview of the ecosystem. In the present study, T-RFLP was used for diversity assessment and a trend was found for higher diversity in group C compared to group OO. The same situation has been described by others (5). Furthermore, as indicated in Figure 1, multivariate analysis confirmed a difference in the intestinal microbiota between group C and group OO.

The exact dietary intake could not be monitored in the present study. However, all children were raised in the same region of Sweden and their days were spent on nursery schools where the children were served meals according to the Swedish Nutrition Recommendations.

In conclusion, the bacterial flora of the gut was shown to differ between preschool children with excessive body weight and normal weight children. The amount of Enterobacteriaceae, for example, was significantly higher in those with excessive body weight. In contrast, A. muciniphila-like bacteria and Desulfovibrio were more abundant in children with normal body weight. There was also a trend for decreased bacterial diversity in children with excessive body weight. Future studies should further investigate the microbial ecosystem and its impact on health during childhood and later in life.

ACKNOWLEDGMENTS

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

The participating children and their parents are greatly acknowledged for their contribution to the study. The research nurses Elna Klemedtson and Ulrika Arvidsson are thanked for organizing sample collection of the OO group and Erik Johansson at Umetrics (Umeå, Sweden) is acknowledged for a discussion about multivariate statistical analysis. The study was financed by grants from the Swedish Research Council Formas (grant no. 222-2007-501), The Crafoord Foundation, The Swedish Institute for Health Sciences (Vårdalinstitutet), and Länsförsäkringar Insurance Alliance Research Foundation.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods and Procedures
  5. Results
  6. Discussion
  7. ACKNOWLEDGMENTS
  8. DISCLOSURE
  9. References
  10. Supporting Information

Supporting Information

FilenameFormatSizeDescription
oby_2760_sm_oby2012110_coi.pdf3000KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.