Obesity and its attendant consequences are a major cause of ill health in developed countries, and a growing problem in the developing world(Reference Friedman1). Obesity can physiologically be attributed to any or all of a combination of increased intake of energy, more efficient absorption of ingested energy, or reduced energy expenditure. The intestine and colon are host to trillions of bacteria, which are largely anaerobic and survive by metabolising unabsorbed dietary constituents(Reference Ramakrishna2). It is estimated that 10–15 % of dietary sugar and starch is not absorbed in the small bowel; fermentation by colonic luminal bacteria to SCFA which are absorbed and metabolised serves to salvage energy(Reference McNeil3). Individuals with a colectomy weigh on average 4 kg less than age- and height-matched healthy individuals with similar energy intake(Reference Behall and Howe4). It is estimated that the colon contributes to 5–10 % of energy requirements in residents of Western countries(Reference Bingham, Cummings and McNeil5, Reference McNeil, Bingham and Cole6). There are theoretical reasons to believe that the colon may contribute significantly more to energy conservation in countries in Asia and Africa where there is a high intake of starchy foods(Reference Cummings and Englyst7).
The intestinal microbiota of obese individuals may be more efficient at extracting energy from a given diet than the flora of lean individuals(Reference Turnbaugh, Ley and Mahowald8, Reference Backhed, Ding and Wang9). Compared with lean mice, obese mice had fewer bacteria belonging to the division Bacteroidetes, and more bacteria belonging to the division Firmicutes(Reference Ley, Backhed and Turnbaugh10). Human studies have consistently reported increases in Firmicutes in obese adults compared with normal individuals(Reference Turnbaugh, Ley and Mahowald8, Reference Zhang, DiBaise and Zuccolo11, Reference Duncan, Lobley and Holtrop12), whereas alterations in Bacteroides–Prevotella were variable. The present study, undertaken in a developing world rural setting, set out to examine the nature of the dominant faecal microbiota in obese children compared with their normal peers.
Methods
The participants were recruited from three private schools. After focus group discussions, children were invited to participate in the study. Obesity was defined as BMI exceeding the 97th percentile for that age using WHO reference growth charts(13). For each obese participant, a non-obese counterpart in the same class was invited to serve as a control. Anyone who had taken antibiotics within the previous month was excluded.
Socio-economic status was graded according to the modified Kuppuswami scale(Reference Kumar, Shekhar and Kumar14). A 24 h diet recall and a FFQ of commonly used foods taken over the previous 3 months were used to calculate macronutrient intake from food composition tables for Indian diets(Reference Gopalan, Rama Sastri and Balasubramanian15). Standard cups and spoons were used to assess meal sizes. Freshly passed specimens of faeces were collected in plastic containers, transported to the laboratory on ice and stored at − 70°C to be processed in batches. Faecal DNA was extracted using the QIAamp DNA stool mini kit (QIAgen GmbH, Hilden, Germany) and quantitative PCR was carried out using genus- and group-specific primers targeting 16S rRNA genes (rDNA) as described in our earlier publications(Reference Balamurugan, Janardhan and George16, Reference Balamurugan, Janardhan and George17).
The present study was conducted according to the guidelines laid down in the Declaration of Helsinki and all procedures involving human participants were approved by the Ethics and Research Committees of the Christian Medical College, Vellore, which reviewed the protocol and the consent forms. Written informed consent was obtained from all participants and their parents.
Statistics
All values are shown as median (interquartile range). Comparisons between groups were done using Mann–Whitney tests and two-tailed P values less than 0·05 were taken as statistically significant.
Results
There were twenty-eight participants included in the study, of whom fifteen (seven female) were obese and thirteen (seven female) non-obese. Their demographic characteristics are shown in Table 1. Dietary intakes of energy and of macronutrients did not differ significantly between the two groups of participants (Table 1).
(Medians and ranges or interquartile ranges (IQR))
* Other than weight and BMI percentile, none of the differences between the two groups was statistically significant.
† BMI-for-age percentile value derived from WHO reference growth charts(13).
‡ Socio-economic scores were used to derive a socio-economic class(Reference Kumar, Shekhar and Kumar14) and was calculated based on the education and occupation of the head of the household and the monthly family income adjusted for 2007. Socio-economic classes I, II and III represent upper, upper-middle and middle classes, respectively, on a scale of I–V.
The quantitative bacterial studies showed no significant differences in the levels of the Bacteroides–Prevotella group, Eubacterium rectale, Bifidobacterium group (Fig. 1) or Lactobacillus acidophilus group (data not shown) between the study groups. However, levels of Faecalibacterium prausnitzii were significantly higher in the obese compared with the non-obese participants (P = 0·0253) (Fig. 1).
Quantitative PCR of different bacterial groups from the faeces of obese (OB) and non-obese (NOB) participants: (a) Bacteroides–Prevotella–Porphyromonas; (b) Bidifobacterium; (c) Eubacterium rectale; (d) Faecalibacterium prausnitzii. Values are shown relative to amplification of a conserved segment of 16S rDNA, and the bars represent medians and interquartile ranges. The only statistically significant difference between the two groups was with respect to F. prausnitzii, which were significantly higher in the faeces of the obese children (P = 0·0253; Mann–Whitney test).
Discussion
Following the demonstration in mice of characteristic alterations in the faecal flora of obese compared with lean animals and the transferable nature of the obese phenotype by transplanting the flora to germ-free mice(Reference Turnbaugh, Ley and Mahowald8), several human studies have been undertaken to confirm the presence and nature of alterations in the faecal microbiota in obese individuals. Studies in adults have consistently reported increases in Firmicutes in obese adults compared with normal or lean individuals(Reference Turnbaugh, Ley and Mahowald8, Reference Zhang, DiBaise and Zuccolo11, Reference Duncan, Lobley and Holtrop12). In a very recent study, Turnbaugh et al. (Reference Turnbaugh, Hamady and Yatsunenko18) examined the faecal microbiota of mono- and dizygotic twin pairs concordant for obesity or leanness, and found that obesity was associated with significantly fewer Bacteroidetes and significantly more Actinobacteria but no significant difference in Firmicutes(Reference Turnbaugh, Hamady and Yatsunenko18). The present study reconfirmed, in as different a setting as possible compared with earlier studies, that there were increases in the population of F. prausnitzii (prominent carbohydrate-fermenting bacteria) in the gut of obese children. At the time of study, energy and macronutrient intake was similar in both groups.
E. rectale–Clostridium coccoides, Bacteroides–Prevotella and F. prausnitzii are the dominant phylogenetic groups in the faecal microbiota(Reference Mueller, Saunier and Hanisch19). Members of the Bacteroides–Prevotella group play important roles in the hydrolysis and fermentation of dietary fibre, producing acetate and propionate(Reference Salyers20). Butyrate, a physiologically important SCFA, is produced by several bacteria, important among which are the E. rectale–C. coccoides group and F. prausnitzii (Reference Suau, Rochet and Sghir21, Reference Louis and Flint22). F. prausnitzii, belonging to the C. leptum group of bacteria, is a Firmicute and a highly functionally active member of the intestinal microbial flora(Reference Lay, Sutren and Rochet23). It has been identified as one of the key functional members of the microbiome that most influence host metabolism(Reference Li, Wang and Zhang24), and is responsible for a significant proportion of fermentation of unabsorbed carbohydrate. In many tropical populations, up to 20 % of dietary carbohydrate, including that in such foods as rice, maize, banana and potatoes, is unabsorbed because of the presence of NSP and amylase-resistant starch(Reference Ramakrishna and Roediger25). It is conceivable that the presence of F. prausnitzii in greater numbers in obese children leads to increased energy salvage from unabsorbed carbohydrate that would not otherwise contribute to dietary energy intake. Interestingly, it has been demonstrated that F. prausnitzii were significantly reduced in frail elderly individuals as well as in patients with chronic idiopathic diarrhoea and malnutrition(Reference van Tongeren, Slaets and Harmsen26, Reference Swidsinski, Loening-Baucke and Verstraelen27). It is of interest that dietary carbohydrate restriction results in reductions of E. rectale and other butyrate-producing Firmicutes in the faeces of obese individuals(Reference Duncan, Lobley and Holtrop12, Reference Duncan, Belenguer and Holtrop28). SCFA such as acetate, propionate and butyrate are ligands for the G-protein-coupled receptors Gpr41 and Gpr43 on colonic epithelial cells, the activation of which result in the release of gut-derived hormones such as peptide YY which affects energy harvest from the diet(Reference Tilg, Moschen and Kaser29). Although the specific microbial classes or genera involved in the effect are not yet known, the establishment of an intestinal microbiota in germ-free mice has been shown to suppress epithelial cell production of fasting-induced adipocyte factor resulting in increased lipoprotein lipase activity in adipocytes and promoting storage of energy as fat(Reference Backhed, Ding and Wang9). Yet another mechanism, involving regulation of systemic inflammation, has been proposed for the connection between intestinal microbiota and obesity. Lipopolysaccharide (LPS) is a key constituent of gut bacteria and plays a central role in innate immune responses in the gut. It is also absorbed systemically presumably through intercellular junctions and this process can be regulated by factors that control intestinal permeability. Obesity has been shown to be associated with increases in intestinal permeability and in plasma LPS in mice fed a high-fat diet; similar increases in plasma LPS were noted in ob/ob mice ingesting normal chow(Reference Cani, Bibiloni and Knauf30). On the other hand, the plasma LPS rise and the metabolic changes associated with obesity were lacking in ob/ob CD14− / − mice that were non-responsive to LPS. Furthermore, antibiotic treatment of the ob/ob mice led to decreases in adipose tissue inflammatory markers and metabolic markers of obesity(Reference Cani, Bibiloni and Knauf30). These microbiota-dependent changes may be mediated via release of glucagon-like peptide(Reference Cani, Possemiers and Van de Wiele31). F. prausnitzii is known to be reduced in the faecal microbiota of patients with Crohn's disease(Reference Sokol, Seksik and Furet32, Reference Willing, Halfvarson and Dicksved33), and is thought to be protective against intestinal inflammation by secreted metabolites able to block NF-κB activation and IL-8 secretion. The intestine normally exists in a state of controlled inflammation, but a relationship between the degree of background mucosal inflammation and obesity has not been explored.
Assigning causality to the association between changes in faecal microbiota and obesity in humans is difficult. One approach is to examine the faecal microbiota at birth and to follow up these children later. Kalliomaki et al. (Reference Kalliomaki, Collado and Salminen34) analysed the faecal microbiota at ages 6 and 12 months in children that were later at age 7 years identified as being overweight or of normal weight(Reference Kalliomaki, Collado and Salminen34). They found that children who were overweight at age 7 years had significantly fewer bifidobacteria and significantly more staphylococci in the stool during infancy. The present study does not make any conclusions regarding causality, and does not indicate whether F. prausnitzii is a mere marker or whether it is causally linked to obesity. Accrual of data from different regional and ethnic groups can possibly improve our understanding of the role of the faecal microbiota in obesity. Studies are also required to identify the mechanisms whereby different constituents of the microbiota regulate body weight. Eventually this understanding may allow the use of therapeutic manipulations of the flora to combat obesity.
Acknowledgements
R. B. and J. K. were supported by Research Fellowships from the Indian Council of Medical Research. The laboratory received grant no. LSI-141/2002 under the Funds for Infrastructure in Science and Technology (FIST) scheme from the Department of Science and Technology, Government of India.
R. B. was responsible for the quantitative analysis and supervision of the study; G. G. was responsible for recruitment, consenting and sample collection; J. K. and A. M. S. C. were responsible for the diversity studies and analysis; J. H. was responsible for dietary analysis; and B. S. R. was responsible for overall conception, supervision and manuscript finalisation.
The authors have no conflicts of interest to declare.
