Interplay Between Weight Loss and Gut Microbiota Composition in Overweight Adolescents

Authors


(yolsanz@iata.csic.es)

Abstract

The aim of this study was to determine the influence of an obesity treatment program on the gut microbiota and body weight of overweight adolescents. Thirty-six adolescents (13–15 years), classified as overweight according to the International Obesity Task Force BMI criteria, were submitted to a calorie-restricted diet (10–40%) and increased physical activity (15–23 kcal/kg body weight/week) program over 10 weeks. Gut bacterial groups were analyzed by quantitative real-time PCR before and after the intervention. A group of subjects (n = 23) experienced >4.0 kg weight loss and showed significant BMI (P = 0.030) and BMI z-score (P = 0.035) reductions after the intervention, while the other group (n = 13) showed <2.0 kg weight loss. No significant differences in dietary intake were found between both groups. In the whole adolescent population, the intervention led to increased Bacteroides fragilis group (P = 0.001) and Lactobacillus group (P = 0.030) counts, and to decreased Clostridium coccoides group (P = 0.028), Bifidobacterium longum (P = 0.031), and Bifidobacterium adolescentis (P = 0.044) counts. In the high weight–loss group, B. fragilis group and Lactobacillus group counts also increased (P = 0.001 and P = 0.007, respectively), whereas C. coccoides group and B. longum counts decreased (P = 0.001 and P = 0.044, respectively) after the intervention. Total bacteria, B. fragilis group and Clostridium leptum group, and Bifidobacterium catenulatum group counts were significantly higher (P < 0.001–0.036) while levels of C. coccoides group, Lactobacillus group, Bifidobacterium, Bifidobacterium breve, and Bifidobacterium bifidum were significantly lower (P < 0.001–0.008) in the high weight–loss group than in the low weight–loss group before and after the intervention. These findings indicate that calorie restriction and physical activity have an impact on gut microbiota composition related to body weight loss, which also seem to be influenced by the individual's microbiota.

Introduction

Obesity is viewed as one of the major current public health problems and its impact is highest in children, contributing to significant morbidity in adulthood (1). The development of metabolic complications, associated with obesity during childhood, has repercussions in adulthood, increasing the risk of type 2 diabetes and premature cardiovascular diseases (2). A link is thought to exist between obesity, chronic low-grade inflammation, insulin resistance, and endothelial dysfunction (3,4). The risk factors for childhood obesity include diet, low socioeconomic status, parental obesity, rapid infancy weight gain, and decreased physical activity (5). Obesity prevention programs based on changes in school and community environments can decrease childhood weight gain to a limited extent (5). Therefore, further studies on dietary and host factors with an impact on energy balance are needed to improve the intervention strategies and measures for obesity control over time.

Recent reports have suggested that gut microbiota is an important factor affecting energy disposal and storage in adipocytes (6,7). The gut microbiota is also known to be involved in modulation of host immunity, and the inflammatory status associated with obesity in mice (8,9). However, the precise mechanisms by which alterations in microbiota affect obesity and associated disorders are still unclear.

It has been reported how diets based on a high protein intake and/or low carbohydrate intake, or high fat intake may alter microbial composition and activity in the large intestine and thus exert an impact on gut health (6,8,9,10). Nevertheless, knowledge of the interactions between energy intake and specific microbial populations, and their influence on body weight, is limited to small-scale clinical trials (7). Specific studies in obese adolescents, who represent a high-risk population group, are lacking.

The objective of this work was to determine the influence of a multidisciplinary obesity treatment program, comprising a calorie-restricted diet and physical activity, on the structure of the fecal microbiota of overweight and obese adolescents and its relation to dietary intake and weight loss by analyzing the main gut bacterial groups and Bifidobacterium species by quantitative real-time PCR.

Methods and Procedures

Subjects and anthropometric measures

Subjects for the study were selected according to their BMI (weight (kg)/(height (m)2). Childhood overweight (including obesity) was defined according to the International Obesity Task Force cutoffs for BMI (11). BMI z-scores were calculated as a function of the subject's obesity degree when compared with BMI local reference standards (12). Body weight (kg) was estimated without shoes and with light clothing, and measured to 0.05 kg by using a standard beam balance. Skinfold thickness was measured on at the left side of the body to the nearest 0.1 mm using a Holtain skinfold caliper at triceps, biceps, subscapular, suprailiac, thigh, and calf, as previously described (12). All the anthropometric variables were measured in order, three times and averaged. For all the anthropometric measurements, intraobserver reliability was >95% and interobserver reliability was >90%. The characteristics of the 36 selected adolescents (18 female and 18 male; mean age: 14.5 years) to be submitted to the obesity treatment program are shown in Table 2. None of the volunteers were treated with antibiotics for at least 1 month before the intervention study and during the study. The study was conducted in accordance with the ethical rules of the Helsinki Declaration (Hong Kong revision, September 1989), following the EEC Good Clinical Practice guidelines (document 111/3976/88 of July 1990) and current Spanish law which regulates clinical research in humans (Royal Decree 561/1993 regarding clinical trials). Informed consent was obtained from all adolescents and their parents, and the study was approved by the local ethics committees.

Table 2.  Clinical characteristics of the studied subjects
inline image

Intervention

Over a 10-week period, the participants were subjected to the intervention based on an energy-restricted diet (a 10–40% reduction) established according to both obesity degree and regular physical activity (13). The maximum energy intake was 1,800 kcal/day for females and 2,200 kcal/day for males. The physical activity was determined by accelerometry and exercise prescribe at least 1 h of moderate-to-vigorous intensity 3 or 5 days per week, depending of the individual physical activity level. The energy expenditure was estimated in metabolic equivalent' values (14) for each activity and the frequency and intensity of the activities of the exercise program (walking, biking, running, swimming, etc.). The energy expenditure range obtained was from 15 to 23 kcal/kg of body weight per week. Diet energy content was set from the resting energy expenditure calculated with the Schofield equation multiplied by 1.3 as physical activity factor (13). Energy restriction was calculated in function of the subject obesity degree: 10% restriction when the subject had a BMI between 0 and 2 s.d. above the mean, 20% with BMI between 2 and 3 s.d. above the mean, 30% between 3 and 4 s.d., and 40% if the subject had a BMI >4 s.d. above the mean according to BMI local reference standards. Macronutrient distribution was 50% of energy from carbohydrates, 30% from fat, and 20% from proteins. Energy distribution during the day was: breakfast: 20% of daily energy; morning snack: 10–15% of daily energy; lunch: 30–35% of daily energy; afternoon snack: 5–10% of daily energy; and dinner: 20–25% of daily energy.

Dietary assessment

Food diary records were kept for 72 h (2 weekdays and 1 weekend day) both before the start of the study (baseline intakes) and after the intervention (week 10). Detailed information on how to record food and drink consumed using common household measures was provided. Food diary records were returned to their dietitian, and analyzed for energy, water, and nutrient contents based on the CESNID food composition database of Spanish foods (15). Starches were defined as complex carbohydrates and fiber was computed as total nondigestible carbohydrates (soluble and nonsoluble). Due to limitations of the food composition database (15) and also the inherent limitation of dietary assessment in free living young populations, no further details are available according to other key nutrients that are proved to serve as substrate for the gut microbiota (i.e., resistant starch, oligosaccharides, or fructans).

Fecal and DNA sample preparation

Fecal samples were kept immediately after collection at −20 °C and stored until analyzed. Samples were diluted 1:10 (wt/vol) in phosphate-buffered saline (pH 7.2), homogenized and one aliquot was used for DNA extraction by using the QIAamp DNA stool Mini kit (Qiagen, Hilden, Germany).

Microbial analysis by quantitative real-time PCR

Specific primers targeting different bacterial genera and species were used to characterize the fecal microbiota by quantitative real-time PCR (Table 1), essentially as described previously (16,17,18,19,20). Briefly, PCR amplification and detection were performed with an ABI PRISM 7000-PCR sequence detection system (Applied Biosystems, Warrington, UK). Each reaction mixture of 25 µl was composed of SYBR Green PCR Master Mix (SuperArray Bioscience, Foster City, CA), 1 µl of each of the specific primers at a concentration of 0.25 µmol/l, and 1 µl of template DNA. Bacterial concentration from each sample was calculated by comparing the Ct values obtained from the standard curves. Standard curves were created using serial tenfold dilution of pure cultures of DNA, corresponding to 102–109 cells from the culture collection as determined by microscopy counts using 4′,6-diamidino-2-phenylindole. The following strains were used as references: Bacteroides fragilis DSMZ 2451, Clostridium coccoides DSMZ 933, Clostridium leptum DSMZ 935, Lactobacillus casei ATCC 393, Escherichia coli CECT 45, Bifidobacterium longum subsp. longum CECT 4503, Bifidobacterium bifidum LMG 11041, Bifidobacterium breve LMG 11042, Bifidobacterium pseudocatenulatum CECT 5776, and Bifidobacterium adolescentis LMG 11037. The strains were obtained from the Spanish Collection of Type Cultures (CECT) and the German Collection of Microorganisms and Cell Cultures (DSMZ).

Table 1.  Oligonucleotide primers used in this study
inline image

Statistical analyses

Statistical analyses were done using the SPSS 11.0 software (SPSS, Chicago, IL). Due to non-normal distribution, microbial data are expressed as medians with interquartile ranges (IQRs) and differences in bacterial populations were determined by applying the Mann–Whitney U-test and the Wilcoxon signed-rank test. Correlations among variables were calculated using the Spearman's correlation test. Differences in clinical and anthropometric data were also determined by applying the Mann–Whitney U-test. Dietary composition (means and standard deviations) was calculated for crude (unadjusted) nutrients from the 72 h dietary registers and data were averaged for the analysis. All dietary variables submitted to log-transformation showed fit normal distribution. Repeated-measures ANOVA adjusted for sex and age was used to examine differences in group mean intake before (baseline) vs. after the intervention. In every case, P <0.050 were considered statistically significant.

Results

Subjects and obesity intervention program

The studied subjects, 50% female (18/36) and 50% male (18/36), were 14.5 years old (13.0–15.0 years) and maintained an apparently good health status during the study. Clinic and anthropometric characteristics did not differ significantly among subjects at recruitment time, particularly regarding weight (P = 0.266), BMI (P = 0.221), and BMI-z-score (P = 0.138) and, therefore, they were comparable (Table 2). The subjects showed marked differences in weight loss after intervention and, accordingly, subdivided into two groups as low weight–loss group (<2.0 kg of weight loss, n = 13) and high weight–loss group (>4.0 kg of weight loss after intervention, n = 23). The median of weight loss after 10 weeks under the intervention program for the first group was of 1.4 (0.75–2.00) kg, corresponding to 1.3% (IQR 0.85–2.25%) of body weight. This group did not showed significant differences in BMI (P = 0.545), weight (P = 0.801), and BMI z-score (P = 0.579) before and after the dietary intervention. In the second group, the median of weight loss after 10 weeks under the intervention program was of 6.8 (4.8–9.0) kg, corresponding to 7.5% (IQR 5.8–9.3%) of body weight, without detecting significant differences between male (P = 0.204) and female (P = 0.083). In this group, significant differences in BMI (P = 0.030) and BMI z-score (P = 0.035) were detected before and after the intervention.

Dietary data before and after the intervention of the low and high weight–loss groups are shown in Table 3. No interaction between time (before and after intervention) per sex or age group was observed. No significant differences in dietary intake of energy, macronutrients, or on food group level were found between groups before and after the intervention program. The consumption of probiotic foods i.e., yogurt was almost one portion per day (0.9 portions in both groups, one portion in Spain is equivalent to 125 g). None of the subjects consumed pre- or probiotics as supplements. The main sources of carbohydrates, in order of increasing intakes per day, were cereals, potatoes, fruits, and diary products. The main fiber sources of this population were vegetables, cereals, fruits, and legumes.

Table 3.  Daily energy and nutrient intake before (baseline) and after the intervention
inline image

In both adolescent groups, the dietary intervention mainly resulted in a significant reduction (P < 0.05) in intake of total energy (63.8% mean reduction; s.d. 1.2) and macronutrients including proteins (74.5% mean reduction, s.d. 27.2), fat (51.8% mean reduction; s.d. 3.8), polyunsaturated fatty acids (PUFAs) (48.7% mean reduction, s.d. 12.5), carbohydrates (71.6% mean reduction, s.d. 3.9), simple carbohydrates (73.3% mean reduction; s.d. 0.8), and complex carbohydrates (70.6% mean reduction; s.d. 7.2). The reduction in complex carbohydrate intake was significantly and negatively correlated (R = −0.334; P = 0.050) to changes in B. fragilis group as a result of the intervention. Likewise, reduction in PUFA intake was almost significantly and negatively correlated (R = −0.313, P = 0.063) to changes in Lactobacillus group counts.

Influence of intervention in fecal bacterial group composition

Interindividual differences on fecal microbiota composition for all studied adolescents were 0.77 (IQR 0.39–1.70) for B. fragilis group, −0.36 (IQR −0.82 to 0.29) for Bifidobacterium, −0.65 (IQR −0.98 to 0.27) for C. coccoides group, 0.02 (IQR −0.50 to 0.45) for C. leptum group, 0.10 (IQR −0.38 to 0.49) for E. coli, and 0.43 (IQR 0.09–0.83) for Lactobacillus group.

The intervention in whole adolescent population (n = 36) resulted in increased counts of B. fragilis group (P = 0.001) and Lactobacillus group (P = 0.030), and decreased counts of C. coccoides group (P = 0.028). No significant differences were found in the other bacterial groups analyzed. B. fragilis group (R = 0.55, P < 0.001) and C. leptum group (R = 0.52, P < 0.001) counts after the intervention significantly correlated with higher weight loss (kg), while the opposite correlations were found for the E. coli (R = −0.26, P = 0.025), C. coccoides group (R = −0.61, P < 0.001), Lactobacillus group (R = −0.40, P = 0.001), and Bifidobacterium (R = −0.37, P = 0.001) counts.

Changes in bacterial counts as a result of the intervention were also evaluated by considering separately the high and the low weight–loss groups (Tables 4 and 5). Significant differences were not found in bacterial counts of any of the analyzed groups before and after intervention in the low weight–loss group (n = 13 and <2.0 kg of weight loss; Table 4), while significant differences were found in the high weight–loss group (n = 23 and >4.0 kg of weight loss; Table 5). In this last group, B. fragilis group and Lactobacillus group counts significantly increased (P = 0.001 and P = 0.007, respectively), while those of the C. coccoides group significantly decreased (P = 0.001) after 10 weeks of intervention. Moreover, the ratio of Bifidobacterium to C. coccoides group counts increased significantly after the intervention (P = 0.022) when compared to the ratio recorded beforehand, while the ratio of Bifidobacterium to B. fragilis group counts decreased (P = 0.001). When subjects of high weight–loss group were classified according to gender, certain significant differences were found between the two groups. In females, B. fragilis group significantly increased (P = 0.002) after the intervention, while C. coccoides group counts decreased (P = 0.023), which was in accordance with the results obtained when considering the total high weight–loss group of adolescents. Lactobacillus group increased but the differences were not statistically significant. In males, Lactobacillus and B. fragilis groups increased significantly (P = 0.001 and P = 0.033, respectively) after the intervention, whereas a significant (P = 0.007) reduction was found in the C. coccoides group, as was detected for the total high weight–loss group of adolescents.

Table 4.  Bacterial counts in fecal samples of low weight–loss (<2.0 kg) group of adolescents, before and after intervention
inline image
Table 5.  Bacterial counts in fecal samples of high weight–loss (>4.0 kg) group of adolescents, before and after intervention
inline image

Significant correlations between bacterial counts after the intervention and weight loss were found in the high weight–loss group (Figure 1). Increased levels of B. fragilis group (R = 0.27, P = 0.055) and Lactobacillus group significantly correlated (R = 0.55, P < 0.001) with weight loss (kg), while the opposite correlation (R = −0.37, P = 0.010) was found for the E. coli (Figure 1). Similar correlations were recorded between Lactobacillus group (R = 0.53, P = 0.008) and B. fragilis group (R = 0. 44, P = 0.036) levels, and body weight-loss percentages. The reductions in BMI z-scores as a result of the intervention were also significantly correlated with increased levels of Lactobacillus group (R = 0.64, P = 0.001) and B. fragilis group (R = 0.46, P = 0.025). Reduced C. coccoides group levels were related to weight loss (R = −0.611, P = 0.001). The correlation between the reduction in Bifidobacterium to C. coccoides group ratio and weight loss was significantly (R = 0.25, P = 0.030), as well as the correlation between the reduction in Bifidobacterium to B. fragilis group ratio and weight loss (R = −0.62, P < 0.001) as a result of the intervention.

Figure 1.

Correlations between fecal bacterial counts and weight loss after intervention in the high weight–loss group (n = 23; >4.0 kg weight loss) of adolescents. Lines showed the Pearson correlation (linear adjustment). (a) Lactobacillus group vs. weight loss, (b) Escherichia coli vs. weight loss, and (c) Bacteroides fragilis group vs. weight loss.

Influence of intervention in bifidobacterium species composition

In the whole adolescent population (n = 36), total Bifidobacterium group counts were similar before and after intervention, while B. longum and B. adolescentis counts were significantly lower after intervention than before (P = 0.031 and P = 0.044, respectively). No significant differences were found in the other Bifidobacterium species analyzed. B. breve (R = −0.56, P < 0.001) and B. bifidum (R = −0.76, P < 0.001) counts after the intervention significantly correlated with lower weight loss (kg), while no correlations were found in the other species.

Changes in Bifidobacterium species counts as a result of the intervention were also evaluated by considering separately the high and the low weight–loss groups (Tables 4 and 5). Bifidobacterium species counts showed significant differences as a result of the intervention in the high weight–loss group (Table 5), while not in the low weight–loss group of adolescents (Table 4). In the high weight–loss group, all Bifidobacterium species analyzed decreased after the dietary intervention, although only the changes in B. longum counts were significant (P = 0.044). Similar trends were found when comparing Bifidobacterium species composition in males or females. However, only B. adolescentis counts decreased significantly after intervention (P = 0.037) in males, whereas no significant differences were found in females. Significant correlations were not detected between Bifidobacterium species counts and either weight loss, BMI, or BMI z-score.

Differences in fecal microbiota composition between the low and high weight–loss groups of adolescents

The differences in fecal microbiota composition between low and high weight–loss groups of adolescents before and after the intervention are shown in Table 6. Before the intervention, total bacteria, B. fragilis group, and C. leptum group counts were significantly higher (P < 0.001, P = 0.004, and P < 0.001, respectively), while those of C. coccoides group, Lactobacillus group, and Bifidobacterium were significantly lower (P < 0.001, P < 0.001, and P = 0.001, respectively) in the high weight–loss group than in the low weight–loss group. The ratio of B. fragilis group to C. coccoides group was also significantly higher (P < 0.001) in the high weight–loss group. The same trend was detected for Bifidobacterium to C. coccoides group ratio but the differences were not significant (P = 0.140). After 10 weeks of intervention, similar differences on microbiota were found between the low and the high weight–loss groups. Total bacteria, B. fragilis group, and C. leptum group counts were significantly higher (P = 0.015, P = 0.001, and P < 0.001, respectively), while counts of the C. coccoides group, Lactobacillus group, and Bifidobacterium were significantly lower (P < 0.001, P < 0.001, and P = 0.008, respectively) in the high weight–loss group than in the low weight–loss group. In addition, B. fragilis group, Bifidobacterium, and Lactobacillus group to C. coccoides group ratios were significantly higher (P < 0.001, P < 0.001, and P = 0.034, respectively) in the high weight–loss group than in the low weight–loss group.

Table 6.  Bacterial counts in fecal samples of low and high weight–loss groups of adolescents, before and after intervention
inline image

In relation to Bifidobacterium species composition, B. breve and B. bifidum group counts were significantly higher in the low weight–loss group than in the high weight–loss group before (P = 0.001 and P < 0.001, respectively) and after intervention (P < 0.001 for both groups), whereas Bifidobacterium catenulatum group levels were higher in high weight–loss group (P = 0.030 and P = 0.036, before and after intervention, respectively).

Discussion

This study shows for the first time that an intervention based on both a reduction in energy intake and an increase in energy expenditure has an important impact on the composition of the gut microbiota of overweight adolescents related to body weight loss. B. fragilis group and Lactobacillus group seem to be the gut bacteria most amenable to dietary intervention on the basis of the relationships established between the shifts of these bacterial counts and complex carbohydrate and PUFA intakes during the intervention. The Bacteroides genus has been shown to have high ability to utilized complex carbohydrates, which may explain the aforementioned correlation (21). A possible correlation between PUFA intake and Lactobacillus group count reductions was also detected, suggesting that PUFA intake may favor the prevalence of Lactobacillus group in the gut microbiota. In previous studies, PUFAs have been shown to be utilized by Lactobacillus, leading to changes in bacterial fatty acids and suggesting a potential role of Lactobacillus as regulators of PUFA absorption in vivo (22). In addition, PUFAs have positively influenced the adhesion of Lactobacillus to the jejunal mucosa of gnotobiotic piglets, indicating that the intake of these fatty acids may influence the intestinal levels of this bacterial group (23). Nevertheless, the extent to which these bacterial group counts may change and influence weight loss do not seem to depend only on the diet because significant differences in bacterial counts, but not in dietary intakes, were detected between the high and the low weight–loss groups during the intervention. Thus, these findings suggest that the individual's gut microbiota is an additional factor contributing together with lifestyle to body weight regulation.

In response to the intervention, levels of the B. fragilis group significantly increased and correlated to weight loss and BMI z-score reductions, while those of the C. coccoides group, which comprises the Clostridium cluster XIVa including members of other genera such as Coprococcus, Eubacterium, Lachnospira, and Ruminococcus (17), decreased and correlated to weight loss in the whole adolescent population and in the high weight–loss group. These findings were in agreement with the results previously obtained in the same population using fluorescent in situ hybridization technique, which showed that proportions of C. histolyticum and Eubacterium rectale–C. coccoides groups dropped and those of the Bacteroides–Prevotella group increased after the intervention in those adolescents that lost >4 kg (24). In other studies, the fecal microbiota of obese adult subjects also showed a significant increase in Bacteroidetes and a proportional decrease in Firmicutes (which included Clostridium genus) after following either a fat- or carbohydrate-restricted low-calorie diet, which led to weight loss over a year (7). Thus, the association between B. fragilis group and C. coccoides group with energy intake and body weight changes confirmed in this short-term intervention study by using different molecular detection techniques resembles that previously established with the broad phyla Bacteroidetes and Firmicutes in a human long-term intervention study (7).

In this study, the ratio of Bifidobacterium to C. coccoides group counts significantly increased as a result of the intervention in the high weight–loss group. A significant reduction of this ratio was also evident in children who developed atopic diseases later, indicating that the relative proportions of these bacterial groups may precede the development of immune-related disorders (25). Thus, a reduction in calorie intake and an increase in energy expenditure may also have a beneficial overall impact on these bacterial populations and their relationship to the proinflammatory status linked to obesity. However, the intervention led to reductions in B. longum and B. adolescentis counts in the whole adolescent population as well as to reductions in B. longum and B. adolescentis counts in the high weight–loss group and in males of this group, respectively. A reduced dietary intake of carbohydrates by obese adult subjects was shown to be associated with reductions in Bifidobacterium counts in previous studies (10), which could also partly explained the reductions of this bacterial groups in the studied adolescents. In fact, genomic and physiological studies have shown that species such as B. longum and B. adolescentis may actively participate in the utilization of complex polysaccharides in the colon (21). In general, beneficial effects have previously been attributed to Bifidobacterium in connection with obesity. In obese mice models fed with a high fat-content diet, increase in Bifidobacterium caused by administering a high fermentable oligosaccharide was positively correlated with the normalization of inflammatory status, improved glucose tolerance, and glucose-induced insulin secretion (8,9). In addition, reductions in Bifidobacterium populations have been shown to precede the development of overweight (26). It is likely that relative proportions of Bifidobacterium to other bacterial groups, like those detected in this study in relation to Clostridium, rather than absolute numbers have a meaning in the context of obesity.

In general, although some of the reported differences in bacterial counts associated to body weight loss were small, from the biological point of view, these differences could be important in the long term by themselves and because they may lead to changes in the relative proportions of other intestinal bacteria competing for the same ecological niche, which may exert a mild but sustained effect on energy metabolism.

Interestingly, significant increases in Lactobacillus group counts in the whole adolescent population and in the high weight–loss group were detected after the intervention, in agreement with the trend previously detected by fluorescent in situ hybridization analyses although the differences were not significant (24). In this study, the increase in Lactobacillus group counts was correlated with weight loss and BMI z-score reductions in the high weight–loss group, pointing to a role for this bacterial group in body weight management. Until now, information about the impact of different diets on Lactobacillus group levels was scarce. In a recent human study, Lactobacillus group levels were not significantly modified after following different diets: high-protein and low-carbohydrate diet or a high-protein and moderate-carbohydrate diet (10). In mice fed with a high fat-content diet, no significant differences were found in Lactobacillus group levels as compared to controls (8,9).

The gut microbiota of adolescents also appeared to be different between subjects showing high and low weight loss during the intervention and, apparently, this feature was not related to significant differences in dietary intakes. The adolescent group, which showed higher counts of total bacteria, B. fragilis group, C. leptum group, and B. catenulatum group, and lower counts of C. coccoides group, Lactobacillus group, Bifidobacterium, B. breve, and B. bifidum in their fecal microbiota, was the one that experienced the highest weight loss under the intervention. In addition, B. fragilis group, Bifidobacterium, and Lactobacillus group to C. coccoides group ratios were higher in the high weight–loss group than in the low weight–loss group. Thus, B. fragilis and C. coccoides group counts of the individual's microbiota seemed to oppositely influence the ability of the host to loss weight under the same dietary intervention in agreement with the detected correlations between these bacterial groups and weight loss. The opposite influences that seem to exert these bacterial groups on body weight are in agreement with previous reports in obese mice models and in a small-scale trial with adult human subjects (6,7). In this context, although increased counts of C. leptum group, which includes certain members of the genera Clostridium, Ruminococcus, Eubacterium, and Faecalibacterium that belong to Clostridium cluster IV (17), also seemed to favor weight loss, this trend was not confirmed when comparing the bacterial counts of this group before and after the intervention in the high weight–loss group. In addition, reduced B. bifidum and B. breve counts and increased B. catenulatum counts seemed to favor weigh loss, but these trends were not confirmed by the changes detected before and after the intervention in the high weight–loss group. Therefore, further studies are needed to draw conclusions about the role of specific Bifidobacterium species in obesity and weight management. In addition, the possibility that the low weight–loss group did not respond to the intervention due to failure to comply completely with the diet cannot be completely disregarded, because it is well recognized that obese patients have difficulty to accurately record their own food intake.

In summary, an association of specific bacterial groups with obesity and body weight loss has been reported in adolescents, pointing to a role played by B. fragilis, Lactobacillus, and C. coccoides groups, as well as by the relative proportions of B. fragilis, Bifidobacterium and Lactobacillus groups to C. coccoides group. The obtained results have also indicated that the interactions between the gut microbiota and body weight may be sensitive to lifestyle intervention to different extent depending on the individual's microbiota structure.

Evasyon Study Group Collaborators

Coordinator: Marcos A. Local clinical treatment teams and researchers; (Principal Investigators are italicized) Campoy C., López-Belmonte G., Delgado M., Martín-Matillas M., Aparicio V., Carbonell A., Agil A., Silva D.R., Pérez-Ballesteros C., Piqueras M.J., Chillón P., Tercedor P., Martín-Lagos J.A., Martín-Bautista E., Pérez-Expósito M., Garófano M., Aguilar M.J., Fernández-Mayorga A., Sánchez P., Molina Font J.A.;Madrid: Marcos A., Wärnberg J., Puertollano M.A., Gómez-Martínez S., Zapatera B., Nova E., Romeo J., Díaz E.L., Pozo T., Morandé G., Villaseñor A., Madruga D., Muñoz R., Veiga O., Villagra A., Martínez-Gómez D., Vaquero M.P., Pérez-Granados A.M., Navas-Carretero S.; Pamplona: Martí A., Azcona C., Moleres A., Rendo T. Marqués M., Miranda M.G., Martínez J.A.; Santander: Redondo-Figuero C., García-Fuentes M., DeRufino P., González-Lamuño D., Amigo T., Sanz R.; Zaragoza: Garagorri J.M., Moreno L.A., Romero P., De Miguel-Etayo P., Rodríguez G., Bueno G., Mesana M.I., Vicente-Rodríguez G., Fernández J., Rey P., Muro C., Tomás C.; Data management and statistical analysis: Wärnberg J., Calle M.E., Barrios L.

Acknowledgments

This work was supported by grants AGL2007-66126-C03-01/ALI and Consolider Fun-C-Food CSD2007-00063 from the Spanish Ministry of Science and Innovation and AP 002/07 from Consejería de Sanidad (Valencia, Spain). The EVASYON study was supported by grants from Spanish Ministry of Health (PI052451, PI050855, PI051080, PI052369, and PI051579). The scholarship from CONACYT (México) to A.S. and I3P Postdoctoral Contract from CSIC (Spain) to M.C.C. are fully acknowledged. The collaborators in the EVASYON study group are also acknowledged and especially the dietitians P. Romero, P. De Miguel-Etayo, T. Rendo, M.J. Piqueras, and B. Zapatera for the dietary analysis included in this article.

Disclosure

The authors declared no conflict of interest.

Ancillary