Correlation of Gut Microbiota with Children Obesity and Weight Loss

Abstract

Children obesity is a serious public health problem drawing much attention around the world. Recent research indicated that gut microbiota plays a vital role in children obesity, and disturbed gut microbiota is a prominent characteristic of obese children. Diet and exercise are efficient intervention for weight loss in obesity children, however, how the gut microbiota is modulated which remains largely unknown. To characterize the feature of gut microbiota in obese children and explore the effect of dietary and exercise on gut microbiota in simple obese children, 107 healthy children and 86 obese children were recruited, and among of the obese children 39 received the dietary-exercise combined weight loss intervention (DEI). The gut microbiota composition was detected by the 16S amplicon sequencing method. The gut microbiota composition was significantly different between obese children and the healthy cohort, and DEI significantly reduced the body weight and ameliorated the gut microbiota dysbiosis. After DEI, the abundance of the Akkermansia muciniphila was increased, while the abundance of the Sutterella genus was decreased in simple obese children. Our results may provide theoretical reference for future personalized obesity interventions based on gut microbiota.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12088-023-01088-3.

Introduction

Children obesity is becoming one important health problem around the world due to the lack of exercise and the intake of high-calorie foods [1]. According to a recent analysis conducted by Global Burden of Disease (GDB) study of the World Health Organization (WHO), the rate of children obesity and overweight have increased by 47% since 1980. Both genetic and environmental factors contribute to obesity development. In recent year, the role of gut microbiota as a main environmental factor that involved in the obesity children development has drawn much attention [2].

Dysbiosis of gut microbiota was observed in obese children, abnormal abundance of Akkermansia muciniphila, Clostridium difficile, Bacteroides plebeius, Bacteroides eggerthii, Enterococci, Blautia, Sutterella and Klebsiella was found in obese children [3–5], and some genera included Akkermansia muciniphila and Desulfovibrio was found to be associated with BMI [6]. In addition, higher ratio of Bacteroides/Firmicutes or lower ratio of Bacteroides to Prevolla was observed [4, 5]. There were dramatic discordances among these studies, and as the gut microbiota is strongly associated with lifestyles, geography, and ethnicities [7, 8], therefore, a localized study is needed to further explore the relations between the gut microbiota and simple obesity children.

Diet and exercise are widely used to control body weight of obese individuals, previous studies indicate that dietary intervention altered the composition of gut microbiota within a day [9], and exercise could induce change of gut microbiota, leading production of more short chain fatty acids [10]. Thus, the gut microbiota should be modulated by weight loss intervention of diet and/or exercise. Previous studies revealed that varies alteration of bacterial such as Lactobacillus, Bifidobacteria and Clostridium cluster XIVa occurred during the weight loss intervention [11–17]. Since varied alterations of gut microbiota were revealed by different studies which may be related to differences in intervention methods, living environments, ethnic backgrounds, and eating habits, it is beneficial to develop personalized obesity interventions by acquisition of localized gut microbiota.

In our study, we recruited 86 obese children as the case group and 107 healthy children as the control group to characterize the composition of gut microbiota, then combined intervention of dietary and exercise was implemented for 39 obese children, and the composition of gut microbiota was explored. Our results may provide clues for future personalized obesity interventions based on gut microbiota.

Information and methods

Study cohort

This study was approved by the Ethics Committee of Hunan Children Hospital. A total of 35 obese children from Hunan Province were enrolled and informed consent was obtained from all participants or the guardians included in the study.

The inclusion and exclusion criteria

The inclusion criteria were followed as below: Aged < 15, BMI meet the criteria for children obesity in China Body Mass Index (BMI) Screening Range of emaciation for school children aged 6–19, and China Body Mass Index (BMI) Screening Range of emaciation, overweight, and obesity for school children, no use of antibiotics and prebiotics or any drug known to alter composition of the microbiota at last month, signed informed consent. Additionally, the exclusion criteria were: (1) suffered from other diseases; (2) refused to sign informed consent;

Research methods

Inclusion method

The design of the study included the potential risks and benefits were clearly explained to the enrolled children or their guardians. Signed informed consents from children or their parents were obtained before starting the trial. The following information was collected, the demographic data such as birthplace and ethnics, and the result of physical examination including body weight, BMI, function of liver and kidney, ultrasonography of major organs, electrocardiogram et al.

Intervention methods

Children were grouped by age and body weight, and received a six-week of dietary and exercise combined intervention (DEI) at the weight loss summer camp. The daily diet, exercise training and accommodation were provided by the summer camp. Each children had a dedicated teacher and management coach, and outing was not allowed during the whole DEI. Interventions designed by professionals were carried out for the children, and the children were trained 6 days a week, with one day off, for 6 weeks, receiving aerobics exercise approximately 4 h per day, 2 h each in the morning and afternoon. Heart rate was used to monitor the intensity during the training. Roughly, high-intensity aerobic trainings such as gymnastics, combating, swimming, outdoor activities (hiking, mountain climbing) and ball games (badminton, football) accounted for 50%-60% of the exercise, while the rest was low-intensity training (30–40%), and the intensity of individual aerobic training was determined to reach a heart rate of about 60% to 80% of the maximal heart rate which was estimated as 220-age. Daily energy consumption was controlled between 1500 and 2000 kcal. All the participants had completed the training under professional guidance. Related adjustment of exercise was made based on the BMI and the heat rates at the time.

All children were subjected to dietary intervention in which all meals are designed and supervised by a professional Nutritionists. Based on the height, body weight and basal metabolic rate of each child, a personized daily diet was provided with energy of 1300–1500 kcal, and carbohydrate, fat and protein constituted 50%, 20 to 25%, 20 to 25% of the total energy respectively. Besides, adequate dietary fiber was added. The dietary adjustments were made based on the exercise intensity and the body weight. The detailed diet was listed in Appendix 1.

Fecal sample processing and DNA extraction

About 10 g of fecal samples were collected by a tube containing 1.0 mL inhibit EX Buffer (Qiagen, Germany). The samples were stored at -20℃ before DNA extraction, and total DNA was extracted by QIAamp Fast DNA Stool Mini Kit (Qiagen, Germany) following the manufacturer’s instruction. DNA concentrations were determined using the Qubit quantification system (Thermo Scientific, Wilmington, DE, US). Extracted DNA was stored at − 20 °C.

16S rRNA amplification and sequencing

The 16S rRNA gene amplification procedure was divided into two PCR steps, in the first PCR reaction, the V3-V4 hypervariable region of the 16S rRNA gene was amplified from genomic DNA using primers 341F(CCTACGGGNGGCWGCAG) and 805R (GACTACHVGGGTA TCTAATCC), and libraries were built as previous described [18]. Amplicon sequencing was performed on the Illuimina MiSeq System (Illumina Inc., CA, USA). The MiSeq Reagent Kits v2 (Illumina Inc.) was used. Automated cluster generation and 2 × 250 bp paired-end sequencing with dual-index reads were performed.

Data processing

Fastq-files were demultiplexed by the MiSeq Controller Software (Illumina Inc.). The sequences were trimmed for amplification primers, diversity spacers, and sequencing adapters, merge-paired and quality filtered by (USEARCH) [19]. UPARSE was used for OTU clustering equaling or above 97%. Taxonomy of the OTUs was assigned and sequences were aligned with RDP classifier. The OTUs were analyzed by phylogenetic and operational taxonomic unit (OTU) methods in the Quantitative Insights into Microbial Ecology (QIIME) [20] software version 1.9.0. α-diversity (Observed OTU number, Shannon index, Simpson index, Chao1 index, PD whole tree index, Goods coverage index) and β-diversity (Bray–Curtis, Unweight UniFrac distances and Weight UniFrac distances) measures were calculated based on the rarefied OTU counts.

Statistical analysis

PERMANOVA was used to test for association between disease status and the overall microbiota composition based on distance matrices (“Adonis” function in R package “vegan”). Significance was assessed by 10,000 permutations. Differential α-diversity analysis was performed using the Wilcoxon rank-sum or Kruskal–Wallis. The software LDA Effect Size (LEfSe) was used to identify taxa characterizing the differences between conditions. False discovery rate (FDR) control was used to correct for multiple testing, statistically significant was confirmed when p ≤ 0.05.

Result

Clinical characteristics of the participants

According to the inclusive and exclusive criteria, swabs samples of feces from 107 control children and 86 obese children were collected. As shown in Table 1, the obese children were elder (9.53 ± 1.78 vs. 8.61 ± 1.41) and higher (135.16 ± 9.92 vs. 143.45 ± 11.61), and there were more boys in obese children (48.6% vs. 75.6%). 39 of the obese children were enrolled for the weight loss interventions, there were 5 girls and 34 boys, and the average age, height, weight, and BMI of the participants before interventions were 10.24 ± 1.59, 148.91 ± 10.06, 61.12 ± 15.05 and 27.33 ± 5.54 (Table 2), respectively.

Table 1.

Physiological data of the obese children and the healthy control, the data was presented as mean ± sd, and the P value was obtained from Wilcoxon test

	Control children	Obese children	P value
Height	135.16 ± 9.92	143.45 ± 11.61	< 0.01
Weight	28.61 ± 7.24	53.17 ± 16.86	< 0.01
BMI	15.48 ± 1.93	25.37 ± 5.86	< 0.01
Age	8.61 ± 1.41	9.53 ± 1.78	< 0.01
Male	52	65	< 0.01
Female	55	21

Table 2.

Physiological data of children enrolled before weight loss intervention, the data was presented as mean ± sd, and the P value was obtained from Wilcoxon test

Height	148.91 ± 10.06
Weight	61.12 ± 15.05
BMI	27.33 ± 5.54
Age	10.24 ± 1.59
Female	5
Male	34

Physiological and sequencing characteristics

We obtained more reads in samples from obese children (Table S1), but there was no significant distribution difference in sequencing quality indicators such as Q20 and Q30. Nevertheless, rarefaction analysis of the data shown that all the sequencing data reached saturation (Figure S1), and the difference of reads had no effects on quantification of the taxonomies.

After intervention, significant weight loss was observed (P < 0.01) as the weight and BMI were 55.46 ± 13.9 versus 61.12 ± 15.05 and 24.16 ± 3.15 versus 27.33 ± 5.54 respectively. The weight losses over time were shown in Fig. 1. The sequencing data indicated that there was no significant difference between the relevant sequencing indicators before and after the intervention, such as total reads, Q20, and Q30 (Table S2). The depth of sequencing before and after the intervention both reached saturation (Figure S2).

Fig. 1.

Weight loss during the intervention. Line Graph was used to trace the weight loss, the points of the same children were linked by dotted line, and “**” indicated the adjusted p value was less than 0.05 or 0.01, when compared with the children’s weight before intervention

Bacterial α- and β-diversity analysis in simple obesity children

α-diversity was calculated based on the quantification of OTU, and Wilcoxon rank-sum test was utilized for differential α-diversity analysis. There was no significant difference of community richness between control and obese children, though the average richness in obese children was lower (Table 3, Fig. 2). β-diversity was carried out based on OTU, and outputted as weighted or unweighted Unifrac distances, and principal coordinate analysis was performed to reveal the difference. No directional effect of obesity was observed with no apparent separation (Fig. 3), however, PERMANOVA analysis revealed that obesity had significant effects on the communities (Table S3).

Table 3.

Community richness of the obese children and the healthy control, the data was presented as mean ± sd, and the P value was obtained from Wilcoxon test

	Control children	Obese children	P value
chao1	756.13 ± 165.3	725.51 ± 197.44	0.49
observed_species	567.46 ± 134.01	531.24 ± 162.01	0.24
PD_whole_tree	40 ± 7.99	38.01 ± 9.78	0.37
shannon	5.21 ± 0.62	5.06 ± 0.73	0.26
simpson	0.93 ± 0.03	0.91 ± 0.06	0.26

Fig. 2.

Phylogenetic diversity of microbiota of control and obese children. Boxplots were drawn for alpha diversities, with the central line marking the median value and upper or lower border represented for the first quartile or the third quartile respectively. The marker “ns” indicated that there was statistic difference between groups

Fig. 3.

Principal component analysis (PCoA) analysis of control and obese children based on unweighted or weighted UniFrac distance. The individual samples are color coded to indicate different groups, and the different gender was plotted at different shape

The DEI induced increase of the community richness (Table 4, Fig. 4). The PCOA analysis based on the weighted and unweighted Unifrac distances indicated that DEI also had no significant directional effect on the gut microbiota, as well as the results of PERMANOVA analysis (Table S4). However, for the same individual, DEI showed a significant impact on the gut microbiota composition (Fig. 5). Our results indicated that the DEI leaded to the alteration of the overall structure of gut microbiota.

Table 4.

Community richness of the children before and after weight loss intervention, the data was presented as mean ± sd, and the P value was obtained from Wilcoxon test

	Before intervention	After intervention	P value
chao1	425.17 ± 154.57	526.73 ± 151.04	0.016
observed_species	320.44 ± 129.01	402.06 ± 132.67	0.028
PD_whole_tree	24.64 ± 8.43	29.96 ± 8.13	0.020
shannon	4.54 ± 0.72	5.02 ± 0.73	0.020
simpson	0.88 ± 0.07	0.92 ± 0.04	0.025

Fig. 4.

Phylogenetic diversity of microbiota before and after weight loss intervention. Boxplots were drawn for alpha diversities, with the central line marking the median value and upper or lower border represented for the first quartile or the third quartile respectively. The points of the same children were linked by dotted line and “*” or “**” indicated that the adjusted p value was less than 0.05 or 0.01 respectively

Fig. 5.

Principal component analysis (PCoA) analysis of samples before and after weight loss intervention based on unweighted or weighted UniFrac distance. The individual samples are color coded to indicate different groups; the points of the same children were linked by dotted line

Dietary-exercise combined intervention modulated the composition of the gut microbiome in simple obesity children

Linear discriminant analysis (LDA) was used to compare the estimated different enriched gut microbiota between obese and healthy children. The relative abundance of Bifidobacterium was higher in healthy children, while Prevotella was overrepresented in the gut microbiota of obese children (Fig. 6). After DEI, abundance of several bacteria was increased including Akkermansia muciniphila (0.00 ± 0.02 vs. 0.02 ± 0.05), Rikenellaceae (0.01 ± 0.01 vs. 0.02 ± 0.03), Enterobacteriaceae (0.03 ± 0.04 vs. 0.06 ± 0.04), while the abundance of Sutterella (0.03 ± 0.03 vs. 0.02 ± 0.01), Alcaligenaceae (0.03 ± 0.03 vs. 0.02 ± 0.01) were decreased (Fig. 7).

Fig. 6.

The differentially enriched bacteria in obese children was determined by Linear discriminant analysis effect size (LefSe) analysis. A Cladogram using the LefSe method indicating the phylogenetic distribution of microbiota exhibited different abundance in control and obese children. B Histogram of the linear discriminant analysis (LDA) scores was displayed for the taxa which showed the significant bacterial difference between the control and obese children

Fig. 7.

The differentially enriched bacteria after weight loss intervention was determined by Linear discriminant analysis effect size (LEfSe) analysis. A Cladogram using the LEfSe method indicating the phylogenetic distribution of microbiota exhibited different abundance before and after weight loss intervention. B Histogram of the linear discriminant analysis (LDA) scores was displayed for the taxa which showed the significant bacterial difference before and after weight loss intervention

Discussion

Gut microbiota imbalance associated with many diseases of children such as constipation, enteritis, obesity, asthma, eczema, autism, and many others [4, 21–24]. Our study investigated the characterizations of obese children and the effects of dietary and exercise combined intervention (DEI) on the gut microbiota, and the results indicated that dysbiosis of gut microbiota was the feature of obese children and DEI could modulate the gut microbiota composition. There was a strong association between gut microbiota and obesity. It further suggested the need for individualized intervention of specific bacteria during weight loss.

In our study, the relative abundance of Bifidobacterium was higher in healthy children, while Prevotella was overrepresented in the gut microbiota of obese children, Bifidobacterium is a widely accepted prebiotic, and is the predominant bacterial colonize the intestinal tract in infants [25]. The abundance of bifidobacterium dropped as the year with age, and administration of Bifidobacterium could reduce body fat, modifies metabolic functions such as improving insulin sensitivity [26–28]. In our cohort, the decreased level of Bifidobacterium may indicate that the metabolic modulation of gut microbiota was attenuated. Prevotella is among the most abundance genera in the bacterial communities, and previous study found that Prevotella was elevated in obese Mexican children [29], while decreased in obese Italian children [3]. When taxonomy of the OTUs were analyzed at species level, Prevotella copri was found to be increased in obese children, since Prevotella copri is considered to be a contributing factor in rheumatoid arthritis [30], and may associated with abnormal immune response [31]. Inadditon, Prevotella copri could activate host chronic inflammatory responses of pigs and increases fat accumulation [32], we speculated that Prevotella copri induced abnormal inflammatory response and promoted fat accumulation in obese children, however, more function assay was needed to determine the detailed mechanism.

39 obese children were recruited for DEI and the changes of gut microbiota before and after the DEI was analyzed. After DEI, the alpha diversity in the obese children was increased, indicating that the low alpha diversity of the gut microbiota in obese children was reversed by DEI, interestingly, which was also in line with previous studies [33]. The PCOA analysis basing on beta diversity revealed that DEI children could not be separated from that of obese children, and the high heterogeneity of gut microbiota may account for it. However, for everyone, obvious separation was observed, suggesting that DEI significantly altered the gut microbiota composition at individual level. Thus, we could still draw the conclusion that DEI modulated the gut microbiota.

To further evaluate the features of gut microbiota associated with DEI, linear discriminant analysis (LDA) was used for excavating the prominent changed organisms before and after the intervention. It was found that the abundance of Akkermansia muciniphila was increased in obese children after the intervention, while the abundance of Sutterella genus was decreased. The abundance of Akkermansia muciniphila in obese individuals is usually found to be reduced [33]. Its abundance is positively correlated with the health status of obese people and high correlated with several physiological indicators [34]. Akkermansia muciniphila administration to obese mice induced significant weight loss. Safety of oral administration of Akkermansia muciniphila was confirmed in human trials, the results showed that the metabolic disorders was ameliorated and the gut microbiota composition was modulated [35]. Akkermansia muciniphila administration decreased food energy efficiency and energy absorption, suggesting its potential application for body weight control [36]. Polyphenols and other functional dietary components can promote the growth of Akkermansia muciniphila [37]. In addition, exercise also increased abundance of Akkermansia muciniphila and inhibits colitis [38]. Therefore, we speculated that polyphenols in diet and exercise synergistically promote the growth of Akkermansia muciniphila in our cohort. Overall, our results suggested that Akkermansia muciniphila may serve as the effector of DEI.

Previous studies have linked the higher abundance of Sutterella genus to a high-fat diet-induced metabolicdisorder [39]. Increased abundance of Sutterella was related with inflammation as revealed in patients with non-alcoholic fatty liver disease [40]. Besides, the abundance of Sutterella genus was also found to be increased in obese children in China [41], suggesting an potential role of Sutterella in obesity. In our present study, it was found that DEI reduced the abundance of Sutterella genus, indicating the decrease the overall inflammation level of the gut. However, the detailed mechanisms needed to be further illustrated.

In summary, we characterized the gut microbiota of obese children, and found that the decreased bifidobacterium and increased Prevotella could be the feature of obese children. DEI was effective to achieve weight loss and ameliorate dysbiosis of gut microbiota. The increased Akkermansia muciniphila and decreased Sutterella were associated with weight loss and may serve as DEI effectors. Our results may provide the evidences for association of gut microbiota and children obesity, and clues for precision modulation of gut microbiota by exercise or diet, and further work was required for the detailed mechanism.

Electronic supplementary material

Below is the link to the electronic supplementary material.

Appendix 1

Breakfast

Oatmeal with fresh fruits and a sprinkle of nuts or seeds.

Steamed or boiled eggs.

Whole-grain toast with a thin spread of natural nut butter.

Green tea or herbal tea without sugar.

Lunch

Grilled chicken or fish.

Steamed or stir-fried vegetables (such as broccoli, bell peppers, and bok choy).

Brown rice or quinoa.

Vegetable soup or clear broth-based soup.

Fresh fruit for dessert.

Dinner

Baked or grilled lean meat (such as chicken breast or fish fillet).

Mixed salad with a variety of vegetables (such as lettuce, cucumber, tomatoes, and carrots) dressed with vinegar or a light vinaigrette.

Steamed or stir-fried vegetables.

Whole-grain noodles or brown rice.

Herbal tea or water.

Author contribution

LJP: Contribution of research design, data collection and writing; YPC: Contribution of obtaining funding and data collection; FQ: Contribution of data analysis; YZ: Contribution of data collection and editing; ZSJ: Contribution of study design and revising.

Funding

This project was supported by the Development and Reform Commission of Hunan Province, (2019) No. 412.

Declarations

Conflict of interest

The authors declare that there is no conflict of interest regarding the publication of this paper.