Exploring the effects of short-course antibiotics on children’s gut microbiota by using 16S rRNA gene sequencing: a case-control study

Abstract

Background

With the widespread use of antibiotics, more attention has been paid to their side effects. We paid extra attention to the impact of antibiotics on children’s bodies. Therefore, we analyzed the characteristic changes in the gut microbiota of children after antibiotic treatment to explore the pathogenesis of antibiotic-associated diseases in more depth and to provide a basis for diagnosis and treatment.

Methods

We recruited 28 children with bronchopneumonia in the western district of Zhuhai, China, and divided them into three treatment groups based on antibiotic type. We took stool samples from children before and 3–5 days after antibiotic treatment. 16S rRNA gene sequencing was used to analyze the effects of antibiotic therapy on the gut microbiota of children. Continuous nonparametric data are represented as median values and analyzed using the Wilcoxon rank-sum test.

Results

While alpha diversity analysis found no significant changes in the mean abundance of the gut microbiota of children after a short course of antibiotic treatment, beta diversity analysis demonstrated significant changes in the composition and diversity of the gut microbiota of children even after a short course of antibiotic therapy. We also found that meloxicillin sulbactam can inhibit the growth of Proteobacteria, Bacteroidetes, and Verrucomicrobia, ceftriaxone inhibits Verrucomicrobia and Bacteroides, and azithromycin inhibits Fusobacteria, Actinobacteria, Proteobacteria, and Verrucomicrobia. We further performed a comparative analysis at the genus level and found significantly different clusters in each group. Finally, we found that azithromycin had the greatest effect on the metabolic function of intestinal microbiota, followed by ceftriaxone, and no significant change in the metabolic process of intestinal microbiota after meloxicillin sulbactam treatment.

Conclusions

Antibiotic treatment significantly affects the diversity of intestinal microbiota in children, even after a short course of antibiotic treatment. Different classes of antibiotics affect diverse microbiota primarily, leading to varying alterations in metabolic function. Meanwhile, we identified a series of intestinal microbiota that differed significantly after antibiotic treatment. These groups of microbiota could be used as biomarkers to provide an additional basis for diagnosing and treating antibiotic-associated diseases.

Introduction

Pathogenic bacteria are a significant cause of infectious diseases in children, such as sepsis, bacterial meningitis, and infectious diarrhea. If not treated properly and timely, it can cause serious consequences [1]. Antibiotics play a vital role in treating bacterial infectious diseases in children. It has also contributed significantly to the reduction of complications and mortality. However, with the wide application of antibiotics, people have found that antibiotics can also cause various harmful effects on the human body, such as antibiotic-associated diarrhea (AAD), an allergic rash, fungal infection, multi-drug-resistant bacteria, and so on [2,3,4]. Researchers also found that antibiotic exposure increases the risk of numerous diseases, such as obesity, diabetes, allergies, asthma, and inflammatory bowel diseases [5].

In addition to killing disease-causing bacteria, antibiotics can affect bacteria that colonize the gut. It will break the original microbial balance of the intestine and cause enteric dysbacteriosis, which is more evident in children [6]. Previous studies have shown that colonization of the gut microbiota begins during the fetal stage and plays a crucial role in the development and maturation of the fetal gut. Similarly, gut microbiota’s role in children’s growth and development is not limited. Gut microbiota can participate in or affect the body’s metabolic and immune processes by maintaining a dynamic balance and producing chemicals [7]. Various studies have recently examined the relationship between disease and gut microbiota [8]. Researchers want to seek different approaches to diagnosis and treatment by uncovering the role of gut microbiota in physiological processes and disease progression.

It is known that antibiotics can cause dysbiosis of the microbiota, inhibiting beneficial bacteria and causing an overgrowth of opportunistic pathogens, resulting in a wide range of clinical manifestations [9]. However, the mechanism has yet to be particularly well known. Most of the gut microbiota in previous studies was cultured by bacteria. Still, most of the culture conditions were only suitable for the growth of some bacteria, so the results were limited. The development of research techniques, particularly at the molecular level and biological information, has provided us with a different perspective on understanding gut microbiota. In addition, previous studies have shown that probiotics are an effective treatment for enteric dysbacteriosis and AAD [10, 11]. However, the mechanism of action needs to be better understood, and it is also critical to note that inappropriate use of probiotics may lead to drug-induced intestinal microbiota disorders.

For these reasons, we sequenced the 16S rRNA V3/ V4 region of stool samples from children treated with different antibiotics. We want to know the changes in the gut microbiome after antibiotic therapy to provide additional evidence to study the mechanism and treatment of intestinal microbiota disorders.

Methods

Human subjects

For the study, we collected 56 stool samples from 28 children, 17 boys, and 11 girls, aged between 5 months and 13 years, in Zhuhai, China. According to different antibiotics, we divided all samples into three research groups, namely research group 1 (RG1, n = 13), research group 2 (RG2, n = 8), and research group 3 (RG3 n = 7). The medicine used in RG1 was meloxicillin sulbactam(Suzhou Erye Pharmaceutical Co. LTD). The medicine used in RG2 was ceftriaxone(Shenzhen Lijian Pharmaceutical Co. LTD), and the RG3 was medicine(Hainan Puli Pharmaceutical Co. LTD). We collected specimens from the three research groups before drug treatment (RGA) and after 3–5 days of treatment (RGB). Key laboratory test data of each research group have been collated in the additional table, including WBC, CRP, PCT, and pathogens (Additional table). All children in our study were treated according to antibiotic use criteria. In our study, we relied on the criteria for antibiotic use: Community-acquired pneumonia diagnosis and treatment standard for children (2019 edition), published by the National Health Commission of the People’s Republic of China and the State Administration of Traditional Chinese Medicine.

Inclusion criteria:

(a) Inclusion age range: children aged 1 month to 14 years; (b) Children with bronchopneumonia; (c) All participants voluntarily joined the study and informed consent from their legal guardians.

Exclusion criteria:

(a) Participants had a history of antibiotic use within four weeks before the study; (b) Participants had a history of gastrointestinal diseases within four weeks before the study, such as abdominal pain, vomiting, diarrhea, constipation, etc.; (c) Participants with a history of probiotics, prebiotics, or any other medications three months before the study that could affect their gut microbiota.

Sample collection and genomic DNA extraction

This study was approved by the Ethics Committee of the Fifth Affiliated Hospital of Zunyi Medical University (Zhuhai). All participants and legal guardians volunteered to participate in this study, and all legal guardians signed informed consent forms. We obtained 56 stool samples from 28 children in three groups. Stool samples were collected and frozen at -80℃ within 15 min. After all the samples are collected, they are transported to the laboratory of the research institution. The microbial community DNA was extracted using MagPure Stool DNA KF kit B (Magen, China) following the manufacturer’s instructions. DNA was quantified with a Qubit Fluorometer using a Qubit dsDNA BR Assay kit (Invitrogen, USA), and the quality was checked by running an aliquot on 1% agarose gel.

Library construction

Variable regions V4 of bacterial 16S rRNA gene was amplified with degenerate PCR primers, 515F (5’-GTGCCAGCMGCCGCGGTAA-3’) and 806R (5’- GGACTACHVGGGTWTCTAAT-3’). Both forward and reverse primers are labeled with Illumina adapter, pad, and linker sequences. PCR enrichment was performed in a 50 µL reaction containing a 30ng template, fusion PCR primer, and PCR master mix. PCR cycling conditions were as follows: 95 °C for 3 min, 30 cycles of 95 °C for 45 s, 56 °C for 45 s, 72 °C for 45 s, and a final extension for 10 min at 72 °C for 10 min. The PCR product was purified using Agencourt AMPure XP beads and eluted in an elution buffer. The Agilent Technologies 2100 Bioanalyzer qualifies the library. The validated libraries were used for sequencing on DNB MGISeq 2000 platform (BGI, Shenzhen, China) following the standard pipelines of DNB and generating 2 × 300 bp paired-end reads.

Sequencing and bioinformatics analysis

Raw reads were filtered to remove adaptors and low-quality and ambiguous bases, and then paired-end reads were added to tags by the Fast Length Adjustment of Short reads program (FLASH, v1.2.11) to get the tags [12]. The tags were clustered into Operational Taxonomic Units (OUTs) with a cutoff value of 97% using UPARSE software (v7 0.0.1090) [13] and chimera sequences were compared with the Gold database using UCHIME (v4.2.40) [14] to detect. Then, OTU representative sequences were taxonomically classified using Ribosomal Database Project (RDP) Classifier v.2.2 with a minimum confidence threshold of 0.6 and trained on the Greengenes database v201305 by QIIME v1.8.0 [15]. The USEARCH global was used to compare all Tags back to OTU to get the OTU abundance statistics table of each sample [16]. The OTU Rank curve was plotted using the R package version 3.1.1. Alpha and beta diversity were estimated by MOTHUR (v1.31.2) [17] and QIIME (v1.8.0) [15] at the OTU level, respectively. The sample cluster was conducted by QIIME (v1.8.0) [15] based on UPGMA. MetaCyc functions were predicted using the PICRUSt software [18]. Principal Coordinate Analysis (PCoA) was performed by QIIME (v1.8.0) [15]. Barplot of different classification levels was plotted with R package v3.4.1 and R package “gplots”, respectively. LEfSe cluster or LDA analysis was conducted by LEfSe. Significant Species or functions were determined by R (v3.4.1) based on Wilcox-test or Kruskal-Test.

Statistical analysis

We use IBM SPSS Statistics 27.0 software for data documentation and statistical analysis. Parametric data of age are expressed as the mean and standard deviation. Continuous nonparametric data are represented as median values and analyzed using the Wilcoxon rank-sum test. P < 0.05 is considered statistically significant.

Results

Study participants feature

We recruited 28 children with bronchopneumonia (male: female, 17: 11; average age 3.93 ± 3.06 years). According to different antibiotics, they were divided into three research groups. RG1 was treated with meloxicillin sulbactam (male: female, 6: 7; average age 3.00 ± 1.78 years), study group 2 was treated with ceftriaxone (male: female, 7: 1; average age 2.75 ± 2.25 years), and study group 3 was treated with azithromycin (male: female, 4: 3; average age 7.00 ± 3.92 years) (Table 1).

Table 1 Characteristics of the study participants

Species sequencing coverage

The rarefaction curves (Fig. 1a) reflect the depth and coverage of sequencing. In this study, the ends of the most rarefaction curves tend to be flat, demonstrating that the current amount of data can reflect the vast majority of the species information in the sample and that the sequencing depth and representation are acceptable. More data will yield only a few new OTUs. The OTU Rank curve (Fig. 1b) had a wide abscissa but a steep slope, indicating that the species richness in the samples was high, but the species composition was not uniform.

Fig. 1

Rarefaction curve and OTU Rank curve (a) The rarefaction curves of sample species. The abscissa is the amount of sample sequencing data, and the ordinate is the actual number of OTUs measured. Blue is for the pre-antibiotic treatment group, and orange is for the post-antibiotic group. (b) OTU Rank curves. The abscissa is ordered according to the number of OTUs, with the ordinate being the relative abundance of OTUs. The different color curves represent different samples, with M for the meloxicillin sulbactam-treated group, X for the ceftriaxone-treated group, and Z for the azithromycin-treated group

Analysis of gut microbiota diversity

We performed diversity analysis separately for each of the three research groups. In alpha diversity, chao1 algorithm results represent species richness within each group and are shown as boxplots (Fig. 2a, b, c). There were no significant differences in mean species richness among the three study groups before and after antibiotic treatment (P = 0.05, P = 0.33, P = 0.80), which might be related to the shorter duration of antibiotic treatment. We would obtain a different result if the course of antibiotic therapy were longer. The beta diversity was analyzed by the unweighted-unifrac algorithm and shown by box plots (Fig. 2d, e, f). There were significant differences in microbiota composition before and after antibiotic treatment in the three study groups (P <​ 0.01, P < 0.01, P = 0.04). The diversity of gut microbiota increased significantly after meloxicillin sulbactam and ceftriaxone treatment. But result decreased substantially after treatment with azithromycin.

Fig. 2

Alpha and beta diversity. (a, b, c) Alpha diversity box plot. The five lines from bottom to top are minimum, first quartile, median, third quartile, and maximum. The abscissa denotes the group, and the ordinate is the Chao index. (d, e, f) beta diversity box plot. The five lines from bottom to top are minimum, first quartile, median, third quartile, and maximum. The abscissa denotes the group; the ordinate is the Unweighted Unifrac index. Different colors indicate different study groups. RG1A is the pre-treatment group of meloxicillin sulbactam, and RG1B is the post-treatment group of meloxicillin sulbactam. RG2A is the pre-treatment group of ceftriaxone, and RG2B is the post-treatment group of ceftriaxone. RG3A is the pre-treatment group of azithromycin, and RG3B is the post-treatment group of azithromycin

​Changes in gut microbiota after antibiotic therapy

The stacked bar chart of species composition shows that at the phylum level, Actinobacteria, Firmicutes, Proteobacteria, Fusobacteria, Verrucomicrobia, and Bacteroidetes were the main compositions of gut microbiota in children (Fig. 3a, b, c). The composition of the gut microbiota differed in the three study groups after antibiotic treatment (Table 2). The relative abundance of Fusobacteria (0.95%, 3.19%) and Actinobacteria (2.01%, 11.71%) increased significantly, while the relative abundance of Proteobacteria (13.61%, 9.24%), Bacteroidetes (44.12%, 37.71%) and Verrucomicrobia (4.33%, 1.79%) decreased significantly after treatment with meloxicillin sulbactam. There was no significant difference in Firmicutes (34.96%, 35.62%). After ceftriaxone treatment, the relative abundance of Proteobacteria (9.43%, 16.67%), Actinobacteria (6.20%, 12.39%), Firmicutes (28.73%, 38.51%), and cyanobacteria (<​ 0.01%, 2.21%) increased significantly. The relative abundance of Verrucomicrobia (8.72%, 2.17%) and Bacteroidetes (45.77%, 26.37%) significantly decreased, and there was no significant difference in Fusobacteria (1.00%, 0.94%). After azithromycin treatment, only the relative abundance of Bacteroidetes (40.19%, 58.67%) increased significantly, and the relative abundance of Fusobacteria (1.28%, 0.17%), Actinobacteria (4.45%, 2.99%), Proteobacteria (15.49%, 2.23%), and Verrucomicrobia (0.55%, 0.28%) decreased significantly. There was no significant difference in Firmicutes (37.92%, 35.54%).

Fig. 3

Bar graph of species composition The abscissa represents the groups, RG1A is the pre-treatment group of meloxicillin sulbactam, and RG1B is the post-treatment group of meloxicillin sulbactam. RG2A is the pre-treatment group of ceftriaxone, and RG2B is the post-treatment group of ceftriaxone. RG3A is the pre-treatment group of azithromycin, and RG3B is the post-treatment group of azithromycin. The ordinate is the proportion of species composition (phylum level). Different colors correspond to different species. Species with abundances less than 0.5% of the sample not annotated at this taxonomic level were combined into Others

Table 2 Proportionality of bacterial community composition

No data indicates

Linear discriminant analysis Effect Size (LEfSe) is used to identify species with significant differences in abundance between different groups. The microbiota abundance of LDA Score > 2 in each group was considered significantly higher than that in the other group, and the larger the score, the more pronounced the difference (P < 0.05). We use the evolutionary clade diagram (Fig. 4a, c, e) and the histogram of the distribution of LDA values (Fig. 4b, d, f) to demonstrate. In this study, 27 microbiota relative abundance increased significantly after meloxicillin sulbactam treatment. The most obvious one is Lactococcus (LDA value 4.17, P < 0.01), 13 microbiota relative abundance decreased significantly, and the most obvious one is Prevotellaceae (LDA value 4.83, P < 0.05) (Fig. 4a, b). After ceftriaxone treatment, the relative abundance of 11 bacterial groups increased significantly, the most obvious one is Actinomycetales (LDA value 4.69, P < 0.05), and 13 bacterial groups decreased significantly, the most obvious one is Bacteroidaceae (LDA value 5.08, P < 0.05) (Fig. 4c, d). After azithromycin treatment, the relative abundance of 6 bacteria groups increased significantly, the most obvious is Bacteroidia (LDA value 4.69, P < 0.05). 17 bacteria groups decreased significantly, and the most obvious is Proteobacteria (LDA value 4.83, P < 0.05) (Fig. 4e, f). These groups of microbiota can be used as biomarkers. They could combine with the microbiota’s biological function to further investigate the mechanisms of antibiotic effects on the children, thus providing additional methods and evidence for diagnosis and treatment.

Fig. 4

Cluster diagram of LEfSe and LDA diagram (a, c, e). LEfSe cluster graph. The nodes with different colors represent microbial communities that play an essential role in the groups. A colored circle represents a biomarker, and the legend in the upper right corner is the name of the biomarker. The diameter of the circle is proportional to the relative abundance. From the inside out, the circles are the species at the level of phylum, class, order, family, and genus. (b, d, f) LDA diagram. It is the distribution map of LDA values of different species, the color represents the corresponding groups, and the length of the bar chart represents the contribution of different species (LDA Score). The figure shows species with significant differences in abundance between different groups under the condition that the LDA Score is greater than the set value (default setting is 2)

Functional difference analysis of metabolic levels

Previous studies have found that gut microbiota participates in the body’s life activities and metabolic processes by producing chemicals. This study confirmed that antibiotics affect the composition ratio and abundance of gut microbiota. After antibiotic treatment, we also analyzed differences in gut microbiota function at the metabolic level to explore how the shift in microbiota affected the body’s metabolic process. As shown in the figure, in the RG1 group, glycan degradation decreased after meloxicillin sulbactam treatment, but the difference was insignificant (Fig. 5a, P = 0.07). In the RG2 group, antibiotic resistance increased significantly after ceftriaxone treatment (Fig. 5b, P = 0.01), while the polymeric compound degradation decreased significantly (Fig. 5b, P < 0.05). In group RG3, nucleoside and nucleotide biosynthesis function, glycolysis function, and secondary metabolite biosynthesis were increased significantly (Fig. 5c, P < 0.01, P = 0.01, P < 0.05), aldehyde and polymeric compound degradation, aldehyde degradation, alcohol degradation, and aromatic compound degradation were decreased significantly (Fig. 5c, P < 0.01, P < 0.05, P < 0.05).

Fig. 5

Analysis of the functional differences. Path difference of the Wilcox test results. Shown on the left is a bar plot showing the relative abundance of the channels for each group. In the middle is the log 2 value of the mean close abundance ratio for the same path in both groups and the right panel shows the p-values and FDR values obtained from the Wilcox test. If the p-value is less than 0.05, the pathway is significantly different between the two groups

Discussion

The side effects of antibiotic treatment on the human body should not be ignored, especially in children. Early exposure to antibiotics can significantly increase the risk of certain diseases, which may be related to a shift in the colonizing microbiota of the child’s gut [19]. Even a short course of antibiotics can take a long time to restore balance among the microbiota and may have long-term effects on colonizing the gut microbiota. This study further seeks to understand antibiotics’ impact on children from a gut microbiota perspective.

From the alpha diversity results, we can see that short-course antibiotic therapy may not significantly affect the mean abundance of gut microbiota, which is the same conclusion reached in other similar studies [20]. It may be related to the course of antibiotics, and the outcome may be different if treatment is prolonged. However, beta diversity analysis showed that antibiotic therapy, even short approaches, significantly affected the composition and homogeneity of gut microbiota. The study showed that gut microbiota diversity increased dramatically after meloxicillin sulbactam and ceftriaxone. However, it decreased significantly after treatment with azithromycin.

From the sample analysis, at the phylum level, the gut microbiota of the children consisted mainly of Actinobacteria, Firmicutes, Proteobacteria, Fusobacteria, Verrucomicrobia, and Bacteroides, in agreement with other studies [21]. At the same time, we found that different antibiotics had different effects on different groups of bacteria. Meloxicillin sulbactam inhibited the growth of Proteobacteria, Bacteroidetes, and Verrucomicrobia, while Fusobacteria and Actinobacteria showed significant increases but had limited effect on Firmicutes. Ceftriaxone had an inhibitory effect on the Verrucomicrobia and Bacteroides, and a substantial increase in Proteobacteria, Actinobacteria, Firmicutes, and Cyanobacteria, with little impact on Fusobacteria. Azithromycin treatment inhibited Fusobacteria, Actinobacteria, Proteobacteria, and Verrucomicrobia, with a significant increase in Bacteroides, but had little effect on Firmicutes. By killing or inhibiting the growth of some bacteria, antibiotics cause a considerable increase in the abundance of others, thus disrupting the original homeostasis of the gut, which may include harmful bacteria or opportunistic pathogens. Cyanobacteria, for example, showed a significant increase in abundance after treatment with ceftriaxone, which may have been acquired by eating seafood. Treatment with ceftriaxone creates an imbalance in the gut microbiota, which may be responsible for the apparent increase in the abundance of Cyanobacteria and other microbiota. Studies have found that Cyanobacteria can produce neurotoxins that may cause neurodegeneration in humans [22].

LEfSe was used to analyze further and compare the microbiota at the genus level of each study group. The microbiota affected by the treatment with meloxicillin sulbactam had the most significant number of species, with 40 groups of microbiota found to be significantly different. Ceftriaxone was followed by 26 species of microbiota that differed significantly. The effects of azithromycin were relatively minor, with 23 groups of microbiota found to be quite different. In terms of bacterial metabolic function, we discovered that azithromycin has the most significant effect on bacterial microbiota’s metabolic process and considerably inhibits the degradation function of amines, aldehydes, aromatic compounds, and other chemical substances. Ceftriaxone promotes antibiotic resistance in the body, while meloxicillin sulbactam has little effect on metabolism. Whether these bacterial imbalances and changes in metabolic function cause clinical symptoms or even have a more significant impact on the body could be a direction for future research.

According to the findings, while most children did not develop significant symptoms after a short course of antibiotics, they significantly altered the composition and function of the microbiota. Clinically, a precise diagnosis is lacking, even when symptoms are present. In our study, the groups of gut microbiota that changed substantially could be used as biomarkers to provide additional evidence for diagnosing antibiotic-associated diseases. In terms of treatment, we can supplement with different probiotics based on changes in specific microbial communities, which play a very significant role in restoring and establishing colonizing bacteria in the child’s gut.

There are still some limitations to our study. The time range of the study was narrow, and only the short-term effects of antibiotic treatment on the gut microbiota of children were analyzed, while the long-term effects were lacking. In addition, the sample size of this study is small, and more samples are needed to prove our findings further. In the future, we may also find specific changes in the microbiota after each antibiotic treatment through more different types of antibiotic studies, thus providing an additional basis for diagnosis and treatment.

Conclusions

Antibiotic treatment significantly affects the diversity of gut microbiota in children, even with short courses of antibiotics. This study confirmed that different classes of antibiotics mainly affected diverse microbiota, resulting in various metabolic function changes. We identified a range of gut microbiota that significantly differed after antibiotic treatment, and they could be used as biomarkers to diagnose and treat antibiotic-associated disease.