Effect of Antibiotics on the Gut Microbiota in Children With Chronic Pancreatitis

OBJECTIVES: Little is known about the effect of antibiotic treatment on the gut microbiota in children with chronic pancreatitis (CCP). Our objective was to identify the main gut microbiota genera and characterize these patients’ functional mutations after using antibiotics. METHODS: The 16S rRNA sequencing method was used to compare the gut microbiota of healthy controls (HCs) with CCP using and not using antibiotics. RESULTS: All CCP demonstrated a signicantly reduced alpha diversity of the gut microbiota (P(cid:0)0.01). The gut microbiota's alpha diversity and the abundance of genera’s beta diversity did not show statistical differences between the non-antibiotics and antibiotics groups. There were 15 altered genera with common abundance in the non-antibiotics and antibiotics groups compared to the HC group. The area under the curve (AUC) of the top three probiotics, i.e., Faecalibacterium, Eubacterium, and Subdoligranulum, was 0.91. Among the 13 genera altered in the non-antibiotics group, the top three genera were not appropriate as biomarkers for cases receiving antibiotics. Compared to these 13 genera, the differences between the genera and the proportion of gram-positive bacteria in the 17 genera altered only in the antibiotics group were not statistically signicant. The Kyoto Encyclopedia of Genes and Genomes (KEGG) analysis indicated that the antibiotics caused alterations in the abundance of certain genera. The enriched functions and the altered gut microbiota in the two groups had their enriched functions. CONCLUSIONS: The use of antibiotics affects the gut microbiota of CCP, but the effect of disease on gut microbiota is still obvious. present higher leukocyte levels, and some hospitals might opt to treat or prevent the infection with antibiotics, and we did not know the specic antibiotics used before admission. [5] Serial studies have revealed that antibiotics alter the abundance and composition of the gut microbiota. The altering capacity depends on the drug class, pharmacokinetics, pharmacodynamics, range of action, dosage, duration, and administration route, causing an increase in the disease's risk, secondary infections allergy, and obesity and the spread of drug-resistant pathogens. [6, 7] However, only a few studies on the gut microbiota of children with CP (CCP) are undergoing antibiotic use within the previous 3 months before admission.


Background
Chronic pancreatitis (CP) is a persistent bro-in ammatory process of the pancreas associated with genetic, environmental, and other pathogenic factors.
This disease eventually led to the pancreas' irreversible injury and increased risk of pancreatic cancer and impaired mental health. [1][2][3] Several evidence has shown that gut microbiota interaction led to the onset and clinical presentation of CP. Participants who received antibiotics or probiotics within 1 month or 3 months before sample collection were excluded. [3,4] However, in some instances of an acute episode of CP, patients often present higher leukocyte levels, and some hospitals might opt to treat or prevent the infection with antibiotics, and we did not know the speci c antibiotics used before admission. [5] Serial studies have revealed that antibiotics alter the abundance and composition of the gut microbiota. The altering capacity depends on the drug class, pharmacokinetics, pharmacodynamics, range of action, dosage, duration, and administration route, causing an increase in the disease's risk, secondary infections allergy, and obesity and the spread of drug-resistant pathogens. [6,7] However, only a few studies on the gut microbiota of children with CP (CCP) are undergoing antibiotic use within the previous 3 months before admission.
The effect of antibiotic treatment on the gut microbiota in the CCP and any difference in the previous treatment results and the antibiotic treatment remains unclear. Therefore, we further analyzed the outcomes in the CCP with antibiotic use by employing the 16S rRNA sequencing method, based on the next-generation sequencing method described in our previous study. [3] Material And Methods

Inclusion and exclusion criteria
The diagnosis of CP was based on the International Study Group of Pediatric Pancreatitis criteria, In Search for a Cure (INSPPIRE), as previously described. [3,8] All cases and healthy controls (HCs) were 4 to 18 years old and were examined from 2014 to 2017. [3] Individuals who used (antibiotics group) and not used (non-antibiotics group) antibiotic(s) or an oral probiotic drink within the previous 3 months were included in the study.
This study was performed in accordance with relevant guidelines and regulations and approved by the Institutional Review Board (IRB) of Shanghai Jiaotong University, and the protocols were approved by the Committee of Human Subjects Protection of the Ruijin Hospital. Informed consent was obtained from the parents of all recruited children. The clinical trial registry number is NCT03809247 (Initial Release Date: 24/12/2018; Last Release Date: 16/01/2019, https://register.clinicaltrials.gov/). [3] Collection of samples, isolation of fecal bacterial DNA, Illumina MiSeq sequencing, processing of sequencing data, and statistical analysis were consistent with our previous paper. [3] Results Basic clinical data The antibiotics group included ten patients, as is shown in Table 1. The information of the non-antibiotic group (n=30) and the HCs group (n=35) have been shown in the previous study. [3] Signi cant differences were not found between the antibiotics and HCs groups in terms of age (8.6 ± 1.1 vs 7.2 ± 0.5 years, P = 0.2181) and sex ratio (50.0% vs 65.7% men, P = 0.366). In addition, signi cant differences were not found between the antibiotics and non-antibiotics groups in terms of age (8.6 ± 1.1 vs 8.3 ± 0.7 years, P = 0.8086), sex ratio (50.0% vs 46.7% men, P = 0.855), and duration of diseases (13.0 ± 19.0 vs 20.0 ± 25.6 years, P = 0.4318).
Alterations of gut microbial diversity in CCP All CCP demonstrated a signi cantly reduced alpha diversity of the gut microbiota (P 0.01; Fig. 1A). An analysis of the beta diversity revealed that the abundance of genera (non-antibiotics group vs. HCs group; antibiotics group vs. HCs group) had distinct clustering ( Fig. 1B and C). The gut microbiota's alpha diversity differences did not reveal statistical differences between the non-antibiotics and antibiotics groups (Fig. 1A). In addition, the analysis of beta diversity demonstrated that the abundance of genera had no distinct clustering between the non-antibiotics and antibiotics groups ( Fig. 1B and C).

Systematic analysis of gene function
Analysis of the Kyoto Encyclopedia of Genes and Genomes (KEGG) database indicated that the gut microbiota of the antibiotics group CCP was enriched in the phosphotransferase system and was depleted in the bacterial motility proteins and porphyrin and chlorophyll metabolism (Fig. 4).

Discussion
Of the CCP patients who used antibiotic(s) or an oral probiotic drink, this is the rst study of the gut microbiota compared to the HCs group. In a recently published study, [3] we compared the gut microbiota of 30 children having CP (non-antibiotics group) with 35 HCs. The participants had not taken any antibiotics or any oral probiotic drink within the previous three months. We identi ed 28 disordered genera in children with CP not treated with antibiotics compared to HC participants. Among the disordered genera, the six genera with the greatest decreases in absolute abundance were selected as CP biomarkers for the pediatric population: Faecalibacterium, Subdoligranulum, Phascolarctobacterium, Bi dobacterium, Collinsella, and Eubacterium. Our ndings further showed that all CCP demonstrated a signi cantly reduced alpha diversity of the gut microbiota. The top three probiotics of the 15 altered genera in the non-antibiotics and antibiotics groups, i.e., Faecalibacterium, Eubacterium, and Subdoligranulum, demonstrated the greatest decreases in the relative abundance and belonged to the six genera as was predicted previously by the biomarkers. [1] These demonstrated excellent diagnostic performance (AUC = 0.91), and there were no differences between the non-antibiotics and antibiotics groups, including the alpha diversity of the gut microbiota or the beta diversity of the abundance of genera (Fig. 1). These ndings show that the main factor affecting gut microbiota is the disease itself and not antibiotics. These ndings further supported the hypothesis that the presence of disordered gut microbiota was related to the pathogenesis of CCP, and some gut microbiota may help to identify new biomarkers or therapeutic targets for CP [3] regardless of the antibiotic use status.
Among the remaining 13 genera only altered in the non-antibiotics group, the top three genera with the greatest decreases in relative abundance were not appropriate as biomarkers for cases receiving antibiotics. The effect of antibiotics on the gut microbiota also existed in the ndings of the depleting bacterial motility proteins, porphyrin, and chlorophyll metabolism, which indicated that in-depth related research still needs to be carried out (Figs. 2 and 3). In the nonantibiotics group, the gut microbiota were enriched in the phosphotransferase system and depleted in the porphyrin and chlorophyll metabolism. Meanwhile, the altered gut microbiota in the non-antibiotics and antibiotics groups also demonstrated speci cally enriched functions. The gut microbiota were speci cally enriched in transporters and depleted in bacterial motility proteins in the antibiotics group, and speci cally depleted in the ribosomes, starch and sucrose metabolism, and aminoacyl tRNA biosynthesis in the non-antibiotic group (Fig. 4). [3] On the other hand, compared to the 17 genera altered only in the antibiotics group, the differences were not statistically signi cant in the genera and the proportion of gram-positive bacteria. These ndings showed that antibiotics' effect on the gut microbiota of the disease exists but is not signi cant. This may be because human microbiotas were remarkably resilient and recovered during antibiotic treatment, with transient dominance of resistant bacteroides and taxa-asymmetric diversity reduction. [9] Our study had some limitations. First, only a small number of patients were included. Studies with more patients and the sequencing and comparison of fecal samples during treatment and follow-up would be more useful. Deep sequencing, including metagenomic sequencing, was also necessary. Second, the details of the antibiotics used in each case before admission were not available. We could only make a preliminary estimate based on experience. The top three antibiotics administered in children patients were cephalosporins (43.8%), penicillins (13.2%), and carbapenems (8.7%), [10] which led to changes of gut microbiota in the study. Detailed pre-admission antibiotic use records are still in progress.
In summary, irrespective of the antibiotic receipt status, all CCP had altered diversity and abundance of the gut microbiota. The three of the six identi ed biomarkers in the previous study (i.e., Faecalibacterium, Eubacterium, and Subdoligranulum) continued to be appropriate for CCP following antibiotic use. Furthermore, the KEGG analysis indicated that the antibiotics caused alterations in the abundance of certain genera and that the enriched functions and the altered gut microbiota in the two groups had their enriched functions. Results showed that the effect of antibiotics on the gut microbiota of the disease does exist, but the effect of disease on gut microbiota is still obvious, which may help further investigation into the pathogenic mechanisms of CP. It reminds us that antibiotics should be used with caution in CCP, even during an acute attack episode, similar to managing pediatric patients with acute pancreatitis. [11] Declarations Funding *, For diagnosis, a patient with abdominal pain and a blood amylase level 3-times higher than the normal reference value or a CT diagnosis of acute pancreatitis was classi ed as having pancreatitis; a patient not meeting these criteria was classi ed as having abdominal pain; a patient who had abdominal pain but no record of blood amylase level or CT examination was classi ed as having abdominal pain 4 .   Figure 1 Alterations of the gut microbiota diversity in the antibiotics, non-antibiotics, and control (healthy adolescent participants) groups. (A) Alpha diversity in patients and control subjects was calculated using the Shannon index, and a two-tailed Wilcoxon rank-sum test was used for comparisons. **P < 0.01. (B and C) Beta diversity in patients and control subjects was calculated using PCoA (principal coordinate analysis) and NMDS (nonmetric multidimensional scaling).   Differences in the functional pathways of the gut microbiota in the antibiotics group (CCP + A) and the control group as analyzed using the Kyoto Encyclopedia of Genes and Genomes (KEGG) orthology (KO). Linear Discriminant Analysis Effect Size (LEfSe) was used for comparison. The predetermined threshold on the logarithmic, linear discriminant analysis (LDA) score for discriminative features was set at >3.0.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.