Correlations between gastrointestinal and oral microbiota in children with cerebral palsy and epilepsy

We here studied the correlation between gut and oral microbiota in children with cerebral palsy and Epilepsy (CPE). We enrolled 27 children with this condition from the social welfare center of Longgang District, collected their oral plaque and stool samples, and analyzed their gut microbiota (GM) and oral microbiota (OM) through 16S rRNA gene sequencing. Taxonomical annotation revealed that the levels of Firmicutes and Bacteroides in the oral cavity were significantly lower in CPE children than in healthy children, whereas the abundance of Actinomycetes increased significantly in CPE children. In addition, Prevotella, Fusobacterium, and Neisseria were the top three abundant genera, representing 15.49%, 9.34%, and 7.68% of the OM and suggesting potential correlations with caries, periodontitis, and malnutrition. For the GM, Bifidobacterium, Bacteroides, and Prevotella were the top three abundant genera in CPE children and probably contributed to the development of chronic inflammation and malnutrition. Furthermore, the OM and GM correlated with each other closely, and the bacterial components of these microbiota in CPE children were remarkably different from those in healthy children, such as Bifidobacterium, Fusobacterium, Bacteroides, and Neisseria. Conclusively, dysbiotic OM can translocate to the intestinal tract and induce GM dysbiosis, suggesting the consistency between OM and GM variations. Altered oral and gut microbial structures have potential impacts on the occurrence of clinical diseases such as periodontitis, caries, and malnutrition.


Introduction
Gut and oral microbiota are the two most abundant microcommunities inhabiting humans, and their impacts on human health are significant. Oral microbiota (OM) are translocated to the gut via sputum swallowing (1), and each adult can swallow 600 ml sputum each day, which contains 10 9 /ml bacterial cells. Similarly, 45% of gut microbiota (GM) are also found among the oral microbiota (OM) (1). As previously reported, the levels of periodontitis-associated Porphyromonas and Fusobacterium nucleatum increased in the GM of liver cirrhosis and colorectal cancer patients, and these bacteria regulated the expression of specific proteins that promote the incidence of these medical conditions (2). In addition, periodontitis and dysbiosis of OM are commonly found to be correlated with inflammatory bowel disease (IBD) (3), and gingival microbiota in liver cirrhosis differed from typical periodontitis-associated microbes (4).
Children with cerebral palsy and epilepsy (CPE) usually have oral health problems, such as periodontitis and dental caries (5,6). Additionally, increased F. nucleatum and Porphyromonas gingivalis (7,8) in the oral cavity can induce GM dysbiosis, thereby causing gastrointestinal dysfunctions such as constipation (9)(10)(11). Clinical studies have observed higher incidence of GM dysbiosis, dental caries, and periodontitis in CPE children than in healthy children, partly owing to a decreased level of specific microbes that improve iron, zinc, calcium, and vitamin D synthesis as well as amino acid absorption to prevent dental caries development (12). Our previous report documented the association of disordered GM in CPE children (13) with malnutrition (14). Moreover, the components of GM in CPE children were affected by diets, and gastrointestinal dysfunction was prevalent in children consuming liquid food (15). Although previous studies have explored the impact of OM and GM on human health, the association between OM and GM as well as the combinational effect of the OM and GM on host health require further investigation.
This study recruited 27 children with CPE to analyze oral and gut microbial structures. Clinical phenotypes were also included to understand the correlation of the OM and GM with periodontitis, dental caries, and malnutrition.

Study design and subject enrollment
We here enrolled 27 CPE children (age: 4-14 years) who received liquid food from the social welfare center of Longgang District. The recruited children had no inherited metabolic diseases, gastrointestinal dysplasia, long-term parenteral nutrition, and antibiotic or probiotic exposure 4 weeks before enrollment. Guardians of the enrolled children provided informed consent. Of the 27 children, children had malnutrition, 26 (96.3%) had periodontitis, 22 (81.48%) had dental caries, and 11 (40.74%) had intractable constipation (Supplementary Table S1).

Sample collection
Oral sampling was performed for children who had not consumed food for at least 2 h. After gargling with sterile saline, two dentists from our hospital collected subgingival microbial samples for the first deciduous or permanent molar at the upper right and stored the samples in buffer (MoBio). Regarding the feces sample, enema was applied if children had not defecated, and then, a caregiver used sterile swabs to collect feces (approximately 5 g). All microbial samples were stored at −80°C within 30 min after sampling.

DNA extraction and sequencing
The bacterial DNA was extracted from the stool and oral samples by using the PowerSoil® DNA Isolation Kit (Mo Bio Laboratories, Carlsbad, USA) and was analyzed through 1% gel electrophoresis. The eligible DNA was amplified with 16S rRNA V3-V4 primers (341-FR: CCTACGGGRBGCASCAG, 806-RR: GGACTACNNGGGTATCTAAT) by using the TruSeq DNA PCR-Free kit (Illumina, San Diego, CA, USA). The qualified amplicon library was prepared for highthroughput sequencing with 250-bp paired-end reads using the Miseq platform (Illumina, USA).

Statistical analysis
The principal component analysis method (ade4 package) was applied to the bacterial relative abundance profile at the genus level, and then, the top two principal components were selected to perform a scatter plot and distinguish samples from different groups. Permutational multivariate analysis of variance (PERMANOVA, vegan package) was used to assess the impact of phenotypes on the microbiota. Significant statistical differences were selected when P was below 0.05. The Pearson's correlation method was applied to analyze the associations between OM and clinical indicators, GM and clinical indicators, and OM and GM. The absolute value of correlation P < 0.05 was considered as a significant correlation.
All the data analysis and visualization in this study were conducted using R software (v3.3.3) and the ggplot2 package.

Characteristics of OM and GM in CPE children at the genus level
We selected the top 15 genera in the OM and GM to perform a comparative analysis. Bacterial composition is different between the OM and GM ( Table 2). Bifidobacterium, Prevotella, and Streptococcus were identified in both OM and GM. The levels of Bifidobacterium, Bacteroidetes, and Prevotella were the highest in the GM, but the top three genera in the OM were Prevotella, Fusobacterium, and Neisseria.

Discussion
Microbial components in the OM and GM are in dynamic homeostasis, which is maintained by various factors such as bile acids in the gut lumen (16)(17)(18). Although millions of microbial cells can be swallowed from the oral cavity into the gastrointestinal tract, the stomach and small intestine function as barriers for these OM and GM because 99% of swallowed microbes are killed at these two sites. Schmidt et al. (19) observed that microbial strains in the gut are translocated from the oral cavity, and a correlation analysis found the oral-gut transmission of opportunistic pathogenic agents as well as dental caries/periodontitis-associated pathogens. For instance, sputum-derived Klebsiella in IBD patients induced intestinal inflammation in gnotobiotic mice, suggesting the role played by microbial transmission in etiology (20). The aforementioned findings implicated the translocation of oral Increasing evidence has demonstrated the association of OM with dental health problems such as periodontitis and dental caries (21)(22)(23). Clinical studies have indicated that 86% of children with cerebral palsy (CP) had mild-to-moderate gingivitis (24) and 52.5%-81% had dental caries (25)(26)(27).
Similarly, we found 96.3% incidence of periodontitis and 81.48% incidence of gingivitis in these children, which can be partly explained by their impaired oral functions, such as occlusion and swallow (25-27). Studies have observed the association of oral health with OM, including Porphyromonas, Treponema, F. nucleatum, Pseudoramibacter, Streptococcus, Prevotella as well as lactic acid-producing bacteria (7, 8, 28- Heatmap for the association of microbiota with clinical phenotypes. The color represents the correlation coefficient: green, negative correlation; red, positive correlation. The asterisk represents the significance of correlation: *, **, ***, and **** represent P = 0.05, 0.01, 0.001, and 0.0001, respectively.
Our studies have found that the abundance of Bacteroides, Faecalibacterium, Lactobacillus, Ruminococcus, Roseburia and other beneficial bacteria is significantly reduced in the gut of CP children, whereas the abundance of harmful bacteria such as Streptococcus, Collinsella, Alistipes, Eggerthella, Enterococcus, and Veillonella is significantly increased (13,15). Neurodegenerative diseases were mainly attributed to Streptococcus (13). Moreover, we found that the abundance of Bifidobacterium was significantly increased in CP children who consumed liquid food, while the abundance of Prevotella increased significantly in CP children who ate normal food (15). Given the strong proof for the association of malnutrition with GM (36), this study also explored GM structures for enrolled children who had malnutrition (40.7% had intractable constipation). As previously reported (13,15), in this study, we observed high levels of pathogens in the recruited children, including Collinsella, Enterococcus, and Streptococcus, and the levels of beneficial microbes such as Bacteroides, Prevotella, Roseburia, and Lactobacillus were depleted (37,38). These beneficial microbes can produce short-chain fatty acids that can function as nutrients, protect gut barriers, and alleviate inflammation (39). Therefore, this is the rationale to speculate the role of GM dysbiosis in malnutrition in our study.
In general, oral microbes are among the sources of GM, and the oral-gut transmission plays a major role in gastrointestinal health (19). As our previous study presented results similar to those of this study, we suspected that disordered OM used various mechanisms to affect gut microenvironments; gut dysfunction can impact gastrointestinal health and disease development. Our findings suggested the potential of probiotics/prebiotics intervention in modulating the OM and GM as well as improving health. However, the lack of healthy children and a small sample size limited the exploration of the association of OM/GM with the health of CPE children. Longitudinal data will also improve the understanding of mechanisms for the correlation between OM/GM and health. Co-occurrence network for oral and gut microbiota. Red lines represent a positive correlation, and blue lines represent a negative correlation.

Data availability statement
The dataset generated for this study is accessible from the NCBI sequence Archive (SRA) database with project number PRJNA853735.

Ethics statement
The studies involving human participants were reviewed and approved by the ethics committee of Longgang District Maternity and Child Healthcare Hospital of Shenzhen city with the registration number of LGFYYXLL-024. Written informed consent to participate in this study was provided by the participants' legal guardian/next of kin.

Author contributions
CH and FW managed the project. CC, YP and NZ collected fecal samples and patients' information. CH and ZY prepared the DNA. CC and ZY were responsible for bioinformatics analysis in this work. JY and FW optimized the graph and statistical analysis. CH and FW explained the analysis results and wrote a paper. All authors contributed to the article and approved the submitted version.

Funding
This work has been strongly supported by Longgang District Science and Technology innovation Bureau (LGKCYLWS2020104).