Changes in oral flora of patients with functional dyspepsia

To explore the changes in oral flora in people with functional dyspepsia (FD). Unstimulated saliva was collected from 21 FD patients diagnosed according to the Rome IV criteria and from 12 healthy controls (HCs) for 16SrRNA sequencing. The pH of saliva samples and community periodontal index (CPI) were tested. The prevalence of small intestinal bacterial overgrowth (SIBO) was obtained by the methane-and hydrogen-based breath test. At the phylum level, FD patients had a higher relative abundance of Spirochaetes and a lower relative abundance of Fusobacteria, TM7 and Proteobacteria than HCs (p < 0.01). In the saliva, Kingella and Abiotrophia genus levels showed significant changes between the FD and HC groups (p < 0.01). Salivary species level marker Intermedia was significantly different between FD and HC groups (p < 0.01). The oral pH of FD patients was higher than that of HCs (p < 0.01). The mean CPI of the FD group was 1.52 and that of the HC group was 0.17 (p < 0.01). Moreover, 71.4% of the FD group was positive for SIBO. The oral flora of FD patients was different from that of HCs. Spirochaetes, Kingella, Abiotrophia, and Intermedia may be diagnostic indicators of FD.

www.nature.com/scientificreports/ treatment were asked to stop treatment ≥ 1 week before sample collection, and people were excluded if they had taken antibiotics in the preceding month. The 12 recruited HCs had to: (1) be in good health; (2) not meet any diagnostic criteria for functional and organic diseases; (3) not have taken drugs that affect oral and intestinal flora in the past month.
Handling and collection of samples. All participants underwent sequencing of 16SrRNA gene amplicons in saliva samples, salivary pH testing, oral CPI testing, and hydrogen-and methane-based breath testing. The specific sample collection methods and test details are as follows: Sequencing of 16SrRNA gene amplicons in saliva samples. To obtain samples of the oral microbiome, individuals were asked to produce 10 mL of saliva after refraining from eating, drinking, and oral-hygiene practice for 8 h. Samples were collected in sterile DNA-and RNA-free 15-mL centrifuge tubes (Precidiag, Sevres, France) and frozen immediately at − 80 °C. De-identified and coded samples were shipped to Exon Bio (Guangzhou, China) on dry ice for extraction and sequencing of DNA.
Salivary pH. Participants were seated comfortably on a chair with their heads bent forward. They were asked to spit into a sterile cup. Saliva was collected after participants had refrained from eating, drinking, and oralhygiene practice for 8 h. Salivary pH was estimated directly using a digital pH meter (Toledo S220-K Seven-Compact pH Meter; Mettler, Geneva, Switzerland) and calibrated against a range of buffers at pH 4.0, 7.0, and 9.21, as described previously 18,19 . Readings were obtained by a recorder blinded to the study protocol. The saliva containers were coded to eliminate the possibility of observer bias.
Distribution of oral diseases. Dental evaluations were undertaken by a single trained dentist to avoid biases. Oral examination was validated using the CPI and DMFT 20 . CPI is to examine gingival bleeding, periodontal pocket depth and dental calculus on index teeth. The scoring criteria that we used was: 0 = gum health; 1 = gingivitis, with bleeding after exploratory examination; 2 = calculus can be detected, but the black part of the probe is exposed outside the gingival pocket; 3 = early periodontal disease, the gingival margin covers part of the black part of the probe, and the gingival pocket has a depth of 4-5 mm; 4 = advanced periodontal disease, the black part of the probe is completely covered by the gingival margin, and the depth of the periodontal pocket is > 6 mm; X = excluded segments (< 2 functional teeth); 9 = unable to check.
Testing of hydrogen and methane in breath. All participants were required to stop using antibiotics, probiotics, lactulose, acid-producing agents, or drugs affecting gastrointestinal dynamics 4 weeks before the examination. An enema had not been given in the past 2 weeks, and a history of acute enteritis was not found. Briefly, 24 h before inspection, participants were instructed to refrain from consuming dairy products, soybean products, wheat, flour products, high-fiber vegetables, and other hydrogen-producing food, and to avoid overeating. Participants were asked to follow a strict, low-residue diet the day before the test and were asked not to eat, drink, or smoke tobacco for the previous 12 h. After obtaining an initial breath sample at baseline, patients ingested 10 g of lactulose dissolved in 50 mL of water. This was followed by measuring the hydrogen and methane in breath samples every 30 min for 150 min 21 . Levels of methane, hydrogen, and carbon dioxide were measured by a Breath Tracker (QuinTron, Santa Maria, CA, USA). Concentrations of hydrogen and methane were corrected for levels of carbon dioxide. Cumulative measurements of hydrogen or methane over time were integrated into a single measurement by the area under the concentration curve (AUC) using the trapezoid method 22 : methane (AUC-CH 4 ) and hydrogen (AUC-H 2 ). Statistical analyses. SPSS v22.0 (IBM, Armonk, NY, USA) was used to analyze data. Clinical and demographic data were compared using the Student's t test or chi-square test as appropriate. Differences in diversity, taxonomic composition, and community function were evaluated using the Kruskal-Wallis or Wilcoxon ranksum test. A p value < 0.05 (two-sided) was considered significant.

Results
Demographic and clinical characteristics. Twenty-one FD patients and 12 HCs were enrolled in this study. Evaluation of demographic and clinical characteristics showed no significant differences in age, sex, or body mass index (BMI) in the FD and HC groups ( Table 1).
Sequencing of 16SrRNA gene amplicons. At the phylum level, Firmicutes, Proteobacteria, Bacteroidetes, Spirochaetes, Actinobacteria, Fusobacteria, and TM7 constituted the vast majority of the predominant salivary microbiota (Fig. 1a). The major genera that contributed to the oral flora in the FD and HC groups was Streptococcus, Prevotella, and Neisseria. Overall, there were significant inter-subject and inter-pair variabilities in microbiota composition (Fig. 1b-d).

FD subjects harbor an altered oral flora. Bray-Curtis principal component analysis and Bray-Curtis
principal coordinate analysis showed that the abundance of oral flora in the HC group was higher than that of the FD group, and there were differences in the phylogenetic structure and composition between the two groups ( Fig. 2a, b). www.nature.com/scientificreports/ Saliva biomarkers can be used to classify FD and HC groups. Further analyses of dysbiosis biomarkers revealed differences in the oral flora of people with FD and in HCs. Genus-level in saliva by LEfSe analysis showed significant changes between the FD group and HCs, including Kingella and Abiotrophia. Species-level biomarkers showed Intermedia to have significant differences between the FD group and HC group (Fig. 3a, b).

Comparison of salivary pH and CPI between the FD group and HC group. Measurement of sali-
vary pH of all participants showed that the mean pH value of the FD group was 6.79 and that of the HC group was 7.48. The oral environment of the FD group was more acidic than that of the HC group (p = 0.002) (Fig. 4a).
The average CPI of the FD group was 1.52 and that of the HC group was 0.17. The association between the FD group and HC group regarding CPI was significant (p = 0.008) (Fig. 4b). DMFT index of FD is 10, DMFT index of HC is 0, there is a significant difference between them (p < 0.05).

SIBO prevalence in FD.
Of the 21 FD patients, 15 (71.4%) had positive results for SIBO according to hydrogen-and methane-based breath tests. In the HC group, only one of 12 (8.3%) patients showed positive results for SIBO (Fig. 5). Significant intergroup differences were noted with respect to SIBO (p < 0.05).

Relationship between salivary pH and microflora in FD.
According to the collected oral pH values of patients with FD, they were divided into four parts: 6.38-6.55, 6.56-6.73, 6.74-6.91, and 6.92-7.20, and the correlation between saliva flora and oral pH values of patients with FD was analyzed. Kruskal-Wallis statistics was used to analyze the difference in the abundance of bacteria with different pH levels, and there was no significant statistical difference in the saliva flora (Firmicutes, Proteobacteria, Bacteroidetes, Spirochetes, Actinobacteria, Fusobacteria, Verrucomicrobia, TM7, SR1, GNO2) of FD patients at different oral pH levels (p > 0.05), which indicated that the content, abundance and proportion of the saliva flora were similar with different oral pH levels, which indicated that the content, abundance and proportion of the saliva flora were similar with different oral pH levels.

Discussion
This is the first study to evaluate changes in the oral flora of FD patients using Rome IV criteria. We showed that the oral flora of people with FD can act as biomarkers. The concentration of genus-level biomarkers in saliva showed significant changes between FD and HC groups, including Kingella and Abiotrophia. The concentration of species-level biomarkers showed Intermedia to have significant differences between FD and HC groups. Deng et al. showed that changes in the composition of oral flora promotes inflammatory reactions, and leads to oral lichen planus disease 23 . Kitamoto et al. 24 reported that periodontal inflammation exacerbates gut inflammation in vivo. Vogtmann et al. 25 found a strong correlation between changes in oral bacteria and the risk of pancreatic cancer. In addition, the intestinal tract and oral cavity share a close relationship and are important components of the human digestive system; moreover, both are major ecological sites of microorganisms in the human body. Nearly all of the saliva produced by adults each day goes into the gastrointestinal tract, so salivary bacteria have good possibilities of colonizing the gut 26 . Sato et al. proposed that orally administered Porphyromonas gingivalis (a representative periodontopathic bacterium) alters the gut microbiome 27 . There is a justified association between microbiomes in the mouth and gut in FD, although current evidence that the microbiome causes FD is far from conclusive. Therefore, when exploring FD pathogenesis, the imbalance in oral flora as an important potential mechanism should be considered.
We showed that the mean pH in the mouth of FD patients was 6.78 and that in the HC group was 7.38. The pH in the mouth of FD patients was acidic, whereas that of the HC group was alkaline. Studies have shown that the oral pH of patients with gastroesophageal reflux disease is acidic 28 . However, the correlation between FD and oral pH has not been reported. A low pH in saliva can cause disorders pertaining to oral flora. In addition, a continuous decline in salivary pH can cause further damage to oral soft tissues, leading to ulcers, erythema, oral lichen planus disease, and other diseases of the oral mucosa. In addition, an acidic oral environment induces more acid production by dental plaques, thereby increasing the acidity of the enamel surface and resulting in local decalcification and greater risk of dental caries. We showed that the oral pH of FD cases was more acidic than that of the HCs, which may increase the risk of oral mucosal diseases and dental caries. Therefore, timely detection of oral pH as well as prevention and treatment of an acidic environment are important. www.nature.com/scientificreports/  www.nature.com/scientificreports/ We demonstrated that the CPI and DMFT in the FD group were significantly higher than that in the HC group. This finding suggests that it is easier for patients with FD to develop oral diseases than HCs. Studies have shown that the higher the value of CPI and DMFT, the more prone to oral diseases 29 . FD patients should, therefore, be treated in time to improve oral acidity and reduce the risk of oral diseases.
Consistent with the literature, we demonstrated that SIBO prevalence in FD patients was higher than that in HCs. Costa et al. 30 showed that SIBO prevalence in FD patients was 56.5%, and that a higher prevalence of SIBO was observed in dyspeptic patients who used proton pump inhibitors compared with that of a control group. Kim et al. 31 also showed that 2-month treatment with ursodeoxycholic acid resulted in FD-symptom improvement and reduced methane values during 90 min on the lactulose breath test, suggesting that methane-producing SIBO were associated with dyspepsia symptoms. In addition, SIBO is believed to be caused by the lateral movement of intestinal flora, which may trigger a mucosal immune response, leading to the release of cytokines www.nature.com/scientificreports/ and infiltration of inflammatory cells. This continuous, low-grade inflammatory state changes gastrointestinal motility and permeability, leading to FD onset 32 . The integrity of the duodenal mucosa in FD patients is damaged and permeability is increased, and these structural changes can cause secondary inflammatory reactions and reduce local antibacterial ability, leading to bacterial translocation and SIBO initiation 33 . We hypothesized