A prospective interventional trial on the effect of periodontal treatment on Fusobacterium nucleatum abundance in patients with colorectal tumours

Fusobacterium nucleatum is associated with the progression of colorectal cancer. Thus, the possibility of preventing colorectal cancer or its progression by targeting F. nucleatum has been explored. As F. nucleatum is associated with periodontitis, we analysed whether treating periodontitis could influence F. nucleatum abundance in the colon. Patients with colorectal tumours who underwent colonoscopy were recruited. Patients diagnosed with periodontitis by a dentist were treated for approximately 3 months. Endoscopic resection of colorectal tumours was performed after periodontitis treatment, and resected tumours were pathologically classified as high-(HGD) or low-grade dysplasia (LGD). Saliva and stool samples were collected before and after the treatment. Of the 58 patients with colorectal tumours, 31 were included in the study, 16 showed improvement in periodontitis, and 11 showed no improvement. Stool F. nucleatum levels before treatment were significantly lower in the LGD group than in the HGD group. A significant decrease in faecal F. nucleatum levels was observed in patients who underwent successful treatment but not in those whose treatment failed. Salivary F. nucleatum levels were not altered in patients despite periodontal treatment. Thus, successful periodontitis treatment reduces stool F. nucleatum levels and may aid research on periodontitis and suppression of colorectal cancer development.

Colorectal cancer is the third most common malignant tumour in the world and is the second most common cause of cancer deaths 1 . The 5-year survival rate for colorectal cancer is approximately 10% for patients in advanced stages with metastasis and approximately 90% for those in the early stage. Therefore, early detection, treatment, and prevention are crucial for the recovery of patients 2 .
In 1982, John Robin Warren and Barry James Marshall discovered Helicobacter pylori, and its link to gastric cancer was extensively investigated 3 . H. pylori eradication significantly contributes to the prevention of gastric cancer. However, in the case of colorectal cancer, the microorganisms that are fundamental to carcinogenesis have not been fully identified. Using quantitative PCR, Castellarin et al. observed that Fusobacterium nucleatum was present in large numbers in colorectal cancer tissues, and that its presence was correlated with lymph node metastasis 4 . When compared with the bacterial flora in normal tissues of healthy subjects, the bacterial flora in normal colon tissues of colorectal cancer patients was more enriched in F. nucleatum; the diversity of intestinal bacteria in colon cancer tissues was reported to be lower than that in normal tissues located at a distance from cancer tissues 5 . According to a study that examined the amount of F. nucleatum DNA in colorectal cancer tissues by digital PCR, the amount of F. nucleatum DNA in colorectal cancer tissues was significantly higher than that in normal tissues. Additionally, the amount of F. nucleatum DNA tended to increase as the disease progressed or at an advanced stage of the disease 6 . Using conventional PCR, Mima et al. showed that patients with a high amount of F. nucleatum DNA in colorectal cancer tissues exhibit a poor prognosis 7 .
The National Health and Nutrition Examination Survey 2009-2014 reported that the incidence of periodontitis in the United States was 42.2% for all adults aged at least 30 years, and 7.8% for those with severe periodontitis 8 . Although daily oral care is considered important in the treatment of periodontitis, self-interruption of treatment by patients leads to its progression. Periodontitis has been considered a risk factor for systemic diseases such as diabetes, atherosclerosis, stroke, and fatty liver [9][10][11][12] . The development of these conditions may be attributed to an increase in the abundance of periodontitis-causing pathogens, which have been reported to cause hyper-endotoxemia 13 .
F. nucleatum is known to be associated with periodontitis. The predominance of oral bacteria in patients with periodontitis is altered by dysbiosis. Core species are bacteria whose proportions remain the same either under normal conditions or in periodontitis. F. nucleatum, one of the core species, activates the pathway of fermentation of lysine to butyrate at the affected periodontal sites. Butyrate levels are negatively correlated with oxygen levels, which are associated with the anaerobic conditions of periodontal pockets [14][15][16] . Therefore, F. nucleatum is one of the bacteria associated with inflammation in the oral cavity and is considered a causative agent of oral cancer 17 . F. nucleatum is thus considered to trigger carcinogenesis.
In our laboratory, we compared the strains of F. nucleatum in saliva and colorectal cancer tissues by arbitrarily primed PCR and identified the same strain of F. nucleatum in more than 40% of our subjects 18 . Analysing the strains is the first step towards identifying the F. nucleatum strains directly associated with colorectal cancer. It is plausible that colorectal cancer-associated F. nucleatum strains are derived from the oral cavity. Based on this study, we hypothesised that oral F. nucleatum and periodontitis are associated with colorectal cancer, and treatment of periodontitis would be an effective approach towards the prevention of colorectal cancer. The present study is a prospective, interventional trial, involving patients with colorectal tumours. We analysed the effect of periodontal treatment on F. nucleatum abundance in the stool and saliva of the subjects. In addition, changes in the gut microflora, occurring because of periodontal treatment, were analysed.

Materials and methods
Trial design and registration. This study was a prospective, interventional trial involving a single centre.
It was conducted at Yokohama City University Hospital from August 2017 to August 2019. Clinical research was conducted in compliance with the regulations established by the Declaration of Helsinki. This study was approved by the Ethics Committee at Yokohama City University Hospital (B161201003) and was registered in the University Hospital Medical Information Network (UMIN) as UMIN000027352 on 16/05/2017. All patients enrolled in the study provided written informed consent.
The number of study subjects was approximately 30 because there are no studies on the extent to which F. nucleatum DNA levels are altered by periodontal treatment.
Eligibility criteria. Patients with colorectal tumours who underwent colonoscopy were included. The age of the subjects ranged between 20 and 80 years. The following patients were excluded: patients who did not wish to undergo endoscopic treatment or periodontal treatment; those with advanced colorectal cancer (primarily because they require early therapeutic intervention); those consuming antibiotics or probiotics (because of their potential effect on the gut microbiota); and those with less than 10 teeth.
Procedures. Periodontal treatments. The inclusion criteria for patients with periodontitis were as follows: patients who had not received any periodontal treatment within 6 months or any antibiotics within 3 months and had at least 10 residual teeth. The diagnosis of periodontitis was based on the JSP Clinical Practice Guideline for the Periodontal Treatment 2015 edited by The Japanese Society of Periodontology 19 . The depth of periodontal pockets (probing pocket depth; PPD) and the presence of bleeding during probing (bleeding on probing; BOP) were assessed. The periodontal examination sites and measurement methods were confirmed prior to the start of the study, and the examinations were performed by two designated dentists using a uniform method throughout the study period. For periodontal pockets, all the pockets were measured and the average value was recorded. Bleeding was recorded as the number of pockets that bled out of the total number of probed pockets. The severity of periodontitis in patients diagnosed based on the presence of attachment loss was classified according to the PPD values (mild: PPD < 4 mm; moderate: 4 mm ≤ PPD < 6 mm; severe: 6 mm ≤ PPD). After evaluation for periodontitis, oral hygiene instructions were provided to each patient followed by scaling and root planing procedures. Scaling was performed using a combination of ultrasonic and hand scalers, and root planing was performed in the subgingival area using ultrasonic and hand scalers under local anaesthesia. These procedures were conducted at the dental office at least once a month. Approximately three months after the first consultation and immediately before endoscopic resection of the colorectal tumours, the patients were evaluated for periodontitis. The 'improvement group' was defined as the one in which the post-treatment mean PPD was lower than the pre-treatment mean PPD and the post-treatment percentage of BOP was lower than or equal to the pre-treatment percentage. The 'non-improvement group' was defined as the one in which the post-treatment mean PPD was higher or equal to the pre-treatment mean PPD or the post-treatment percentage of BOP was higher than the pre-treatment percentage.
Colonoscopy and endoscopic tumour resection. The size and location of all tumours observed during colonoscopy were recorded. One of the colorectal tumours was biopsied, and a portion of the specimen was used for www.nature.com/scientificreports/ DNA extraction. Three months after colonoscopy, the tumour was endoscopically resected to the extent possible, and all resected tumours were evaluated by a pathologist. The tumour, identical to the colon tumour biopsied prior to periodontal treatment, was biopsied and used for DNA extraction. Tumours were classified into highgrade dysplasia (HGD) and low-grade dysplasia (LGD).
Collection of saliva and stool samples. Stool and saliva samples were collected before and after periodontal treatment. The patient was asked to gargle 10 mL saline for 1 min. The saliva specimen was centrifuged at 9100g for 3 min; the precipitate was stored at − 80 °C. Stool specimens were collected and stored at − 18 °C in a freezer at home. In our laboratory, they were stored at − 30 °C. Bacterial flora analysis. DNA extraction was performed as previously described; the obtained DNA was stored at − 80 °C until further use. Analysis of the V3-V4 region of bacterial 16S rRNA was performed using a published protocol with minor modifications [21][22][23] . Briefly, the amplicons containing the V3-V4 region of 16S rRNA and unique indices incorporated by an Illumina Nextera XT Index kit v2 (Illumina. K., Japan) were purified using AMPure XP beads (Beckman Coulter, Inc., Brea, CA, USA). The purified barcoded DNA library was diluted to 4 nmol/L using 10 mmol/L Tris-HCl (pH 8.0), and the same volume of each sample was pooled for multiplex sequencing. The multiplexed library pool (6 pmol/L) was spiked with 5% PhiX control DNA (6 pmol/L) and was sequenced using a 2 × 300 bp paired-end run on a MiSeq platform using a MiSeq Reagent Kit v3 (Illumina). All quality-approved, trimmed, and filtered sequences were processed using a custom script based on the QIIME software suite (http:// qiime. org/) 24 .

Outcomes.
Diversity analysis: Alpha diversity was applied to analyse the complexity and species diversity of samples via two indexes: Chao1 and Shannon index. These indices were calculated with QIIME and displayed with the package 'diversity' in the R software (Version 3.6.1). Beta-diversity analysis was used to evaluate the differences in species diversity of samples. To calculate the beta-diversity values, cluster analysis was preceded by principal coordinate analysis (PCoA) using the R Software Version 3.6.1 (https:// www.r-proje ct. org/). Digital PCR. The extracted DNA was diluted with RNAase-free water. The reaction mixture included 33 ng of DNA, RNAase-free water, 7.5 µL QuantStudio™ 3D Digital PCR Master Mix v2 (Applied Biosystems), and primers (made up to a volume of 14.5 µL) and was applied to the QuantStudio™ 3D Digital PCR 20 K Chip Kit v2 (Applied Biosystems). The primer sequences are as follows: forward primer, 5′-AAG CGC GTC TAG GTG GTT ATGT-3′; reverse primer, 5′-TGT AGT TCC GCT TAC CTC TCCAG-3′; and probe, 5′-FAM-CAC GCA ATA CAG TTG AGC CCT GCA TT-3′ (Applied Biosystems). The ProFlex PCR System (Applied Biosystems) thermocycler was used for amplification. DNA was amplified by initial denaturation at 96 °C for 10 min, followed by 39 cycles of 56 °C for 2 min, 98 °C for 30 s, 60 °C for 2 min, and a final hold at 10 °C. The chips were read using the QuantS-tudio™ 3D Digital PCR instrument (Applied Biosystems) and analysed using the QuantStudio 3D Analysis Suite Software (Applied Biosystems).
Statistical analysis. All statistical analyses were performed using JMP ® 15 (SAS Institute Inc., Cary, NC, USA). Data are shown as means ± standard error. A paired Student's t-test was performed to compare the two groups before and after treatment (Figs. 2, 3b,d, 4a,c, S1b, and S2). Shapiro-Wilk's test was used to determine if the data followed a normal distribution; the Wilcoxon rank-sum test was used for data types that followed a non-normal distribution (Figs. 3a and S1a). Fisher's exact tests were performed for categorical data (Table 1). A p-value less than 0.05 was considered significant. For the gut microbiome analysis, we calculated the false discovery rate (FDR) using the Benjamini and Hochberg method. An FDR value of less than 0.1 was defined as statistically significant.

Results
Study flow and patient characteristics. Written informed consent was obtained from 58 patients who underwent colonoscopy and were diagnosed with colorectal tumours. Thirty-one patients were evaluated as the following groups: improvement (n = 16), non-improvement (n = 11), and healthy subjects (n = 4). Twenty patients were excluded after participation in the study: eight patients did not visit the dentist; one patient was found to be consuming probiotics, and three patients withdrew their consent. Four patients were excluded after the screening by the dentist because they had less than 10 teeth remaining. Of the 37 patients diagnosed with periodontitis, seven patients were excluded: two patients whose stool specimens were not collected, two patients who did not visit the dentist, two patients who did not appear for follow-up consultations, and one patient with an inappropriate follow-up period. Of the 30 patients who were examined by a dentist after periodontal treatment, two patients in the improvement group and one patient in the non-improvement group were excluded because stool specimens could not be collected from them. In the healthy subject group, one patient who showed periodontitis after 3 months of screening was excluded (Fig. 1). The clinical characteristics of patients are shown in Table 1.

Changes in the mean value of PPD and the percentage of BOP after treatment of periodontitis.
Although PPD values appeared low even in the presence of periodontitis, we calculated the average value of PPD for all pockets, which resulted in a low average value of PPD. The mean value of PPD and the percentage of BOP in patients significantly decreased after treatment (p < 0.01, p < 0.01, respectively) in the improvement group. In the non-improvement group, the mean value of PPD and the percentage of BOP were not significantly different after treatment (Fig. 2).
F. nucleatum DNA levels in stool and tumours before periodontal treatment. F. nucleatum DNA levels in stool were measured separately for patients with HGD and patients with only LGD (no HGD, LGD group). Figure 3a shows a comparison of F. nucleatum DNA levels in stool between the LGD and HGD groups; the LGD group showed significantly lower amounts of F. nucleatum DNA in their stool (p < 0.001) (Fig. 3a). F. nucleatum DNA levels in colorectal tumours were measured before and after periodontal treatment, but no significant changes were observed in either group (Supplementary Fig. S1).
Changes in F. nucleatum DNA levels in stool samples and gut microflora analysis before and after periodontal treatment. F. nucleatum DNA levels in stool before and after periodontal treatment were quantified by digital PCR (Fig. 3b). In the improvement group, the amount of F. nucleatum DNA was significantly decreased in the stool of patients after treatment (p < 0.05). However, in the non-improvement group, there was no significant change in F. nucleatum DNA levels before and after treatment; in fact, several samples showed an increase in F. nucleatum DNA amount. As the amount of F. nucleatum DNA in stool is associated with the status of periodontal treatment, it may serve as a sensitive marker for the success or failure of the treatment. Furthermore, next-generation sequencing of stool samples revealed no significant changes in bacterial composition before and after treatment (Fig. 3c). There were no significant changes in the Chao1 index and Shannon index before and after treatment. According to PCoA, no clustering shifts occurred before and after treatment (Fig. 3d). These results indicated that periodontal treatment did not change the bacterial composition of the stool.
Changes in F. nucleatum abundance and analysis of salivary microflora before and after treatment of periodontitis. The bacterial flora was analysed using DNA extracted from the saliva of patients before and after periodontal treatment. Digital PCR analysis showed no difference in the amount of F. nucleatum www.nature.com/scientificreports/ DNA before and after treatment in both the improvement and non-improvement groups (Fig. 4a). No significant change in bacterial flora was observed before and after treatment in both the groups (Fig. 4b). There were no significant changes in Chao1 and Shannon indices before and after treatment. PCoA revealed that no clustering shift occurred (Fig. 4c). It was suggested that the periodontal treatment that we performed may not be able to alter the salivary flora and F. nucleatum abundance and that changes in the microbial flora and F. nucleatum abundance in saliva may not be directly linked to a decrease in F. nucleatum abundance in the stool.

Discussion
In this study, we show that successful periodontal treatment could reduce F. nucleatum abundance in the stool. In contrast, there was no decrease in the abundance of F. nucleatum in the saliva and no significant change in the gut microbiota despite periodontal treatment.
According to the guidelines of the European Federation of Periodontology, the endpoint of intervention is improvement in PPD and BOP 25 . In addition, it has been reported that the bacterial load of periodontal pathogens, including F. nucleatum, is correlated to PPD and BOP 26 . Therefore, we considered it appropriate to use PPD and BOP as indicators of periodontal therapy. In several patients in this study, the treatment for periodontitis was unsuccessful. The treatment of periodontitis included voluntary routine oral care by the patients. Two factors prevented many cases of periodontitis from improving: (1) the short period of time (3 months) between the initiation of treatment and the resection of the colorectal tumour, and (2) oral self-care by the patient in addition to professional care. However, if periodontitis is successfully treated, it is possible to reduce F. nucleatum abundance in the stool.
Yachida et al. showed that F. nucleatum abundance increased with the progression of colorectal cancer and that Atopobium parvulum and Actinomyces odontolyticus abundance increased during the carcinogenesis stage and decreased as the disease progressed. This finding suggests that A. parvulum and A. odontolyticus may play a role in the mechanism of carcinogenesis of colorectal cancer and that F. nucleatum is associated with its progression 27 . PCR analysis revealed a higher level of F. nucleatum DNA in tissues from patients with colorectal cancer and Screening at the dentist after the periodontal treatment (n=30) Screening at the dentist (n=5) No stool specimen collected (n=2)

Healthy subject group (n=4)
Patient diagnosed with periodontitis (n=1) Figure 1. Flow diagram of the study. www.nature.com/scientificreports/ those with high-grade adenomas than in the control, but there was no significant increase in the amount of F. nucleatum DNA in tissues from patients with tubulovillous adenomas or low-grade adenomas 28 . This suggests that F. nucleatum may be associated with the initiation of cancer. It has been reported that Fap2 and Fad-A, which are expressed in F. nucleatum, are involved in the growth of colorectal cancer and inflammatory responses 29,30 . Fap2 of F. nucleatum binds to Gal-Gal-NAc, which is overexpressed in colorectal cancer tissues. Fad-A proteins expressed in F. nucleatum adhere to cells and invade cells via E-cadherin, a cell adhesion molecule. Fad-A promotes the growth of colorectal cancer cells by activating β-catenin signalling and promoting the expression of Wnt genes and oncogenes. Therefore, reducing the population of F. nucleatum in the colon may contribute to suppressing the progression or initiation of colorectal cancer. In the current study, F. nucleatum levels in colorectal tumours were not significantly changed upon periodontitis treatment. There were several patients with LGD in this study, and there may be a weak correlation between the presence of F. nucleatum and tumour tissues in LGDs. A larger number of HGD tissues may be analysed in future studies to determine whether periodontal treatment influences F. nucleatum levels in tumours.
In the present study, notably, F. nucleatum abundance in the stool of patients with LGDs was low. Flanagan et al. reported that faecal F. nucleatum abundance was higher in patients with HGD than in patients with LGD and control patients, and there was no relationship between the amount of F. nucleatum DNA in stool and that in tumour tissue samples from the same patients 28 . Amitay et al. reported that F. nucleatum abundance in the stool of patients with colorectal cancer was significantly higher than that in patients with no signs of neoplasm, non-advanced adenoma, and advanced adenoma 31 . We used digital PCR because of its high sensitivity, which makes it a more suitable tool for the quantification of F. nucleatum. Cases of LGDs are frequently detected and the amount of F. nucleatum DNA in the stool of patients with LGD is often low 28 . Digital PCR has the advantage of detecting target DNA with high sensitivity and provides absolute quantification. In a report on the detection of Salmonella typhimurium using both digital PCR and real-time PCR, digital PCR detected lower amounts of DNA and functioned in the presence of inhibitors 32 .
F. nucleatum abundance and bacterial composition of the saliva of patients remained unaltered in the present study, similar to observations in previous studies 33,34 . Although the bacterial composition of saliva did not change after periodontal treatment, the abundance of Fusobacterium in the supragingival plaque has been reported to decrease after treatment 33 . It is possible that changes occurred in subgingival or supragingival plaques; however, this was not analysed in the present study because our previous research led us to believe that F. nucleatum is transferred to the colon by the swallowing of saliva 18 . It has been reported that the salivary flora has a circadian  www.nature.com/scientificreports/ rhythm with a 24-h cycle 35 . In our study, the saliva was collected immediately before the endoscopies to avoid a large difference between the composition at the starting time of the first and second endoscopies. Nevertheless, the time was not exactly the same due to the time required to finish bowel preparation. It is possible that this is the reason why no changes in salivary F. nucleatum levels were observed. If F. nucleatum is transferred to the colon via the gastrointestinal tract and the amount of F. nucleatum swallowed per day was altered by periodontal treatment, it could have been observed as a change in F. nucleatum in the stool. Several patients with ulcerative colitis and Crohn's disease also exhibit periodontitis 36 . F. nucleatum is abundantly detected in the colonic mucosa of patients with ulcerative colitis and has been reported to affect the activity and clinical course of ulcerative colitis, promote mucosal injury, and increase the expression of inflammatory cytokines 37 . In addition, inflammatory bowel disease is influenced by pro-inflammatory cytokines in the gingiva 38 , and improvement in the periodontitis status may alleviate inflammation in the colon. It is possible that F. nucleatum levels decrease as colonic inflammation improves, or vice versa. It is also plausible that changes in the bacterial flora of the gingival plaque affect the colon. According to a previous study, the abundance of F. nucleatum in supragingival plaques decreases after periodontal treatment 33 , and this reduction may be associated with a decrease in F. nucleatum abundance in the stool. In a previous study, blood samples obtained from patients with periodontitis after brushing were analysed by PCR. The results revealed a higher incidence of bacteraemia in these patients 39 . In a mouse model of colorectal cancer, F. nucleatum was injected into the tail vein of mice. F. nucleatum localised to mouse tumour tissues in large numbers via Fap2, indicating that F. nucleatum reached colorectal tumour tissues via a haematogenous route 29 . These findings suggest that brushing may cause dissemination of F. nucleatum into the bloodstream, allowing it to access and bind to colorectal cancer tissues. If F. nucleatum is transferred to the colon via the bloodstream from periodontal pockets, it is possible that periodontal intervention decreased subgingival and colonic F. nucleatum abundance. In addition, biofilms produced by oral bacteria influence colon health and cause chronic inflammation, which is thought to be one of the mechanisms underlying the development of colorectal cancer 40 . The effect of periodontal treatment on the bacterial composition of saliva remains limited, as the composition is strongly influenced by the entire upper respiratory tract, including the pharynx, and not the oral cavity alone.
The limitation of this study is that the subjects did not include patients in advanced stages of colorectal cancer. F. nucleatum abundance in the stool of patients included in this study was lower than that observed in patients with advanced cancer 28 . The benefit of periodontal treatment for patients with colorectal cancer in the advanced www.nature.com/scientificreports/ stages could not be examined. Additionally, plaque analysis could not be performed because oral plaque samples were not collected. Furthermore, although the amount of F. nucleatum DNA in colorectal tumours did not change significantly before and after periodontal treatment, the possibility of contamination by stool adhering to the tumour tissue surface could not be ruled out. Thus, accurate analyses may not have been possible. Finally, this study is not a randomised controlled trial; it is an exploratory trial, because no previous studies have reported the extent to which F. nucleatum DNA levels are altered by periodontal treatment. Therefore, the number of subjects was low. Additionally, the healthy subject sample size was small because we intended to focus our analysis on patients who underwent periodontal interventions.
In summary, we show that periodontal treatment reduces F. nucleatum abundance in the stool of patients with colorectal tumours. Although it is unclear whether the reduction in F. nucleatum abundance in stool can prevent the initiation and progression of colorectal cancer, we believe that this study contributes to the development of research on the possible effects of periodontal treatment on colorectal cancer. Additional studies are needed to verify the mechanism by which F. nucleatum levels in the stool are reduced.

Data availability
The data that support the findings of this study are available from the corresponding author, upon reasonable request.

Funding
This research was supported by KAKENHI to TH (18K07950) and Japan Agency for Medical Research and Development (AMED) to TH (JP16ck0106199).