Roles of oral microbiota and oral-gut microbial transmission in hypertension

Graphical abstract


Introduction
As a global public health concern, hypertension (HTN) is a significant risk factor for cardiovascular diseases and increases allcause morbidity and mortality worldwide [1]. Although plenty of efforts have been made in investigating the epidemiology and pathophysiology, the etiology of HTN remains complex [2]. Accumulating evidence has indicated gut microbiota as an essential environmental factor involved in the development and progression of HTN [3]. However, it is much less clear whether other microbial flora, for example, oral microbiota, also plays a role in HTN, although the oral cavity is the second largest microbial habitat in human bodies [4].
In fact, very little is known regarding the function of oral microbiota in HTN, although the association between periodontitis (PD)a common periodontal disease closely related to oral microbial dysbiosis -and HTN has been established. Cross-sectional studies have demonstrated that patients with PD have higher blood pressure (BP) and that more elevated BP usually occurs in patients with more severe PD [5]. Intensive periodontal treatment of PD has been shown to decrease BP more effectively than control periodontal treatment in a randomized controlled trial [6]. Interestingly, the BP reduction correlates with the decrease of periodontal pocket depth, an important indicator for improvement of periodontal status, implying a causal relationship between PD and HTN [6]. Growing evidence has suggested that dysbiotic oral microbiota not only causes PD by undermining periodontal supporting tissues but is also adverse to systematic diseases [7]. A study in older women has suggested associations between oral microbiota and BP [8]. However, the function of oral microbiota dysbiosis in HTN has remained incompletely understood.
The importance of oral-gut microbial transmission in systematic diseases has been increasingly appreciated. Unlike the conventional view considering translocations of oral microbes to the gut rare events because of physiological segregation, recent data indicate that it is common and extensive for oral microbes to translocate to and then colonize in the intestine and that the oral cavity is regarded as an endogenous reservoir for gut microbiota [9]. Moreover, elevated ectopic colonization of oral microbes in the gut has been linked to various disease conditions, including cirrhosis [10], rheumatoid arthritis [11], inflammatory bowel disease [12], Alzheimer's Disease, [13] and even coronavirus disease 2019 [14].
However, it is unknown whether oral-gut microbial transmission plays a role in HTN.
In this study, we set out to explore the functions of oral/gut microbiota and oral-gut microbial transmission in HTN. First, we studied the relationship between PD and HTN in both crosssectional and follow-up cohorts. We also investigated the alterations of oral and gut microbiota in the cross-sectional cohort using 16S rRNA gene sequencing; explored the correlations between oral/gut microbiota and BP/other clinical parameters; surveyed communications between oral and gut microbiota at the genus level. We further analyzed the oral/gut microbiota of this cohort using metagenomic sequencing and explored the oral-gut microbial transmission at the species level. Subsequently, we studied the alterations of oral/gut microbiota and oral-gut microbial transmission in the follow-up cohort. Finally, we investigated the causal link between oral-gut microbial transmission and HTN in mice.

Materials and methods
Detailed methods are available in the Supplementary Material. Raw sequences are available in the Sequence Read Archive of NIH with accession numbers PRJNA764503, PRJNA765566 and PRJNA774166.

Ethics statement
All experiments involving human and animals were conducted according to the ethical policies and procedures approved by the Institutional Review and Ethics Board of Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine (Approval no. SH9H-2018-T66-3).

Study cohorts
In the cross-sectional cohort, 95 participants with HTN and 39 controls without HTN were recruited. A follow-up cohort consisted of 52 HTN participants and 26 controls. HTN diagnosis was based on the criteria of 2018 ESC/ESH Guidelines [2], and periodontitis (PD) was defined according to the gold standard of Centers for Disease Control and Prevention/American Academy of Periodontology case definitions [15]. The detailed criteria for diagnosis of HTN were systolic blood pressure (SBP) ! 140 mmHg and/or diastolic blood pressure (DBP) ! 90 mmHg. BP of each participant was measured for 3 times with 5-min intervals by a physician using Omron electronic sphygmomanometer, and the average was recorded. The detailed criteria for diagnosis of PD were at least 2 interproximal sites had clinical attachment loss (CAL) ! 3 mm and probing depth (PD) ! 4 mm. Participants with any of these conditions were excluded: 1. Pregnancy; 2. Smoking; 3. Diseases including peripheral artery disease, autoimmune disease, heart failure, renal failure, cancer, irritable bowel syndrome, inflammatory bowel disease, and recurrent aphthous oral ulcers; 4. Being treated with antibiotics or probiotics, or undergone oral/gut surgeries/treatments within the last 2 months; 5. Having<8 natural teeth. All medical data were collected according to standard clinical procedures. The HTN group had hypertension for an average of 12.3 years. There was no change in medications during the 6-month follow-up period. The body weights of the participants in the follow-up did not change significantly either.

Sample size calculation
Sample size was estimated using a previously established method [16]. To achieve a minimum correlation coefficient of 0.25 (r = 0.25) [17], 5% significance level (a = 0.05), and 80% test power (b = 0.2) in correlation analyses, a sample size of 123 was required.

Animal experiments
Ten-week-old male C57BL/6J mice were pretreated with antibiotics cocktail (ABX) to deplete gut microbiota and then treated with sterilized water, saliva from participants with or without HTN by oral gavage. The mice were then subcutaneously implanted with minipumps containing vehicle (saline) or angiotensin II (Ang II). BP was measured using Tail-cuff.

Microbiota sequencing
Oral and gut samples were collected and DNA was extracted. Both 16S rRNA gene sequencing and metagenomic sequencing were performed at Personal Bio Inc. (Shanghai, China) using Illumina platforms.

Statistical analysis
Student's t-test, two-way ANOVA followed by Sidak's multiple comparison test, and Mann-Whitney U rank-sum (MW) test were performed using Prism (GraphPad Software). Kruskal-Wallis rank-sum test and Permutational multivariate analysis of variance were used to analyze alpha diversity and Principal coordinate analysis (PCoA) distance matrix respectively by R software (Version 4.0.2). The Pearson's chi-squared test was used for statistical analysis of sex, drinkers, and follow-up rate between no HTN and HTN. Adonis test was also performed by R software. Kruskal-Wallis rank-sum test was used to analyze the abundance of microbiota. Spearman's correlations among microbiota, clinic parameters and metabolites were tested and visualized by corrplot R package. Mann-Whitney U test and Wilcoxon rank-sum test were performed to test the single nucleotide variant (SNV) distance of oral-gut sample pair between inter-individual and intraindividual. The Linear discriminant analysis effect size was used to identify different taxa/pathway enrichment between no HTN and HTN.

PD is associated with higher BP in both cross-sectional and follow-up study
To study the impacts of PD and oral/gut microbiota on BP, we recruited 134 participants, 39 of which were without HTN (no HTN) and 95 had HTN (Table 1). We first analyzed the association between PD and HTN at baseline. When the whole cohort was analyzed together, participants with PD manifested significantly higher systolic BP (SBP) than those without PD (no PD) (Fig. 1A). When non-hypertensive and hypertensive participants were analyzed separately, non-hypertensive participants with PD showed a strong trend to have higher SBP than no PD (Fig. 1B), and hypertensive participants with PD showed significantly higher SBP than no PD (Fig. 1C). Participants with PD were also inclined to have higher diastolic BP (DBP) than no PD either among the whole cohort or being divided into non-hypertensive and hypertensive groups, although no statistical difference was detected ( Fig. 1D-F). These results are largely consistent with previous reports [18,19].
In addition to these cross-sectional analyses at baseline, we conducted a 6-month follow-up study. Both SBP and DBP were substantially higher in participants with PD than no PD throughout the follow-up (Fig. 1G, H). Together these results demonstrated that participants with PD had persistently higher BP than no PD.

Altered diversity of oral and gut microbiota in participants with HTN
We identified 368 genera under 38 phyla in saliva, 265 genera under 31 phyla in subgingival plaques, and 113 genera under 28 phyla in feces of our study population using 16S rRNA gene sequencing (Table S1). Analyses of alpha (a)-diversity demonstrated that microbiota of HTN participants tended to have lower Chao1 richness, Faith's phylogenetic diversity, Shannon diversity index, and Pielou's Evenness index in all 3 sample types compared to no HTN participants, although none of the differences reached statistical significance ( Fig. 2A). Shannon rarefaction curves also supported the trend of decrease in a-diversity in the HTN group, especially for gut microbiota (Fig. 2B), consistent with the results   previously reported [20]. PCoA was performed to assess beta (b)diversity of the microbiota (Fig. 2C). Permutational multivariate analysis of variance based on Bray-Curtis distance demonstrated statistically significant differences in b-diversity of subgingival microbiota and gut microbiota between no HTN and HTN group (Fig. 2D). The difference in b-diversity of salivary microbiota also showed a statistical trend (p = 0.056) between the two groups ( Fig. 2D). Results of Adonis test based on Bray-Curtis did not show significant impact of diabetes or antihypertensive treatment on microbiota (Table S2).

The composition of oral and gut microbiota shifts in participants with HTN
Stacked bar plots of relative abundances at the phylum level demonstrated evident differences in oral and gut microbiota between no HTN and HTN (Fig. 3A). Firmicutes, Bacteroidetes, Proteobacteria, Actinobacteria, and Fusobacteria were predominant phyla in saliva and subgingival plaques, and Firmicutes and Bacteroidetes were predominant phyla in feces of both no HTN and HTN (Fig. S1, relative abundance > 10%). Firmicutes significantly A B D C Fig. 2. Shifts of microbiota diversity in oral cavity and intestine of participants with HTN. A, Chao1, Faith's phylogenetic diversity (Faith's), Shannon index, and Pielou's evenness (Pielou's) of oral (saliva and subgingival plaques) microbiota and gut (feces) microbiota in participants with or without HTN. B, Rarefaction curves of oral and gut microbiota in participants with or without HTN. C, PCoA of oral and gut microbiota in participants with or without HTN. D, Bray-Curtis distance of oral and gut microbiota in participants with or without HTN. All microbiota was analyzed using 16S rRNA gene sequencing. n = 39:94 for saliva, 39:93 for subgingival plaques, and 24:52 for feces. Kruskal-Wallis rank sum test was used for statistical analysis in A and permutational multivariate analysis of variance in D. decreased in HTN compared to no HTN in all 3 types of samples and Proteobacteria substantially increased in saliva and subgingival plaques of HTN (Fig. 3B). Moreover, in saliva and subgingival plaques, Bacteroidetes were enriched in no HTN, whereas Actinobacteria and Fusobacteria were enriched in HTN group; in feces, Bacteroidetes, Proteobacteria, Actinobacteria and Fusobacteria were enriched in HTN group, although these differences did not reach statistical significance (Fig. 3B). We further analyzed the composition of microbiota at the genus level. Among the top 50 genera, there were 14 genera in salivary microbiota, 15 genera in subgingival microbiota, and 10 genera in gut microbiota that showed substantial differences between no HTN and HTN (p < 0.1) (Fig. 3C).

Associations between oral/gut microbiota and BP and other clinical parameters
To evaluate the clinical significance of the alterations of oral and gut microbiota in HTN, Spearman's correlation analyses were performed to assess the associations between clinical parameters and the relative abundances of the top 50 genera in saliva, subgingival plaques, and feces. We first investigated the associations between microbiota and BP. SBP had a significant positive correlation with Actinobacillus and negative correlations with Catonella, Megasphaera, Prevotella, Veillonella, and Given the importance of inflammation in HTN [21], we then analyzed the associations between inflammatory markers and microbiota. Interleukin-6 (IL-6) had significant correlations with 7 salivary, 10 subgingival, and 6 fecal genera ( Fig. 4A-C). Procalcitonin had significant correlations with 2 salivary and 12 subgingival genera (Fig. 4 A-B). C-reactive protein (CRP) had significant correlation with 6 salivary, 9 subgingival, and 6 fecal genera (Fig. 4-A-C). Notably, PD-associated genera such as salivary Porphyromonas, subgingival Porphyromonas, Fusobacterium and Treponema had significant positive correlations with plasma IL-6, and salivary Porphyromonas had significant positive correlation with plasma CRP (Fig. 4A-B).
We further analyzed the associations between immunological parameters and microbiota, focusing on the red-or bluehighlighted microbes. Leukocyte counts and leukocyte ratios manifested similar pattern of correlations with abundances of oral/gut microbiota ( Fig. 4A-C). Interestingly, Veillonella in each sample type significantly correlated with at least one leukocyte count/ratio ( Fig. 4A-C). We also identified correlations between oral/gut microbiota and other clinical parameters, including anthropometric measurements, immunoglobulins, red blood cells, platelets, liver and kidney function, metabolic traits, and electrolytes (Fig. S2).
Taken together, these results revealed close correlations between oral/gut microbiota and BP/other clinical parameters, and suggested that microbiota could directly affect BP or influence BP by modulating blood pressure-related clinical parameters.

Communications between oral and gut microbiota in HTN
The importance of oral-gut microbial transmission has been recognized in different diseases [14,22]. We next extended our analysis to investigate the potential communications between oral and gut microbiota in HTN. Microbial abundance co-correlation networks illustrated that the HTN-enriched networks had more interconnections than no HTN-enriched ones in all 3 sample types (Fig. S3A).
Comparison of the microbiota among different types of samples using the Venn diagram identified 121 shared genera that coexisted in saliva, subgingival plaques, and feces (Fig. 5A). The relative abundance and prevalence of the top 30 shared genera showed reciprocal relationships between oral samples and gut samples (high in oral samples and low in gut samples, vice versa) (Fig. 5B). Genera such as Veillonella, Porphyromonas, Streptococcus and Prevotella that are usually found in the oral cavity were more prevalent or abundant in the gut of HTN than no HTN participants (Fig. 5B). The most pronounced change was the sharp increase of the relative abundance of Veillonella in the gut of HTN versus no HTN participants (Fig. 5B). These data together suggested that orally-originated genera such as Veillonella ectopically colonized in the gut and that HTN promoted such ectopic colonization.
We then analyzed the correlations of the top 30 shared genera between oral and gut sample types and paid special attention to Veillonella. The relative abundance of gut Veillonella was significantly correlated with salivary Gemmiger, Faecalibacterium, Porphyromonas, and Streptococcus (Fig. 5C). However, there was no significant correlation between gut Veillonella and subgingival microbiota (Fig. 5D). These results suggested that the potential ectopic colonization of Veillonella more likely involved the salivary rather than subgingival genus. SourceTracker analysis showed that HTN tended to increase the saliva-originated proportion while decrease the subgingival plaques-originated proportion in gut microbiota (Fig. S3B), indicating that the saliva-gut communication of Veillonella was associated with HTN. We further inspected the relationship between Veillonella and the above-mentioned 4 genera (Gemmiger, Faecalibacterium, Porphyromonas, and Streptococcus) within each sample type (Fig. 5E-G). The relationship between Streptococcus and Veillonella caught our close attention because only the correlations between these two genera within each sample type (saliva, subgingival plaques, and feces) were significant and positive (Fig. 5E-G). Previous reports have considered Streptococcus and Veillonella common colonizers in the gastrointestinal tract [23,24], and Streptococcus is able to promote the growth of Veillonella by producing lactic acid [25]. It was plausible that salivary Streptococcus and Veillonella co-colonized ectopically in the gut to affect BP. Therefore, these data suggested potential communications between oral and gut microbiota in HTN.
Analyses of the correlations of the top 30 shared genera between saliva and subgingival plaques unveiled 183 significant correlations, most strikingly among which the significant positive correlations between 10 genera in saliva and 9 genera in subgingival plaques clustered together, suggesting strong communications of microbiota between saliva and subgingival plaques as expected (Fig. S3C).

Oral-gut transmission of microbiota in HTN at species level
To further investigate the communications between oral and gut microbiota, we performed shotgun metagenomic sequencing of oral and gut samples from 24 no HTN participants and 36 HTN participants. Co-correlation networks of predominant species illustrated that the HTN-enriched networks had more interconnections than no HTN-enriched ones in all 3 sample types (Fig. S4). Linear discriminant analysis effect size (LEfSe) revealed differentially enriched microbial species between HTN and no HTN (Fig. S5A-C). Interestingly, PD-associated species (Porphyromonas gingivalis, Tannerella forsythia, Treponema denticola, and Porphyromonas endodontalis) were significantly enriched in subgingival plaques of HTN participants (Fig. S5B). LEfSe also revealed differentially enriched microbial functional modules between HTN and no HTN (Fig. S6A, B). The majority of the microbial gene functions in oral samples overlapped with those in gut samples, indicating shared functional modules between oral and gut microbiota (Fig. S6C). Multiple microbial functional modules significantly correlated with BP and other clinic parameters (Fig. S7).
We detected 461 species prevalent in oral samples (saliva and subgingival plaques combined), 1095 species prevalent in feces, and 213 prevalent in both oral and gut samples (Fig. 6A). Previous reports have suggested that tracking microbial communities at the level of strains rather than species is more reliable for establishing and quantifying microbiota transmission [26,27]. We therefore analyzed the SNVs in our metagenomic sequencing data. Based on Manhattan distance obtained from the SNV distance matrix of oral-gut sample pairs, 16 out of the 213 species shared in oral and gut samples were identified as oral-gut transmitters, 4 of which (Streptococcus equinus, Streptococcus parasanguinis, Veillonella parvula, and Veillonella atypica) were categorized as frequent transmitters and the other 12 as occasional transmitters according to their oral-gut transmission scores (Fig. 6B). These 16 oral-gut transmitters belonged to 5 genera (Veillonella, Streptococcus, Haemophilus, Prevotella, and Megasphaera) under 3 phyla (Firmcutes, Proteobacteria, and Bacteroidetes) and most of them were more abundant in oral samples (especially saliva) than feces (Fig. 6-C-F). Similar to the negative correlation between Streptococcus and Veillonella at the genus level (Fig. 5C-D), most of the associations between oral Streptococcus spp. and gut Veillonella spp. were negative correlations (Fig. 6C).
The impacts of HTN on the relative abundances of these 16 species in feces were largely consistent with those at the genus level, except for Megasphaera micronuciformis (and its belonging genus Megasphaera) (Fig. 6F, S8). Importantly, all 4 species under the genus of Veillonella (Veillonella parvula, Veillonella atypica, Veillonella dispar, and Veillonella sp. oral taxon 158) were increased in HTN compared to no HTN group (Fig. 6F). Therefore, the 16 oralgut transmitting species, particularly Veillonella spp, may play important roles in HTN.

Sustained oral-gut transmission of microbiota in HTN
Both oral and gut microbiota fluctuate over time [28,29]. We conducted a follow-up study in 52 HTN participants and 26 controls to address whether the alterations of microbiota in HTN and            the oral-gut transmission sustained. Different from the lower adiversity between HTN and no HTN group 6 months ago, 16S rRNA gene sequencing revealed that both salivary and subgingival microbiota of HTN subjects had higher Chao1 richness, Faith's phylogenetic diversity, and Shannon diversity index (Fig. 7A). These 3 indexes in gut microbiota and Pielou's Evenness index for all sample types maintained the similar trend as before (Fig. 7A). PCoA and permutational multivariate analysis of variance based on Bray-Curtis distance demonstrated a significant difference in bdiversity of gut but not oral microbiota between no HTN and HTN group (Fig. 7B, S9). The difference in diversity between these two time points suggested that microbiota in the gut had less fluctuation than in the oral cavity. SourceTracker analysis in the follow-up cohort suggested an increase of oral-gut communication in HTN, particularly manifested by the significantly increased proportion of salivaoriginated microbes in gut microbiota (Fig. 7C). The composition of the top 50 genera in each sample type was similar to that 6 months ago, although the significantly different genera between no HTN and HTN varied (Fig. 7D). Among the 5 genera that the 16 oral-gut transmitting species belong to, the relative abundance of Veillonella significantly increased in gut microbiota of HTN participants (Fig. 7E), consistent with the results 6 months ago (Fig. S8). Together these data indicated the stability of the oral-gut transmission and ectopic colonization of oral microbiota in HTN.

A causal relationship between oral-gut microbial transmission and HTN
To investigate the potential causal link between oral-gut transmitters and HTN, we performed saliva microbiota transplantation experiments in ABX-pretreated recipient mice infused with vehicle (saline) or Ang II infusion (Fig. 8A). Tail-cuff BP measurements revealed that mice received saliva from HTN participants (HTNsaliva) had significantly higher SBP and/or DBP than those received saliva from no HTN participants (no HTN-saliva) or water after Ang II infusion (Fig. 8B). Ang II-induced hypertrophy and fibrosis of aortas were significantly increased in mice received HTN-saliva compared to those received no HTN-saliva or water (Fig. 8C-D). Moreover, Ang II-induced expression of collagen-I in aortas and ANP and BNP in hearts was markedly higher in mice received HTN-saliva ( Fig. 8E-F). Mesenteric arteries isolated from mice received HTN-saliva manifested stronger contraction in response to phenylephrine and Ang II (Fig. S10). These results demonstrated that the ectopic colonization of HTN-saliva in mouse intestine exacerbated angiotensin II-induced HTN.
To validate the importance of the oral-gut transmitters identified in our clinical study, we performed 16S rRNA gene sequencing on human saliva and feces of mice received human saliva. PCoA demonstrated that the gut microbial community composition between mice received no HTN-saliva and those received HTNsaliva were completely separated but more similar after Ang II infusion (Fig. 8G). ABX pretreatment depleted all oral-gut transmitting genera (Veillonella, Streptococcus, Prevotella, Haemophilus, and Megasphaera) in mouse gut microbiota (Fig. S11). Although the human saliva samples contained all transmitting genera (4 genera in Fig. 8H, Megasphaera was not shown due to low abundance), only Veillonella successfully colonized and presented in the intestine of all saliva-treated mice (Streptococcus was only detected in two mice, Fig. 8H). Importantly, Veillonella was more enriched (Fig. 8I) and more abundant in the gut microbiota of mice received HTN-saliva and infused with Ang II (Fig. 8J). Furthermore, the importance of Veillonella ranked the highest according to random forest regression analysis of gut microbiota of mice received human saliva and infused with Ang II (Fig. 8K).
Taken together, the results of these animal experiments are largely consistent with those of our clinical study. Mice transplanted with HTN-saliva had significantly elevated BP. A higher abundance of saliva-derived Veillonella in mouse gut may be a causal link between oral-gut microbial transmission and HTN.

Discussion
PD is an oral microbiota-related disease and has been linked to HTN [19,30,31]. However, few studies have focused on direct relationships between oral microbiota and HTN or oral-gut microbiota transmission in HTN. Herein, we for the first time reported a comprehensive analysis of oral and gut microbiota between no HTN and HTN by 16S rRNA gene sequencing and metagenomic sequencing. Besides substantiating the association between PD and HTN, our data revealed the differences in diversity and composition of oral and gut microbiota between no HTN and HTN, as well as established correlations between oral/gut microbiota and clinical parameters. We identified oral-gut microbial transmitters at both genus and species levels. Most notably, Veillonella spp., among 16 oral-gut transmitting species, may exert crucial functions in HTN.
Our results have strengthened the association between PD and HTN. According to a recent systematic review and meta-analysis, patients with PD exhibit 4.5 and 2.0 mmHg higher SBP and DBP respectively than those without PD [32]. Our results confirmed such association between PD and BP in both hypertensive and normotensive participants. The differences in BP between no PD and PD in this study were greater than those reported before [19], likely because of the difference in study populations. More importantly, our data for the first time demonstrated that participants with PD had higher BP not only in cross-sectional but also in followup studies, suggesting that PD might affect both onset and progression of HTN. We also observed a trend of decrease in BP for both PD and no PD participants during the follow-up, likely because of change in seasons (from winter to summer in this case) and/or ease of anxiety of the participants over time [33,34]. Our study and previous studies together strongly support that PD is an important risk factor for HTN and needs to be monitored continuously.
We have provided comprehensive evidence to support strong associations between oral microbiota and HTN. We observed a host of oral genera that were significantly altered in HTN and/or significantly correlated with BP. Among these genera, the relative abundances of oral Streptococcus and Prevotella were significantly decreased in HTN participants, consistent with the results of a previous study using subgingival plaques from older women [8]. It has been reported that Veillonella is an important nitrate-reducing genus in the oral cavity linked to lower BP [35]. Interestingly, diet- The relative abundance of Veillonella in mouse gut microbiota. K. Random forest regression analysis of mouse gut microbiota at genus level. The heatmap represents relative abundances of microbes. n = 3:3:3 for H2O + Vehicle vs no HTN-saliva + Vehicle vs HTN-saliva + Vehicle; n = 5:5:4 for H2O + Ang II vs no HTN-saliva + Ang II vs HTNsaliva + Ang II. ## p < 0.01 for no HTN-saliva + Ang II vs HTN-saliva + Ang II at day 35 in (B); *p < 0.05, **p < 0.01, *** p < 0.001. ary nitrates have beneficial blood pressure-lowering effects, and oral microbiome plays an important role in that the commensal bacteria are required to convert nitrate to nitrite, the latter of which may be further converted to nitric oxide to lower blood pressure [36,37]. Consistently, the relative abundance of oral Veillonella was markedly decreased in HTN and negatively correlated with BP in our study, although other nitrate-reducers such as Neisseria, Haemophilus and Rothia were increased in HTN. The relative abundances of oral Burkholderia, Lautropia, and Ralstonia were significantly increased in HTN participants of our study. Interestingly, these 3 genera have been shown to correlate with pulmonary arterial HTN [38,39]. Moreover, we established new associations between HTN and oral microbes, including Actinobacillus, Aggregatibacter, Atopobium, Bulleidia, Cupriavidus, Desulfomicrobium, Eikenella, Euzebya, Kingella, Moraxella, Olsenella, Pasteurella, Pelomonas, and Selenomonas. Our results also revealed that PDassociated pathogens such as Porphyromonas, Fusobacterium and Treponema were more abundant in subgingival plaques of HTN participants, and positively correlated with IL-6 and/or CRP. These 3 PD-associated pathogens and Prevotella have been reported to play a pro-inflammatory role, which may be mediated by the microbial toxins such as gingipain derived from Porphyromonas gingivalis and lipopolysaccharide derived from Prevotella intermedia [40][41][42][43]. These associations further substantiated the importance of oral microbiota in HTN.
The most novel finding of this study is the identification of important oral-gut transmitting microbes in HTN. Our overall findings of gut microbiota in HTN participants were consistent with those previously reported (e.g., similar a-diversity and genus enrichment) [20]. We further explored the influence of oral-gut transmitting microbes and discovered oral-gut microbial transmitters that might participate in BP regulation. SourceTracker analyses in both cross-sectional and follow-up cohorts suggested increases of transmission between oral and gut microbiota in HTN participants. Correlation analyses at genus level indicated communications between oral and gut microbiota, and orally-originated genera such as Veillonella and Streptococcus might co-colonize ectopically in the gut to affect BP. Further analysis at species level identified 16 oral-gut transmitting species under 5 genera (Veillonella, Streptococcus, Haemophilus, Prevotella, and Megasphaera) that may play important roles in HTN. It is worth to point out that all these 5 genera are the core oral taxa with high relative abundances [44]. Intriguingly, except for and Megasphaera, the rest 4 genera fall into the oral-gut microbial transmission category as reported before [9]. Given the pivotal role of gut microbiome on sympathetic nerves, oral-derived gut bacteria could also influence BP via the brain-gut axis after colonization [45].
Veillonella has earned an eminent place among the oral-gut microbial transmitters in HTN. Relative abundances of fecal Veillonella significantly increased in HTN participants of both crosssectional and follow-up cohorts. Among the 16 oral-gut transmitting species, all 4 Veillonella species were enriched in HTN participants. Ectopic colonization of oral microbiota in the intestine is usually considered a hallmark of diseases [12,46,47]. Orallyderived gut microbes including Veillonella have been used to establish a diagnostic model to distinguish and predict colorectal cancer with a high efficacy [48]. Veillonella is also a potential pathobiont expanded in autoimmune hepatitis and associated with disease status [49]. In addition, oral microbiota such as Veillonella atypical and Veillonella dispar may translocate to the gut of patients with schizophrenia [50]. Our results confirmed that Veillonella was a strong oral-gut transmitter. Particularly, the colonization of saliva-derived Veillonella in the gut of mice transplanted with human saliva suggested that the oral cavity was a reservoir for the ectopic colonization of Veillonella. Saliva-derived Veillonella was more enriched in Ang II-infused mice, indicating that it was more likely to colonize under HTN status. More importantly, higher abundance of Veillonella colonization was accompanied by significantly higher BP in mice transplanted with saliva from HTN participants, suggesting that the ectopic colonization of oral Veillonella in the gut was an important factor influencing the development and progression of HTN. Our study used male mice only, which was a limitation given the sex differences in hypertensive animal models [51].
Taken all together, this study has reinforced the association between PD and HTN, established strong correlations between oral and gut microbiota, between oral/gut microbiota and HTN/HTNassociated clinical parameters, as well as identified HTN-related oral-gut transmitting microbes. These data have comprehensively demonstrated the roles of PD and oral microbiota in HTN, particularly revealing the importance of oral-gut transmission of microbes such as Veillonella spp. These findings support joint control of PD and HTN, and the identification of orally-derived microbes in the gut may provide novel strategies for prevention, diagnosis and therapy of HTN.

Conclusions
PD is constantly associated with HTN. Dysbiosis of oral and gut microbiota highlights the association between PD and HTN. Oralgut transmission of microbes, particularly Veillonella spp., is an important mechanism contributing to HTN. Regular monitoring PD and targeting oral-gut microbial transmission may become effective strategies to improve the prevention and treatment of HTN.

Compliance with ethics requirement
The study protocol was approved by the Institutional Review and Ethics Board of Shanghai Ninth People's Hospital, Shanghai Jiao Tong University School of Medicine (SH9H-2018-T66-3). Informed consent was signed by all subjects before enrollment.