Periodontal disease severity is associated to pathogenic consortia comprising putative and candidate periodontal pathogens

Abstract Based on a holistic concept of polymicrobial etiology, we have hypothesized that putative and candidate periodontal pathogens are more frequently detected in consortia than alone in advanced forms of periodontal diseases (PD). Objective To correlate specific consortia of periodontal pathogens with clinical periodontal status and severity of periodontitis. Methodology Subgingival biofilm was obtained from individuals with periodontal health (113, PH), gingivitis (91, G), and periodontitis (209, P). Genomic DNA was purified and the species Aggregatibacter actinomycetemcomitans (Aa), Aa JP2-like strain, Porphyromonas gingivalis (Pg), Dialister pneumosintes (Dp), and Filifactor alocis (Fa) were detected by PCR. Configural frequency and logistic regression analyses were performed to correlate microbial consortia and PD. Results Aa + Pg in the presence of Dp (phi=0.240; χ2=11.9, p<0.01), as well as Aa JP2 + Dp + Fa (phi=0.186, χ2=4.6, p<0.05) were significantly more associated in advanced stages of P. The consortium Aa + Fa + Dp was strongly associated with deep pocketing and inflammation (p<0.001). The best predictors of disease severity (80% accuracy) included older age (OR 1.11 [95% CI 1.07 – 1.15], p<0.001), Black/African-American ancestry (OR 1.89 [95% CI 1.19 – 2.99], p=0.007), and high frequency of Aa + Pg + Dp (OR 3.04 [95% CI 1.49 – 6.22], p=0.002). Conclusion Specific microbial consortia of putative and novel periodontal pathogens, associated with demographic parameters, correlate with severe periodontitis, supporting the multifactorial nature of PD.


Introduction
Periodontal disease (PD) is a biofilm-driven chronic inflammatory condition that eventually leads to the destruction of the supporting periodontal tissues. Such disease results from an imbalance of the dynamic interaction between the microbiota and the host response, modulated by environmental, systemic, and genetic factors. 1,2 Shifts in the oral microbiome are implicated in the onset and/or progression of PDs, 3,4 and although over 400 species colonize the subgingival biofilm, 5,6 only a fraction of them has been implicated in destructive diseases. [7][8][9] Thus, PDs are not the "one pathogen, one infection" type of disease, since they are associated with an unbalanced consortia of members of the normal microbiota, including pathobionts and commensals. 2,3,10 Advances in biotechnology for more rapid and accurate identification of microorganisms in oral specimens have provided insights about the role of specific bacteria within microbial complexes in the etiology and pathogenesis of severe form of periodontitis. 11 Although Aggregatibacter actinomycetemcomitans (Aa) and Porphyromonas gingivalis are the main cultivable putative periodontal pathogens recognized worldwide, 12,13 novel taxa are currently being considered as potential periodontal pathobionts. [14][15][16] For instance, evidence suggests strong associations between Filifactor alocis, Dialister pneumosintes, and even uncultivated taxa of Tannerella and Saccharibacteria with advanced forms of periodontitis and peri-implantitis. [15][16][17][18][19][20] The potential mechanisms of interactions among members of the oral microbiota and between these communities and the periodontal environment are endless and not fully understood. As previously described for the microbial complexes of the subgingival plaque, 8

Subgingival biofilm sampling
In patients with PH, subgingival biofilm samples Detection of putative and potential periodontal pathogens PCR amplification of target sequences was performed using two sets of primers to each bacterial species and one set of universal primers for the 16S rRNA gene (Supplementary Table 1 conducted to determine the best microbial consortia and non-microbial parameters as predictors of periodontitis and/or disease severity. The stepwise Wald method was selected and ORs with 95% CI were reported. The significance level determined was 5%. All analyses were performed using a commercial software package (IBM SPSS ® Statistics, IBM Brasil, SP, Brazil).

Results
Clinical and socio-demographic features of the study population Table 1 and Figure 1 show the characteristics of the

Frequency of detection of putative and candidate periodontal pathogens
The putative periodontal pathogens Aa and P.
gingivalis and the species D. pneumosintes were significantly more prevalent in periodontal diseases than health, whereas F. alocis predominated in G group (p<0.05, χ 2 test; Figure 2A). Among P patients, however, F. alocis was significantly more increased in advanced stages of disease (p<0.05, χ 2 test; Figure   2B). The highly leukotoxic strain Aa JP2 was detected in only six patients, one with G and five with P stages III/IV. Due to differences in socio-demographic features among clinical groups, associations between periodontal bacteria and these parameters were explored for further adjustments (Supplementary Symbols refer to significant differences among groups (*Kruskal-Wallis and †χ 2 tests, p<0.05). χ 2 test). Nevertheless, when we remove individuals with low income or schooling level from the analysis, the prevalence of these species in the clinical groups present a similar distribution observed in the whole sample (data not shown).

Figure 1-Clinical data of the study population (n=413).
Box plots represent the distribution in quartiles of mean % of sites with probing pocket depth (PPD) of 5-6 mm; PPD >6 mm; clinical attachment level (CAL) of 4-5 mm; CAL >5 mm; visible supragingival plaque (PL), marginal gingival bleeding (GI), bleeding on probing (BOP) and calculus (CA). Whiskers indicate minimum and maximum values; circles indicate outliers. All parameters differed significantly among groups (Kruskal-Wallis test, p<0.05). *Refers to significant differences between periodontal health and gingivitis; #between gingivitis and periodontitis and $between periodontal health and periodontitis groups (Mann-Whitney test, p<0.05).

Associations among periodontal species within consortia
Since only a few patients presented simultaneously >3 of the species evaluated in the subgingival biofilm sample, analyses were limited to consortia composed by two or three different species. Associations among species were first examined by comparing the expected and observed frequencies of microorganisms in combination or as single isolates in each clinical condition. For the PH group, we failed to observe significant differences in the frequency of detection between single and combined species. However, the strongest associations of microbial pairs in P patients were seen for Aa + F. alocis (phi=0.498; χ 2 =51.3), Aa + P. gingivalis (phi=0.319; χ 2 =21.2) and Aa + D. pneumosintes (phi=0.290; χ 2 =17.6; p<0.001).
Likewise, Aa JP2 clone + D. pneumosintes in the presence of F. alocis was modestly correlated in stages III and IV of P, but not in initial stages of the disease (phi=0.186, χ 2 =4.6; p<0.05; data not shown).

Relation between consortia and clinical parameters
The consortia of bacteria were associated with    As expected, Aa, P. gingivalis, and D. pneumosintes were more prevalent in individuals with periodontitis than in healthy patients. 8,12,15,[18][19][20][25][26][27] However, the high frequency of F. alocis in G group compared to PH and P groups was surprising. Such species is highly prevalent in periodontitis, peri-implantitis, and endodontic infections, and its high values in our G patients compared to P disagrees with data reported by other authors. 9 pneumosintes and T. forsythia, but not between D. pneumosintes and P. gingivalis in subgingival samples of P patients. Similar to our findings, these species increased with patients' age. Sha and Chen 38 (2020) studied biofilm formation using various strains of Aa in co-culture with P. gingivalis and D. pneumosintes.

Discussion
They reported a synergistic relation between Aa + D.
pneumosintes for most of the Aa strains evaluated, but failed to find significant differences between the amounts of single-species and dual-species Aa + P. gingivalis biofilms. In another biofilm study, 39 a P.
gingivalis strain with high biofilm-formation ability and proteolytic activity, presented a competitive advantage against Aa during biofilm formation.
Although D. pneumosintes has been considered a relevant pathogen in oral infections for over 20 years, information on its putative virulence factors related to the pathogenesis of periodontitis is scarce. 40 A recent systematic review suggested a relation of Dialister spp.
with several types of cancer. 41 One of their arguments is that metabolic end products of Dialister spp., such as acetate and lactate, may play a role in carcinogenesis. Yost,et al. 42 (2018)  virulence during PD onset and progression may be driven at the strain or clone levels, and not limited to species. 38,39,43,44 Our PCR method was not designed to identify specific strains or clones of these oral species except for Aa JP2.
An interesting finding was the potential antagonism found in D. pneumosintes and P. gingivalis or F. alocis in biofilm samples from G but not from P patients. In the non-destructive inflammatory state of gingivitis, D. pneumosintes seems to compete with either P. gingivalis or F. alocis, since they are more frequently Despite the significant correlations between specific consortia and disease severity, our data must be considered within its methodology limitations.
Although many individuals were analyzed, we targeted only four periodontal pathogens among several species that compose the subgingival biofilm. Moreover, our aim was not to evaluate the levels, viability, or virulence of these species and strains. Therefore, other pathogenic consortia correlated to disease severity may predominate in distinct populations presenting PD. Studying the mechanisms that occur in biofilm shifts from a healthy state to PD is not an easy task.
Several distinct in vitro and in vivo approaches are needed to investigate the interactions among a wide range of species, even novel taxa not yet recognized.
Nevertheless, our data reinforce the potential role of specific microbial communities in PD, indicating that the consortium comprising the putative pathogens Aa and P. gingivalis, and the candidate pathogen D. pneumosintes, associated with older age and African ancestry, increases the likelihood of severe periodontitis. Under the clinical point of view, this information may contribute to discriminate individuals with greater risk of having severe disease, and to develop a more personalized approach focusing on arresting disease onset and progression by targeting these key consortia.