Assessment of the role of Trichomonas tenax in the etiopathogenesis of human periodontitis: A systematic review

Objective This systematic review was to assess the presence of Trichomonas tenax in patients with periodontitis and to elucidate its potential role in the onset and development of this disease. Method Systematic review was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines and by consulting the five databases: Medline, Science Direct, Web of Science, Dentistry and Oral Science Sources and Cochrane Central Register of Controlled Trials. Following Koch’s postulates revisited by Socransky as PICO framework, this collection data was only including full text of clinical trials concerning patients with periodontitis, case-reports and in vitro research published between 1960 and March 2019. Results On the 376 studies identified, only 25 fulfilled our eligible criteria. Most of these studies were in vitro research articles designed to evaluate potential virulence factors, and others were clinical trials (case-control studies, randomized controlled trial) and case-reports. The analysis of these papers has shown that i) Trichomonas tenax is more frequently detected in dental biofilm from sites with periodontitis than in healthy sites; ii) this live flagellate seems capable of producing diverse enzymes that could participate in periodontal breakdown and has the capacity to adhere to epithelial cells, its lysed form could induce the synthesis of IL-8 from macrophage cell lines; iii) the impact of non-surgical treatment of periodontitis have not been thoroughly evaluated on the presence of T. tenax Conclusions This systematic review has reported the presence of T. tenax more frequently in diseased than healthy sites and the capacity of this flagellate to synthesis enzymes which could participate to the degradation of periodontal tissues. Nevertheless, these data do not meet all the postulates and are not enough to provide firm conclusions about the role of T. tenax in the etiopathogenesis of periodontitis.


Introduction
The global burden of oral disorders and their associated sequelae have been assessed at various dates (1990,2005 and 2010) by Marcenes et al. [1]. These authors concluded that DALYS (Disability-Adjusted Life Years) induced by oral conditions had increased by 20.8% between 1990 and 2010 due to population growth, the ageing process and severe periodontitis. Severe periodontitis is the sixth most prevalent disease affecting millions of people worldwide and because of its dental and systemic impacts should therefore be a public health priority [2].
Periodontitis is recognized as an inflammatory disease, mainly induced by pathobionts such as Porphyromonas gingivalis in synergy with other putative pathogens present in a complex and diversified biofilm. These periodontopathogens cause an impaired immune response leading to the loss of periodontal tissue [3]. But neither immune deficiencies nor the virulence of periodontal pathogens can explain why, despite the identification of periodontal pathogenic bacteria in the saliva, only a few teeth are severely affected by periodontitis in patients, or why one tooth presented alveolar bone and soft tissue destruction while the periodontium of its neighbour was barely affected.
In order to explain these clinical observations, but also the failure in stabilizing periodontitis in certain patients, the authors attempted to explore other potential etiologies of periodontal diseases. The oral cavity harbours a rich and complex biofilm that includes bacteria, but also viruses, fungi and protozoans [4]. All these microorganisms could be implicated directly or indirectly in the development of periodontitis. A systematic review with meta-analysis evaluating the microbiota from patients with periodontitis has shown a significant presence of some herpesviruses and other authors have identified fungi (Candida genus), both of which could have a potential pathological role in this disease [5,6]. Moreover, two main protozoans have been identified in the oral cavity: Entamoeba gingivalis belonging to the rhizopod class and Trichomonas tenax to the zoomastigophora one. As regards Trichomonas tenax, some authors suggested its implication in periodontal diseases [7,8]. A recent review concerning T. tenax and periodontal diseases has been published in which the authors concluded that there was a correlation between the presence of the flagellate and the periodontal diseases and that this protozoan was involved in inflammatory process of gingivitis [9]. Nevertheless, this review did not follow the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines and the PICO (Patient Intervention Comparison Outcome) framework as recommended by the Centre for Evidence-Based Medicine [10,11]. In this review, 47 clinical studies about prevalence of the protozoan in periodontal diseases were not classified according to population (adults, children), presence of systemic disease, previous antibiotic consumption, and the keywords used for the literature research on periodontal diseases assessment were not detailed.
Therefore, we propose here the first systematic review of literature to investigate the presence of T. tenax in periodontitis in adult patients and to assess its potential role in the etiopathogenesis of periodontitis, according to Koch's postulates revisited by Socransky [12]: (i) postulate "Association" (elevated level of microorganism in lesions of periodontitis), (ii) postulate "Elimination" (microorganism elimination resulted in successful therapy), (iii) postulate "Host response" (microorganism induces immune response: elevated serum and local antibodies, cytokines production), (iv) postulate "Virulence factors" (microorganism produces enzymes inducing a periodontal degradation) and (v) postulate "Animal studies" (microorganism induces disease in gnotobiotic animal).
Firstly, we investigate through literature data the presence of T. tenax in diseased sites and its absence in healthy sites in adult patients with periodontitis, secondly the impact of non-surgical periodontal treatment on the decrease or elimination of this flagellate in healed sites and thirdly the potential virulence factors of this protozoan which could cause the destruction of periodontal tissue.

Protocol and study design
This systematic review was elaborated by following the PRISMA guidelines [10] (S1 Table).

Eligibility criteria
In this review, the focused questions addressed using the PICO framework, aimed to evaluate the potential pathogenicity of Trichomonas tenax in the development of periodontitis according to Koch's postulates revisited by Socransky [12]: 1. Is T. tenax (I) identified in high numbers in sites with periodontitis (P) and in lesser numbers or not at all in healthy sites? (postulate "Association") 2. Does the non-surgical treatment of periodontitis induce the decrease or the absence of T. tenax in the healed periodontal pocket and/or periodontal pocket with depth reduction? (C) (postulate "Elimination") 3. Does T. tenax produce virulence factors (adhesion capacity and tissue invasion, immunomodulating molecules, toxicity) and can it induce similar periodontal diseases in an appropriate animal model? (O) (postulates "Virulence factors", "Host response" and "Animal studies").

Study selection
Only full text articles published in the English, French, Italian and Spanish/Portuguese languages were considered.
All studies relating to Trichomonas tenax were considered eligible for inclusion if they were original articles in peer-reviewed journals conducted in humans and were: • Cross sectional, case-control studies, controlled trials • Case reports or case series (accepted for research pathogenicity factors).
• In vitro studies with evaluation of potential virulence factors and animal models of periodontitis.
All reviews of literature were excluded from the selected articles because they give redundant informations and are not detailed enough.
Inclusion criteria. Studies that were selected for this analysis include the following criteria: • Adult population (� 18 years) • Patients diagnosed with periodontitis using one or more of the following terms: pocket > 3 mm and/or CAL > 2 mm (Clinical Attachment Level), bone loss, grade of the periodontal tissue degradation (moderate or severe periodontitis, Armitage classification, 1999) [13].
• Patients diagnosed with periodontitis with no surgical treatment in the last 3 months (to evaluate the prevalence of T. tenax in periodontal pockets) • Patients diagnosed with periodontitis treated by non-surgical treatment (to evaluate the impact of the therapy on protozoan in the healed sites).
• Studies with collection of dental biofilm (supra and/or subgingival) • In vitro studies with detailed protocol Exclusion criteria. Articles were excluded from the selection because of the presence of: • In vitro studies with other trichomonads species than T. tenax.

Data selection process and elements
Data were collected independently by two authors (CB, SMD) and definitively included by these two authors, and all disagreements were resolved after discussion and/or with the intervention of a third author (MM). Reasons of exclusion of the articles were fully detailed in S1 Table, and quality appraisal of included clinical studies was described in S2 Table. If the study was published twice, the article presenting the most complete data set was selected.
All publications meeting the eligibility criteria were included, processed for data extraction and categorized according to the PICO framework: Population: adult patients with periodontitis Intervention: presence of trichomonads in sites with periodontitis Comparison: absence or reduction of the flagellates in diseased sites after non-surgical treatment Outcome: potential virulence factors of the protozoan (adhesion, tissue invasion, cytokines, antibody, toxicity).

Risk of bias in individual studies and across studies
Risk of bias was assessed through the Newcastle-Ottawa Scale (NOS) for case-control studies and modified NOS for cross-sectional studies [14]. The following criteria were assessed: • Representativeness of the exposed cohort

Study selection
As shown in Fig 1, a total of 376 articles were identified from the previously described electronic database search. After duplicate removal and language exclusion, a total of 156 articles were obtained. Finally, after the removal of non-full text and non-relevant papers, 25 articles were deemed appropriate to be included. The main reasons for exclusion were: the absence of periodontal status, lack of precise periodontal parameters, the identification of trichomonads in animal studies but not in human studies, the studies included patients with medication modifying microbiota, the diagnosis of periodontitis as a manifestation of systemic disease, other trichomonads species than T. tenax, the absence of a T. tenax factor of virulence. Some clinical studies were discarded because essential details were missing: 1. One study included patients with periodontitis as well as patients with periodontitis and diabetes and did not specify the proportion of patients T. tenax-positive among those with diabetes.
2. one study includes both children and adults but did not specify the T. tenax positivity according to the population age.
3. In other studies, some included patients had already been treated with antibiotics, without anyone knowing the proportion of patients T. tenax positive among those receiving or not antibiotics.
4. Finally, in some papers the origin of the sample was unknown such as "rinsing sample" and/or samples from saliva, mucosa, decay.

Characteristics of included studies
• Presence of T. tenax in periodontitis or gingivitis sites and in healthy sites from controls ( Table 1). Three case-control studies were selected to evaluate the presence of T. tenax in periodontitis ( Table 1). Two of them have used the microscope to detect the flagellates, and only one the molecular biology tools. Whatever the technique used, the authors showed a higher frequency of this flagellate in patients with periodontitis (from 42.5 to 32.3%) than those with gingivitis (from 32.3% to 0%). The controls with healthy periodontium had positivity to T. tenax from 28.2% to 7.1%. The frequency of the flagellate in the control group doubled depending on authors, but its frequency in periodontitis sites was similar [7,8,15].
• Impact of non-surgical treatment of flagellate in healed sites (Table 1). Only one randomized controlled trial evaluated the impact of scaling and root planing (SRP) on the frequency of T. tenax in periodontitis sites [16]. The authors selected patients with moderate to severe periodontitis (CAL > 5 mm) and randomly assigned them to the experimental and control  group. Before receiving an oral hygiene instruction, a sample of dental biofilm was collected from the patients in each group. Three weeks after SRP, a second biofilm sample was collected from patients in the experimental group. The authors concluded that the SRP could reduce the number of T. tenax in saliva but not in the dental plaque of patients with moderate to severe periodontitis. No detailed description of PPD (Periodontal Pocket Depth) and CAL before and after SRP was reported.
• Potential virulence factors from protozoans ( Table 2). Authors have detected the capacity of T. tenax ATCC 30207 to adhere to diverse cells and induce cytotoxicity [17], while others studies showed no adherence or cytotoxic property [18][19]. Moreover, flagellates produce numerous enzymes, such as cysteine protease, cathepsin B like protein, haemolysins, metalloproteinase and other proteinases that could participate in the breakdown of collagen, gelatin observed in periodontal diseases [20][21][22][23][24][25]. Live trophozoites of the Hs-4 strain of T. tenax failed to induce secretion of IL-1β, IL-8, IL-10 or TNFα from macrophages. Lysed trophozoites stimulated synthesis of IL-8 from phagocytic cells [26]. The intravenous injection of T. tenax antigen and/or live flagellates in the rabbit elicits the production of antibodies [27,28]. The observation by transmission electron microscopy revealed the presence of bacteria inside the T. tenax cytoplasm: some were dividing while others were dead.
• Research on the potential pathogenicity and invasive power of T. tenax in extra-oral location (Table 3). Flagellates have been isolated from sputum, pleural sample, bronchoalveolar lavage, cervical lymph nodes, vaginal swabs, pharyngeal samples, urines and subhepatic abscess pus [29][30][31][32][33][34][35]. Generally, the identification was a microscope observation with staining or PCR. The authors concluded an aspiration of the parasite from the oral cavity to the lung location, and autoinoculation via hands or through oral sex for vaginal and urine identification. However, in two case reports, T. tenax could have migrated via the lymphatic vessels to the cervical lymph node and via bloodstream to subhepatic abscess. Some authors suggested a potential pathogenic role and/or flagellate co-infection agent in various infections. But most of these publications are case reports presenting infectious disease in immunocompromised patients.

Risk of bias and quality assessment of selected studies and across studies
• Regarding the case-control studies (Table 1). For some studies, the absence of detailed protocol of collection of the microbiota makes it difficult to draw any conclusions on the localization of the periodontal colonisation of this protozoan: in sub-gingival and/or supragingival biofilm? Even if no study has selected an appropriate sample size of patients to provide sufficient statistical power to detect the flagellate prevalence in periodontitis, studies showing its presence in periodontal pocket could not be discarded. Two of three case-control studies used the microscope to detect T. tenax which may be biased because its phenotypic characteristics are quite similar to those of T. vaginalis and T. hominis, but these species are very rarely described in the mouth [32]. Quality assessment of clinical studies is detailed in S2 Table. • Regarding the randomized controlled trials (Table 1). No periodontal parameters have been established and measurable before and after SRP, periodontal healing cannot be evaluated according to the protozoan decrease. The authors only specified the degree of periodontal degradation with CAL parameter but not the infected periodontal status of the patients (no PPD data). The researcher did not discuss the credibility of their findings with the   healing or not of sites with periodontitis, and only reported differences in the frequency of flagellate detection before and after the treatment in both saliva and biofilm sample. S2 Table  shows the details of the quality assessment.
• As regards the laboratory studies (Table 2). Numerous publications have explored the potential pathogenicity of T. tenax. However, the authors generally used an ATCC strain that could not fully reflect the virulence of clinical isolates and used cell lines for cytotoxicity testing and cytokine production. Certainly, cell lines can react differently from primary cell lines that constitute the cellular model with the closest similarity to the patient cells. Moreover, the methodological heterogeneity of protocols, the few T. tenax strains available and the xenic culture medium used in some studies make it difficult to evaluate results. No animal model of periodontitis has been found with T. tenax.
Medium used in some studies make it difficult to evaluate results. No animal model of periodontitis has been found with T. tenax.
• As regards the extra-oral location of T. tenax to research the pathogenic potential and invasive power (Table 3). Some studies identified flagellates under the microscope, which is not the most appropriate technique to confirm the trichomonads species. The authors postulated the inhalation of flagellates from the oral cavity to the respiratory tract without any dental sample analysis [30,31,33]. The majority of publications were case reports or case report series; further investigations are required to conclude of any T. tenax pathogenicity in an extra-oral location.

Discussion
This systematic review has evaluated the role of T. tenax in the etiopathogenesis of periodontitis using Koch's postulates revisited by Socransky as PICO framework. The principle findings were discussed for each postulate and concerned in particular two of them: "Association" and "Virulence factors". Among the various postulates and causal models that could provide evidence of pathogenicity for a putative periodontal pathogen, Koch's postulates revisited by Socransky is the most well-known and used. Unfortunately, the implementation of these postulates is limited to cultivable microbiota and has never be adapted to uncultivable microorganisms identified by molecular biology that could also be suspected of initiating or participating in the pathogenesis of the disease.
These Koch's postulates must be adapted to new technologies and could be based on the detection of nucleic acid sequence as suggested by Fredricks and Relman [36]. But these last postulates cannot be applied to this literature review, because in most of the clinical studies the authors identified T. tenax using microscopy.
It is increasingly clear that periodontitis is the result of the synergy of microorganisms embedded in a biofilm [37]. But among all the detected microorganisms the Socransky modified postulates remains a way to discriminate pathogenic bacteria in periodontitis. In this review, these postulates are applied to protozoan to compare their virulence factors to those of well-known periodontopathogens.

The prevalence of Trichomonas tenax in periodontitis
Periodontitis is an inflammatory disease resulting from dysbiotic microbial communities of pathobiont and keystone pathogens, which develop virulence factors inducing periodontal tissue degradation. The dysbiosis of periodontal microbiota results in the breakdown of the bacterial balance in favour of periodontopathogens and to the detriment of commensal bacteria [38]. The pathogens are present in relative low abundance in the healthy sites and their quantity is proportional to the PPD in diseased sites.
The real prevalence of T. tenax in periodontitis could not be established for two reasons: (i) the patients sample size seems to be insufficient to achieve a statistical significance, (ii) the periodontal parameters were not detailed enough. Nevertheless, this protozoan has been found in healthy sites as well as sites with gingivitis and periodontitis, from 7.1 to 28.2%, 0 to 32.3% and 32.3 to 42.5% respectively. No indication on the quantity of flagellates in the various sites is known. The implementation of the dysbiosis concept for T. tenax cannot be totally excluded as the flagellate is detected in both pathological and healthy sites as indicated for periodontal pathogens [7,8,15,39]. The identification method (direct observation or observation of the culture after the incubation time) and the heterogeneity of the sampling protocol make it difficult to conclude on the specific identification of the protozoan in periodontitis due to the morphological similarities with T. hominis and T. vaginalis. Nevertheless, one case-control study used the PCR assay method and confirmed the presence of trichomonad species in samples from both patients with or without periodontitis. Some clinical studies have identified the flagellate with PCR, but they could not be included in this review because authors have selected patients with periodontal diseases as manifestation of systemic disease and/or patients without any description of periodontium [40,41,32].

The postulate "Elimination"
The postulate "Elimination" evaluates the impact of non-surgical periodontal therapy on the decrease or elimination of a putative pathogen in a healed site. Only one randomly controlled trial evaluates the impact of scaling and root planing (SRP) on the frequency of protozoan identification in patients with periodontitis (CAL > 5 mm) [16]. Surprisingly, the SRP treatment induced a decrease of T. tenax in saliva but not in dental plaque. Some authors found a high frequency of the flagellate in patients with poor hygiene [42].
One might expect that improved hygiene and a debridement of the periodontal pocket would reduce the number of flagellates, as demonstrated for bacteria [43]. Only one periodontopathogen was not eliminated after SRP: Aggregatibacter actinomycetemcomitans [44]. This pathobiont presents numerous virulence factors and a tissue and cell invasiveness capacity that could explain "its resistance" to non-surgical treatment, but these properties have not been described for flagellates. In the absence of periodontal parameters description before and after treatment and in the absence of dental sample location (subgingival and/or supragingival), it is difficult to draw solid conclusions from this study.

The postulate "Virulence factors"
Numerous authors have evaluated the factors that may respond to the postulate "Virulence factors" in relation to T. tenax. Cellular adhesion is the first step in tissue colonisation for any microorganism. The adhesion and colonisation capacity has clearly been described for T. vaginalis but not for T. tenax [18,19]. In 1983, Ribaux et al. described fibronectin-like protein located on plasma membrane and protozoan axostyle, which could potentially have a role in T. tenax adhesion to eukaryotic cells and in the phagocytosis of bacteria. These authors demonstrated a particularly intense labelling localised in the contact zone between bacteria and two strains of T. tenax by indirect immunostaining, one isolated from a patient and the other from a laboratory strain, cultured in xenic medium [45]. This aspect was confirmed by Ribeiro et al. who demonstrated the ability of T. tenax ATCC 30207 to adhere to both cell lines and primary cells (patient epithelial cells) [17].
The explanation of the inconsistencies between the Alderete and Ribeiro adhesion results could be attributed to the different times of protozoan cell interaction: thirty minutes (no contact between the flagellates and cells) and 2 hours (visible interaction between the flagellates and the epithelial cells) respectively [17][18][19].
Moreover, Ribeiro has also shown the ability of T. tenax to induce cytotoxic effects such as damage and disruption of cell monolayers. The oral protozoan appears to phagocytose membrane portions of MDCK (Mardin-Darby Canine Kidney) epithelial cells and induces membrane blebbing and apoptotic bodies in HeLa cells. The epithelial tissue is the first barrier protecting the periodontium. The potential capacity of T. tenax to induce damage in cell monolayers could promote and facilitate its invasion into the deepest connective tissue. This hypothesis should be confirmed in animal models and with the analysis of gingival tissue of patients with periodontitis. This cytotoxicity would prove the pathogenic role of T. tenax and its potential participation in the cell breakdown described in periodontal diseases [46]. Likewise, Alderete observed no measurable cytotoxicity due to the too short contact of T. tenax with epithelial cell lines [18].
Numerous articles dealt with the capacity of T. tenax to produce enzymes. The authors have identified some of them and assessed their properties. In 1991, Bózner and Demes, and more recently El Sibaei, identified a temperature-dependant proteinase capable of degrading types I, III, IV and V collagen, that would be cysteine proteinases or metallo-proteinases [20,24,25]. These enzymes are cell-associated and extracellular proteinases released from alive T. tenax [25]. Moreover, Yamamoto et al. have shown that an ATCC T. tenax strain can produce a specific enzyme that hydrolyses acid soluble type I collagen as well as gelatin, described as cathepsin B [21]. Numerous publications have described one of the main virulence factors in periodontitis: cysteine proteinases, mainly produced by P. gingivalis but also by Tannerella forsythia, are capable of directly degrading host tissues, activating the host proenzymes or neutralizing the host immune system [47,48]. The collagenolytic activities of T. tenax proteases could also, as indicated for bacterial cysteine proteases, participate in host matrix protein damage and play a role in the etiopathogenesis of periodontal diseases [49]. But Bózner et al. concluded that it would be of interest to purify T. tenax enzymes and test their effect on extracellular matrix proteins in the oral cavity [25]. In addition to these enzymes, T. tenax seems to be able to produce haemolysins, which are also synthesised by many pathobionts such as P. gingivalis, Prevotella intermedia, A. actinomycetemcomitans and Treponema denticola [50][51][52][53]22].

The postulate "Host responses"
All bacteria regarded as periodontal pathogens can trigger a "Host response" to both innate and adaptative immunity. The only study assessing the response of macrophages to contamination with live ATCC T. tenax strain failed to show pro-inflammatory cytokine production [26]. Nevertheless, interleukine-8 is the only macrophage-produced cytokine after 6 hours of incubation with flagellate lysates. The absence of cytokine synthesis with live flagellate could be the result of the low ratio of trichomonads: macrophage, 1:5. Certainly, Ribeiro's cytotoxicity assay used an inverted ratio (5:1). T. tenax seems capable of inducing antibody production as shown in the blood analysis of haemodialysis patients [27]. Moreover, Kott and Adler detected two types of antibodies: one targeting the T. tenax flagella and another an agglutinin, but this result is not relevant to show the potential pathogenicity of T. tenax [28]. Indeed, intravenous injection of a commensal bacteria could induce the production of systemic antibodies [54]. Ioli et al. have detected antibodies against T. tenax in the blood of haemodialysis patients without controlling the presence of bone loss around teeth of these patients. So, the conclusion of the flagellate involvement in periodontitis should be cautiously considered. All of these in vitro studies provided some possible responses to the question of the pathogenicity of T. tenax. But these in vitro cellular models generally used cell lines and ATCC T. tenax strain that cannot fully replicate the in vivo interaction between the T. tenax wild-type strain and primary human gingival cells.

Potential pathogenicity in extra-oral location of Trichomonas tenax
T. tenax is described in the literature as an inhabitant of the oral cavity, but some studies have shown this flagellate in extra-oral locations such as the vagina, the respiratory system, the liver and the lymph node of an immunocompromised patient with severe infectious disease. Most of these articles are case reports which have the lowest level of evidence of all clinical studies and conclude that T. tenax has no invasive power (via bloodstream) or pathogenicity. The authors hypothesized that the flagellate moved from the oral cavity to the lung or was self-inoculated via the hands or through oral sex to the vagina and was only a co-factor aggravating the underlying infection. Duboucher et al. could not explain the lymphatic colonization of the flagellate, although an invasion via the lymphatic vessels, and postulated a putative role in the cervical adenopathy of the patient as this protozoan was the only microorganism identified in the lymph node [34]. Even if three weeks later a culture of the excised lymph node showed Mycobacterium tuberculosis, the authors suggest that the observed caseous necrosis could be attributed not only to M. tuberculosis but also to T. tenax [34]. Jakobsen speculated about a possible migration through the blood to explain the presence of T. tenax in subhepatic abscess [35]. Only Brosh-Nissimov et al. have reported the presence of the flagellate in extra-oral locations in immunocompetent patients with urogenital infection. These authors concluded that T. tenax could be one of the potential etiologies of this dysuria [29].

Strengths and limitations
As far as is known, this review is the first one to investigate the role of T. tenax in periodontitis using PRISMA guidelines and Koch's postulates revisited by Socransky. These results shown an interesting evaluation of the role of this flagellate in the periodontitis of human adult without any systemic nor genetic disease. Nevertheless, this review presents some limitations: (i) language restriction, (ii) limited number of papers with precise periodontal parameters detecting T. tenax, (iii) majority of the included articles identified this trichomonads by microscope which is not the most accurate technique, (iv) regarding research articles, authors used either ATCC or patient strain which may not be representative of all wild strains in patients with periodontitis (v) and the use of different protocols (MOI, time of incubation, cell lines) makes comparison difficult and does not allow a firm conclusion.

Conclusions
Even if T. tenax appears to be identified in patients with periodontitis, there are still many unanswered questions. The cellular model with flagellate has demonstrated its acknowledged adhesion, cytotoxicity and enzyme production that could contribute to periodontal tissue breakdown brought about by periodontopathogens. Its direct pathogenic role in periodontitis but also in extra-oral infections remains unclear and requires further investigations. Periodontal microbiota dysbiosis is not only an imbalance in the relative abundance of certain microbial species found in diseased sites, but also involves polymicrobial synergy leading to immune dysfunctions that result in alveolar bone and soft tissue destruction. In the absence of direct pathogenicity, T. tenax could be indirectly involved in the etiopathogenesis of periodontitis. Certainly, the bacteria phagocytized by the flagellate may be either lysed or alive inside its cytoplasm and be protected from immune cells and antibiotics action [55]. In this new context, the roles of individual bacteria and their interaction with host and also with other microorganisms such as protozoan must be evaluated.
Supporting information S1