Influence of species composition and cultivation condition on peri-implant biofilm dysbiosis in vitro

Influence of species composition and cultivation condition on peri-implant biofilm dysbiosis in vitro | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Influence of species composition and cultivation condition on peri-implant biofilm dysbiosis in vitro Nils Heine, Kristina Bittroff, Szymon P. Szafrański, Maya Duitscher, and 12 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6573834/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Background Changes in bacterial species composition within oral biofilms, known as biofilm dysbiosis, are associated with the development of severe oral diseases. To better understand this process and help establish early detection systems, models are needed which replicate oral biofilm dysbiosis in vitro – ideally by also mimicking natural salivary flow conditions. Methods For this purpose, the present study cultivated two different combinations of oral commensal and pathogenic strains – Streptococcus oralis, Actinomyces naeslundii, Veillonella dispar/parvula, Fusobacterium nucleatum and Porphyromonas gingivalis – comparatively within an established flow chamber model on the implant material titanium, and statically in 6-well plates for 21 days. Biofilm morphology, species distribution, and bacterial metabolism were analyzed by fluorescence microscopy, molecular biological methods, and metabolic interaction prediction. Results Biofilm growth and composition were strongly influenced by bacterial species selection, and to a more minor extent, by cultivation conditions. Within the model containing V. dispar and a laboratory P. gingivalis strain, a diversification of commensal species was observed over time along with a significantly reduced pH-value. In contrast, the model containing V. parvula and the clinical isolate P. gingivalis W83, a dysbiotic shift with increased pathogen levels, pH-value, and virulence factors was achieved. Conclusion Within the present study, different in vitro oral multispecies biofilm models were successfully developed. Depending on bacterial species selection, these models were able to depict the infection-associated dysbiotic shift in species composition under flow conditions solely by intrinsic interactions and without the use of external stimuli. dental plaque dysbiosis dental implants microbiological techniques dynamic cultivation Figures Figure 1 Figure 2 Figure 3 Figure 4 Key findings In this study, different in vitro oral multispecies biofilm models were developed, which resemble an infection-associated dybiotic shift dependent on initial species and strain selection, and were also influenced by presence or absence of laminar flow. Plain Language Summary The human mouth hosts a large variety of bacterial species, which form communities embedded in a self-produced matrix of extracellular polymeric substances, called biofilms. Species associated with disease are called pathogenic. A change in the species composition within oral biofilms towards a higher share of pathogenic bacteria is called dysbiosis. To improve understanding of this process and help establish systems for early detection of dysbiosis, models which replicate it in vitro are required. Since oral biofilms are constantly exposed to salivary flow, ideally these models should simulate the flow as well. In this study, two different combinations of five oral bacterial species were cultivated to form biofilms for 21 days with and without flow conditions. Changes in species composition were heavily dependent on the initial selection of species and strain combinations. One combination (“commensal”) resulted in a diversification of non-pathogenic species over time without an increase of pathogenic bacteria. The other combination (“dysbiotic”) resulted in an increase of species associated with disease. Furthermore, the composition was significantly influenced by the presence or absence of flow. These different in vitro oral multispecies models can now be used to study the dysbiosis process and establish early detection systems. Introduction Bacterial biofilms of the oral cavity – also known as dental plaque – are associated with the development and progression of multiple oral diseases. These biofilms are formed by a multitude of oral bacterial species that adhere both to surfaces and to each other. These bacteria successfully protect themselves within an extracellular matrix, resulting in drastically increased tolerance towards the immune system and antibiotic treatment. In this regard, biofilm formation on dental implants in particular is closely linked to progressive diseases, since the implant lacks an innate immune response and other protective anatomical features [ 1 ]. Peri-implant mucositis and peri-implantitis – analogous to gingivitis and periodontitis on natural teeth – are associated with severe inflammatory reactions that can lead to subsequent soft- and hard-tissue destruction. From the microbiological perspective, the progression of peri-implantitis/periodontitis is often accompanied by a notable shift in the biofilm species composition and activity [ 2 ]. Whereas the initial commensal biofilm is dominated by mitis-group streptococci, Actinomyces and Veillonella species, advanced disease state biofilms frequently exhibit larger amounts of Prevotella and Peptostreptococacae species [ 2 ]. This process of disease-associated changes in bacterial species composition is called bacterial dysbiosis. To prevent the onset of peri-implantitis, early detection of bacterial dysbiosis would allow for a timely treatment that also circumvents tolerance development. However, the establishment of dysbiosis sensors – e.g., based on spectroscopy or chemometrics – requires reproducible in vitro models to serve as test systems. Oral multispecies biofilm in vitro models typically contain up to ten characteristic bacterial species that are either sampled from volunteers or commercially available type strains [ 3 – 5 ]. These biofilms are grown on various materials (including implant-grade titanium) for several days or weeks under either static or salivary shear force-mimicking dynamic conditions [6;7]. One example is the Hannoverian Oral Multispecies Biofilm Implant Flow Chamber (HOBIC) model developed in our group, which contains the oral commensals Streptococcus oralis , Actinomyces naeslundii , Veillonella dispar as well as the oral pathogen Porphyromonas gingivalis [ 8 ]. This four-species biofilm is grown on titanium discs in custom-made flow chambers designed for non-invasive microscopic readout. Within the incubation time of 24 hours, reproducible biofilms of commensal composition are formed. In contrast, Siddiqui et al. have reported on a similar six-species biofilm model on titanium that was cultivated under static conditions for 21 days [ 6 ]. Over time, a clear shift in bacterial species composition towards the increase of pathogenic species could be detected. However, this model lacks the naturally existing flow shear forces. The aim of the present study was to advance the HOBIC model and reproduce the bacterial dysbiosis associated with peri-implantitis in vitro . For this purpose, two different five-species combinations were grown comparatively under both static and dynamic conditions over 21 days and analyzed for bacterial growth (optical density and pH development), biofilm morphology (live/dead fluorescence staining with confocal microscopy and digital image analysis) and species distribution (quantitative real-time PCR and fluorescence in situ hybridization). By this, the research hypotheses that (I) a bacterial shift can be introduced solely by increasing cultivation time as well as that (II) the shift depends on species composition and (III) cultivation conditions were addressed. Material and Methods Bacterial strains and culture conditions Bacteria were routinely stored as glycerol stocks at -80°C. Veillonella dispar DSM 20735 (German Collection of Microorganisms and Cell Cultures GmbH, DSMZ, Braunschweig, Germany), Veillonella parvula ATCC® 17745™ (American Type Culture Collection, ATCC, Manassas, VA, USA), Fusobacterium nucleatum DSM 15643, Porphyromonas gingivalis DSM 20709 and Porphyromonas gingivalis ATCC W83, were streaked out on fastidious anaerobe agar (Lab M Ltd., Heywood, UK) plates supplemented with 5% defibrinated sheep blood (Thermo Fisher Scientific Inc., Waltham, MA, USA) and incubated at 37°C under anaerobic conditions, which were achieved using AnaeroGen™ bags (Thermo Fisher Scientific Inc.), for three days. Afterwards, colonies from the agar plates were transferred to liquid medium and cultured overnight in brain heart infusion medium (BHI, Oxoid Deutschland GmbH, Wesel, Germany) supplemented with 10 mg/L vitamin K (Carl Roth GmbH + Co. KG, Karlsruhe, Germany) (BHI + VitK, V. dispar/parvula ) or in fastidious anaerobe broth (Lab M Ltd., F. nucleatum and P. gingivalis ) at 37°C under anaerobic conditions. Streptococcus oralis ATCC 9811™ and Actinomyces naeslundii DSM 43013 were cultured overnight in BHI + VitK at 37°C under anaerobic conditions. Static biofilm growth in well plates Bacterial overnight cultures were adjusted to an optical density at 600 nm (OD 600 ) of 0.05 in BHI + VitK with 5 mg/L hemin (BHI + VitK/Hem, Sigma Aldrich, St. Louis, MO, USA), and mixed in two different five-species combinations: S. oralis , A. naeslundii and F. nucleatum were combined either with V. dispar and P gingivalis DSM 20709 (commensal model) or V. parvula and P. gingivalis ATCC W83 (dysbiotic model). 5 mL per well of the mixed suspension were added to 6-well plates and incubated for 1, 3, 6, 10, 15, or 21 days at 37°C under anaerobic conditions. Every other day, half of the medium was replaced with fresh medium. Before analysis, biofilms were washed once with phosphate buffered saline (PBS, Sigma Aldrich). Biofilm growth in the adaptive HOBIC model The setup of the flow chamber system is shown in Fig. 1 A, and is based on the previously described “Hannoverian Oral Multispecies Biofilm Implant Flow Chamber (HOBIC)” model [ 8 ] with the following modifications: In this case, pH-sensitive flow-through cells connected to optical fibers (FTC-SU-LG1-S, PreSense Precision Sensing GmbH, Regensburg, Germany) were integrated behind the flow chambers. Grade 4 titanium discs (12 mm diameter, 1.5 mm height, R a = 0.31 µm) were submerged in artificial saliva (850 mg/L mucin, 10 µg/mL lysozyme, 1 mg/mL α-amylase, 40 µg/mL albumin) during sterile assembly of the chambers. Then, the chambers were integrated into the system and 2.1 mL per strain with OD 600 of 0.05 were added to 1.5 L BHI + VitK/Hem in the bioreactor. In addition to the previously described HOBIC model [ 8 ], F. nucleatum was added as fifth bacterium to the inoculum for both combinations and V. dispar and P. gingivalis 20709 were replaced by V. parvula and P. gingivalis W83 as “dysbiotic” strain combination. After 24 hours of cultivation with 100 µL/min at 37°C under anaerobic conditions, the components up to the bubble trap (Fig. 1 A) were replaced with new sterile parts, except for the OD measuring bypass. The system was then run for 20 additional days with sterile 1:2 diluted medium. On day 1, 3, 6, 10, 15, and 21, chambers were washed with PBS for 30 min at 100 µL/min before subjected to further analysis. Live/Dead staining and microscopic analysis Static and flow chamber biofilms were stained using SYTO®9 and propidium iodide (PI) of the LIVE/DEAD® BacLight™ Bacterial Viability Kit (Life Technologies, Darmstadt, Germany) at a concentration of 1:2000 of the stock solutions in PBS, followed by fixation with 2.5% glutardialdehyde. For staining of flow chamber biofilms, the dye and fixation solutions were pumped through the system as previously described [ 8 ]. Likewise, confocal laser-scanning microscopy was done using established protocols [ 8 ]. For the HOBIC systems, from day 6 onwards, images were taken from the cover slip downwards rather than directly on the titanium surface, since the laser could not reach through the thick biofilm. From all images, biofilm volume and live/dead distribution were quantified using the software Imaris (v8.4.1, Bitplane AG, Zurich, Switzerland). PMA treatment, DNA extraction, and qRT-PCR After microscopy, the biofilm inside the chambers was harvested and either directly frozen or subjected to additional PMA treatment [ 9 ]. Following washing with PBS, bacteria were incubated with 0.2 mM PMAxx™ (Biotium, Inc., Fremont, CA, USA) for 10 min at 4°C, and then again for 20 min in the PMA-Lite™ LED Photolysis Device (Biotium, Inc.). DNA was extracted using a customized protocol which deploys a combination of enzymatic lysis, mechanical disruption, and column-based DNA isolation. Bacterial sample material was initially treated with 450 µL lysozyme solution (20 mg/mL lysozyme (Merck, Darmstadt, Germany) in 20 mM Tris HCl, pH 8.0; 2 mM EDTA; 1,2% Triton) for two hours at 37°C. After addition of 50 µL Proteinase K and 500 µL AL buffer (both Qiagen, Hilden, Germany), treatment was extended for 30 min at 56°C and 15 min at 95°C. The complete sample material was then transferred to Lysing Matrix E tubes (MP Biomedicals, Eschwege, Germany) and mechanically disrupted in three cycles of 6500 rpm for 30 s in a Precellys 24 homogenizer (Bertin Technologies, Frankfurt am Main, Germany), punctuated with cooling on ice for five minutes in between cycles. Finally, beads and debris were sedimented by centrifugation (5 min, 14000 x g), and the cleared supernatant was mixed 1:1 with 100% ethanol. Subsequent steps were performed with the QIAamp Mini Kit (Qiagen) according to the manufacturer’s protocol “DNA Purification from Blood or Body Fluids” – starting with the application to the spin columns. To reduce the risk of contamination, a new collection tube was used after each of the kit-specific wash steps. DNA was eluted with 50 µL of PCR-grade water and stored at -20°C until further usage. Quantitative real-time PCR and calculation of respective cell numbers and relative species distribution were performed as described by Kommerein et al. [ 9 ] using the SYBR Green reaction mix (Bio-Rad Laboratories GmbH, Feldkirchen, Germany) and the LightCycler 96 (Roche Holding GmbH, Grenzach-Wyhlen, Germany). Primer pairs, reaction components, cycle conditions and genome weight per cell are all given in the Supplementary Tables S1-4, respectively. Fluorescence- In-Situ -Hybridization To prepare HOBIC samples for fluorescence in situ hybridization (FISH), 50% (v/v) ethanol was pumped through the system for 20 min with a flow rate of 250 µL/min, after which the ethanol filled chambers were removed from the system and then stored at 4°C overnight to fixate the bacteria. The chambers were then opened, and the titanium specimen were transferred to a 6-well plate and left to dry under sterile conditions. FISH staining and CLSM analysis were performed as previously reported [8;10]. Briefly, 1 g/L lysozyme (Merck) treatment at 37°C for 10 min was used to disrupt cell membranes. Lysis was stopped with pure ethanol, samples were dried and then stained with six fluorescently labeled 16S rRNA probes (Supplementary Table S5) in hybridization buffer at 46°C for 30 min. F. nucleatum was targeted by two probes that shared the same nucleotide sequence, but were labeled with different dyes – resulting in co-localized blue and red fluorescence. Afterwards, samples were washed several times and analyzed by CLSM. A 630-fold magnification was used to take image stacks with an xy-size of 185 x 185 µm² and a 2 µm z-step-size. Scanning was done sequentially per frame. The first sequence used a 405 nm and a 552 nm laser for excitation, and detected blue and yellow signals in the wavelength ranges 413–477 nm and 576–648 nm, respectively. During the second sequence, a 488 nm and a 638 nm laser were used to excite the samples, and emission detection was done in the wavelength ranges 509–576 nm and 648–777 nm for green and red signals, respectively. Gingipain-Specific Enzyme-Linked Immunosorbent Assay (ELISA) Supernatants of the HOBIC model flow chambers were collected, frozen, and used for gingipain protein quantification using the human P. gingivalis- specific IgG antibody ELISA kit (Huangshi INS Biological Technology Co., LTD, Huangshi, China). Analysis was done according to the manufacturer’s protocol – but with the samples being additionally incubated for 1 hour at room temperature followed by 3 washing steps before addition of the detection antibody. Literature-Based Metabolic Interaction Prediction The potential of biofilm members to engage in metabolic and enzyme-based interspecies interactions was inferred using a custom database [ 11 ]. Interaction data were sourced from the literature and subjected to manual curation [12;13]. Custom-designed graphs were employed to visualize the interaction networks. Statistical Analysis Data presentation and statistical analysis were done using GraphPad Prism 8.4 (GraphPad Software Inc., San Diego, CA, USA). Statistical test details can be found in the respective figure captions. Family-wise significance level was defined with α = 0.05. Results Time-dependent biofilm growth and viability Initial bacterial growth in the bioreactor was monitored using inline optical density measurement, and showed typical bacterial growth curves with significantly increased growth of the dysbiotic model (Fig. 1 A, B). Subsequent biofilm growth on titanium discs was then analyzed by fluorescence staining and confocal microscopy. Within both commensal models, the biofilm volume significantly decreased after one day, and then re-established until day 10 (Fig. 2 A, Supplementary Table S6). This development was more pronounced in the commensal HOBIC model. In contrast, the biofilm volume of both dysbiotic models significantly increased after day 1, reaching a plateau at day 6 (HOBIC system) and day 15 (static system), respectively (Fig. 2 A, Supplementary Table S6). Biofilm viability – analyzed by fluorescence-based membrane integrity – was observed to significantly decrease over time for all cultivation conditions except for the static commensal model (Fig. 2 A, B, Supplementary Table S7). Viability development thereby replicates the respective biofilm volume pattern. Cultivation condition-dependent species composition Time-dependent species composition (both, viable and total count) was analyzed by DNA isolation and qRT-PCR as well as FISH staining, and these analyses revealed clear differences between the commensal and dysbiotic model (Fig. 3 A, B, Supplementary Figure S1). The changes in bacterial species distribution were more pronounced for viable cells (Fig. 3 ) than for the total count (Supplementary Figure S1). Within the commensal HOBIC model, viable S. oralis was the initially dominant species, although its amount significantly decreased over time from approx. 70% to merely 35% (Supplementary Table S8). Within the commensal static model, this decrease was also observed – but only from day 6. Initially, V. dispar was the dominant species, but was then gradually replaced by S. oralis up until day 6. Afterwards, V. dispar ’s distribution remained on average stable at 30% within both commensal models. In parallel, A. naeslundii established itself with prolonged incubation to approx. 30%. F. nucelatum and P. gingivalis were almost undetectable in both commensal models. In contrast, total and viable species distribution of the dysbiotic model differed remarkably (Fig. 3 A, B, Supplementary Figure S1). For the static system, V. parvula remained the dominant species independent of incubation time (50–60%), followed by P. gingivalis (25%), F. nucleatum (20%), and only a very low amount S. oralis and A. naeslundii . In the HOBIC system, the initially dominant V. parvula significantly decreased from approx. 95–20% (Supplementary Table S8), while A. naeslundii , F. nucleatum and P. gingivalis successively increased from day 6, 10, and 21, respectively, up to 15–30% (Fig. 3 B). Species composition-dependent biofilm metabolism Within the commensal and dysbiotic HOBIC models, pH-values and P. gingivalis gingipain protein concentration over time were determined by optical fiber measurement and ELISA, respectively, and showed clear differences (Fig. 4 A, B). For the commensal model, pH-values initially dropped below pH 6.0, sharply increased upon medium change at day 1, and then established itself at approx. pH 6.3. In contrast, for the dysbiotic model, pH-values only dropped to pH 6.3 and then gradually increased to pH 6.9 by day 21. These higher pH-values showed a tendency to explicitly increase the growth of the dysbiotic model’s P. gingivalis strain (Supplementary Figure S2A). In line with these observations, the amount of gingipain protein significantly increased over time only within the dysbiotic model (Fig. 4 B). To account for these differences, literature-based prediction of metabolic interactions between the six bacterial species was performed (Fig. 4 C). Based on the available nutrients from the culture medium, the species were found to engage in multiple food chain and enzyme sharing behaviors, with peptides, glucose, vitamins and other growth factors produced by S. oralis , A. naeslundii and V. dispar/parvula and then utilized by F. nucleatum and P. gingivalis . Whereas V. dispar and V. parvula generally exhibited similar metabolisms, only V. parvula was capable of de novo thiamine (vitamin B1) synthesis. However, although the dysbiotic model’s P. gingivalis strain showed increased overall growth compared to the commensal model’s strain, no growth difference in Veillonella-preconditioned medium could be detected (Supplementary Figure S2B). Discussion Early detection of oral biofilm dysbiosis on dental implants can prevent the development of severe infections like peri-implantitis. To help establish dysbiosis sensors, however, reliable in vitro models must first be developed for use as test systems. Within the present study, the existing HOBIC model was successfully adapted to reproduce bacterial dysbiosis for this purpose. With regard to the initial research hypotheses, this model helped to confirm that the dysbiotic shift depended both on the selected bacterial species as well as on the cultivation conditions. The bacterial species selected for this study were S. oralis, A. naeslundii, V. dispar or V. parvula, F. nucleatum , and P. gingivalis. All genera but Fusobacterium were already included in the previous HOBIC model [8;9]. S. oralis and A. naeslundii are among the dominant primary oral colonizers associated with oral health [ 2 ]. V. dispar and V. parvula are part of the core microbiome, co-aggregate with S. oralis and A. naeslundii , and metabolize lactate produced by them [2;14]. P. gingivalis is part of the red complex bacteria and associated with periodontitis and peri-implantitis [ 2 ]. As the biofilm of the initial HOBIC model mainly consisted of S. oralis, A. naeslundii , and V. dispar , and contained only a small proportion of P. gingivalis , it depicts an early commensal oral biofilm [8;9]. To support the dysbiotic shift, F. nucleatum was added as fifth bacterium in this study. The species is considered a central bridging bacterium between commensal and dysbiotic strains as it co-aggregates with bacteria of both groups [ 2 ]. The species selection closely matched those of other commensal and dysbiotic oral biofilm models [5–7;15]. They can therefore be considered to collectively constitute a reliable representation of the oral microbiome. However, since minor differences in species selection in other studies lead to both similar and different results [6;15], it bears noting that all models remain simplifications of the natural microbiome complexity and all results generated by models should therefore be interpreted with this limitation in mind. Biofilm growth of the selected bacteria was done in full medium supplemented with vitamin K and hemin to support P. gingivalis growth (according to preliminary experiments and other studies [ 6 ]) with all species inoculated once at the same time. In addition, titanium surfaces were pre-conditioned using artificial saliva. Within this experimental setup, the dysbiotic shift was induced by intrinsic bacterial interactions only, closely replicating natural oral conditions. In contrast, some other models induced dysbiosis by changing medium and oxygen level, as well as inoculating additional bacterial species [4;16]. Whereas these approaches offer the possibility of external control, the setup selected here is inherently more beneficial for observing intrinsic processes – a pre-requirement for the development of dysbiosis sensors. Over the incubation time of 21 days, biofilm volume increased for the dysbiotic model only, reaching a plateau after 6 days. This growth pattern has also been observed for other models with a similar dysbiotic species composition and cultivation time of more than 7 days, as well as for in situ grown biofilms on implant healing abutments [3;6;11]. In comparison to the commensal model, the initial planktonic growth in the bioreactor of the HOBIC model, and the growth of V. parvula and P. gingivalis W83 at different pH (Supplementary Figure S2A) was found to be significantly higher. This makes the increased growth of these individual species the most likely reason for the elevated biofilm volume of the dysbiotic model. The parallel decrease of biofilm viability independently of the species composition is also in line with previous in vitro and in situ results [5;7;11], and most likely due to limited nutrient availability in deeper layers of the maturated biofilm. The most obvious difference between the commensal and dysbiotic models lies in their divergent species composition. During the first days, the commensal model was dominated by S. oralis after a short initial establishing phase. In contrast, the dysbiotic model was initially dominated by V. parvula . Differing proportions of S. oralis and V. parvula have already been described in previous oral biofilm models – with studies showing an initial dominance of S. oralis [ 5 ], an initial dominance of V. parvula [ 7 ] or equal amounts of both species [3;6]. Their establishment is probably significantly influenced by the (pre-culture) cultivation conditions; however, the details of these alterations remain to be analyzed in future studies. The different proportions of S. oralis are also the most probable reason for the differences in pH-values measured in the HOBIC model. As shown in the predicted metabolic interactions, S. oralis utilizes carbohydrates from the medium to produce lactate via fermentation, causing strong medium acidification. In contrast, Veillonellaceae metabolize nutrients (including lactate) to less acidic acetate and propionate, resulting in higher pH-values in the medium. With prolonged cultivation time, a diversification of commensal species was observed in the commensal model over time. In contrast, the dysbiotic model showed a notable increase in pathogenic species with reduced proportions of commensal strains. This observation is further supported by the increase in gingipain proteins, which are trypsin-like cysteine proteases that are among the most important P . gingivalis virulence factors [ 17 ]. Since these models only differed in the Veillonella species and the P. gingivalis strain, their contribution seems to be crucial for the dysbiotic shift – at least within the limited setting of an in vitro experiment. Further evidence to support this observation can also be found in the literature: In a co-association study of Streptococcus mutans, Veillonella dispar and Veillonella parvula within the context of root caries, only V. parvula was found to support S. mutans growth in vitro [ 18 ]. More importantly, a metatranscriptomic analysis of different oral in vitro biofilm models found that P. gingivalis W83 – a virulent strain which was also used for the dysbiotic model of this study – had a greater effect on biofilm dysbiosis than a lower virulent type strain by specifically influencing genes related to metabolic pathways and quorum sensing of several commensal species [ 15 ]. An effect on the strain level would also be supported by the species-level metabolic interaction prediction of this study, where only a minor difference in thiamine (vitamin B1) production solely by V. parvula could be identified. Even though P. gingivalis utilizes vitamins produced by Veillonellaceae, they exhibit the enzymes for thiamine metabolism themselves according to KEGG-pathway analysis, and thus probably would not rely on cross-feeding by V. parvula . On the other hand, although P. gingivalis W83 showed increased growth compared to P. gingivalis 20709 , Veillonella-conditioned medium had no different effect on both strains. In summary, these results strongly underscore the importance of further analyzing metabolic interaction within oral microbial communities that seem to significantly contribute to dysbiosis development within basic research approaches. The models developed here can be used for initial insights as well as validation for this purpose. Aside from strain selection, cultivation conditions also influenced bacterial species composition – albeit to a more minor extent. For both the commensal and dysbiotic model, changes in species composition were more pronounced in the HOBIC than the static system. During static cultivation (even though medium was exchanged every other day), metabolites accumulated within the biofilm, probably making metabolite-based changes of species composition slower or less pronounced. In contrast, during cultivation in the dynamic HOBIC system, medium is constantly replaced, thus preventing the metabolites from accumulating to a greater degree which might very well induce changes in species composition. Conclusion Within the present study, different in vitro oral multispecies biofilm models were successfully developed. Depending on bacterial species selection, these models were able to depict the infection-associated dysbiotic shift in species composition solely by intrinsic interactions, and without any deployment of or reference to external stimuli. The different results between cultivation conditions offer the possibility for a number of different future application: For the direct comparison between commensal and dysbiotic biofilms, straightforward static cultivation can be used with species composition being already different after 24 hours. In contrast, for the observation of the bacterial shift over time (for example by novel sensor systems), the dysbiotic HOBIC model is to be preferred. The results of this study also point towards the (current) limitations of in vitro models and the need for further in vivo studies that focus on examining how metabolic interactions on the strain level influence bacterial species composition. For the validation of these in vivo observations, we believe that the presented biofilm models will serve as a valuable tool. Declarations Acknowledgements This work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under the Collaborative Research Center SFB/TRR-298-SIIRI – Project ID 426335750. The authors would like to additionally thank Wiebke Smolinski, Hanna Lena Thoms and Teresa Lea Ngyuen for excellent technical assistance and Dr. Andreas Winkel for providing the BioRender image of Figure 1A. The authors declare no conflict of interest. Author contribution Conceptualization: AH, TS, MS, KDN; Methodology: NH, WB, CM, NK, MS, KDN; Investigation: NH, KB, SPS, MD, CV, ND, KF, KDN; Formal Analysis: NH, KDN; Writing – Original Draft Preparation: NH, SPS, WB, KDN; Writing – Review & Editing: NH, SPS, KB, SPS, MD, WB, CV, CM, NK, ND, KF, AH, TS, MLTM, JB, MS; Supervision: CM, NK, AH, TS, MLTM, JB, MS, KDN; Funding Acquisition: AH, TS, MS, KDN. Conflict of interest The authors declare no conflict of interest. Data availability The data are available upon request from the corresponding authors. References Belibasakis GN (2014) Microbiological and immuno-pathological aspects of peri-implant diseases. Arch Oral Biol 59:66–72. 10.1016/j.archoralbio.2013.09.013 Colombo APV, Tanner ACR (2019) The role of bacterial biofilms in dental caries and periodontal and peri-implant diseases: A historical perspective. J Dent Res 98:373–385. 10.1177/0022034519830686 Thurnheer T, Bostanci N, Belibasakis GN (2016) Microbial dynamics during conversion from supragingival to subgingival biofilms in an in vitro model. Mol Oral Microbiol 31:125–135. 10.1111/omi.12108 Dalwai F, Spratt DA, Pratten J (2006) Modeling shifts in microbial populations associated with health or disease. Appl Environ Microbiol 72:3678–3684. 10.1128/AEM.72.5.3678-3684.2006 Sanchez MC, Llama-Palacios A, Blanc V, Leon R, Herrera D, Sanz M (2011) Structure, viability and bacterial kinetics of an in vitro biofilm model using six bacteria from the subgingival microbiota. J Periodontal Res 46:252–260 Siddiqui DA, Fidai AB, Natarajan SG, Rodrigues DC (2022) Succession of oral bacterial colonizers on dental implant materials: An in vitro biofilm model. Dent Mater 38:384–396. 10.1016/j.dental.2021.12.021 Blanc V, Isabal S, Sanchez MC, Llama-Palacios A, Herrera D, Sanz M et al (2014) Characterization and application of a flow system for in vitro multispecies oral biofilm formation. J Periodontal Res 49:323–332 Kommerein N, Doll K, Stumpp NS, Stiesch M (2018) Development and characterization of an oral multispecies biofilm implant flow chamber model. PLoS ONE 13:e0196967 Kommerein N, Stumpp SN, Musken M, Ehlert N, Winkel A, Haussler S et al (2017) An oral multispecies biofilm model for high content screening applications. PLoS ONE 12:e0173973 Debener N, Heine N, Legutko B, Denkena B, Prasanthan V, Frings K et al (2024) Optically accessible, 3D-printed flow chamber with integrated sensors for the monitoring of oral multispecies biofilm growth in vitro. Front Bioeng Biotechnol 12:1483200. 10.3389/fbioe.2024.1483200 Dieckow S, Szafranski SP, Grischke J, Qu T, Doll-Nikutta K, Steglich M et al (2024) Structure and composition of early biofilms formed on dental implants are complex, diverse, subject-specific and dynamic. NPJ Biofilms Microbiomes 10:155–153. 10.1038/s41522-024-00624-3 Giacomini JJ, Torres-Morales J, Dewhirst FE, Borisy GG, Mark Welch JL (2023) Site specialization of human oral veillonella species. Microbiol Spectr 11:e0404222–e0404222 Epub 2023 Jan 25. 10.1128/spectrum.04042-22 Goodfellow M, Kämpfer P, Busse H, Trujillo ME, Suzuki et al (2012) &., Ken-ichiro,. Bergey's manual of systematic bacteriology (3–5 ed.) Springer New York. rg/10.1007/978-0-387-68233-4 Kolenbrander PE (2011) Multispecies communities: Interspecies interactions influence growth on saliva as sole nutritional source. Int J Oral Sci 3:49–54 Zhang Y, Shi W, Song Y, Wang J (2019) Metatranscriptomic analysis of an in vitro biofilm model reveals strain-specific interactions among multiple bacterial species. J Oral Microbiol 11:1599670. 10.1080/20002297.2019.1599670 Thurnheer T, Gmur R, Guggenheim B (2004) Multiplex FISH analysis of a six-species bacterial biofilm. J Microbiol Methods 56:37–47 Guo Y, Nguyen K, PotemJan (2010) Dichotomy of gingipains action as virulence factors: From cleaving substrates with the precision of a surgeon's knife to a meat chopper-like brutal degradation of proteins. Periodontol 2000 54:15–44. 10.1111/j.1600-0757.2010.00377.x Abram AM, Szewczyk MM, Park SG, Sam SS, Eldana HB, Koria FJ et al (2022) A co-association of streptococcus mutans and veillonella parvula/dispar in root caries patients and in vitro biofilms. Infect Immun 90:e0035522–e0035522 Epub 2022 Sep 21. 10.1128/iai.00355 – 22 Additional Declarations The authors declare no competing interests. Supplementary Files 250404SupportingInformationAdaptivesHOBICmodelNH.docx Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6573834","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":450769103,"identity":"6fb0e47f-8d4e-4a42-9270-daf6fa846ae4","order_by":0,"name":"Nils Heine","email":"","orcid":"https://orcid.org/0009-0003-2144-0355","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Nils","middleName":"","lastName":"Heine","suffix":""},{"id":450769104,"identity":"f4c10b96-bc19-4945-8f43-1c287403f97c","order_by":1,"name":"Kristina Bittroff","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Kristina","middleName":"","lastName":"Bittroff","suffix":""},{"id":450769105,"identity":"1be55b17-54bd-4ec6-ba17-0377870fa568","order_by":2,"name":"Szymon P. Szafrański","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Szymon","middleName":"P.","lastName":"Szafrański","suffix":""},{"id":450769106,"identity":"7e8be3d4-e655-41b8-bba9-1ee89b7a3420","order_by":3,"name":"Maya Duitscher","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Maya","middleName":"","lastName":"Duitscher","suffix":""},{"id":450769107,"identity":"98954624-7a70-48c6-8abe-ed5c2346d552","order_by":4,"name":"Wiebke Behrens","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Wiebke","middleName":"","lastName":"Behrens","suffix":""},{"id":450769108,"identity":"225a97b5-42f4-499f-a056-8f9cf5c3438d","order_by":5,"name":"Clarissa Vollmer","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Clarissa","middleName":"","lastName":"Vollmer","suffix":""},{"id":450769109,"identity":"30279d9d-a4d2-4a2c-bfbb-76d727a671ff","order_by":6,"name":"Carina Mikolai","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Carina","middleName":"","lastName":"Mikolai","suffix":""},{"id":450769110,"identity":"9612ae30-8b55-41a3-acb7-8877136de210","order_by":7,"name":"Nadine Kommerein","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Nadine","middleName":"","lastName":"Kommerein","suffix":""},{"id":450769111,"identity":"bdd670e5-98d5-4a79-a1e7-98f744538623","order_by":8,"name":"Nicolas Debener","email":"","orcid":"","institution":"Institute of Technical Chemistry, Leibniz University Hannover","correspondingAuthor":false,"prefix":"","firstName":"Nicolas","middleName":"","lastName":"Debener","suffix":""},{"id":450769112,"identity":"fb40cf96-c938-4337-b0f5-d7edf800ba00","order_by":9,"name":"Katharina Frings","email":"","orcid":"","institution":"Institute of Quantum Optics, Leibniz University Hannover","correspondingAuthor":false,"prefix":"","firstName":"Katharina","middleName":"","lastName":"Frings","suffix":""},{"id":450769113,"identity":"e0d878de-2d85-4265-90b4-b9260ec50466","order_by":10,"name":"Alexander Heisterkamp","email":"","orcid":"","institution":"Institute of Quantum Optics, Leibniz University Hannover","correspondingAuthor":false,"prefix":"","firstName":"Alexander","middleName":"","lastName":"Heisterkamp","suffix":""},{"id":450769114,"identity":"030ce762-04b0-4b28-a2b1-230d2013f55b","order_by":11,"name":"Thomas Scheper","email":"","orcid":"","institution":"Institute of Technical Chemistry, Leibniz University Hannover","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Scheper","suffix":""},{"id":450769115,"identity":"a9a18b8a-0df6-4e00-a4bb-b1077d7c5ec3","order_by":12,"name":"Maria L. Torres-Mapa","email":"","orcid":"","institution":"Institute of Quantum Optics, Leibniz University Hannover","correspondingAuthor":false,"prefix":"","firstName":"Maria","middleName":"L.","lastName":"Torres-Mapa","suffix":""},{"id":450769116,"identity":"52ad0b96-5649-4b12-a799-846ab55e7359","order_by":13,"name":"Janina Bahnemann","email":"","orcid":"","institution":"Institute of Physics, University of Augsburg","correspondingAuthor":false,"prefix":"","firstName":"Janina","middleName":"","lastName":"Bahnemann","suffix":""},{"id":450769117,"identity":"9277edb4-1b6b-49b5-a526-8c3605e7c9ba","order_by":14,"name":"Meike Stiesch","email":"","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":false,"prefix":"","firstName":"Meike","middleName":"","lastName":"Stiesch","suffix":""},{"id":450769118,"identity":"041c6886-153e-4fc8-8a7e-7a4e0bf904da","order_by":15,"name":"Katharina Doll-Nikutta","email":"data:image/png;base64,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","orcid":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","correspondingAuthor":true,"prefix":"","firstName":"Katharina","middleName":"","lastName":"Doll-Nikutta","suffix":""}],"badges":[],"createdAt":"2025-05-01 18:56:49","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-6573834/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6573834/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":81961500,"identity":"3fd78b5e-2e68-4508-a8c1-a4b63e42f82d","added_by":"auto","created_at":"2025-05-05 10:57:35","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":890870,"visible":true,"origin":"","legend":"\u003cp\u003eBacterial growth in the Hannoverian Oral Multispecies Biofilm Implant Flow Chamber (HOBIC) model. (A) Schematics of the flow chamber system with indicated components. Blue arrows indicate flow direction with 100 µL/min. Created with BioRender, Winkel. A. (2025) https://BioRender.com/ d73y389. (B) Bacterial growth curves (mean ± standard deviation) of commensal and dysbiotic species composition within the first 24 hours in the bioreactor measured by an inline photometer at λ= 600 nm. Optical densities at 6, 12, 18, and 24 hours were compared using 2-way ANOVA with Śidák’s correction for multiple comparison (N = 5) and revealed significant differences (*) between the commensal and dysbiotic model after 18 and 24 hours with p ≤ 0.05.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6573834/v1/6d1b4faaef5289ef68fcd9f8.png"},{"id":81961501,"identity":"b45cfd67-fa86-4b0f-98f1-869cf391cf06","added_by":"auto","created_at":"2025-05-05 10:57:35","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":3466208,"visible":true,"origin":"","legend":"\u003cp\u003eBiofilm volume and viability development over time of the different oral multispecies models. (A) Tukey box plots of biofilm volume per image (each left) and mean ±standard deviation of membrane-based biofilm viability (each right) of the commensal and dysbiotic models during static and HOBIC cultivation over time analyzed by fluorescence staining and CLSM. For statistical analysis (at least N = 15 individual images per condition), biofilm volume data were tested for normal distribution using D’Agostino \u0026amp; Pearson Omnibus Normality test followed by Kruskal-Wallis test with Dunn’s multiple comparison correction. Biofilm viability data were tested using 2-way ANOVA with Tukey’s multiple comparison test. For all statistical comparisons, family-wise significance level was set to α = 0.05. Results are given in Supplementary Tables S6 and S7. (B) Representative 3D reconstructed CLSM images of the commensal and dysbiotic HOBIC model at different time points. Green fluorescence indicates viable cells with intact membrane, whereas red/yellow fluorescence indicates cells with damaged membrane.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6573834/v1/9638d6ea32b896451b1658ee.png"},{"id":81961503,"identity":"32c04113-9c60-4958-a746-66942e0cfe49","added_by":"auto","created_at":"2025-05-05 10:57:35","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":3620109,"visible":true,"origin":"","legend":"\u003cp\u003eViable bacterial species distribution over time of the different oral multispecies biofilm models. (A) Tukey box plots of individual species distributions of the commensal and dysbiotic models during static and HOBIC cultivation over time quantified by qRT-PCR with PMA pre-treatment. Statistical comparisons for individual species development over time (N = 9 replicates per condition) was done using 2-way ANOVA with Dunnett’s test for multiple comparison to family-wise α= 0.05. Results are given in Supplementary Table S8. (B) Representative FISH images of the commensal and dysbiotic HOBIC model at different time points. Color coding is similar to those of (A).\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-6573834/v1/cc1b1acbe1b54ac5649a0bdb.png"},{"id":83302761,"identity":"8ce20c78-30ac-4a24-adbb-7a5ee1c6fa7e","added_by":"auto","created_at":"2025-05-22 15:29:37","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":221096,"visible":true,"origin":"","legend":"\u003cp\u003eBacterial metabolism of the different oral multispecies biofilm models. Mean ±standard deviation of (A) pH values and (B) gingipain protein concentrations over time in the commensal and dysbiotic HOBIC models. The peak in pH value after day 1 results from the change to fresh, sterile medium. Statistical comparison between the models at individual time points was done using 2-way ANOVA with Śidák’s correction for multiple comparison (N = 5 for pH, N = 9 for gingipain) and statistically significant differences with p ≤ 0.05 are marked with (*). (C) Potential metabolic interactions between the different bacterial species interfered using a custom-made database [11]summarizing curated phenotypic information [12;13]. Nodes representing species were placed on an arbitrary circle.\u003c/p\u003e","description":"","filename":"grafik.png","url":"https://assets-eu.researchsquare.com/files/rs-6573834/v1/4fd0fed111e48101048be2d0.png"},{"id":83304108,"identity":"8d52930e-25b9-43d7-a879-a4a66bce9348","added_by":"auto","created_at":"2025-05-22 15:54:25","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":8183155,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6573834/v1/0c6a1048-ab60-4d66-804e-16b6bfc81dea.pdf"},{"id":81961704,"identity":"3426e97c-7779-4362-80ca-0d7532e89a3d","added_by":"auto","created_at":"2025-05-05 11:05:35","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":728660,"visible":true,"origin":"","legend":"","description":"","filename":"250404SupportingInformationAdaptivesHOBICmodelNH.docx","url":"https://assets-eu.researchsquare.com/files/rs-6573834/v1/c0764556baaad1676a6bb045.docx"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003eInfluence of species composition and cultivation condition on peri-implant biofilm dysbiosis \u003cem\u003ein vitro\u003c/em\u003e\u003c/p\u003e","fulltext":[{"header":"Key findings","content":"\u003cp\u003eIn this study, different \u003cem\u003ein vitro\u0026nbsp;\u003c/em\u003eoral multispecies biofilm models were developed, which resemble an infection-associated dybiotic shift dependent on initial species and strain selection, and were also influenced by presence or absence of laminar flow.\u003c/p\u003e"},{"header":"Plain Language Summary","content":"\u003cp\u003eThe human mouth hosts a large variety of bacterial species, which form communities embedded in a self-produced matrix of extracellular polymeric substances, called biofilms. Species associated with disease are called pathogenic. A change in the species composition within oral biofilms towards a higher share of pathogenic bacteria is called dysbiosis. To improve understanding of this process and help establish systems for early detection of dysbiosis, models which replicate it \u003cem\u003ein vitro\u003c/em\u003e are required. Since oral biofilms are constantly exposed to salivary flow, ideally these models should simulate the flow as well. In this study, two different combinations of five oral bacterial species were cultivated to form biofilms for 21 days with and without flow conditions. Changes in species composition were heavily dependent on the initial selection of species and strain combinations. One combination (\u0026ldquo;commensal\u0026rdquo;) resulted in a diversification of non-pathogenic species over time without an increase of pathogenic bacteria. The other combination (\u0026ldquo;dysbiotic\u0026rdquo;) resulted in an increase of species associated with disease. Furthermore, the composition was significantly influenced by the presence or absence of flow. These different\u0026nbsp;\u003cem\u003ein vitro\u0026nbsp;\u003c/em\u003eoral multispecies models can now be used to study the dysbiosis process and establish early detection systems.\u003c/p\u003e"},{"header":"Introduction","content":"\u003cp\u003eBacterial biofilms of the oral cavity \u0026ndash; also known as dental plaque \u0026ndash; are associated with the development and progression of multiple oral diseases. These biofilms are formed by a multitude of oral bacterial species that adhere both to surfaces and to each other. These bacteria successfully protect themselves within an extracellular matrix, resulting in drastically increased tolerance towards the immune system and antibiotic treatment. In this regard, biofilm formation on dental implants in particular is closely linked to progressive diseases, since the implant lacks an innate immune response and other protective anatomical features [\u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e]. Peri-implant mucositis and peri-implantitis \u0026ndash; analogous to gingivitis and periodontitis on natural teeth \u0026ndash; are associated with severe inflammatory reactions that can lead to subsequent soft- and hard-tissue destruction. From the microbiological perspective, the progression of peri-implantitis/periodontitis is often accompanied by a notable shift in the biofilm species composition and activity [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. Whereas the initial commensal biofilm is dominated by mitis-group streptococci, \u003cem\u003eActinomyces\u003c/em\u003e and \u003cem\u003eVeillonella\u003c/em\u003e species, advanced disease state biofilms frequently exhibit larger amounts of \u003cem\u003ePrevotella\u003c/em\u003e and \u003cem\u003ePeptostreptococacae\u003c/em\u003e species [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. This process of disease-associated changes in bacterial species composition is called bacterial dysbiosis.\u003c/p\u003e \u003cp\u003eTo prevent the onset of peri-implantitis, early detection of bacterial dysbiosis would allow for a timely treatment that also circumvents tolerance development. However, the establishment of dysbiosis sensors \u0026ndash; e.g., based on spectroscopy or chemometrics \u0026ndash; requires reproducible \u003cem\u003ein vitro\u003c/em\u003e models to serve as test systems. Oral multispecies biofilm \u003cem\u003ein vitro\u003c/em\u003e models typically contain up to ten characteristic bacterial species that are either sampled from volunteers or commercially available type strains [\u003cspan additionalcitationids=\"CR4\" citationid=\"CR3\" class=\"CitationRef\"\u003e3\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e]. These biofilms are grown on various materials (including implant-grade titanium) for several days or weeks under either static or salivary shear force-mimicking dynamic conditions [6;7]. One example is the Hannoverian Oral Multispecies Biofilm Implant Flow Chamber (HOBIC) model developed in our group, which contains the oral commensals \u003cem\u003eStreptococcus oralis\u003c/em\u003e, \u003cem\u003eActinomyces naeslundii\u003c/em\u003e, \u003cem\u003eVeillonella dispar\u003c/em\u003e as well as the oral pathogen \u003cem\u003ePorphyromonas gingivalis\u003c/em\u003e [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. This four-species biofilm is grown on titanium discs in custom-made flow chambers designed for non-invasive microscopic readout. Within the incubation time of 24 hours, reproducible biofilms of commensal composition are formed. In contrast, Siddiqui et al. have reported on a similar six-species biofilm model on titanium that was cultivated under static conditions for 21 days [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]. Over time, a clear shift in bacterial species composition towards the increase of pathogenic species could be detected. However, this model lacks the naturally existing flow shear forces.\u003c/p\u003e \u003cp\u003eThe aim of the present study was to advance the HOBIC model and reproduce the bacterial dysbiosis associated with peri-implantitis \u003cem\u003ein vitro\u003c/em\u003e. For this purpose, two different five-species combinations were grown comparatively under both static and dynamic conditions over 21 days and analyzed for bacterial growth (optical density and pH development), biofilm morphology (live/dead fluorescence staining with confocal microscopy and digital image analysis) and species distribution (quantitative real-time PCR and fluorescence \u003cem\u003ein situ\u003c/em\u003e hybridization). By this, the research hypotheses that (I) a bacterial shift can be introduced solely by increasing cultivation time as well as that (II) the shift depends on species composition and (III) cultivation conditions were addressed.\u003c/p\u003e"},{"header":"Material and Methods","content":"\u003cp\u003eBacterial strains and culture conditions\u003c/p\u003e \u003cp\u003eBacteria were routinely stored as glycerol stocks at -80\u0026deg;C. \u003cem\u003eVeillonella dispar\u003c/em\u003e DSM 20735 (German Collection of Microorganisms and Cell Cultures GmbH, DSMZ, Braunschweig, Germany), \u003cem\u003eVeillonella parvula\u003c/em\u003e ATCC\u0026reg; 17745\u0026trade; (American Type Culture Collection, ATCC, Manassas, VA, USA), \u003cem\u003eFusobacterium nucleatum\u003c/em\u003e DSM 15643, \u003cem\u003ePorphyromonas gingivalis\u003c/em\u003e DSM 20709 and \u003cem\u003ePorphyromonas gingivalis\u003c/em\u003e ATCC W83, were streaked out on fastidious anaerobe agar (Lab M Ltd., Heywood, UK) plates supplemented with 5% defibrinated sheep blood (Thermo Fisher Scientific Inc., Waltham, MA, USA) and incubated at 37\u0026deg;C under anaerobic conditions, which were achieved using AnaeroGen\u0026trade; bags (Thermo Fisher Scientific Inc.), for three days. Afterwards, colonies from the agar plates were transferred to liquid medium and cultured overnight in brain heart infusion medium (BHI, Oxoid Deutschland GmbH, Wesel, Germany) supplemented with 10 mg/L vitamin K (Carl Roth GmbH\u0026thinsp;+\u0026thinsp;Co. KG, Karlsruhe, Germany) (BHI\u0026thinsp;+\u0026thinsp;VitK, \u003cem\u003eV. dispar/parvula\u003c/em\u003e) or in fastidious anaerobe broth (Lab M Ltd., \u003cem\u003eF. nucleatum\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e) at 37\u0026deg;C under anaerobic conditions. \u003cem\u003eStreptococcus oralis\u003c/em\u003e ATCC 9811\u0026trade; and \u003cem\u003eActinomyces naeslundii\u003c/em\u003e DSM 43013 were cultured overnight in BHI\u0026thinsp;+\u0026thinsp;VitK at 37\u0026deg;C under anaerobic conditions.\u003c/p\u003e \u003cp\u003eStatic biofilm growth in well plates\u003c/p\u003e \u003cp\u003eBacterial overnight cultures were adjusted to an optical density at 600 nm (OD\u003csub\u003e600\u003c/sub\u003e) of 0.05 in BHI\u0026thinsp;+\u0026thinsp;VitK with 5 mg/L hemin (BHI\u0026thinsp;+\u0026thinsp;VitK/Hem, Sigma Aldrich, St. Louis, MO, USA), and mixed in two different five-species combinations: \u003cem\u003eS. oralis\u003c/em\u003e, \u003cem\u003eA. naeslundii\u003c/em\u003e and \u003cem\u003eF. nucleatum\u003c/em\u003e were combined either with \u003cem\u003eV. dispar\u003c/em\u003e and \u003cem\u003eP gingivalis\u003c/em\u003e DSM 20709 (commensal model) or \u003cem\u003eV. parvula\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e ATCC W83 (dysbiotic model). 5 mL per well of the mixed suspension were added to 6-well plates and incubated for 1, 3, 6, 10, 15, or 21 days at 37\u0026deg;C under anaerobic conditions. Every other day, half of the medium was replaced with fresh medium. Before analysis, biofilms were washed once with phosphate buffered saline (PBS, Sigma Aldrich).\u003c/p\u003e \u003cp\u003eBiofilm growth in the adaptive HOBIC model\u003c/p\u003e \u003cp\u003eThe setup of the flow chamber system is shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, and is based on the previously described \u0026ldquo;Hannoverian Oral Multispecies Biofilm Implant Flow Chamber (HOBIC)\u0026rdquo; model [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e] with the following modifications: In this case, pH-sensitive flow-through cells connected to optical fibers (FTC-SU-LG1-S, PreSense Precision Sensing GmbH, Regensburg, Germany) were integrated behind the flow chambers. Grade 4 titanium discs (12 mm diameter, 1.5 mm height, R\u003csub\u003ea\u003c/sub\u003e = 0.31 \u0026micro;m) were submerged in artificial saliva (850 mg/L mucin, 10 \u0026micro;g/mL lysozyme, 1 mg/mL α-amylase, 40 \u0026micro;g/mL albumin) during sterile assembly of the chambers. Then, the chambers were integrated into the system and 2.1 mL per strain with OD\u003csub\u003e600\u003c/sub\u003e of 0.05 were added to 1.5 L BHI\u0026thinsp;+\u0026thinsp;VitK/Hem in the bioreactor. In addition to the previously described HOBIC model [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e], \u003cem\u003eF. nucleatum\u003c/em\u003e was added as fifth bacterium to the inoculum for both combinations and \u003cem\u003eV. dispar\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e\u003csub\u003e\u003cem\u003e20709\u003c/em\u003e\u003c/sub\u003e were replaced by \u003cem\u003eV. parvula\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e\u003csub\u003e\u003cem\u003eW83\u003c/em\u003e\u003c/sub\u003e as \u0026ldquo;dysbiotic\u0026rdquo; strain combination. After 24 hours of cultivation with 100 \u0026micro;L/min at 37\u0026deg;C under anaerobic conditions, the components up to the bubble trap (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA) were replaced with new sterile parts, except for the OD measuring bypass. The system was then run for 20 additional days with sterile 1:2 diluted medium. On day 1, 3, 6, 10, 15, and 21, chambers were washed with PBS for 30 min at 100 \u0026micro;L/min before subjected to further analysis.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eLive/Dead staining and microscopic analysis\u003c/p\u003e \u003cp\u003eStatic and flow chamber biofilms were stained using SYTO\u0026reg;9 and propidium iodide (PI) of the LIVE/DEAD\u0026reg; BacLight\u0026trade; Bacterial Viability Kit (Life Technologies, Darmstadt, Germany) at a concentration of 1:2000 of the stock solutions in PBS, followed by fixation with 2.5% glutardialdehyde. For staining of flow chamber biofilms, the dye and fixation solutions were pumped through the system as previously described [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. Likewise, confocal laser-scanning microscopy was done using established protocols [\u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e]. For the HOBIC systems, from day 6 onwards, images were taken from the cover slip downwards rather than directly on the titanium surface, since the laser could not reach through the thick biofilm. From all images, biofilm volume and live/dead distribution were quantified using the software Imaris (v8.4.1, Bitplane AG, Zurich, Switzerland).\u003c/p\u003e \u003cp\u003ePMA treatment, DNA extraction, and qRT-PCR\u003c/p\u003e \u003cp\u003eAfter microscopy, the biofilm inside the chambers was harvested and either directly frozen or subjected to additional PMA treatment [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e]. Following washing with PBS, bacteria were incubated with 0.2 mM PMAxx\u0026trade; (Biotium, Inc., Fremont, CA, USA) for 10 min at 4\u0026deg;C, and then again for 20 min in the PMA-Lite\u0026trade; LED Photolysis Device (Biotium, Inc.). DNA was extracted using a customized protocol which deploys a combination of enzymatic lysis, mechanical disruption, and column-based DNA isolation. Bacterial sample material was initially treated with 450 \u0026micro;L lysozyme solution (20 mg/mL lysozyme (Merck, Darmstadt, Germany) in 20 mM Tris HCl, pH 8.0; 2 mM EDTA; 1,2% Triton) for two hours at 37\u0026deg;C. After addition of 50 \u0026micro;L Proteinase K and 500 \u0026micro;L AL buffer (both Qiagen, Hilden, Germany), treatment was extended for 30 min at 56\u0026deg;C and 15 min at 95\u0026deg;C. The complete sample material was then transferred to Lysing Matrix E tubes (MP Biomedicals, Eschwege, Germany) and mechanically disrupted in three cycles of 6500 rpm for 30 s in a Precellys 24 homogenizer (Bertin Technologies, Frankfurt am Main, Germany), punctuated with cooling on ice for five minutes in between cycles. Finally, beads and debris were sedimented by centrifugation (5 min, 14000 x g), and the cleared supernatant was mixed 1:1 with 100% ethanol. Subsequent steps were performed with the QIAamp Mini Kit (Qiagen) according to the manufacturer\u0026rsquo;s protocol \u0026ldquo;DNA Purification from Blood or Body Fluids\u0026rdquo; \u0026ndash; starting with the application to the spin columns. To reduce the risk of contamination, a new collection tube was used after each of the kit-specific wash steps. DNA was eluted with 50 \u0026micro;L of PCR-grade water and stored at -20\u0026deg;C until further usage.\u003c/p\u003e \u003cp\u003eQuantitative real-time PCR and calculation of respective cell numbers and relative species distribution were performed as described by Kommerein et al. [\u003cspan citationid=\"CR9\" class=\"CitationRef\"\u003e9\u003c/span\u003e] using the SYBR Green reaction mix (Bio-Rad Laboratories GmbH, Feldkirchen, Germany) and the LightCycler 96 (Roche Holding GmbH, Grenzach-Wyhlen, Germany). Primer pairs, reaction components, cycle conditions and genome weight per cell are all given in the Supplementary Tables S1-4, respectively.\u003c/p\u003e \u003cp\u003eFluorescence-\u003cem\u003eIn-Situ\u003c/em\u003e-Hybridization\u003c/p\u003e \u003cp\u003eTo prepare HOBIC samples for fluorescence \u003cem\u003ein situ\u003c/em\u003e hybridization (FISH), 50% (v/v) ethanol was pumped through the system for 20 min with a flow rate of 250 \u0026micro;L/min, after which the ethanol filled chambers were removed from the system and then stored at 4\u0026deg;C overnight to fixate the bacteria. The chambers were then opened, and the titanium specimen were transferred to a 6-well plate and left to dry under sterile conditions. FISH staining and CLSM analysis were performed as previously reported [8;10]. Briefly, 1 g/L lysozyme (Merck) treatment at 37\u0026deg;C for 10 min was used to disrupt cell membranes. Lysis was stopped with pure ethanol, samples were dried and then stained with six fluorescently labeled 16S rRNA probes (Supplementary Table S5) in hybridization buffer at 46\u0026deg;C for 30 min. \u003cem\u003eF. nucleatum\u003c/em\u003e was targeted by two probes that shared the same nucleotide sequence, but were labeled with different dyes \u0026ndash; resulting in co-localized blue and red fluorescence. Afterwards, samples were washed several times and analyzed by CLSM. A 630-fold magnification was used to take image stacks with an xy-size of 185 x 185 \u0026micro;m\u0026sup2; and a 2 \u0026micro;m z-step-size. Scanning was done sequentially per frame. The first sequence used a 405 nm and a 552 nm laser for excitation, and detected blue and yellow signals in the wavelength ranges 413\u0026ndash;477 nm and 576\u0026ndash;648 nm, respectively. During the second sequence, a 488 nm and a 638 nm laser were used to excite the samples, and emission detection was done in the wavelength ranges 509\u0026ndash;576 nm and 648\u0026ndash;777 nm for green and red signals, respectively.\u003c/p\u003e \u003cp\u003eGingipain-Specific Enzyme-Linked Immunosorbent Assay (ELISA)\u003c/p\u003e \u003cp\u003eSupernatants of the HOBIC model flow chambers were collected, frozen, and used for gingipain protein quantification using the human \u003cem\u003eP. gingivalis-\u003c/em\u003especific IgG antibody ELISA kit (Huangshi INS Biological Technology Co., LTD, Huangshi, China). Analysis was done according to the manufacturer\u0026rsquo;s protocol \u0026ndash; but with the samples being additionally incubated for 1 hour at room temperature followed by 3 washing steps before addition of the detection antibody.\u003c/p\u003e \u003cp\u003eLiterature-Based Metabolic Interaction Prediction\u003c/p\u003e \u003cp\u003eThe potential of biofilm members to engage in metabolic and enzyme-based interspecies interactions was inferred using a custom database [\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e]. Interaction data were sourced from the literature and subjected to manual curation [12;13]. Custom-designed graphs were employed to visualize the interaction networks.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eStatistical Analysis\u003c/h2\u003e \u003cp\u003eData presentation and statistical analysis were done using GraphPad Prism 8.4 (GraphPad Software Inc., San Diego, CA, USA). Statistical test details can be found in the respective figure captions. Family-wise significance level was defined with α\u0026thinsp;=\u0026thinsp;0.05.\u003c/p\u003e \u003c/div\u003e"},{"header":"Results","content":"\u003cp\u003eTime-dependent biofilm growth and viability\u003c/p\u003e \u003cp\u003eInitial bacterial growth in the bioreactor was monitored using inline optical density measurement, and showed typical bacterial growth curves with significantly increased growth of the dysbiotic model (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eA, B). Subsequent biofilm growth on titanium discs was then analyzed by fluorescence staining and confocal microscopy. Within both commensal models, the biofilm volume significantly decreased after one day, and then re-established until day 10 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, Supplementary Table S6). This development was more pronounced in the commensal HOBIC model. In contrast, the biofilm volume of both dysbiotic models significantly increased after day 1, reaching a plateau at day 6 (HOBIC system) and day 15 (static system), respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, Supplementary Table S6). Biofilm viability \u0026ndash; analyzed by fluorescence-based membrane integrity \u0026ndash; was observed to significantly decrease over time for all cultivation conditions except for the static commensal model (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eA, B, Supplementary Table S7). Viability development thereby replicates the respective biofilm volume pattern.\u003c/p\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003eCultivation condition-dependent species composition\u003c/p\u003e \u003cp\u003eTime-dependent species composition (both, viable and total count) was analyzed by DNA isolation and qRT-PCR as well as FISH staining, and these analyses revealed clear differences between the commensal and dysbiotic model (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, B, Supplementary Figure S1). The changes in bacterial species distribution were more pronounced for viable cells (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) than for the total count (Supplementary Figure S1). Within the commensal HOBIC model, viable \u003cem\u003eS. oralis\u003c/em\u003e was the initially dominant species, although its amount significantly decreased over time from approx. 70% to merely 35% (Supplementary Table S8). Within the commensal static model, this decrease was also observed \u0026ndash; but only from day 6. Initially, \u003cem\u003eV. dispar\u003c/em\u003e was the dominant species, but was then gradually replaced by \u003cem\u003eS. oralis\u003c/em\u003e up until day 6. Afterwards, \u003cem\u003eV. dispar\u003c/em\u003e\u0026rsquo;s distribution remained on average stable at 30% within both commensal models. In parallel, \u003cem\u003eA. naeslundii\u003c/em\u003e established itself with prolonged incubation to approx. 30%. \u003cem\u003eF. nucelatum\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e were almost undetectable in both commensal models. In contrast, total and viable species distribution of the dysbiotic model differed remarkably (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eA, B, Supplementary Figure S1). For the static system, \u003cem\u003eV. parvula\u003c/em\u003e remained the dominant species independent of incubation time (50\u0026ndash;60%), followed by \u003cem\u003eP. gingivalis\u003c/em\u003e (25%), \u003cem\u003eF. nucleatum\u003c/em\u003e (20%), and only a very low amount \u003cem\u003eS. oralis\u003c/em\u003e and \u003cem\u003eA. naeslundii\u003c/em\u003e. In the HOBIC system, the initially dominant \u003cem\u003eV. parvula\u003c/em\u003e significantly decreased from approx. 95\u0026ndash;20% (Supplementary Table S8), while \u003cem\u003eA. naeslundii\u003c/em\u003e, \u003cem\u003eF. nucleatum\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e successively increased from day 6, 10, and 21, respectively, up to 15\u0026ndash;30% (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eB).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSpecies composition-dependent biofilm metabolism\u003c/p\u003e \u003cp\u003eWithin the commensal and dysbiotic HOBIC models, pH-values and \u003cem\u003eP. gingivalis\u003c/em\u003e gingipain protein concentration over time were determined by optical fiber measurement and ELISA, respectively, and showed clear differences (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eA, B). For the commensal model, pH-values initially dropped below pH 6.0, sharply increased upon medium change at day 1, and then established itself at approx. pH 6.3. In contrast, for the dysbiotic model, pH-values only dropped to pH 6.3 and then gradually increased to pH 6.9 by day 21. These higher pH-values showed a tendency to explicitly increase the growth of the dysbiotic model\u0026rsquo;s \u003cem\u003eP. gingivalis\u003c/em\u003e strain (Supplementary Figure S2A). In line with these observations, the amount of gingipain protein significantly increased over time only within the dysbiotic model (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eB). To account for these differences, literature-based prediction of metabolic interactions between the six bacterial species was performed (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eC). Based on the available nutrients from the culture medium, the species were found to engage in multiple food chain and enzyme sharing behaviors, with peptides, glucose, vitamins and other growth factors produced by \u003cem\u003eS. oralis\u003c/em\u003e, \u003cem\u003eA. naeslundii\u003c/em\u003e and \u003cem\u003eV. dispar/parvula\u003c/em\u003e and then utilized by \u003cem\u003eF. nucleatum\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e. Whereas \u003cem\u003eV. dispar\u003c/em\u003e and \u003cem\u003eV. parvula\u003c/em\u003e generally exhibited similar metabolisms, only \u003cem\u003eV. parvula\u003c/em\u003e was capable of \u003cem\u003ede novo\u003c/em\u003e thiamine (vitamin B1) synthesis. However, although the dysbiotic model\u0026rsquo;s P. gingivalis strain showed increased overall growth compared to the commensal model\u0026rsquo;s strain, no growth difference in Veillonella-preconditioned medium could be detected (Supplementary Figure S2B).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eEarly detection of oral biofilm dysbiosis on dental implants can prevent the development of severe infections like peri-implantitis. To help establish dysbiosis sensors, however, reliable \u003cem\u003ein vitro\u003c/em\u003e models must first be developed for use as test systems. Within the present study, the existing HOBIC model was successfully adapted to reproduce bacterial dysbiosis for this purpose. With regard to the initial research hypotheses, this model helped to confirm that the dysbiotic shift depended both on the selected bacterial species as well as on the cultivation conditions.\u003c/p\u003e \u003cp\u003eThe bacterial species selected for this study were \u003cem\u003eS. oralis, A. naeslundii, V. dispar\u003c/em\u003e or \u003cem\u003eV. parvula, F. nucleatum\u003c/em\u003e, and \u003cem\u003eP. gingivalis.\u003c/em\u003e All genera but \u003cem\u003eFusobacterium\u003c/em\u003e were already included in the previous HOBIC model [8;9]. \u003cem\u003eS. oralis\u003c/em\u003e and \u003cem\u003eA. naeslundii\u003c/em\u003e are among the dominant primary oral colonizers associated with oral health [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. \u003cem\u003eV. dispar\u003c/em\u003e and \u003cem\u003eV. parvula\u003c/em\u003e are part of the core microbiome, co-aggregate with \u003cem\u003eS. oralis\u003c/em\u003e and \u003cem\u003eA. naeslundii\u003c/em\u003e, and metabolize lactate produced by them [2;14]. \u003cem\u003eP. gingivalis\u003c/em\u003e is part of the red complex bacteria and associated with periodontitis and peri-implantitis [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. As the biofilm of the initial HOBIC model mainly consisted of \u003cem\u003eS. oralis, A. naeslundii\u003c/em\u003e, and \u003cem\u003eV. dispar\u003c/em\u003e, and contained only a small proportion of \u003cem\u003eP. gingivalis\u003c/em\u003e, it depicts an early commensal oral biofilm [8;9]. To support the dysbiotic shift, \u003cem\u003eF. nucleatum\u003c/em\u003e was added as fifth bacterium in this study. The species is considered a central bridging bacterium between commensal and dysbiotic strains as it co-aggregates with bacteria of both groups [\u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2\u003c/span\u003e]. The species selection closely matched those of other commensal and dysbiotic oral biofilm models [5\u0026ndash;7;15]. They can therefore be considered to collectively constitute a reliable representation of the oral microbiome. However, since minor differences in species selection in other studies lead to both similar and different results [6;15], it bears noting that all models remain simplifications of the natural microbiome complexity and all results generated by models should therefore be interpreted with this limitation in mind.\u003c/p\u003e \u003cp\u003eBiofilm growth of the selected bacteria was done in full medium supplemented with vitamin K and hemin to support \u003cem\u003eP. gingivalis\u003c/em\u003e growth (according to preliminary experiments and other studies [\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e]) with all species inoculated once at the same time. In addition, titanium surfaces were pre-conditioned using artificial saliva. Within this experimental setup, the dysbiotic shift was induced by intrinsic bacterial interactions only, closely replicating natural oral conditions. In contrast, some other models induced dysbiosis by changing medium and oxygen level, as well as inoculating additional bacterial species [4;16]. Whereas these approaches offer the possibility of external control, the setup selected here is inherently more beneficial for observing intrinsic processes \u0026ndash; a pre-requirement for the development of dysbiosis sensors.\u003c/p\u003e \u003cp\u003eOver the incubation time of 21 days, biofilm volume increased for the dysbiotic model only, reaching a plateau after 6 days. This growth pattern has also been observed for other models with a similar dysbiotic species composition and cultivation time of more than 7 days, as well as for \u003cem\u003ein situ\u003c/em\u003e grown biofilms on implant healing abutments [3;6;11]. In comparison to the commensal model, the initial planktonic growth in the bioreactor of the HOBIC model, and the growth of \u003cem\u003eV. parvula\u003c/em\u003e and \u003cem\u003eP. gingivalis\u003c/em\u003e\u003csub\u003e\u003cem\u003eW83\u003c/em\u003e\u003c/sub\u003e at different pH (Supplementary Figure S2A) was found to be significantly higher. This makes the increased growth of these individual species the most likely reason for the elevated biofilm volume of the dysbiotic model. The parallel decrease of biofilm viability independently of the species composition is also in line with previous \u003cem\u003ein vitro\u003c/em\u003e and \u003cem\u003ein situ\u003c/em\u003e results [5;7;11], and most likely due to limited nutrient availability in deeper layers of the maturated biofilm.\u003c/p\u003e \u003cp\u003eThe most obvious difference between the commensal and dysbiotic models lies in their divergent species composition. During the first days, the commensal model was dominated by \u003cem\u003eS. oralis\u003c/em\u003e after a short initial establishing phase. In contrast, the dysbiotic model was initially dominated by \u003cem\u003eV. parvula\u003c/em\u003e. Differing proportions of \u003cem\u003eS. oralis\u003c/em\u003e and \u003cem\u003eV. parvula\u003c/em\u003e have already been described in previous oral biofilm models \u0026ndash; with studies showing an initial dominance of \u003cem\u003eS. oralis\u003c/em\u003e [\u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e5\u003c/span\u003e], an initial dominance of \u003cem\u003eV. parvula\u003c/em\u003e [\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e] or equal amounts of both species [3;6]. Their establishment is probably significantly influenced by the (pre-culture) cultivation conditions; however, the details of these alterations remain to be analyzed in future studies. The different proportions of \u003cem\u003eS. oralis\u003c/em\u003e are also the most probable reason for the differences in pH-values measured in the HOBIC model. As shown in the predicted metabolic interactions, \u003cem\u003eS. oralis\u003c/em\u003e utilizes carbohydrates from the medium to produce lactate via fermentation, causing strong medium acidification. In contrast, Veillonellaceae metabolize nutrients (including lactate) to less acidic acetate and propionate, resulting in higher pH-values in the medium.\u003c/p\u003e \u003cp\u003eWith prolonged cultivation time, a diversification of commensal species was observed in the commensal model over time. In contrast, the dysbiotic model showed a notable increase in pathogenic species with reduced proportions of commensal strains. This observation is further supported by the increase in gingipain proteins, which are trypsin-like cysteine proteases that are among the most important \u003cem\u003eP\u003c/em\u003e. \u003cem\u003egingivalis\u003c/em\u003e virulence factors [\u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]. Since these models only differed in the \u003cem\u003eVeillonella\u003c/em\u003e species and the \u003cem\u003eP. gingivalis\u003c/em\u003e strain, their contribution seems to be crucial for the dysbiotic shift \u0026ndash; at least within the limited setting of an \u003cem\u003ein vitro\u003c/em\u003e experiment. Further evidence to support this observation can also be found in the literature: In a co-association study of \u003cem\u003eStreptococcus mutans, Veillonella dispar\u003c/em\u003e and \u003cem\u003eVeillonella parvula\u003c/em\u003e within the context of root caries, only \u003cem\u003eV. parvula\u003c/em\u003e was found to support \u003cem\u003eS. mutans\u003c/em\u003e growth \u003cem\u003ein vitro\u003c/em\u003e [\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]. More importantly, a metatranscriptomic analysis of different oral \u003cem\u003ein vitro\u003c/em\u003e biofilm models found that \u003cem\u003eP. gingivalis\u003c/em\u003e\u003csub\u003e\u003cem\u003eW83\u003c/em\u003e\u003c/sub\u003e \u0026ndash; a virulent strain which was also used for the dysbiotic model of this study \u0026ndash; had a greater effect on biofilm dysbiosis than a lower virulent type strain by specifically influencing genes related to metabolic pathways and quorum sensing of several commensal species [\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]. An effect on the strain level would also be supported by the species-level metabolic interaction prediction of this study, where only a minor difference in thiamine (vitamin B1) production solely by \u003cem\u003eV. parvula\u003c/em\u003e could be identified. Even though \u003cem\u003eP. gingivalis\u003c/em\u003e utilizes vitamins produced by Veillonellaceae, they exhibit the enzymes for thiamine metabolism themselves according to KEGG-pathway analysis, and thus probably would not rely on cross-feeding by \u003cem\u003eV. parvula\u003c/em\u003e. On the other hand, although \u003cem\u003eP. gingivalis\u003c/em\u003e\u003csub\u003e\u003cem\u003eW83\u003c/em\u003e\u003c/sub\u003e showed increased growth compared to \u003cem\u003eP. gingivalis\u003c/em\u003e\u003csub\u003e\u003cem\u003e20709\u003c/em\u003e\u003c/sub\u003e, Veillonella-conditioned medium had no different effect on both strains. In summary, these results strongly underscore the importance of further analyzing metabolic interaction within oral microbial communities that seem to significantly contribute to dysbiosis development within basic research approaches. The models developed here can be used for initial insights as well as validation for this purpose.\u003c/p\u003e \u003cp\u003eAside from strain selection, cultivation conditions also influenced bacterial species composition \u0026ndash; albeit to a more minor extent. For both the commensal and dysbiotic model, changes in species composition were more pronounced in the HOBIC than the static system. During static cultivation (even though medium was exchanged every other day), metabolites accumulated within the biofilm, probably making metabolite-based changes of species composition slower or less pronounced. In contrast, during cultivation in the dynamic HOBIC system, medium is constantly replaced, thus preventing the metabolites from accumulating to a greater degree which might very well induce changes in species composition.\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eWithin the present study, different \u003cem\u003ein vitro\u003c/em\u003e oral multispecies biofilm models were successfully developed. Depending on bacterial species selection, these models were able to depict the infection-associated dysbiotic shift in species composition solely by intrinsic interactions, and without any deployment of or reference to external stimuli. The different results between cultivation conditions offer the possibility for a number of different future application: For the direct comparison between commensal and dysbiotic biofilms, straightforward static cultivation can be used with species composition being already different after 24 hours. In contrast, for the observation of the bacterial shift over time (for example by novel sensor systems), the dysbiotic HOBIC model is to be preferred. The results of this study also point towards the (current) limitations of \u003cem\u003ein vitro\u003c/em\u003e models and the need for further \u003cem\u003ein vivo\u003c/em\u003e studies that focus on examining how metabolic interactions on the strain level influence bacterial species composition. For the validation of these \u003cem\u003ein vivo\u003c/em\u003e observations, we believe that the presented biofilm models will serve as a valuable tool.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003eAcknowledgements\u003c/p\u003e\n\u003cp\u003eThis work was funded by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under the Collaborative Research Center SFB/TRR-298-SIIRI \u0026ndash; Project ID 426335750. The authors would like to additionally thank Wiebke Smolinski, Hanna Lena Thoms and Teresa Lea Ngyuen for excellent technical assistance and Dr. Andreas Winkel for providing the BioRender image of Figure 1A.\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e\n\u003cp\u003eAuthor contribution\u003c/p\u003e\n\u003cp\u003eConceptualization: AH, TS, MS, KDN; Methodology: NH, WB, CM, NK, MS, KDN; Investigation: NH, KB, SPS, MD, CV, ND, KF, KDN; Formal Analysis: NH, KDN; Writing \u0026ndash; Original Draft Preparation: NH, SPS, WB, KDN; Writing \u0026ndash; Review \u0026amp; Editing: NH, SPS, KB, SPS, MD, WB, CV, CM, NK, ND, KF, AH, TS, MLTM, JB, MS; Supervision: CM, NK, AH, TS, MLTM, JB, MS, KDN; Funding Acquisition: AH, TS, MS, KDN.\u003c/p\u003e\n\u003cp\u003eConflict of interest\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e\n\u003cp\u003eData availability\u003c/p\u003e\n\u003cp\u003eThe data are available upon request from the corresponding authors.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBelibasakis GN (2014) Microbiological and immuno-pathological aspects of peri-implant diseases. Arch Oral Biol 59:66\u0026ndash;72. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.archoralbio.2013.09.013\u003c/span\u003e\u003cspan address=\"10.1016/j.archoralbio.2013.09.013\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eColombo APV, Tanner ACR (2019) The role of bacterial biofilms in dental caries and periodontal and peri-implant diseases: A historical perspective. J Dent Res 98:373\u0026ndash;385. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1177/0022034519830686\u003c/span\u003e\u003cspan address=\"10.1177/0022034519830686\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThurnheer T, Bostanci N, Belibasakis GN (2016) Microbial dynamics during conversion from supragingival to subgingival biofilms in an in vitro model. Mol Oral Microbiol 31:125\u0026ndash;135. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/omi.12108\u003c/span\u003e\u003cspan address=\"10.1111/omi.12108\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDalwai F, Spratt DA, Pratten J (2006) Modeling shifts in microbial populations associated with health or disease. Appl Environ Microbiol 72:3678\u0026ndash;3684. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1128/AEM.72.5.3678-3684.2006\u003c/span\u003e\u003cspan address=\"10.1128/AEM.72.5.3678-3684.2006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSanchez MC, Llama-Palacios A, Blanc V, Leon R, Herrera D, Sanz M (2011) Structure, viability and bacterial kinetics of an in vitro biofilm model using six bacteria from the subgingival microbiota. J Periodontal Res 46:252\u0026ndash;260\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSiddiqui DA, Fidai AB, Natarajan SG, Rodrigues DC (2022) Succession of oral bacterial colonizers on dental implant materials: An in vitro biofilm model. Dent Mater 38:384\u0026ndash;396. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1016/j.dental.2021.12.021\u003c/span\u003e\u003cspan address=\"10.1016/j.dental.2021.12.021\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlanc V, Isabal S, Sanchez MC, Llama-Palacios A, Herrera D, Sanz M et al (2014) Characterization and application of a flow system for in vitro multispecies oral biofilm formation. J Periodontal Res 49:323\u0026ndash;332\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKommerein N, Doll K, Stumpp NS, Stiesch M (2018) Development and characterization of an oral multispecies biofilm implant flow chamber model. PLoS ONE 13:e0196967\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKommerein N, Stumpp SN, Musken M, Ehlert N, Winkel A, Haussler S et al (2017) An oral multispecies biofilm model for high content screening applications. PLoS ONE 12:e0173973\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDebener N, Heine N, Legutko B, Denkena B, Prasanthan V, Frings K et al (2024) Optically accessible, 3D-printed flow chamber with integrated sensors for the monitoring of oral multispecies biofilm growth in vitro. Front Bioeng Biotechnol 12:1483200. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.3389/fbioe.2024.1483200\u003c/span\u003e\u003cspan address=\"10.3389/fbioe.2024.1483200\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDieckow S, Szafranski SP, Grischke J, Qu T, Doll-Nikutta K, Steglich M et al (2024) Structure and composition of early biofilms formed on dental implants are complex, diverse, subject-specific and dynamic. NPJ Biofilms Microbiomes 10:155\u0026ndash;153. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1038/s41522-024-00624-3\u003c/span\u003e\u003cspan address=\"10.1038/s41522-024-00624-3\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGiacomini JJ, Torres-Morales J, Dewhirst FE, Borisy GG, Mark Welch JL (2023) Site specialization of human oral veillonella species. Microbiol Spectr 11:e0404222\u0026ndash;e0404222 Epub 2023 Jan 25. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1128/spectrum.04042-22\u003c/span\u003e\u003cspan address=\"10.1128/spectrum.04042-22\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoodfellow M, K\u0026auml;mpfer P, Busse H, Trujillo ME, Suzuki et al (2012) \u0026amp;., Ken-ichiro,. \u003cem\u003eBergey's manual of systematic bacteriology\u003c/em\u003e (3\u0026ndash;5 ed.) Springer New York. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003erg/10.1007/978-0-387-68233-4\u003c/span\u003e\u003cspan address=\"rg/10.1007/978-0-387-68233-4\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKolenbrander PE (2011) Multispecies communities: Interspecies interactions influence growth on saliva as sole nutritional source. Int J Oral Sci 3:49\u0026ndash;54\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang Y, Shi W, Song Y, Wang J (2019) Metatranscriptomic analysis of an in vitro biofilm model reveals strain-specific interactions among multiple bacterial species. J Oral Microbiol 11:1599670. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1080/20002297.2019.1599670\u003c/span\u003e\u003cspan address=\"10.1080/20002297.2019.1599670\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eThurnheer T, Gmur R, Guggenheim B (2004) Multiplex FISH analysis of a six-species bacterial biofilm. J Microbiol Methods 56:37\u0026ndash;47\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGuo Y, Nguyen K, PotemJan (2010) Dichotomy of gingipains action as virulence factors: From cleaving substrates with the precision of a surgeon's knife to a meat chopper-like brutal degradation of proteins. Periodontol 2000 54:15\u0026ndash;44. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1111/j.1600-0757.2010.00377.x\u003c/span\u003e\u003cspan address=\"10.1111/j.1600-0757.2010.00377.x\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAbram AM, Szewczyk MM, Park SG, Sam SS, Eldana HB, Koria FJ et al (2022) A co-association of streptococcus mutans and veillonella parvula/dispar in root caries patients and in vitro biofilms. Infect Immun 90:e0035522\u0026ndash;e0035522 Epub 2022 Sep 21. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003e10.1128/iai.00355\u0026thinsp;\u0026ndash;\u0026thinsp;22\u003c/span\u003e\u003cspan address=\"10.1128/iai.00355\u0026thinsp;\u0026ndash;\u0026thinsp;22\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[{"identity":"956d861a-c8e8-4cda-b126-4166713da174","identifier":"10.13039/501100001659","name":"Deutsche Forschungsgemeinschaft","awardNumber":"426335750","order_by":0}],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"Department of Dental Prosthetics and Biomedical Materials Science, Hannover Medical School","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"dental plaque, dysbiosis, dental implants, microbiological techniques, dynamic cultivation","lastPublishedDoi":"10.21203/rs.3.rs-6573834/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6573834/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003ch2\u003eBackground\u003c/h2\u003e \u003cp\u003eChanges in bacterial species composition within oral biofilms, known as biofilm dysbiosis, are associated with the development of severe oral diseases. To better understand this process and help establish early detection systems, models are needed which replicate oral biofilm dysbiosis \u003cem\u003ein vitro\u003c/em\u003e \u0026ndash; ideally by also mimicking natural salivary flow conditions.\u003c/p\u003e\u003ch2\u003eMethods\u003c/h2\u003e \u003cp\u003eFor this purpose, the present study cultivated two different combinations of oral commensal and pathogenic strains \u0026ndash; \u003cem\u003eStreptococcus oralis, Actinomyces naeslundii, Veillonella dispar/parvula, Fusobacterium nucleatum\u003c/em\u003e and \u003cem\u003ePorphyromonas gingivalis\u003c/em\u003e \u0026ndash; comparatively within an established flow chamber model on the implant material titanium, and statically in 6-well plates for 21 days. Biofilm morphology, species distribution, and bacterial metabolism were analyzed by fluorescence microscopy, molecular biological methods, and metabolic interaction prediction.\u003c/p\u003e\u003ch2\u003eResults\u003c/h2\u003e \u003cp\u003eBiofilm growth and composition were strongly influenced by bacterial species selection, and to a more minor extent, by cultivation conditions. Within the model containing \u003cem\u003eV. dispar\u003c/em\u003e and a laboratory \u003cem\u003eP. gingivalis\u003c/em\u003e strain, a diversification of commensal species was observed over time along with a significantly reduced pH-value. In contrast, the model containing \u003cem\u003eV. parvula\u003c/em\u003e and the clinical isolate \u003cem\u003eP. gingivalis\u003c/em\u003e W83, a dysbiotic shift with increased pathogen levels, pH-value, and virulence factors was achieved.\u003c/p\u003e\u003ch2\u003eConclusion\u003c/h2\u003e \u003cp\u003eWithin the present study, different \u003cem\u003ein vitro\u003c/em\u003e oral multispecies biofilm models were successfully developed. Depending on bacterial species selection, these models were able to depict the infection-associated dysbiotic shift in species composition under flow conditions solely by intrinsic interactions and without the use of external stimuli.\u003c/p\u003e","manuscriptTitle":"Influence of species composition and cultivation condition on peri-implant biofilm dysbiosis in vitro","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-05-05 10:57:30","doi":"10.21203/rs.3.rs-6573834/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"efc4f302-f4ea-4123-a04b-b9f04efba47c","owner":[],"postedDate":"May 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[],"tags":[],"updatedAt":"2025-05-05T10:57:30+00:00","versionOfRecord":[],"versionCreatedAt":"2025-05-05 10:57:30","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6573834","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6573834","identity":"rs-6573834","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}