High resolution profiling of oral biofilms: from toolbox development to novel treatment analysis for periodontitis

Periodontitis is a multifactorial chronic inflammatory disease in the oral cavity associated with the accumulation of microorganisms in a biofilm. Studying these complex communities under controlled conditions require in vitro biofilm model systems that mimic the natural environment as close as possible. The aim of this research project is the establishment of a multispecies periodontal model that is representative for periodontitis, mimics the continuous flow of nutrients at the air-liquid interface in the oral cavity, and can be used to generate mechanistic understanding of novel treatment approaches (particularly pre- and probiotics) by applying spatiotemporal profiling.

The multispecies periodontal community consists of five key species associated to periodontitis of which four are fluorescently labeled; Streptococcus gordonii GFPmut3*, Streptococcus oralis GFPmut3*, Streptococcus sanguinis pVMCherry, Fusobacterium nucleatum, and Porphyromonas gingivalis SNAP26. They were characterized in individual and coculture bioreactor experiments in terms of growth and metabolism. Significant effort was spent on the establishment of a reliable analysis for free and protein-bound amino acids as well as hydrogen peroxide production potential. All data were used to create an interaction map pointing out competitive, co-operative and inhibitory interactions between the community members. Interspecies interactions were reflected in a continuous bioreactor culture, reaching quasi steady state concentrations for all community members, except for Porphyromonas gingivalis SNAP26 due to competition for peptides.

The periodontal community was grown in an re-engineered drip flow biofilm reactor in order to mimic the continuous flow of nutrients at the air-liquid interface in the oral cavity. The design enabled real-time characterization. The biofilm in the reactor developed into a heterogeneous, spatially uniform, dense and metabolically active biofilm with relative cell abundances similar to those in a healthy individual. Metabolic activity, structural features and bacterial composition of the biofilm remained stable from 3 to 6 days. As a proof of concept for our periodontal model, the 3-day grown, steady biofilm was exposed to a prebiotic treatment with L-arginine. Multifaceted effects of L-arginine on the oral biofilm were validated in this model setup. Additionally, L-arginine was only temporarily exposed to the steady biofilm as well as supplemented to an early-developing biofilm. Although L-arginine consistently inhibited growth and incorporation of the pathogenic species, biofilm thickness and volume reductions were only observed for treatment on steady biofilms, and the biofilm quickly recovered after treatment. As such, we showed that the impact of L-arginine on the periodontal biofilm is only temporal and dependent on a complex interplay between the state of the biofilm and the available nutrients.

Particle tracking is a more novel microrheological tool in biofilm research and was used in this work to characterize the mechanical structure of the biofilm and the impact of L-arginine on it. In the drip flow biofilm reactor, the periodontal biofilm developed quickly into a rigid, dense biofilm, characterized by a large abundance of immobile particles. Although L-arginine seem to have the potential to destabilize the mechanics and reduce amounts of extracellular polymeric substances, effects were minimal and thus conflicting with existing literature. This was likely caused by the formation of a biofilm under shear stress, inducing stiff biofilms which are more difficult to intervene with. As such, this work stressed the importance of working with flow systems when mechanics matter.

To extend the pool of characterization techniques, a high-throughput flow cytometry enumeration approach was elaborated on. In self-assembled and real community samples, the combination of flow cytometry and CellScanner accurately predicted individual bacterial numbers in the community. By using sonication to break up bacterial chains, the quantification results for the streptococcal strains were significantly improved. Since sonication had no impact on the fluorescence characteristics of the streptococcal strains, we proposed a workflow to combine sonication, flow cytometry and CellScanner for future community experiments.

Overall, this work showed the importance and relevance of the established periodontal biofilm model in studying dynamic treatment responses. A set of characterization techniques was translated to the model setup to provide spatiotemporal information. The combination of the model and characterization toolbox can be further exploited to increase the general understanding of periodontitis and prove the effectiveness of novel treatment approaches, such as pre- and probiotics, that focus on maintaining a healthy state.