A long-term in vitro 3D gingival tissue model to investigate host-pathogen interactions.

Introduction:: Oral host-pathogen interactions are known to be fundamental in maintaining the local and systemic health of the human body, although they are challenging to study in both clinical and in vitro applications. Particularly, in vitro, despite the efforts, decoding host-pathogen interactions cannot be achieved by traditional tissue engineering strategies (i.e., 2D or 3D planar models). Currently, the main setbacks are due to the difficulties of 1) recapitulating the anatomical architecture and physiological features of the host-pathogen niche and 2) co-culturing mammalian cells and microbes for long-term studies. Rather, the success of an in vitro engineered host-microbiome interface lies in the cyto-anatomical features (i.e., cells, oxygen, and pH gradients) and microenvironmental signals (i.e., salivary flow, drinking, chewing) that force both players to constantly adapt and shape the oral physiology. Therefore, there is a need for in vitro models that can support long-term investigations of host-pathogen homeostasis and imbalances, with the goal of understanding disease trajectories (i.e., gingivitis or periodontitis) and developing intervention methods to restore periodontal health, especially given the prevalence of the disease- 45-50% in the U.S. population. The aim of the proposed work was to recapitulate the physiological oral conditions (human cells and microbiome, oxygen gradients, physiological shear stress, saliva buffering) to study long-term host-bacterial interactions in vitro and assess the tissue response to therapeutic exposure.


Materials and Methods:: The 3D dynamic gingival tissue model consisted of a customized bioreactor (chamber, peristaltic pump, and pH and O₂ sensors), a biocompatible scaffold, dental-resin teeth, human gingival cells, and artificial saliva. The host-pathogen niche was mimicked by casting a replica mold of an adult human mandibular (lower jaw) gingiva, employing silk protein from B. mori silkworm’s cocoons, to mirror the healthy depth (0.69 mm) of the human dentogingival junction. After scaffold fabrication, primary human stromal, embedded in Collagen I, and epithelial gingival cells were seeded in the anatomical construct and cultured in an optimized artificial saliva formulation for seven days to ensure tissue maturation. Subsequently, healthy patient-derived microbiomes isolated from subgingival plaque (h-plaque d0) (N-Patients=8) were inoculated into the gingival pocket, and the complete construct was then cultured under dynamic regime (1 mL/min) for another week (Figure). During culture, oral hygiene was simulated via rinsing with alcohol-free mouthwash for 10 seconds on day 3 and day 5 post-microbiome inoculum. Host-pathogen interactions were assessed via molecular, morphological, and functional readouts (lactate dehydrogenase-LDH, transepithelial electrical resistance-TEER, Scanning Electron Microscopy-SEM, pro/anti-inflammatory response- cytokine/chemokine Magnetic Bead Panel (MilliPlex), anti-microbial peptide (hBD-2) release via enzyme-linked immunosorbent assay (ELISA), continuous O₂ and end-point pH monitoring, microbial live/dead assay (Syto9/Propidium Iodide (PI)) and 16S rDNA gene sequencing- V1-V3 regions). Statistics: Student’s T-Test, Two-Way ANOVA Repetitive Measurements, Kruskal-Wallis H Test (*p< 0.05). N=5.


Results, Conclusions, and Discussions:: To support our claim of long-term, sustained host-pathogen interactions within the model, we performed molecular and functional oral homeostasis assessments for both the host and microbiome. Compared with a not-inoculated anatomical model, the functional barrier (TEER) was maintained and LDH (cytotoxicity) trends were similar, demonstrating the resistance of epithelial barrier to microbial challenge. After host assessments, we explored both host-microbial interactions and polymicrobial dynamics within the microbiome. A major feature of the proposed model is the high-fidelity replication of the anatomical geometry of the sulcus, which determines the formation of a vertical oxygen gradient that allows the survival of both aerobic and anaerobic species. SEM micrographs of the upper, medium, and lower regions of the construct, which were previously identified through oxygen content, confirmed the attachment and the integration of the microbial communities to the model. Moreover, SEM validated the presence of epithelium in the construct, which appeared still stratified, thus corroborating the TEER results. The viability of the microbiome, instead, was proven by live/dead staining, which showed low PI expression in all anatomical regions. Lastly, we performed bacterial sequencing to characterize which species were under or over-represented in the model. Comparison of 16S rDNA between the original sample (h-plaque d0) and the microbiome grown in the anatomical model (d7-regions combined) showed that the model continued to support the viability and growth of the microbiome, maintaining reasonable diversity and richness within the community, although other characteristics of microbiome (Relative Abundance, Venn analysis and a and b-diversity indices) were decreased. Specifically, the analysis (cladogram) showed that the model was able to preserve the population of commensal bacterial taxa, as Gemella morbillorum, and colonizers, as Streptococcus salivarius, which are critical in the regulation of immune homeostasis and the host inflammatory response (Figure). In conclusion, we have developed a dynamic tissue model that mimics in vivo gingival features and can support long-term in vitro cultures of host and oral polymicrobial communities. This model will be useful as a platform to study early events (dysbiosis) of host-pathogen interactions and evaluate tissue response to therapeutic exposure.


Acknowledgements (Optional): :
NIH NIDCR R03DE030224, and ORAU.


References (Optional): :