Periodontitis is a leading cause of tooth loss worldwide. The Gram-negative anaerobe, Porphyromonas gingivalis, has been implicated in the initiation and cyclical progression of this inflammatory disease, which may be associated with its ability to invade oral epithelial cells. The majority of studies investigating P. gingivalis invasion have utilised monolayer cultures of epithelial cells. However, these do not represent the oral mucosa due to the lack of a multi-layered epithelium and fibroblast-embedded connective tissue. Therefore, a fibroblast-containing, connective-tissue collagen scaffold was used to create three-dimensional oral mucosal models (OMM). These were constructed using oral fibroblasts and either the oral keratinocyte cell line (H357) or normal oral keratinocytes (NOK) isolated from healthy patients. OMM were raised to the air-to-liquid interface allowing keratinocyte stratification and differentiation (gingival/buccal OMM) or completely submerged resulting in epithelium consisting of 2-3 cell layers (junctional epithelial OMM). Both models resembled normal oral tissue in terms of immunohistochemical staining for several cytokeratin markers, laminin 5 and E-cadherin.
A standard antibiotic protection assay was optimised for OMM and percentage invasion was shown to be similar to that of monolayer cultures. The optimal method was an incubation period of 3-6 hours of OMM with P. gingivalis in an aerobic atmosphere and release of intracellular P. gingivalis by homogenisation. Using these optimised conditions, a range of parameters of P. gingivalis invasion were investigated.
At diseased periodontal sites there is an increase in the level of haemin and pocket temperature due to inflammation. The culture of P. gingivalis in both a haemin-rich and high temperature environment resulted in an increase in invasion, suggesting that active periodontal sites may preferentially support bacterial internalisation. Additionally, it was shown that following invasion, P. gingivalis can leave epithelial cells after as little as three hours, which may contribute to the periods of progression and remission commonly observed with this disease. Furthermore, the concentration of environmental haemin has previously been shown to influence the expression of P. gingivalis gingipains and it was thought that this may also influence invasion. Indeed, percentage invasion was shown to increase with loss of gingipain activity, particularly Arg-gingipain. This suggested that the degradation of epithelial cell receptors by gingipains may contribute to a decrease in the ability of this bacterium to invade. Candidate host receptors were the complement receptor CD46, tetraspanin family members and the integrin α5β1. These receptors were blocked using antibodies or cells transfected with siRNA to inhibit their function. A small effect on invasion was seen using anti-α5β1 but the antibodies to other molecules did not influence the invasion of P. gingivalis suggesting that there may be some redundancy in the uptake system exploited by the bacteria.
Finally, the response of epithelial cells to invasion by P. gingivalis in terms of cytokine release and expression was determined. Using a semi-quantitative cytokine array, there was a decrease in the majority of cytokines tested in the presence of P. gingivalis when compared with TNF-stimulated control cells which was assumed to be due to the proteolytic action of P. gingivalis gingipains. Due to the conflicting nature of the literature regarding the modulation of CXCL8 by P. gingivalis, this chemokine was selected for further quantification using monolayer cultures. ELISA and quantitative PCR indicated that, in the presence of P. gingivalis, CXCL8 protein concentration decreased in a gingipain-dependent manner, whereas mRNA expression of CXCL8 increased following stimulation by P. gingivalis, suggesting post-transcriptional and/or post-translational modification of CXCL8 by gingipains. No change in protein concentration or mRNA expression was observed following stimulation of OMM which may reflect the multi-layered nature of this model. Differences between monolayer and OMM indicate a role for OMM to investigate bacterial invasion and resultant cytokine release due to its comparability with the oral mucosa.
The work presented in this thesis has described the development, characterisation and optimisation of OMM to investigate invasion by P. gingivalis. Invasion was shown to be influenced by environmental changes and P. gingivalis protease expression. Although P. gingivalis degrades key surface molecules including CD46, tetraspanins and α5β1, blocking experiments with antibodies could not explain the protease-dependent effects on invasion. Modulation of cytokine production, particularly CXCL8, by P. gingivalis gingipains, may contribute to a disruption in leukocyte recruitment resulting in a dysregulated inflammatory response. Future development of OMM in terms of including an immune cell element and endothelial component to extend the study of P. gingivalis-host cell interactions will add value to this model. The data presented here indicate that P. gingivalis invasion of the epithelium is likely to be an important contributor to periodontal disease progression.
