Discovery and Validation of Periodontitis Biomarkers

Periodontitis is a major public health issue in China: it is responsible for loss of masticatory function in 60 million older adults, and 400-500 million adults are on the same disease trajectory. In addition, gingivitis and early-stage periodontitis are highly prevalent in all age groups. The Lancet 2021 burden of disease study provides worrying projections for China's oral health, with a 47.8% increase in advanced-stage periodontitis and a 217% increase in edentulism by the year 2050. The numbers are not manageable by the Chinese health system unless a series of coordinated actions are implemented: i) health education promoting oral hygiene in school and the workplace; ii) effective AI-based self-detection strategies and accurate identification of high-risk subjects; iii) efficient treatment modalities; and iv) reorganization of the health system.

We have developed, patented, and validated a self-detection AI-based screening test for the general population through an app. It is based on a few validated questions and the performance of a lateral flow immunoassay to detect activated matrix metalloproteinase 8 (aMMP8). The algorithm enables accurate self-detection of severe periodontitis. The system, however, cannot identify subjects without clinically evident periodontitis (subjects who present with superficial inflammation consistent with gingivitis and incipient periodontitis) who will develop the disease, which, therefore, should be the target of early interventions. This limitation is due to insufficient knowledge of the process that turns superficial inflammation (gingivitis) into periodontitis. This limitation is apparent in the recently published NIH-sponsored American diagnostic trial results to detect periodontitis onset biomarkers (and progression). In their study, Teles et al. (2024) show that almost 24% of gingivitis subjects progress to periodontitis over a 12-month period but failed to identify salivary or serum biomarkers. Similarly, our recently completed study (Li et al. in preparation) did not identify highly accurate biomarkers for disease onset and progression. Importantly, the American and our study have tested putative biomarkers identified based on the current crude knowledge of the disease process. Gaps in fundamental knowledge are now apparent and limit our ability to detect periodontitis early. In addition, the current crude differential diagnosis based on clinical examination with a periodontal probe with millimeter markings cannot accurately differentiate gingivitis from early-stage periodontitis, complicating the ground truth definition (gold standard).

In the current study, we propose implementing a multi-omics approach to test the ability to discriminate a mixed population of clinically undifferentiable gingivitis and stage I periodontitis into two or more clusters. In this biomarker discovery phase, we plan to use multiple state-of-the-art methods: i) laser scanning microdissection proteomics of tissue biopsies, ii) conventional salivary proteomics, iii) tissue biopsy transcriptomics, and iv) shotgun microbiome analysis. The methods will be applied in an agnostic approach to test the following hypotheses:

1. It is possible to identify two or more clusters of subjects from a mixed population of gingivitis and stage I periodontitis subjects.
2. The clusters differ based on host-derived biomarkers and/or microbiome factors and the risk of progression to periodontitis.
3. The biomarker pathways and microbial virulence factors among subjects identified according to the different approaches used to explore disease biology are generally consistent.
4. It is possible to identify a limited set of biomarkers that can be used to predict periodontitis onset and thus target early interventions for this high-risk population.