| 其他摘要 | Caries is one of the most common infections in children and adult world-wide, however preventive intervention remains difficult. An accurate caries risk assessment method can identify patients at high caries risk for caries onset and thus help to deliver preventive therapies. Previous work in our research group showed that the spatial and temporal variation of oral microbiota can be employed for prediction of caries onset in children. However, the use of human cohorts have resulted in significant difficulties in mechanistic understanding of such predictive models, as in human cohorts it is difficult to precisely control the microenviroment and behavior of the individual subjects. The rodent model of dental caries, rats in particular, not only can simulate the onset and progression of caries in natural environments, but allows direct application and controls of the various etiological factors of caries, and thus greatly facilitate the longitudinal tracking of the microbiota dynamics during caries onset and progression. These advantages have made the rat model an excellent animal model for mechanistic study of the microbiota-based prediction of caries onset. However, previously there have been no reports that profile microbiome dynamics during caries development in rats.
Here we simultaneously tracked the longitudinal development of microbiota of 57 three-week-old rats for 12 weeks, during which 27 stayed healthy and 30 transited from health into cariogenesis. The techniques were mainly 16S rRNA gene amplicon-based pyrosequencing technology coupled with varies multivariate statistical analyses. Firstly, we found that the factors of age and status exert the key impact on the overall composition of the rat oral microbiota, and difference in the taxa level used had a strong influence on the analysis of variation of oral microbiota. Secondly, a host-aging correlated microbiota development pattern apparent in healthy stage was retarded by caries onset. Moreover, oral microbiota during caries onset was significantly more correlated with changes in disease severity than that during caries progression. Thirdly, by distinguishing between aging- and disease-associated OTUs and exploiting the distinct microbiota dynamics between the onset and progression phases of caries, Microbial indicators of Caries (MiC) based on Random Forests algorithm was proposed, which diagnosed caries from healthy samples with 99% accuracy, and furthermore predicted future new caries-onsets for those samples presently clinically perceived as healthy with 77% accuracy. Finally, we analyzed the link and distinction in microbiota development underlying cariogenesis between rat and human, and found that the temporal variation of the structure and microbial metabolism of the rat oral microbiota is consistent with human oral microbiota during caries and health, and the predictive model for caries onset based on Random Forests algorithm is robust in both human and rat oral microbiota. Interestingly, bacterial composition of oral microbiota at the genus level was different between human and rat, and the divergent evolution of human oral microbiota and rat oral microbiota was driven by status (caries) or age.
In summary, the findings in this M.S. thesis have several implications. Firstly, via the animal model of rat, this study further validates the notion that caries onset can be predicted via oral microbiota, and lays a theoretical foundation for future preventive intervention experiments based on such predictive model of diseases. Secondly, this study suggests that the microbiota data can be analyzed from microbiota structural changes and microbial metabolism, to analyze the experimental data from the microbiota of the animal models that are used to study how microbiota predicts disease. Finally, this study suggests that the Random Forests algorithm has huge potential for the analysis of the experimental data from the microbiota of the animal models and for the construction of disease predictive models. |
修改评论