A Uniquely Altered Oral Microbiome Composition Was Observed in Pregnant Rats With Porphyromonas gingivalis Induced Periodontal Disease

Abstract

Porphyromonas gingivalis is an anaerobic bacterium commonly found in the oral cavity and associated with the development of periodontal disease. P. gingivalis has also been linked to several systemic vascular and inflammatory diseases including poor pregnancy outcomes. Little is known about the changes in the oral flora during pregnancy in connection to P. gingivalis infection. This pilot study aims to explore changes in the oral microbiome due to P. gingivalis inoculation and pregnancy in an in vivo rat model of periodontal disease. A metagenomic sequencing analysis targeting seven of the 16S rRNA gene variable regions was performed for oral samples collected at the following time points: baseline control (week 0), P. gingivalis inoculated (week 11), P. gingivalis inoculated and pregnant rat at necropsy (week 16). A second set of animals were also sampled to generate a sham-inoculated (week 11) control group. We found that the rat oral microbiome profiles were more similar to that of the human oral cavity compared to previous reports targeting one or two 16S variable regions. Overall, there appears to be a relatively stable core microbiome in the oral cavity. As expected, P. gingivalis induced periodontal disease resulted in oral microbiome dysbiosis. During pregnancy, some aspects of the oral microbiome shifted toward a more baseline-like profile. However, population analyses in terms of dissimilarity measures and especially metagenomic based predictions of select characteristics such as cell morphology, oxygen requirement, and major metabolite synthesis showed that pregnancy did not restore the composition of the oral microbiome. Rather, a uniquely altered oral microbiome composition was observed in pregnant rats with pre-established periodontal disease.

Keywords: microbial metabolites; oral microbiome; oxygen requirement; periodontal disease; pregnancy; rat model.

Figure 1

All P. gingivalis inoculated animals had significantly greater alveolar bone loss than aged-matched sham-inoculated controls (baseline) confirming that these animals had periodontal disease. Morphometric assessment of mandibular and maxillary bone loss in pregnant control and PG-inoculated rats. Values in each graph represent the extent of horizontal bone loss, labeled as CEJ-ABC area measurements (mm²), on the lingual (A) and buccal (B) sides from each mandible. (C) Representative images of the lingual aspect of the rat mandible illustrating the extent of horizontal bone loss. Yellow lines demarcate the CEJ-ABC junction that was used to determine extent of bone loss. Representative images of maxillary interdental papilla from control (D) and PG-inoculated rats (E) demonstrating how CEJ and ABC measurements were taken for histomorphometry (F). Bars in all graphs show the mean ± SD. Data were analyzed by unpaired student's t test.

Figure 2

We also measured P. gingivalis specific IgM and IgG before (baseline), after inoculation, and during pregnancy to assess microbial exposure.

Figure 3

Alpha diversity was evaluated by four different statistical analyses including: Observed Species, Simpson Index, Shannon Index, Chao1 index.

Figure 4

Bray Curtis PCoA plot of beta diversity measures of dissimilarity shows at the species level, baseline, PG-inoculated, and pregnant groups clearly clustered within their group along PC2 not PC1.

Figure 5

Bray Curtis PCoA plot of beta diversity measures of dissimilarity shows at the species level, baseline, PG-inoculated, and pregnant groups clearly clustered within their group along PC2 and PC4 but not PC3.

Figure 6

Bray Curtis PCoA plot of beta diversity measures of dissimilarity shows that at the genus level, the baseline, PG-inoculated, and pregnant groups clustered within their groups along PC3 at the (A) genus level and (B) family level.

Figure 7

Bray Curtis PCoA plot of beta diversity measures of dissimilarity shows, when the data sets were sorted by infection status, they clustered within their groups along PC5 at the species level (A), PC6 at the genus level (B), and PC9 at the family level (C).

Figure 8

Collective consensus microbiome at the family, genus, or species level and plotted as four data points in PCoA plots using (A) Bray Curtis or (B) Eucledian of beta diversity measures of dissimilarity.

Figure 9

Population proportion of identified community members in terms of predicted Gram stain classification and cell morphology for baseline, PG-inoculated, and pregnant groups.

Figure 10

Population proportion of identified community members in terms of predicted (A) oxygen requirements alone, or (B) oxygen requirements in context of Gram stain classification, for baseline, PG-inoculated, and pregnant groups.

Figure 11

Population proportion of identified community members in terms of predicted metabolic characteristics as a major generator of target microbial metabolites for baseline, PG-inoculated, and pregnant groups.

Figure 12

Population proportion of identified community members in terms of predicted metabolic characteristic as a major generator of acetate for baseline, PG-inoculated, and pregnant groups.

Figure 13

Population proportion of identified community members in terms of predicted ability to synthesize riboflavin for baseline, PG-inoculated, and pregnant groups.