Preliminary functional analysis of the subgingival microbiota of cats with periodontitis and feline chronic gingivostomatitis

Abstract

The subgingival microbial communities of domestic cats remain incompletely characterized and it is unknown whether their functional profiles are associated with disease. In this study, we used a shotgun metagenomic approach to explore the functional potential of subgingival microbial communities in client-owned cats, comparing findings between periodontally healthy cats and cats with naturally occurring chronic periodontitis, aggressive periodontitis, and feline chronic gingivostomatitis. Subgingival samples were subjected to shotgun sequencing and the metagenomic datasets were analyzed using the MG-RAST metagenomic analysis server and STAMP v2.1.3 (Statistical Analysis of Metagenomic Profiles) software. The microbial composition was also described to better understand the predicted features of the communities. The Respiration category in the level 1 Subsystems database varied significantly among groups. In this category, the abundance of V-Type ATP-synthase and Biogenesis of cytochrome c oxidases were significantly enriched in the diseased and in the healthy groups, respectively. Both features have been previously described in periodontal studies in people and are in consonance with the microbial composition of feline subgingival sites. In addition, the narH (nitrate reductase) gene frequency, identified using the KEGG Orthology database, was significantly increased in the healthy group. The results of this study provide preliminary functional insights of the microbial communities associated with periodontitis in domestic cats and suggest that the ATP-synthase and nitrate-nitrite-NO pathways may represent appropriate targets for the treatment of this common disease.

Figure 1

Relative abundance at domain level in subgingival metagenomes from healthy cats and cats diagnosed with aggressive periodontitis (AP_), chronic periodontitis (CP_), and feline gingivostomatitis. (A) Bar graph representing the relative distribution (mean relative frequency (%) and standard error of the mean) at domain level among healthy and diseased cats. (B) Heatmap plot illustrating abundance (%) at domain level in each metagenomic sample analyzed. (A) Generated using GraphPad Prism 9 (GraphPad Software LLC, La Jolla, CA, https://www.graphpad.com). (B) Generated using STAMP v2.1.3 (Statistical Analysis of Metagenomic Profiles) software (https://beikolab.cs.dal.ca/software/STAMP).

Figure 2

Distribution of phylum in subgingival metagenomes from healthy cats and cats diagnosed with aggressive periodontitis (AP), chronic periodontitis (CP) and feline gingivostomatitis (FCGS). (A) Bar graph representing the relative distribution (mean relative frequency (%) and standard error of the mean) of the most abundant phyla, minimum of 1% of abundance, in subgingival samples from cats. (B) Box and whiskers plots illustrating the mean (star), median (line within the box), quartiles (IQR; 75th to 25th of the data), maximum and minimum (whiskers extend to the most extreme value) proportion of sequences (%) of the phylum found with significant statistical difference in the metagenomic datasets studied (false discovery rate, FDR P-value < 0.05). The results were filtered using a P-value of 0.05 in STAMP v2.1.3 (Statistical Analysis of Metagenomic Profiles) software. (C) Extended error bar plots representing abundance of Actinobacteria that showed significant higher abundance in healthy cats compared to diseased cats using Tukey–Kramer with Benjamin-Hochberg FDR correction (P-value < 0.05). (A) Generated using GraphPad Prism 9 (GraphPad Software LLC, La Jolla, CA, https://www.graphpad.com). (B, C) Generated using STAMP v2.1.3 (Statistical Analysis of Metagenomic Profiles) software (https://beikolab.cs.dal.ca/software/STAMP).

Figure 3

Comparison of the microbial populations in subgingival sites of healthy cats and cats diagnosed with aggressive periodontitis (AP), chronic periodontitis (CP) and feline chronic gingivostomatitis (FCGS). (A) Distribution of families found be statically different among groups of subgingival samples using ANOVA and Tukey–Kramer with Benjamin-Hochberg FDR correction (P-value < 0.05). (B) Overall distribution of the thirty most abundant bacterial genera identified in subgingival metagenomic samples from healthy cats and diseased cats. Cells plots were built in JMP 15 (SAS Institute Inc., Cary, NC, https://www.jmp.com/en_us/home.html) using mean relative frequency (%) calculated in STAMP v2.1.3 (Statistical Analysis of Metagenomic Profiles) software (https://beikolab.cs.dal.ca/software/STAMP).

Figure 4

Distribution of functional annotations by SEED Subsystems of subgingival metagenomic data of healthy cats (H) and cats diagnosed with aggressive periodontitis (AP), chronic periodontitis (CP) and feline chronic gingivostomatitis (FCGS). (A) Heatmap dendrogram in which features at level 1 were ranked by abundance and metagenomic samples grouped by similarity (unweighted pair group method with arithmetic mean, UPGMA). (B) Extended error bar plot showing abundance of predicted functional category, level 1, with significant statistical difference between healthy and diseased cats. (C) Extended bar plot illustrating the mean proportion of predict functional features at level 3. (D) Extended bar plot showing the mean proportion of predict functional features at function level. The results were filtered using a P-value of 0.05 according to ANOVA and Tukey–Kramer with Benjamin-Hochberg FDR correction (P-value < 0.05). All graphs were generated using STAMP v2.1.3 (Statistical Analysis of Metagenomic Profiles) software (https://beikolab.cs.dal.ca/software/STAMP).

Figure 5

Box and whiskers plots illustrating the mean (star), median (line within the box), quartiles (IQR; 75th to 25th of the data), maximum and minimum (whiskers extend to the most extreme value) proportion of sequences (%) of the gene found with significant statistical difference in the metagenomic datasets studied (false discovery rate, FDR P-value < 0.05). The results were filtered using a P-value of 0.05; the data was computed using the MG-Rast metagenomics analysis server using KEGG Orthology (KO) database. This figure was generated using STAMP v2.1.3 (Statistical Analysis of Metagenomic Profiles) software (https://beikolab.cs.dal.ca/software/STAMP).