Longitudinal analysis reveals characteristically high proportions of bacterial vaginosis-associated bacteria and temporal variability of vaginal microbiota in northern pig-tailed macaques (Macaca leonina)

Abstract

The complex and dynamic vaginal microbial ecosystem is critical to both health and disease of the host. Studies focusing on how vaginal microbiota influences HIV-1 infection may face limitations in selecting proper animal models. Given that northern pig-tailed macaques (Macaca leonina) are susceptible to HIV-1 infection, they may be an optimal animal model for elucidating the mechanisms by which vaginal microbiota contributes to resistance and susceptibility to HIV-1 infection. However, little is known about the composition and temporal variability of vaginal microbiota of the northern pig-tailed macaque. Here, we present a comprehensive catalog of the composition and temporal dynamics of vaginal microbiota of two healthy northern pig-tailed macaques over 19 weeks using 454-pyrosequencing of 16S rRNA genes. We found remarkably high proportions of a diverse array of anaerobic bacteria associated with bacterial vaginosis. Atopobium and Sneathia were dominant genera, and interestingly, we demonstrated the presence of Lactobacillus-dominated vaginal microbiota. Moreover, longitudinal analysis demonstrated that the temporal dynamics of the vaginal microbiota were considerably individualized. Finally, network analysis revealed that vaginal pH may influence the temporal dynamics of the vaginal microbiota, suggesting that inter-subject variability of vaginal bacterial communities could be mirrored in inter-subject variation in correlation profiles of species with each other and with vaginal pH over time. Our results suggest that the northern pig-tailed macaque could be an ideal animal model for prospective investigation of the mechanisms by which vaginal microbiota influence susceptibility and resistance to HIV-1 infection in the context of highly polymicrobial and Lactobacillus-dominated states.

Keywords: Bacterial vaginosis; Macaque; Networks; Temporal dynamics; Vaginal microbiome.

Figure 1

Clustering analysis of vaginal bacterial communities A: Ward linkage hierarchical cluster dendrogram of the Bray-Curtis distances between bacterial communities in 27 samples from two northern pig-tailed macaques (Macaca leonina). B: Color bar indicating vaginal bacterial CST wherein each sample was assigned to one of four CSTs (CST A, CST S, CST L, and CST D). C: Lower colored blocks indicate the subject from which the sample was obtained. D: Shannon diversity indices calculated for the 27 samples (color key is indicated in the middle left side). E: Vaginal pH measurements for the 27 samples (color key is indicated in the middle left side). F: Heatmap of the relative abundances of the 14 most abundant phylotypes (mean relative abundance of each phylotype across all samples of the two subjects was above 1%). Column Z-score indicates differences between samples in terms of relative abundances of phylotypes associated with the samples. Individual cells are color-coded according to Z-scores to show the normalized abundance of a phylotype in one sample relative to the mean abundance across all phylotypes of the column. Relative intensity of the colors indicates how many standard deviations the observed phylotype abundance is above or below the mean. White color indicates relative abundance of phylotypes with column average. Blue color indicates relative abundances less than average abundance. Red color indicates relative abundance above column average. Phylotypes are listed in order of dominance, with the most dominant on the top. *represents unclassified genera of order/family.

Figure 2

Temporal dynamics of vaginal bacterial communities A: Vaginal pH measurements and Shannon diversity indices calculated for each of the 13 samples of subject A and 14 samples of subject B (color key is indicated in the top right corner). B: Profiles of CST for the two northern pig-tailed macaques over time. Each cell represents one sample in the time series. The CSTs in the time series for each macaque are color-coded (lower colored blocks indicate CST wherein each sample was assigned to one of the four CSTs). C: For each macaque, a heatmap was constructed from the relative abundances of phylotypes that comprised at least 1% of the sequences found in that macaque (color key is indicated in the middle right side). Phylotypes are listed in order of dominance, with the most dominant on the bottom. *represents unclassified genera of order/family. D: For each macaque, the bar plot represents the relative abundances of phyla that comprised at least 1% of the total sequences found in that macaque. Phyla that comprised less than 1% of the total sequences found in that macaque were grouped into the “Others” category.

Figure 3

Integration of vaginal pH with microbial co-occurrence networks using time-series data of each northern pig-tailed macaque Note, the nodes in the correlation network correspond to OTUs (ellipse nodes) and vaginal pH (round rectangle nodes), whereas the edges represent strong and significant evidence for correlation between nodes. Size of the nodes indicates relative abundances of OTUs (mean relative abundance of each OTU across all samples of each subject was above 0.1%) observed within a subject over time.