Local Innate Markers and Vaginal Microbiota Composition Are Influenced by Hormonal Cycle Phases

Abstract

Background: The female reproductive tract (FRT) mucosa is the first line of defense against sexually transmitted infection (STI). FRT environmental factors, including immune-cell composition and the vaginal microbiota, interact with each other to modulate susceptibility to STIs. Moreover, the menstrual cycle induces important modifications within the FRT mucosa. Cynomolgus macaques are used as a model for the pathogenesis and prophylaxis of STIs. In addition, their menstrual cycle and FRT morphology are similar to women. The cynomolgus macaque vaginal microbiota is highly diverse and similar to dysbiotic vaginal microbiota observed in women. However, the impact of the menstrual cycle on immune markers and the vaginal microbiota in female cynomolgus macaques is unknown. We conducted a longitudinal study covering three menstrual cycles in cynomolgus macaques. The evolution of the composition of the vaginal microbiota and inflammation (cytokine/chemokine profile and neutrophil phenotype) in the FRT and blood was determined throughout the menstrual cycle.

Results: Cervicovaginal cytokine/chemokine concentrations were affected by the menstrual cycle, with a peak of production during menstruation. We observed three main cervicovaginal neutrophil subpopulations: CD11b^high CD101⁺ CD10⁺ CD32a⁺, CD11b^high CD101⁺ CD10^- CD32a⁺, and CD11b^low CD101^low CD10^- CD32a^-, of which the proportion varied during the menstrual cycle. During menstruation, there was an increase in the CD11b^high CD101⁺ CD10⁺ CD32a⁺ subset of neutrophils, which expressed higher levels of CD62L. Various bacterial taxa in the vaginal microbiota showed differential abundance depending on the phase of the menstrual cycle. Compilation of the factors that vary according to hormonal phase showed the clustering of samples collected during menstruation, characterized by a high concentration of cytokines and an elevated abundance of the CD11b^high CD101⁺ CD10⁺ CD32a⁺ CD62L⁺ neutrophil subpopulation.

Conclusions: We show a significant impact of menstruation on the local environment (cytokine production, neutrophil phenotype, and vaginal microbiota composition) in female cynomolgus macaques. Menstruation triggers increased production of cytokines, shift of the vaginal microbiota composition and the recruitment of mature/activated neutrophils from the blood to the FRT. These results support the need to monitor the menstrual cycle and a longitudinal sampling schedule for further studies in female animals and/or women focusing on the mucosal FRT environment.

Keywords: blood compartment; cytokines; female reproductive tract (FRT); inflammation; menstrual cycle; neutrophils; vaginal microbiota.

Figure 1

Experimental design and progesterone level. (A) Nine female cynomolgus macaques were included in the study. Blood, cervicovaginal fluids (Weck-cel^®) and swabs, as well as cervicovaginal cells, were collected once a week for three months. Samples were not collected during weeks 5 and 6. Image created in BioRender.com. (B) Progesterone concentrations were quantified in plasma once a week for all individuals. The graphical representation of progesterone concentration including all animals was obtained using interpolated and aligned data (left). On the contrary, the graphical representation of progesterone concentration for each individual was obtained using the raw data (right). Red arrows represent menstruation.

Figure 2

Cytokine and chemokine expression in the plasma of female cynomolgus macaques (n = 9). (A) Heatmap representing the mean fold change in the expression of cytokines and chemokines in the plasma of each animal (n = 9). The fold change was calculated based on the mean expression of each cytokine/chemokine for all females and time points. One way ANOVA was performed to compare the total fold change value to those of each animal for each cytokine. Asterisks indicate p values considered to be statistically significant (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001). (B) Heat map representing the mean fold change in the expression of cytokines and chemokines in the plasma for all animals according to time (n = 9). The fold change was calculated based on the mean expression of each cytokine/chemokine in each female. Numbers 1 to 14 refers to weeks. Red arrows represent menstruation.

Figure 3

Cytokine and chemokine expression in cervicovaginal fluids of female cynomolgus macaques (n = 9). (A) Heatmap representing the mean fold change in the expression of cytokines and chemokines in the cervicovaginal fluids of each animal (n = 9). The fold change was calculated based on the mean expression of each cytokine/chemokine for all females. One way ANOVA was performed to compare the total fold change value to those of each animal for each cytokine. (B) Heat map representing the mean fold change in the expression of cytokines and chemokines in the cervicovaginal fluids of all animals (n = 9). The fold change was calculated based on the mean expression of each cytokine/chemokine for each female. Numbers 1 to 14 refers to weeks. Red arrows represent menstruation (C) Samples clustered into three groups based on the progesterone level or menstruation and each cytokine/chemokine concentration was plotted. A Kruskal-Wallis test with Dunn’s test to adjust the p value was performed. Asterisks indicate p values considered to be statistically significant (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001).

Figure 4

Neutrophil subpopulations in cervicovaginal cytobrushes and blood of female cynomolgus macaques (n = 9). (A) Percentage of neutrophils among CD45+ cells in cervicovaginal cytobrushes and blood for all animals. One dot represents one sample of one animal. Blood samples are shown in red and cervicovaginal samples in blue. (B) Percentage of neutrophil subpopulations among CD66+ Lin- CD14- CDw125- in blood (left) and cervicovaginal cytobrushes (right) for all females. Each animal is represented by one symbol. (C) Percentage of CD62L+, HLA-DR+, and PD-L1+ surface expression for all neutrophil subpopulations in blood (left) and for the main neutrophil subpopulations in cervicovaginal cytobrushes (right).

Figure 5

Variation of neutrophil subpopulations in cervicovaginal cytobrushes and blood during the menstrual cycle. (A) Main neutrophil subpopulations among CD66+ Lin- cells in both compartments according to the time of collection for all animals (mean of all animals, left) or for each animal (right). Blood (B) and cervicovaginal cytobrush (C) samples were clustered into three groups based on progesterone levels or menstruation and the percentage of each neutrophil subpopulation was plotted. A Kruskal-Wallis test followed by Dunn’s test to adjust the p value was performed. Asterisks indicate p values considered to be statistically significant (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001, ****p ≤ 0.0001). Only significant results are presented for (B) blood and (C) cervicovaginal cytobrushes.

Figure 6

Kinetics of the vaginal microbiota composition of female cynomolgus macaques (n = 9). Relative abundance of bacterial taxa at the phylum level (A) in each animal according to time. Relative abundance of the nine most represented genera (B) or families (C) for all females at all time points and each female according to time. The other families and genera are shown in black (other).

Figure 7

Variation of bacterial taxa during each phase of the menstrual cycle. Percentage of the mean relative abundance of the nine most represented genera in the high-progesterone, low-progesterone, and menstruation groups is represented in a pie chart for each female. Other genera are shown in grey (other).

Figure 8

Vaginal microbiota diversity and differentially abundant bacterial taxa in each phase of the menstrual cycle. (A) The Shannon index was calculated for each female for the high-progesterone and low-progesterone phases and during menstruation and plotted together. A Kruskal-Wallis test followed by Dunn’s test to adjust the p value was performed (**p ≤ 0.01). Analyses were performed using the FitZIG algorithm at the family (B) or genus (C) level. Heatmaps representing the p values were generated and plus and minus signs were added to visualize increases or decreases in the abundance of each bacterial taxa. The sign is associated with the group in bold text. As an example: the abundance of Christensenellaceae was higher during the high-progesterone phase than the low-progesterone phase.

Figure 9

Heatmap representing bacterial taxa (family level), neutrophil subpopulations, and cytokine/chemokine levels that varied during the hormonal cycle. The heatmap represents the log10 fold change for each parameter (cytokines, neutrophil subpopulations, bacterial taxa, indicated on the right). Increased values are shown in red and decreased in blue. Each animal is represented by a color code, as well as the hormonal phases (high progesterone in green, low progesterone in orange, menstruation in red). Hierarchical clustering divided the samples (sample ID in the X-axis) into two large clusters (A, B). Each cluster was then separated into subclusters (A1, A2 and B1, B2).