A Systems Approach to Characterize Drivers of Vaginal Microbiome Composition and Bacterial Vaginosis Treatment Efficacy
Lee, Christina
2023
Abstract
The vaginal microbiome (VMB) is critical to female reproductive health, with Lactobacillus spp. dominance associated with health and a diverse, anaerobic community composition associated with a variety of adverse reproductive outcomes and increased susceptibility to sexually transmitted infections. This dysbiosis, known as bacterial vaginosis (BV), impacts nearly 30% of reproductive-age women. Therapies to treat this condition have largely remained the same over the last three decades despite the high frequency of recurrent cases. Difficulty in treating BV stems from several challenges such as there is not a single microbe causing the condition, complex ecological interactions between microbial species and the microenvironment dictate community composition and stability, and lack of adequate preclinical models to assess therapies. Ordinary differential equation (ODE)-based modeling can help capture these complexities and model the dynamics and stability of multi-species communities to better characterize key drivers of composition shifts to BV-associated communities and in response to therapies aimed to re-establish optimal composition. Here, we established a series of mathematical models to help address key aspects of modeling the vaginal microbiome and the impact of BV therapies. In the first aim, a novel multi-species model was developed to replicate shifts between BV-associated bacteria and optimal composition (Lactobacillus spp. dominance) after treatment with a standard-of-care antibiotic, metronidazole. This model captured microbial metabolism by the target microbial species, G. vaginalis, and sequestration of metronidazole by the non-target species, Lactobacillus iners. Using this model, sensitivity analyses indicated that a key driver in antibiotic efficacy was the pre-treatment ratio of G. vaginalis relative to L. iners. Counterintuitively, the model associated higher amounts of L. iners relative to G. vaginalis with worse treatment outcomes due to L. iners sequestering metronidazole to an extent that reduced the amount that could inhibit G. vaginalis growth. This association was validated with in vitro co-cultures and in two small clinical cohorts. In the second aim, we used a model to understand the stability of VMB community state types and to identify specific microbial interactions that drive variability in stability. For this approach, we used a generalized Lotka-Volterra model, which predicts community composition as a function of pairwise interspecies interactions, microbial growth rates, and carrying capacities. Physiological parameter ranges were defined and used to generate virtual populations that predicted the frequency of mono- and multi-stable states observed across two clinical cohorts. Virtual populations were refined and then validated for community composition changes during and after menses and antibiotic therapies in separate clinical cohorts. The virtual population emphasized the importance of how BV-associated bacteria interact with Lactobacillus spp. for maintaining or reestablishing Lactobacillus spp. dominance. Lastly, in aim 3 a similar approach using virtual populations was applied to the assessment of optimal probiotic characteristics for treating BV. The model explored two dosing regimens, an acute probiotic regimen and a regimen described for a phase 2b clinical trial of Lactin-V. Simulations supported that both resident Lactobacillus spp. and BV-associated bacteria interactions with probiotics could significantly impact probiotic efficacy, and thus should be an additional probiotic screening criterion. Model predictions exhibited similar population-level recurrence frequencies as observed in the Lactin-V trial and allowed for exploration of the relationship between probiotic strain characteristics and dosing regimens. Overall, these three aims describe new frameworks to help explore and design new BV therapies that otherwise would be difficult to study in vitro or preclinically.Deep Blue DOI
Subjects
Vaginal Microbiome Systems Biology
Types
Thesis
Metadata
Show full item recordCollections
Remediation of Harmful Language
The University of Michigan Library aims to describe library materials in a way that respects the people and communities who create, use, and are represented in our collections. Report harmful or offensive language in catalog records, finding aids, or elsewhere in our collections anonymously through our metadata feedback form. More information at Remediation of Harmful Language.
Accessibility
If you are unable to use this file in its current format, please select the Contact Us link and we can modify it to make it more accessible to you.