Abstract
Antibiotic resistant microbial infections pose an urgent threat to human health. In 2019, 4.95 million people died worldwide from antimicrobial resistant infections, which is projected to increase to 10 million deaths each year by 2050. Antibiotic resistant bacterial infections are associated with biofilms, which are multicellular microbial communities encased in a self-produced polysaccharide matrix. Heterogeneous bacterial physiological states exist within biofilms that promote infection and decrease antibiotic susceptibility. Gaps in understanding the connected, multiscale mechanisms controlling biofilm formation can result in undesirable treatment outcomes, such as the development of antibiotic resistance. There is an urgent demand for informed, alternative treatment strategies that mitigate the development of antibiotic resistance. In this work, we applied quantitative analysis using a dynamic criterion and developed a computational framework to characterize and predict biofilm outcomes in response to intervention strategies.
In Chapter 2, chemorepellent-loaded polymeric nanoparticles were designed to inhibit the initial biofilm formation stage of Escherichia coli transport to a surface. A burst chemorepellent release profile promoted localized interference of bacterial transport and inhibited biofilm formation for up to five hours. The biofilm control outcome was analyzed using a semi-quantitative dynamic criterion that compared competing biofilm formation and nanoparticle release processes. While chemorepellent delivery inhibited initial biofilm formation stages, delivery of chemorepellents alone may not be suitable to remove established, non-motile, heterogenous biofilms.
To understand and inform strategies to treat complex, established biofilms, a novel multiscale computational framework, Multi-scale model of Metabolism in Cellular Systems (MiMICS), was constructed in Chapter 3. To mechanistically predict heterogenous biofilm metabolism, MiMICS spanned scales of intracellular metabolism, individual cell behavior, and continuum-level metabolite transport. Guided by a published spatial transcriptomics dataset, MiMICS was used to connect emergent metabolic microenvironments to heterogeneous metabolic states in a Pseudomonas aeruginosa biofilm. In Chapter 4, MiMICS was further applied to probe nitric oxide and quorum-sensing mediated stochastic transcription mechanisms controlling microscale spatial patterns of metabolic states in the P. aeruginosa biofilm. In silico gene knockouts were leveraged to identify potential metabolic targets that promoted biofilm cell death induced by endogenously produced nitric oxide. Furthermore, simulations predicted biofilm regions with heterogenous susceptibility to exogeneous application of cytotoxic nitric oxide. Altogether, this work provided informed, quantitative insight into key transport and metabolic mechanisms of competing bacterial biofilm formation and intervention strategies, which can guide future development of biofilm treatment strategies.
