UC Berkeley

Listeria monocytogenes (Lm) is a Gram positive, facultative intracellular pathogen that infects its hosts upon ingestion of contaminated foods. Exposure is generally innocuous or followed by mild self-limiting gastroenteritis, but in certain contexts it can lead to more severe outcomes. In immunocompromised individuals, Lm can infect the central nervous system (CNS) and cause permanent brain damage or even be lethal, while infections in pregnant women often lead to high bacteremia in the placenta and loss of the fetus. In addition to being a public health issue, Lm is also a model system used to establish fundamental understandings about microbiology and cell biology, and a tool used to develop cancer-immunotherapies. Dissecting the dynamics of pathogen dissemination in a variety of contexts allows for the identification of bottlenecks and trafficking pathways that could be targeted for therapies and tool development both with Lm and other microbes.

In our study we first assessed the dynamics of Lm colonization and replication within mouse brains using barcoded bacteria. Upon intravenous inoculation, the CNS was seeded from multiple independent invasion events from the blood which led to a polyclonal CNS, yet one clone usually dominated within the brain. Sequential infections and intracranial inoculations established that timing of colonization, rather that entry site or access to a certain niche within the brain, was the main driver of clonal dominance. Since Lm is a food-borne disease, we then designed a new model of oral inoculation-induced cerebral listeriosis using immunocompromised mice fed with a piece of bread soaked in bacteria and pre-treated with streptomycin. In this context, rare colonization events led to a highly infected yet monoclonal CNS. This stringent dissemination bottleneck likely arose from pathogen transit into the blood, rather than directly from the blood to the brain and did not limit bacterial replication within the brain tissue. Collectively, our findings provided an in-depth analysis of the Lm population dynamics that lead to cerebral listeriosis and established a framework for studying the bacterial spread with other neuropathogens.In order to assess the host bottlenecks that affect pathogen dissemination to the brain, we next investigated the role of the innate immune receptor STING in bacterial trafficking. Our goal was to establish whether activation of STING would control bacterial replication and limit spread of Lm to the brain, or if STING activity would instead facilitate bacterial dissemination by recruiting monocytes into circulation and providing Lm with “Trojan horses”. Upon oral inoculation, Lm accumulated more efficiently within the CNS of STING-deficient mice (Goldenticket, Gt) than wild-type mice, suggesting that STING activity was involved in controlling host-wide infection more than promoting bacterial dissemination. STING-derived Type 1 interferon was not involved in the heightened Gt susceptibility, but signaling via the STING C-terminal tail was involved in STING-depended Lm control. Oral infection of Rag1KO-Gt double mutant mice suggested that STING was involved in controlling early stages of infection rather than bacterial accumulation in the brain over time. Finally, we expanded our study to assess bacterial dissemination to the placenta in the context of STING-deficiency. Intravenous infection of pregnant mice led to mild protection of Gt placentas, while oral infection led to increased sensitivity in Gt mice. Collectively, our results provided a new understanding on the limited impact of STING-activity during bacterial dissemination to the brain and established a framework for further analysis of the intriguing role of STING during placenta infections.

We then investigated Lm infection as a beneficial therapy in the context of tumor-bearing mice using barcoded attenuated Lm. Our preliminary work assessed the impact of the route of inoculation on bacterial spread and established that Lm colonized tumors within 4 hours, but only to a limited degree following intravenous (IV) inoculation, then dramatically expended locally with little to no further dissemination between tissues. Intratumorally (IT) inoculated mice experienced Lm spread beyond the tumor quite efficiently, but again Lm migration between tissues was limited by 3 days post injection. Combination therapy with IV vaccination followed by IT challenge one week later, led to Lm spread into systemic tissues from the tumor during the first 4 hours post inoculation, but this was followed by a massive clearance event. In cases where bacteria persisted in the tumor microenvironment, bacterial burden was generally high with a mostly clonal population. This finding suggested that host pressures cleared most of the early colonizing Lm, but a few bacteria were able to evade clearance and persist over time. Overall, our work provided a more granular picture of the landscape of pressures faced by Lm during dissemination in tumor-bearing mice and helped better understand the mechanisms of tumor clearance in IV+IT combination immunotherapy.