Engineering of programmable anti-infective gene circuits in mammalian cells

Abstract

The fusion between cell biology and engineering disciplines founded the field of synthetic biology, which aims to functionalize cells with desirable non-natural capabilities. Resulting engineered cells have served unmet medical or biotechnological needs and first applications of living therapeutics for cancer immunotherapy are revolutionizing clinical practice. In the introduction of this thesis, we therefore give an overview of available genetic circuits that tackle major disorders by autonomously correcting disease states. The spread of serious bacterial infections due to the evolution of antibiotic resistance jeopardizes the safety of patients and underscores the need for next-generation infection control strategies. Treatment failures are worsened by antibiotic tolerance caused by dormant, non- dividing persisters escaping elimination by antimicrobials. The misuse of antibiotics not only kills bacteria, but also selects for resistant as well as tolerant populations and has thus significantly increased the risk of a global health threat by multi-drug-resistant bacteria. Antibiotic resistance arises from spontaneous genetic mutations or from transfer of resistance genes between microbes and can be intensified by bacterial communication that synchronizes bacterial population behavior like biofilm formation. Antiinfectives that circumvent the selection for resistant clones and simultaneously silence bacterial communication are expected to mitigate the risk of resistance. Here we have focused on the design of anti-infective genetic programs operated in human cells and made the concept a reality. In the first approach, we demonstrated that microbial infections caused by diverse human pathogens can be attenuated with mammalian designer cells. The strategy used in this study was inspired by living therapeutics made from mammalian cells, currently being examined for a diverse range of metabolic and oncological disorders. To this end, cells were genetically equipped with infection-specific sensors that recognize by-products of bacterial protein translation, anaerobic metabolism or bacterial communication signals. The implemented broad range anti-infective genetic circuit detected bacterial or fungal pathogens and subsequently modulated microbial communication known as quorum sensing (QS) by timing the production of a universal bacterial signal molecule autoinducer-2 (AI-2). The synthetic sensor system recognized microbial signal peptides as a proxy signal for infection and activated the intended signaling cascade. This, in turn, resulted in the adjusted expression of AI- 2 biosynthetic enzymes and hence the production of AI-2 to interfere with QS-controlled virulence or biofilm formation of responsive pathogens. Because bacteria can alternatively be tackled by small molecule quorum sensing inhibitors, AI-2-engineered human cell lines were subsequently employed to identify and characterize antivirulent compounds. After benchmarking the assay with known inhibitors of AI- 2 biosynthesis and miniaturizing the assay to the nanoliter scale, we challenged cells with a library of natural products and chemical compounds in an automated screen and discovered compounds that block AI-2 activity with nanomolar affinity. Among those, the antineoplastic nucleoside analogue 5-Fluorouracil (5-FU) substantially decreased cellular AI-2 production, while lacking observable cytotoxicity or antibiotic activity. This discovery helped to explain how 5-FU and antibiotics accomplish synergism. For tackling Pseudomonas aeruginosa, a major source of healthcare-associated infections, human cells were instead equipped with a cytosolic autoinducer-inducible transcription factor to monitor bacterial intra-species communication. The biosensor was able to coordinate quorum-quenching via the release of an anti-infective triad, which accomplished hydrolysis of acyl-homoserine lactones (autoinducers) and biofilm disruption. Since the programmable therapeutic payload comprises engineered secreted catalytic effectors to silence bacterial communication, it led to reduced virulence. Noteworthy, the secretion-engineered effectors dissolved glycosidic biofilm shields of P. aeruginosa, thereby potentiating the killing efficacy of a traditional antibiotic. Efficient cell-based prevention and dispersal of notoriously difficult-to-treat biofilms was observed around the developed quorum-quencher cells. Last, we introduced the design and implementation of a mammalian lactate-responsive transgene control network that could serve as a sensor for sepsis as well as to score accumulation of this metabolite under hypoxic conditions. By rerouting a lactate sensor derived from human adipocytes with a synthetic adapter protein for the activation of transgene expression, we could link environmental lactate levels to a stable diagnostic readout. Beyond this, the genetic circuit might potentially monitor serious metabolic imbalances or could be exploited for control of bioprocessing of difficult-to-produce proteins.