Design of synthetic promoter-based gene circuits for the surveillance of cellular state transitions

Abstract

In recent years, synthetic biology and biomolecular computing have joined forces to create programmable synthetic gene networks resembling naturally occurring signaling or regulatory pathways. Among them, ‘cell classifiers’ autonomously assess the presence or absence of multiple endogenous bimolecular inputs, integrate them according to user-defined logic programs and elicit customizable responses in a cell state-specific manner. Several in vivo studies have demonstrated that such devices hold great potential to herald a new era of gene-therapeutic modalities that will eventually replace the prevalent ‘one-target – one-drug’ paradigm and revolutionize the treatment of complex multifactorial diseases, such as cancer. However, despite considerable progress in the design and clinical translation of gene circuits, the dynamic nature of cellular states has not yet received adequate attention. Conventional classifiers have primarily been employed for distinguishing between static ‘state snapshots’ of different cells, e.g., healthy vs. cancerous. Conversely, the use of multi-input biomolecular computing systems for monitoring and, if necessary, manipulating dynamic cell state transitions, a process otherwise known as surveillance, remains unexplored. In this thesis, I describe the rational design and implementation of a synthetic surveillance gene circuit capable of continuously monitoring the state of an individual cell and triggering a programmable response only when the cell transitions to another predefined state. As a prototypical state transition, I use the epithelial-to-mesenchymal transition (EMT), a reversible and evolutionarily conserved developmental program with major physiological and pathological relevance. To recapitulate EMT in vitro, I employ a well-established cell culture model, wherein A-549 lung adenocarcinoma cells are treated with transforming growth factorß1 (TGF-ß1). Starting from global RNA-sequencing (RNA-seq) analysis, I devise an unbiased systematic approach for constructing and testing promoter-based state detectors that utilize endogenous gene expression levels as inputs. Further, I show that two state detectors with opposing state specificity can readily be combined in multi-input gene circuits implementing ‘M-state’ AND NOT (‘E-state’) logic. This circuit architecture significantly enhances targeting precision and, given its high modularity, facilitates tunable control over output expression strength. Upon lentiviral integration, the surveillance circuit robustly responds to EMT induction with minimal false positive activation in epithelial cells and maintains stable expression over extended periods of time. Finally, substituting the fluorescent output with a suicide gene, commonly used for cancer gene therapy, enables the highly selective killing of cells that undergo EMT. These findings demonstrate that the circuit’s surveillance capacity can effectively be translated into a biologically relevant effect, providing the opportunity to purposefully interfere with cellular behavior. Ultimately, I discuss how the current circuit design could be further optimized and propose potential future clinical applications for the EMT surveillance gene circuit developed in this thesis.