Improving bladder cancer immunotherapy by tailoring protein secretion in mycobacteria

Bladder cancer is the 4th to 10th most common cancer worldwide, with non-muscle invasive bladder cancer (NMIBC) counting for 70-75% of cases. Standard treatment of NMIBC includes resection of the tumor followed by intravesical treatment with Mycobacterium bovis BCG (BCG). BCG is a live but attenuated vaccine strain well-known for its use against tuberculosis, an infectious disease caused by Mycobacterium tuberculosis. BCG bacteria induce a local inflammatory response that not only clears the bacterium but also eradicates the remaining tumor cells. Although effective, BCG treatment may still show up to 30% recurrences. The research described in this thesis aims for improvement of bladder cancer treatment. First, since secretion and surface display of (tumor) antigens might improve BCG anti-tumor efficacy, the aim was to enable heterologous secretion in mycobacteria. Mycobacteria use specialized secretion systems to secrete proteins over their unique cell envelope, known as Type VII secretion systems. For heterologous secretion we focused on one of these systems, the ESX-5 system. Model protein OVA was fused to the PE and linker domain of ESX-5 substrate LipY. Screening of an M. marinum error-prone PCR mutant library showed that two mutations in OVA or deletion of the linker domain lead to increased secretion. Thus, the ESX-5 secretion system is suitable for heterologous secretion in mycobacteria but optimization is required for each desired heterologous protein. Second, we focused on the function of ESX-5. Studying the processing of LipY in more detail lead to the discovery of PecA, an ESX-5 dependent secreted aspartic protease that cleaves the PE domain of LipY, other PE_PGRS proteins and itself. Furthermore, deletion of pecA resulted in an attenuated phenotype in zebrafish larvae. Next, we studied the role of ESX-5 in fatty acid uptake. In tuberculosis mycobacteria reside in the lungs inside structures known as granulomas, in which lipids form the main energy source. By screening for transposon mutants showing increased growth on Tween 80 as sole carbon source, two PE proteins, MMAR_0369 and MMAR_4321, were identified to be able to cleave fatty acids. These cutinase-like proteins harbor an α/β hydrolase domain, a domain also predicted for multiple other proteins in M. tuberculosis and M. marinum. Together, these findings demonstrated that PE proteins can have distinctive functions depicted by their C-terminal domains and not by the PE domain by which they are usually categorized. The main aim was to improve bladder cancer immunotherapy. To study anti-tumor activity in vivo we used an orthotopic bladder cancer mouse model. We showed that combining multimodal imaging techniques, e.g. high ultrasound imaging, photoacoustic imaging and bioluminescent imaging, enables longitudinal studies on tumor development in the bladder. The integration of individual strengths enabled sensitive and improved quantification and understanding of tumor biology and ultimately can aid in the discovery and development of new therapeutics. Finally, we tested the hypothesis that anti-tumor activity can be improved by enhancing bacterial binding to the bladder wall. Recently, Salmonella enterica Typhi Ty21a (Ty21a), an oral vaccine strain against typhoid fever, has been shown to be potentially more effective against bladder cancer than BCG. Labeling of Ty21a with adhesion protein FimH resulted in a ~5 fold increased binding to the mouse bladder in vivo. Furthermore, a single bladder cancer mouse study showed a modest increase in median survival for mice treated with FimH-labeled Ty21a. In conclusion, this thesis presents a study on mycobacterial protein secretion in order to improve bladder cancer immunotherapy. We identified novel functions for 3 PE proteins providing more insight into the function of ESX- 5 and its role in virulence. Furthermore, we showed that labeling Ty21a with FimH shows promise for improving bladder cancer treatment.