Preview

PDF, English Download (6MB) | Terms of use

Abstract

In recent years, the adeno-associated virus (AAV) gained considerable attention mainly due to the approval of the first AAV-based gene therapy treatment in the Western hemisphere in 2012, named Glybera®. It not only conveyed the feasibility of utilizing this parvovirus to introduce healthy gene copies but simultaneously reinforced further interest in developing more specific and efficient synthetic vectors by capsid engineering approaches such as DNA family shuffling or random peptide display. However, the characterization of lead candidates resulting from these directed evolution strategies is labor-intensive and therefore excludes the possibility to validate multiple promising variants. Therefore, a comprehensive high-throughput capsid validation pipeline was established in this work adapting a previously reported approach in which a DNA barcode-comprising AAV genome is assigned to a chosen capsid variant during virus production. Thus, the identification of the respective capsid in the complex physiological environment of living animals is enabled by solely detecting the barcode sequence via next generation sequencing. The principle was further improved by placing the barcode into the 3’UTR of a CMV promoter-driven eyfp transgene permitting tracking on the DNA and RNA level. Hence, next to information about transduction efficiency, the especially crucial transcriptional activity in a certain tissue was measured. Using this design, three barcoded AAV libraries were generated comprising up to 157 variants including 12 commonly used serotypes, >70 peptide-displaying mutants based on these naturally occurring wild types and several published benchmarks such as AAVDJ, AAV9_PHP.B and AAVAnc80L65. After intravenously injecting the library into C57BL/6J mice and analyzing the RNA and DNA data from >20 collected tissues, prior observations for the literature variants could be confirmed thus validating the workflow. Most impressively, a peptide display mutant previously created in our laboratory exhibited drastically improved efficiencies in the diaphragm, heart and skeletal muscles in comparison to AAV9wt on the cDNA and protein level while in addition demonstrating pronounced muscle specificity. In conclusion, in the course of this PhD thesis a highly robust barcode-based capsid screening pipeline was established that facilitates and accelerates the identification of promising candidates for gene therapies, best exemplified by the discovery of the muscle-tropism of our lead candidate.