9. rAAV Designs Harboring DNA Secondary Structures with High Thermal Stabilities Produce Heterogenic Viral Genome Populations

Abstract

It has been known for more than a decade that AAV genomes in ssAAV particles are heterogeneous, and can contain smaller than unit-length molecules to varying degrees. However, the cause underlying this phenomenon remained unknown. Our earlier work demonstrated that short DNA hairpin (shDNA) sequences designed into rAAV vectors, resulted in truncated genomes through a template-switching mechanism during AAV genome replication. Importantly, the knowledge we have gained about shDNA-mediated genomic truncation has helped to improve shRNA-cassette design. Specifically, we have found a correlation between the thermal stability of shDNA structures with the prevalence of truncation events. By introducing point mutations into the passenger strand of shRNA sequences to create DNA bulges within shDNAs in scAAV vectors, we found that lowering the thermal stability of DNA hairpins lowered the proportion of truncated to complete genomes. In addition, we found that scAAV vectors incorporated with pri-miRNA transgenes, which contain natural bulges in their stem-loop structures, also produced fewer truncated genomes, relative to shDNA-rAAV constructs. By embedding the guide strand of small RNAs into pri-miRNA scaffolds, we have defined critical improvements to the genomic integrity of rAAV vectors expressing small RNAs. Furthermore, we developed a method, named AAV-GPSeq (AAV genome populations sequencing), to directly sequence whole vector genome populations from purified rAAVs using the PacBio platform for high-throughput sequencing. We discovered through this methodology that truncation events can originate from palindromic sequences and inverted repeats that reside in the expression cassette elements of widely used ssAAV and scAAV vector designs (i.e. transcriptional regulatory elements, transgene sequences, and post-transcriptional regulatory regions). The resulting diversity of genomic truncations produces populations of packaged virions with variable transgene efficacies, and gives us the first insights into the phenomenon of rAAV heterogeneity. By examining the genomes from rAAV vectors harboring shDNAs, pri-miRNA fragments, and palindromic/inverted repeat sequences, we now highlight the importance of DNA secondary structure on vector design and genomic heterogeneity in packaged viral vectors. Improvements upon quality control standards are thus necessary and critical to efficacious and safe clinical uses for rAAV as a biomedicine. Further implications for our novel findings towards understanding AAV replication, and new considerations for future therapeutic rAAV vector designs will be discussed. To improve the homogeneity of clinical rAAV vectors, we are optimizing the vector production procedure, testing non-palindrome promoters, changing the codon usage in transgenes to lower the thermostability of rAAV genomes, and modifying rAAV packaging cell lines to minimize template-switching events to produce more intact rAAV genomes.