Design and Characterization of FLT210, a Potent Next Generation AAV-hFVIII Vector Candidate

Background: Several first-generation gene therapies are currently in clinical trials for Haemophilia A. These trials have to date exhibited varying results including issues with large patient to patient variation in FVIII levels and lack of durability. It is hypothesized that these challenges may, in part, be caused by the large size of the F8 transgenes used, resulting in genome sizes surpassing the optimal packaging capacity of AAV (~4.7kbp). We hypothesised that the oversize AAV vectors contain truncated genomes which cause batch to batch variations and decrease the quality of the resulting drug product.

Aim: Armed with clinical success with our proprietary AAVS3 capsid in Haemophilia B (Nathwani et al, 2018), we designed novel, next generation potent AAV-F8 vectors with genomes of less than 4.9kbp to increase the quality of the resulting product and the predictability of FVIII expression.

Methods: We deployed several methods to achieve our goal, including optimisations to the 'F8-SQ' coding sequence and optimisations of components of the expression cassette from the perspective of balancing size and potency of the resulting vectors. Over 20 optimized gene sequences, 10 different signal peptides, more than 60 short liver-specific promoters with a size of less than 270bp, and several shorter FVIII coding sequences were designed. We evaluated the resulting constructs for potency in human liver cell line Huh7 cells as well as in C57BI/6 mice. In addition, a range of different deletion lengths encompassing the B-domain were assessed for their in vitro and in vivo efficiency. Selected potent constructs were further evaluated in haemophilia A mice, and their haemostatic ability was demonstrated in the classical tail clip model.

Results: All elements of the expression cassette were optimized to generate a construct with more than 70-fold higher expression than the wild-type FVIII construct as evaluated in the human liver cell line Huh7. Importantly, we were able to develop a series of short promoters (<150 bp) that significantly shortened construct length while maintaining high expression levels and tissue specificity. A select few of the promising AAVS3-FVIII constructs were produced using a small-scale version of our manufacturing platform and all showed superior quality (i.e. packaged host cell and residual plasmid DNA, full-to-empty ratio) compared to larger AAVS3-FVIII-SQ vectors. When evaluated in haemophilia A mice, our top constructs had similar expression levels to previously developed constructs and were capable of controlling bleeding in the classical 'tail clip' bleeding assay. At the ASH 2019 meeting, we will present FLT210, our second generation AAV-hFVIII vector which we believe possesses the qualities and attributes required to potentially become the best in class vector to treat Haemophilia A patients.

Conclusion: We have developed several highly potent lead candidates with sizes of less than 4.9kbp, some being less than 4.75kbp in length. These candidates all have vector genomes markedly shorter in length than AAV-FVIII candidates currently in development and amongst these, we have selected the top best candidate, FLT210, for further development.