ZFN Mediated Targeting Of Albumin “Safe Harbor” Results In Therapeutic Levels Of Human Factor VIII In a Mouse Model Of Hemophilia A

Genome editing utilizing engineered zinc finger nucleases (ZFNs) is a promising approach to achieve long-term expression of therapeutic genes in vivo. We have previously demonstrated in vivotargeting of the endogenous murine albumin locus as a “safe harbor” for high levels of protein production, resulting in sufficient Factor IX to correct the disease phenotype in hemophilia B mice. Targeted insertion of the donor sequence into the genome offers multiple advantages. First, we are able to exploit the high transcriptional activity of the native albumin enhancer/promoter. Second, by obviating the need for these regulatory elements within the donor, we expand the effective carrying capacity of adeno-associated viral (AAV) vectors to enable delivery of larger transgenes that may not package efficiently, such as coagulation factor 8.

B-Domain Deleted Factor VIII (BDD-F8) cDNA is approximately 4.4kb. Inclusion of required enhancer/promoter elements results in a construct that exceeds the ideal packaging limitations of rAAV vectors. However, since these regions are not required for our gene editing approach the promoterless hBDD-F8 donor remained below the AAV packaging capacity. Importantly, intravenous delivery of 5e¹¹ vg of AAV8-mAlb-ZFN and 5e¹¹ vg of AAV8-BDD-F8-Donor to hemophilia A mice resulted in 54.6% (±4.1%) FVIII activity in the blood 2 weeks following administration - confirming the potential of the albumin locus to express high levels of the targeted transgene. To further optimize in vivo ZFN-based genome editing with a view toward the ultimate clinical use of this technology, we sought methods to further increase ZFN potency (and thus limit the dose of AAV necessary for function). Of several strategies pursued (e.g. codon optimization and inclusion of a intron in the expression cassette) the most successful was to deploy separate vectors expressing each individual ZFN rather than a single vector encoding a dual expression cassette carrying both ZFNs separated by a 2A fusion peptide. Using next generation sequencing (Illumina’s MiSeq) technology to quantify insertions and deletions indicative of DNA cleavage and repair, we observed a >3-fold increase in ZFN potency in vivo by transitioning from the dual expression vector to two individual ZFN vectors at equivalent total vector doses. Given the encouraging results obtained in mice, we next sought to examine the effectiveness of targeting the albumin locus in non-human primates (NHPs). Importantly, a single intravenous co-injection of two individual AAV vectors encoding each of the NHP targeted albumin-specific ZFNs resulted in persistent levels of gene modification in liver biopsies from treated Rhesus macaques - demonstrating successful in vivocleavage in a large animal model.

These data support the use of ZFN technology in the targeting of endogenous loci with large therapeutic transgenes that are not ideally suited for episomal AAV based expression (such as F.VIII). Together our results support the further investigation of genome editing at the albumin locus as a novel method for in vivo protein replacement.