Phenotypic Correction of a Mouse Model of Hemophilia B by In Vivo Genetic Correction of the F9 Gene

Abstract LBA-5

Inherited hematologic disorders have the potential to be effectively treated by gene therapy, with recent successes reported for several genetic disorders using viral vector-mediated gene transfer (ADA-SCID, NEJM 2009; β-thalassemia, Nature 2010). However, these trials and others illustrate some of the disadvantages and risks of using viral vector-based gene addition strategies, including loss of endogenous gene regulation and random insertion leading to potential for insertional mutagenesis. An alternative approach is gene correction, where in situ correction of a gene mutation allows endogenous gene regulation and decreases risks related to random integration. Gene correction is based on gene targeting, the therapeutic utility of which has historically been limited to mouse embryonic stem cells due to low homologous recombination rates in other cell types. However, a recently developed class of fusion proteins, zinc finger nucleases (ZFNs), have been shown to increase targeting efficiency 2–3 logs by inducing site-specific DNA double strand breaks at the intended targeting site. ZFNs have permitted high efficiency therapeutic gene targeting in a variety of cultured cells previously thought intractable to these processes, but ZFN-mediated gene correction has yet to be successfully achieved in vivo in an animal model of disease. Here we show ZFN-mediated therapeutic gene targeting of a mutated F9 gene in vivo, resulting in phenotypic correction of a mouse model of hemophilia B (HB). We first generated ZFNs targeting intron 1 of the human F9 gene (F9 ZFNs). We hypothesized the F9 ZFNs would mediate insertion of a wild-type F9 exons 2–8 minigene into intron 1 via gene targeting, thus bypassing the 95% of F9 mutations that occur in exons 2–8. We next generated a humanized HB mouse model with a deletion of the mouse F9 gene and knock-in (at the ROSA 26 locus) of a catalytic domain-deleted human F9 mini-gene (hF9mut) transgene. Adeno-associated viral (AAV) vector delivery of the F9 ZFNs to hF9mut mouse liver resulted in cleavage of the intron 1 target site in 45% of hepatocytes. We then generated an AAV donor vector containing a w.t. exons 2–8 insert flanked by arms of homology. Co-delivery of the AAV-ZFN and AAV-donor vectors to neonatal hF9mut mice (n=16) resulted in circulating F.IX levels of 120–350 ng/mL (2-7% of normal), whereas mice receiving AAV-ZFN alone (n=17) or AAV-mock & AAV-donor (n=15) had no detectable F.IX expression (detection limit 15 ng/mL), or <25 ng/mL F.IX, respectively. PCR analysis of liver DNA from ZFN+donor mice demonstrated genomic evidence of gene targeting at a rate of 2–7% of alleles. F.IX expression in ZFN+donor mice was shown to be stable after 5 months, with follow-up ongoing. In addition, there was no loss of expression following partial hepatectomy, which causes loss of expression from non-integrated episomes upon subsequent hepatocyte proliferation. F.IX expression was also shown to be specific, as opposed to resulting from random integration, as mice lacking the hF9mut gene averaged less than 30 ng/mL after receiving AAV-ZFN and AAV-donor. hF.IX RT-PCR on 10 different tissues confirmed liver-specific expression. To assess phenotypic correction, we performed aPTTs on mice that received ZFN+donor or mock+donor, as well as wild-type (WT) mice and HB mice. WT mice averaged 36 seconds, ZFN+donor mice averaged 44 seconds, mock+donor mice averaged 60 seconds, and HB mice averaged 67 seconds. There was no significant difference in aPTT between WT and ZFN+donor, or mock+donor and HB (p = 0.086 and 0.11, respectively). However, the aPTT for ZFN+donor mice was significantly shortened compared to mock+donor mice (p=0.0014), demonstrating phenotypic correction of the defect in clot formation in HB mice. To our knowledge this is the first demonstration of ZFN-driven gene correction in vivo, and the first demonstration of the in vivo use of ZFNs to correct an animal model of human disease. These results establish a novel paradigm for in vivo gene correction as a method for treating inherited hematologic diseases.