Mice that could be highly repopulated with human hepatocytes would have many potential uses in drug development and research applications. The best available model of liver humanization, the uroplasminogen-activator transgenic model, has major practical limitations. To provide a broadly useful hepatic xenorepopulation system, we generated severely immunodeficient, fumarylacetoacetate hydrolase (Fah)-deficient mice. After pretreatment with a urokinaseexpressing adenovirus, these animals could be highly engrafted (up to 90%) with human hepatocytes from multiple sources, including liver biopsies. Furthermore, human cells could be serially transplanted from primary donors and repopulate the liver for at least four sequential rounds. The expanded cells displayed typical human drug metabolism. This system provides a robust platform to produce high-quality human hepatocytes for tissue culture. It may also be useful for testing the toxicity of drug metabolites and for evaluating pathogens dependent on human liver cells for replication.The liver is the principal site for the metabolism of xenobiotic compounds, including medical drugs. Because many hepatic enzymes are species specific, it is necessary to evaluate the metabolism of candidate pharmaceuticals using cultured primary human hepatocytes 1,2 . Today, hepatocytes are isolated primarily from cadaveric organs and then shipped to the location where testing will be performed. The condition (viability and state of differentiation) of hepatocytes from cadaveric sources is highly variable, and many cell preparations are of marginal quality. The availability of high-quality human hepatocytes is further hampered by the fact that they cannot be substantially expanded in tissue culture 3,4 . Hepatocytes from readily available mammalian species, such as the mouse, are not suitable for drug testing because they have a different complement of metabolic enzymes and

Site-specific gene addition can allow stable transgene expression for gene therapy. When possible, this is preferred over the use of promiscuously integrating vectors, which are sometimes associated with clonal expansion1 and oncogenesis2. Site-specific endonucleases that can induce high rates of targeted genome editing are finding increasing applications in biological discovery and gene therapy3. However, two safety concerns persist: (1) endonuclease-associated adverse effects, both on4 and off-target5,6; and (2) oncogene activation caused by promoter integration, even without nucleases7. Here, we perform recombinant adeno-associated virus (rAAV) mediated promoterless gene targeting without nucleases and demonstrate amelioration of the bleeding diathesis in haemophilia B mice. In particular, we target a promoterless human coagulation factor IX (hF9) gene to the liver-expressed albumin (Alb) locus. hF9 is targeted, along with a preceding 2A-peptide coding sequence, to be integrated just upstream to the Alb stop codon. While hF9 is fused to Alb at the DNA and RNA levels, two separate proteins are synthesized by way of ribosomal skipping. Thus, hF9 expression is linked to robust hepatic albumin expression without disrupting it. We injected an AAV8-hF9 vector into neonatal and adult mice and achieved on-target integration into ~0.5% of the albumin alleles in hepatocytes. We established that hF9 was produced only from on-target integration, and ribosomal skipping was highly efficient. Stable hF9 plasma levels at 7–20% of normal were obtained, and treated factor IX deficient mice had normal coagulation times. In conclusion, transgene integration as a 2A-fusion to a highly expressed endogenous gene may obviate the requirement for nucleases and/or vector-borne promoters. This method may allow for safe and efficacious gene targeting in both infants and adults by greatly diminishing off-target effects while still providing therapeutic levels of expression from integration.

Glycogen storage disease type II or Pompe disease is a severe neuromuscular disorder caused by mutations in the lysosomal enzyme, acid α-glucosidase (GAA), which result in pathological accumulation of glycogen throughout the body. Enzyme replacement therapy is available for Pompe disease; however, it has limited efficacy, has high immunogenicity, and fails to correct pathological glycogen accumulation in nervous tissue and skeletal muscle. Using bioinformatics analysis and protein engineering, we developed transgenes encoding GAA that could be expressed and secreted by hepatocytes. Then, we used adeno-associated virus (AAV) vectors optimized for hepatic expression to deliver the GAA transgenes to Gaa knockout (Gaa−/−) mice, a model of Pompe disease. Therapeutic gene transfer to the liver rescued glycogen accumulation in muscle and the central nervous system, and ameliorated cardiac hypertrophy as well as muscle and respiratory dysfunction in the Gaa−/− mice; mouse survival was also increased. Secretable GAA showed improved therapeutic efficacy and lower immunogenicity compared to nonengineered GAA. Scale-up to nonhuman primates, and modeling of GAA expression in primary human hepatocytes using hepatotropic AAV vectors, demonstrated the therapeutic potential of AAV vector–mediated liver expression of secretable GAA for treating pathological glycogen accumulation in multiple tissues in Pompe disease.

Existing recombinant adeno-associated virus (rAAV) serotypes for delivering in vivo gene therapy treatments for human liver diseases have not yielded combined high-level human hepatocyte transduction and favorable humoral neutralization properties in diverse patient groups. Yet, these combined properties are important for therapeutic efficacy. To bioengineer capsids that exhibit both unique seroreactivity profiles and functionally transduce human hepatocytes at therapeutically relevant levels, we performed multiplexed sequential directed evolution screens using diverse capsid libraries in both primary human hepatocytes in vivo and with pooled human sera from thousands of patients. AAV libraries were subjected to five rounds of in vivo selection in xenografted mice with human livers to isolate an enriched human-hepatotropic library that was then used as input for a sequential on-bead screen against pooled human immunoglobulins. Evolved variants were vectorized and validated against existing hepatotropic serotypes. Two of the evolved AAV serotypes, NP40 and NP59, exhibited dramatically improved functional human hepatocyte transduction in vivo in xenografted mice with human livers, along with favorable human seroreactivity profiles, compared with existing serotypes. These novel capsids represent enhanced vector delivery systems for future human liver gene therapy applications.