The following are the supplementary data related to this

The following are the supplementary data related to this article.
Acknowledgments We thank Kersti Jääger and Angelika Fatkina for their help in flow cytometry experiments. We are very thankful to Dr. Sirje Rüütel Boudinot, Dr. Marko Piirsoo and Jekaterina Kazantseva for valuable comments and discussions. We are indebted to Jelena Arshavskaja for technical assistance in ELISA experiments and Grete Rullinkov for isolation of peripheral blood mononuclear cells. This study was supported by the Estonian Science Foundation grant MJD266 to AP.
Introduction The liver plays a central role in multiple functions of the human body including detoxification of drugs and storage of nutrients. Animal models exhibit different metabolic profiles from humans and are thus not directly applicable to humans. Human primary hepatocytes are valuable for assessing drug toxicity and testing drug kinetics (Davila et al., 2004) but these applications are impeded from limited availability and differences between samples. In addition to their use in drug development, hepatocytes are expected to be an alternative therapeutic approach for producing transplantable hepatocytes since severe liver dysfunctions are treatable only by liver transplantation in many cases. Embryonic stem (ES) cells are derived from the inner cell mass of the fertilized egg, which is pluripotent, and can be cultured indefinitely in an undifferentiated state and have the potential to differentiate into various cell types derived from the three germ layers (Evans and Kaufman, 1981; Kaufman et al., 1983; Suemori et al., 2001; Thomson et al., 1998). Induced pluripotent stem (iPS) cells (Takahashi and Yamanaka, 2006; Takahashi et al., 2007, 2009; Belmonte et al., 2009) resemble ES cells but raise fewer ethical concerns because they can be derived from somatic cells. ES cells and iPS cells represent powerful tools for studying basic developmental biology and are potential sources of hepatocyte generation for applications in regenerative medicine and drug development (Davila et al., 2004). ES and iPS cells have been shown to recapitulate developmental processes and differentiate into hepatic lineages in vitro following the sequential addition of growth factors. Recently, human ES cells were directed to differentiate in a stepwise manner into the definitive endoderm, hepatic lineages, and hepatocytes following the addition of Activin/Nodal, fibroblast growth factor (FGF), BMP, retinoic Cy5.5 hydrazide (RA), Wnt3a and dimethyl sulfoxide (DMSO), hepatocyte growth factor (HGF), and oncostatin M (OSM) (Touboul et al., 2010; Shiraki et al., 2008a, 2008b; Sullivan et al., 2010; Hay et al., 2008a, 2008b; Cai et al., 2007; Zaret and Grompe, 2008; Si et al., 2010; Katsumoto et al., 2010). We have previously established a procedure in which ES cells are sequentially induced in vitro into the mesendoderm, definitive endoderm, and regional specific digestive organs such as the pancreas and liver (Shiraki et al., 2008a, 2008b, 2009). Approximately 80% of the hES that were cultured on a monolayer of the M15 mesonephric-derived cell line and given growth factors differentiated into α-fetoprotein (AFP)-positive hepatic precursor cells and approximately 10% of cells later became albumin (ALB)-positive cells and secreted a substantial level of ALB protein (Shiraki et al., 2008b). Recently, we found that optimization of the culture condition and culturing cells on a synthesized basement membrane (sBM) resulted in an improvement of ALB secretion (Shiraki et al., 2011). Modification of endogenous genes by homologous recombination is well established and practiced in mouse ES cells. However, although several reports have described gene targeting by electroporation in hES cells (Zwaka and Thomson, 2003; Urbach et al., 2004; Di et al., 2008; Ruby and Zheng, 2009; Sakurai et al., 2010), the efficiencies of stable transfection (random integration) were extremely low (approximately 1×10 per cell) and the ratio of targeted to random chromosomal integration was also low (approximately 2%). In contrast, the helper-dependent adenovirus vector (HDAdV) was a reportedly efficient and versatile gene targeting vector in mouse ES cells and allowed the insertion of a long homologous sequence of up to 35kb (Ohbayashi et al., 2005). HDAdV was originally developed to overcome host immune responses against E1D AdVs in vivo (Palmer and Ng, 2005). Because all of the viral genes are removed from the vector genome, HDAdVs offer additional advantages such as decreased cytotoxicity and expanded cloning capacity, permitting insertion of longer segments of homologous DNA for homologous recombination. These features are advantageous for obtaining highly efficient site-specific integration into host chromosomes through homologous recombination. The HDAdV system was shown to be a powerful tool for genetic manipulation in hES and human iPS (hiPS) cells. When HDAdVs were used to target several loci, the recombination efficiencies were 3–55% (Suzuki et al., 2008; Aizawa et al., 2012). Recently, targeted gene correction in patient-specific iPS cells was reported using HDAdVs (Liu et al., 2011; Li et al., 2011). Taken together, gene transfer mediated by HDAdVs is a powerful technology for genetic manipulation in hES and hiPS, both for biological studies and therapeutic applications.