• Potential conflict of interest: Nothing to report.


We thank Moriguchi et al. for their interest in our study1 recently published in Hepatology. We share the view that tumor risk evaluation for induced pluripotent stem (iPS) cells is important and that clinical applications of human iPS cells require safe iPS generation. Our article is the first to focus on the feasibility of deriving human iPS cells from cells of human endoderm origin (i.e., human primary hepatocytes that are very short lived in vitro and thus are considered to be hard to reprogram into iPS cells) and redirecting these endoderm-derived iPS cells into multistage hepatic cells.

Because we are fully aware of the importance of tumor risk evaluation for iPS cells and their derivatives, in vivo transplantation studies with our human endoderm iPS cells have been performed. So far, we have not detected any malignant transformation with hepatic cells differentiated from iPS cells when they have been transplanted into immunodeficient mouse models of chronic liver diseases.

More importantly, we have also derived virus-free and integration-free human iPS cells (Fig. 1A) with an established nonviral approach,2 and we have been able to differentiate these virus-free iPS cells into hepatic endoderm cells (Fig. 1B). Therefore, we have demonstrated that the hepatic differentiation protocol that we used in our previous study1 is also applicable to the hepatic differentiation of virus-free and integration-free human iPS cells (Fig. 1).

Figure 1.

(A) Representative immunofluorescence analysis of one of the integration-free human iPS cell lines (iNF5) growing on Matrigel. The clear expression of embryonic stem cell surface antigens SSEA4 and TRA-1-60 and nuclear transcription factors OCT4 and NANOG can be observed. (B) Differentiation of iNF5 cells into hepatic cells: efficient endoderm induction of integration-free human iPS cells. Fluorescence-activated cell sorting analysis showed that approximately 80% of the induced cells expressed the definitive endoderm marker CXCR4 on day 4 after activin A treatment. Abbreviations: CXCR4, chemokine (C-X-C motif) receptor 4; DAPI, 4′,6-diamidino-2-phenylindole; OCT4, octamer 4; SSEA, stage-specific embryonic antigen.

Although we have observed reliable iPS generation with the nonviral method (Fig. 1), it is less efficient than the viral protocol; the reprogramming process has been significantly longer with the nonviral method (17-35 days for the detection of iPS colonies) versus the viral method that we reported (6-14 days for the detection of iPS colonies).1 Therefore, novel reprogramming technologies that can efficiently generate safe iPS cells from mature cell types will be highly beneficial to the field.

Hua Liu*, Yonghak Kim*, Saul Sharkis*, Zhaohui Ye*, Yoon-Young Jang*, * Johns Hopkins University School of Medicine, Baltimore, MD.