Human induced pluripotent stem cells (iPSCs), similar to human embryonic stem cells (ESCs), are capable of unlimited proliferation and have the potential to differentiate into all cell types of the body 
. These cells, thus, have applications in basic biology, disease modeling, drug development, and transplantation therapies. By expressing a defined set of reprogramming factors, iPSCs have been generated from many cell types of different species 
. Initial methods for iPSC generation employed genome-integrating retroviral or lentiviral vectors 
. These approaches could produce tumorigenic insertional mutations, and residual or reactivation of transgene expression during iPSC differentiation could affect lineage choice and the functionality of iPSC derivatives 
. To overcome these problems, various methods were developed to derive iPSCs free of exogenous DNA (footprint-free), including repeated treatments with reprogramming factors (plasmids, minicircle DNA, non-integrating adenoviral vectors and proteins), transposons and RNA viral vectors 
. However, these methods suffer one or more of the following limitations: the unacceptable low reprogramming efficiency; the labor-intensive removal of reprogramming factors from iPSCs; the requirement for viral packaging or feeder cells. Thus, there is a need to develop a simple and efficient feeder-free method to enable the routine derivation of footprint-free iPSCs from many human donor samples, and eventually the derivation of clinical-grade human iPSCs.
A recent report described the efficient derivation of footprint-free human iPSCs from fibroblasts using synthetic modified mRNA 
. Compared to viral and DNA-based reprogramming methods, the mRNA-mediated transgene delivery offers a safer approach for the derivation of clinical-grade human iPSCs. The requirement for repeated transfections, however, limits the application of this method to cells types that are easily transfectable such as skin fibroblasts. It remains to be seen whether this method can be readily adapted to cells that are relatively resistant to lipid-mediated transfections, such as cells of hematopoietic lineages. In addition to the mutations arising during reprogramming, somatic mutations present in the donor cells may also significantly affect the safety of human iPSCs. Therefore, the selection of appropriate donor cell types will likely be important for the derivation of clinical-grade human iPSCs. A reprogramming method that is applicable to different cell types will be highly desirable to address this question. Additionally, recent data suggest the retention of donor cell epigenetic memory in early passage iPSCs 
, which influences their in vitro
differential capacity. It remains to be seen whether this is affected by the specific methods employed in the derivation of iPSCs. Thus alternative methods are needed for the efficient derivation of human footprint-free iPSCs.
We have previously generated footprint-free human iPSCs using oriP/EBNA-1 (Epstein-Barr nuclear antigen-1) episomal vectors to deliver reprogramming genes (OCT4
. Compared to other methods, this approach has several advantages. First, the oriP/EBNA-1 vectors have a wide host cell range, enabling the application of this method to many human cell types. Second, it does not require viral packaging. Third, no repeated treatments with reprogramming factors are needed. A single transfection of episomal vectors is sufficient for the derivation of human iPSCs. Moreover, higher transfection efficiency can be achieved with these vectors due to the oriP/EBNA-1-mediated nuclear import and retention of vector DNA 
. Fourth, the oriP/EBNA-1 vectors replicate once per cell cycle and are generally present at low copy number per cell, thus minimizing DNA rearrangement and genome integration 
. Last, the removal of episomal vectors from human iPSCs can be accomplished by simple cell culture without any additional manipulation, due to the silencing of the viral promoter driving EBNA-1 expression in iPSCs, and the inherent instability of oriP/EBNA-1 episomal state - stably established episomes are lost from cells at a rate of ~5% per cell generation due to defects in vector synthesis and partitioning 
. However, despite these advantages, our original episomal method yielded low reprogramming efficiency (~3 iPSC colonies from ~1×106
input human foreskin fibroblasts), and used mouse embryonic fibroblast (MEF) feeder cells, which seriously limit the industrial and therapeutic applications of this method.
In this report, we have made significant improvement of the episomal reprogramming method. Using chemically defined media, we have established a small molecule-aided feeder-free reprogramming condition for the efficient derivation of footprint-free human iPSCs from skin fibroblasts, adipose tissue derived cells and cord blood cells. This method can be readily adapted to the derivation of clinical-grade human iPSCs. Of particular interest, iPSC derivation with this method appeared to progress through a distinct intermediate stage. It will be interesting to find out how this small-molecule aided episomal reprogramming method compares to other reprogramming methods in terms of the quality of iPSCs they generate.