For the clinical use of products derived from human pluripotent stem cells, it is essential to improve both the efficacy and safety of the final product. In order to develop safe hiPSC-based treatments, the hurdle of tumorigenicity arising from undifferentiated cells must be overcome 
. To address the issue of tumorigenicity, some recent publications have advocated the development of protocols for the derivation of hiPSC 
and have outlined methods for the elimination of residual hESCs 
. However, to date, more than several hundred cells are necessary for human pluripotent stem cells to form a tumor in immunocompromised mice 
. Therefore, highly sensitive tumorigenicity assays and their standardization are necessary for detecting a small population of residual undifferentiated cells in products derived from human pluripotent stem cells.
In the present study, we evaluated three methods for detection of residual undifferentiated hiPSCs in hiPSCs-derived differentiated cells: soft agar colony formation assay, flow cytometry and qRT-PCR. summarizes the advantages and disadvantages of the assays associated with product tumorigenicity. The soft agar colony formation assay is known to be more sensitive in the detection of certain types of tumorigenic cells, compared to in vivo
methods using immunocompromised mice 
. However, we found the soft agar colony formation assay unsuitable for detecting residual undifferentiated hiPSCs, presumably attributable to the dissociation-induced apoptosis of hiPSCs 
. On the other hand, flow cytometry and qRT-PCR assays were found to be able to detect a trace amount of undifferentiated cells. These two assays have been exploited for characterization of stem cell-based products, as well as undifferentiated pluripotent stem cells, but the present study is the first analytical and quantitative approach designed to evaluate the detection of residual undifferentiated cells in products derived from human pluripotent stem cells. The advantage of the flow cytometry assay is that it is able to identify undifferentiated cells. Unfortunately, the results are greatly affected by gating, and only the cells expressing the marker protein are detectable. On the other hand, the advantages of qRT-PCR are its rapidity, quantitativity and high sensitivity, whereas its disadvantage is that only the cells expressing the marker gene are detectable. Although the in vivo
tumorigenicity assay is costly and time-consuming, it can directly analyze tumor formation in a specific microenvironment where the product is implanted (eg. retina). Therefore, a combination of relevant in vitro
and in vivo
assays would be necessary to ensure the safety of a product derived from human pluripotent stem cells. The rationale for the choice of specific assays would be justified, based on their characteristics shown in .
Comparison of the tumorigenicity-associated assays.
We have demonstrated that the qRT-PCR assay can successfully detect 0.002% residual undifferentiated hiPSCs in hiPSC-induced RPE cells using Lin28 as a target gene (). To the best of our knowledge, this qRT-PCR assay, using solely 250 ng of total RNA obtained from approximately 5×104
cells, is the most sensitive of the previously reported methods in detecting undifferentiated pluripotent stem cells. Lin28 is known to specifically inhibit the processing of let-7 miRNAs, which are involved in cell-fate decisions 
. Interestingly, the aberrant expression of Lin28 transcripts has been recently reported in human germ-cell tumors 
, suggesting that Lin28 is a useful marker of germ-cell malignancy as well as of pluripotency of hiPSCs. Lin28 mRNA gradually decreased in the differentiation process and was completely diminished by passage 4 (). These observations suggest that Lin28 transcripts are also available for presenting degree of differentiation in hiPSC-derived products because detection of residual Lin28 confirms the contamination of undifferentiated cells even at a late stage of differentiation. Needless to say, the distinct expression of Lin28 could possibly be observed in other normal somatic cells. However, Lin28 is, at least in part, one of the potent markers for detecting incompletely differentiated cells contained in RPE cells derived from pluripotent stem cells.
A great deal of international research is currently being directed at developing regenerative medicine using pluripotent stem cells. Until now, however, little attention has been paid to developing methods to detect undifferentiated cells in pluripotent stem cell-based products. Here, we have revealed that a qRT-PCR method targeting Lin28 can effectively detect a trace amount of hiPSCs in hiPSC-induced RPE cells and shows potential as an in vitro tumorigenecity assay of hiPSC-derived cells. We expect our findings to contribute to the process of validation and quality control of hiPSCs-based cell therapy products and to promote the application of regenerative medicine in the treatment of a wide variety of diseases in the near future.