Induced pluripotent stem cells from different sources behave differently, even if all have passed stringent criteria for pluripotency, according to work published in Nature Biotechnology.

To assess differences in safety, Shinya Yamanaka and colleagues at Kyoto University took and subjected 36 induced pluripotent stem cell lines to a neural differentiation program. More specifically, they cultured the cells to form neurospheres, balls of neural tissue that can be made to form in culture. Then they disaggregated the neurospheres and induced their formation a second time. Next, they tested whether the neurospheres could form teratomas.

They compared ES cells with iPS cells generated from mouse embryonic fibroblasts, tail-tip fibroblasts, gastric epithelial cells, and liver cells (hepatocytes). The iPS cell lines were generated with and without the use of c-Myc, an oncogene which also boosts the number of cells that become iPS cells. The team found that the presence or absence of cMyc did not seem to affect whether those cells from neurospheres would form a teratoma.

However, the source of the cells did.1 Over four-fifths of the 55 mice transplanted with neurospheres derived from one of 11 lines derived from tail-tip fibrobasts became weak or died from a teratoma. So did one-third of the 36 mice transplanted with one of 7 lines derived from hepatocytes, as did quarter of the 8 mice transplanted with one of two lines derived from stomach epithelial cells. This was true for just under a tenth of mice transplanted with cells from mouse embryonic fibroblasts or mouse embryonic stem cell lines. (9 of 100 mice with one of 12 lines and 3 of 34 mice with one of 3 lines respectively.)

The researchers also looked at undifferentiated cells in the neurospheres and saw significant differences depending on the tissue of origin;the persistence of undifferentiated cells correlated with teratoma formation. Thus, tail-tip fibroblasts seem far more resistant to differentiation than other cell types examined. Still, there is further work to be done. Every iPS cell line assessed was generated by using viruses to insert pluripotency genes into the cells, but this process creates lines with different numbers of insertions at different places in the genome. Further work will evaluate iPS cells made with methods that limit that kind of variation, and even more work will be required to understand the potential sources of variation for human cells. When presenting this work at a recent conference, Yamanaka was reluctant to speculate about what exactly these results might mean for the use of iPS cells as cell therapies. “Humans do not have tails,” he said. The only thing this shows is that much more work must be done to evaluate iPS cells.