Are ruddy cheeks a sign of health or a symptom of sickness? New work from Mickie Bhatia and colleagues at McMaster University suggests that, when it comes to embryonic stem cells, the very qualities researchers use to pick out a robust cell line may in fact be bestowed by precancerous transformations. “Current measurements are not capable of distinguishing the difference between great stem cells and cancer stem cells in vitro,” says Bhatia.

The researchers identified two variants of the common human embryonic stem (hES) cell line H9; they consistently looked different from other populations from the same line. Colonies of the cells had ill-defined borders; the cells did not generate the usual number of support cells. However, according to the tests generally used to assess an ES cell line, they appeared perfectly normal. Typically for these assessments, a handful of cells are examined to make sure they have the expected number of chromosomes. Bhatia's team looked at 500 cells per variant and found no abnormalities.

Further characterization of the variants showed they had high levels of the markers usually associated with pluripotency, proliferated more quickly than other ES cells (but not as quickly as carcinoma cells), and they could get by on less of the growth factors that most hES cells require to grow robustly. All these are traits that Bhatia believes would make them easily mistaken for high quality ES cells.

When injected into mice, the variant cells formed the teratomas typically used to assess pluripotency. In fact, these variants generated teratomas at more than 20 times the rate of other cells. But while the teratomas formed from these cells still attracted blood vessels and still formed all three of the germ layers typical of a developing embryo, compared to teratomas generated from other ES cells, cells from the different layers appeared less distinct. Moreover, unlike teratomas generated from standard hES cells, teratomas from the variants still contained pockets of cells that expressed Oct4, a telltale marker for pluripotency as well as growth teracarcinoma growth transformation. Techniques to differentiate the variant ES cells into more-specialized blood and neural cell types were less successful than with the normal strains. Although these variant cells are not malignant, they are not ideal for developing into therapies. Bhatia suspects that many researchers do not realize that the cells are problematic until they run into problems completing lengthy differentiation protocols.

Ultimately, Bhatia and his team used DNA sequencing assays far more sensitive than those typically used to find differences between the variant strains and standard strains; these can generally find changes on the order of 5-10 megabases, whereas Bhatia used a technique that could find so-called copy-number variants (CNVs), or duplications of genomic regions, about one-thousandth of that length. In one variant cell line, his team found small amplifications on chromosome 20 in a slightly different region from what has previously been reported on other stem cell lines. The other variant seemed to contain at least two kinds of cells; all had a small deletion in chromosome 5, and a very small subset had an extra copy of chromosome 12.

Bhatia feels cataloguing every genetic change that might transform ES cells is not the best way to assess qualities. “We know there are lots of ways, outside CNVs, to get variants,” he says. The functional assays can take two or three months, “but right now, this is all we have, and it is the only way to tell the great stem cells from the transformed ones.” The best way for labs to detect bad lines, he says, is to check whether the cells continue to grow without the usual growth factors and and whether they fail to create an in vitro niche by generating their own support cells.

“This paper shows that human ES lines with submicroscopic genetic abnormalities can display altered growth and differentiation properties suggestive of premalignant change,” says Martin Pera who studies embryonic stem cells at the University of Southern California in Los Angeles. “In other words, a normal karyotype is not necessarily a guarantee of a normal genetic makeup within a cell line.” Though Bhatia's team did not show that the genetic abnormalities caused the changes in the cells' behavior, he says, other specific recurrent genetic lesions suggest that changes at certain loci somehow provide an advantage for altered cells to grow in culture. One of the “major challenges to the field” is developing techniques that can detect rare, abnormal cells, particularly if the transformations are not due to changes in gene sequence.

Also important, he says, is figuring out just how and when such cells might be dangerous. “Ultimately it may be difficult or impossible to rule out with certainty that a given culture is totally free of abnormal cells.”

Related articles

How to assess a stem cell genome Stem cell lines accrue cancer mutations