A cell biologist looks at the risk and promise of a new insight into stem cells and cancer.

I study both stem and tumour cells, and am fascinated by their close relationship. Both exhibit pluripotency — the capacity to develop into any cell type — and the ability to cause cancer. Even some apparently normal stem cells can cause tumours, whereas others, sometimes from the same culture, lack this power. It seems that not all stem cells are created equal — even in the same dish.

A recent paper from Mickie Bhatia's group (T. E. Werbowetski-Ogilvie et al. Nature Biotechnol. 27, 91–97; 2009) is the first to directly address this heterogeneity in human embryonic stem cell (ESC) cultures. The team found that individual human ESC lines contain significant subpopulations that vary in a number of ways, including in tumorigenicity.

Variant human ESC lines were about 20 times more tumorigenic than the cultures they had been derived from and showed small changes in chromosome structure. These could be identified by array-based comparative genomic hybridization (aCGH), but were not detectable by standard karyotyping. Thus for 'normal' stem cells being considered for use in regenerative medicine, karyotyping is not enough. Screening should also include aCGH, and perhaps an analysis of gene-expression patterns.

This previously covert diversity has implications for both tumour biology and medical applications involving stem cells. It may shed light on the 'locked in' self-renewal that is emerging as an important feature of many sorts of tumour and tumour stem cell.

The heterogeneity of human ESC cultures represents an additional hurdle in terms of producing safe stem-cell-based transplants. At the same time, it may offer a valuable bonus: the chance to purify variant human ESC sub-lines that are less tumorigenic.

Discuss this paper at http://blogs.nature.com/nature/journalclub