Primed (left) and naive (right) human pluripotent stem cells. Image courtesy of J. Hanna.

Embryonic stem cells are derived from a transient population of cells in the early embryo, the inner cell mass (ICM), which gives rise to the various lineages of the developing organism. Cells of the ICM are not meant to stay pluripotent for very long. But this is precisely what researchers would like embryonic stem cells in the culture dish to do, in order to both study them as an in vitro model of development and use them as a source of other cell types in a myriad of applications in the laboratory and clinic. The trick, however, is in working out what the best culture conditions might be.

Work in the mouse in recent years has shown that culture in LIF-2i conditions (a combination of leukemia inhibitory factor and inhibitors of GSK-3β and ERK1/2) yields cells that are in a developmentally naive 'ground state' of pluripotency. These cells, it is thought, are the closest in vitro equivalent to the in vivo ICM. Conventional human embryonic stem cells (hESCs), however, are morphologically, molecularly and functionally quite distinct from these naive mouse cells. The question has remained open whether conditions can be found in which human cells can be maintained stably in a naive-like state.

In a recent paper, Jacob Hanna and colleagues at the Weizmann Institute in Israel describe such conditions. They started not with embryonic stem cells but with cells they had reprogrammed to pluripotency from fibroblasts, using doxycycline-inducible versions of the standard Yamanaka reprogramming factors. Previously they had reported that, in the continued presence of doxycycline, which maintains expression of reprogramming factors, they could grow naive-like human pluripotent cells in LIF-2i (conventional human pluripotent cells, by contrast, differentiate in LIF-2i). Although the researchers identified small molecules that could keep the cells naive even upon withdrawal of doxycycline, they could not stably maintain these without also ectopically expressing at least two reprogramming factors. In their present work, Hanna and colleagues searched for conditions that maintain naive-like human cells indefinitely in LIF-2i, in the absence of additional genetic manipulation.

After an initial screen of 16 factors that modulate a variety of signaling pathways, followed by much combinatorial optimization, the researchers settled on a cocktail they term naive human stem cell medium (NHSM). The resulting cells, molecularly characterized in fine detail and reported to be very similar to naive mouse cells, can be derived by reprogramming from somatic cells, from already established hESC lines and directly from blastocysts. The naive human cells have gene expression profiles more similar to the ICM than do conventional human pluripotent stem cells and are less heterogenous in their gene expression as well.

Perhaps most notably, when the human cells are injected into mouse morulae and the chimeric embryos allowed to develop for up to 8 days in the mouse in vivo, human cells can still be detected in the embryos by confocal microscopy at embryonic day 10.5, in contrast to previous reports with conventional hESCs. To what extent these cells actually contribute to developing tissues in this interspecies system remains unknown.

“One way that people deal with heterogeneity [in stem cells] is to accept it, to say let's profile the cells and predict their bias” says Hanna. “But there's another approach, which I prefer, and that's to say let's make them more uniform.” He predicts that naive human cells will prove more adept at differentiating along various lineages. The experiments are undoubtedly being done, and time will tell.