Continuously renewing tissues, such as the epidermis, are maintained by stem cells that slowly proliferate and remain in the tissue for life. Although it has been known for decades that epithelial stem cells can be identified as label-retaining cells (LRCs) by long term retention of a nuclear label, isolating a pure population of stem cells has been problematic. Using a Hoechst and propidium iodide dye combination and specifically defined gating, we sorted mouse epidermal basal cells into three fractions, which we have now identified as stem, transient amplifying (TA), and non-proliferative basal cells. More than 90% of freshly isolated stem cells showed a G0/G1 cell cycle profile, while greater than 20% of the TA cells were actively dividing. Both stem and TA cells retained proliferative capacity, but the stem cells formed larger, more expandable colonies in culture. Both populations could be transduced with a retroviral vector and used to bioengineer an epidermis. However, only the epidermis from the stem cell population continued to grow and express the reporter gene for 6 months in organotypic culture. The epidermis from the transient amplifying cell fraction completely differentiated by 2 months. This novel sorting method yields pure viable epithelial stem cells that can be used to bioengineer a tissue and to test permanent recombinant gene expression.

The epidermis is a continuously renewing tissue maintained by undifferentiated stem cells. For decades it has been assumed that epidermal stem cells (ESCs) were held in the G0 phase of the cell cycle and that they only entered the cell cycle when needed. Previously, we showed that ESCs retained nuclear label for long periods, indicating that these cells did not proceed through the cell cycle at the same rate as the other proliferative basal cells. However, their exact cell-cycle profile has not been determined because a pure population of ESCs has not been available. In this study, we sorted stem and transient amplifying (TA) cells from murine neonatal back skin, and adult ear, footpad, and back skin, using our recently developed method. We found that neonatal back skin had two times the number of ESCs as the adult tissues. Despite the age and anatomical difference, these ESC populations exhibited similar cell cycle profiles with approximately 96% in G0/G1 and 4% in S-G2/M. The cell cycle profiles of the TA cells from neonatal back skin and adult footpad also showed a profile similar to each other (85% in G1 and 15% in S-G2/M). Examination of genes on a cell cycle chip showed that proliferation associated genes and only p57 were upregulated in the TA cell and ESC population, respectively. We found BrdU positive and cyclin B1 positive cells in all groups, confirming that both ESCs and TA cells were cycling. These data demonstrate that there are more TA cells dividing than ESCs, that the cell cycle profile of adult TA cells is related to the proliferative state of the tissue in which they reside, and that ESC proceed through the cell cycle.