If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Epidermal label-retaining cells (LRC) can be identified by giving neonatal mice repeated injections of 3H-thymidine or 5-bromo-2′-deoxyuridine and then finding the cells that are still labelled in adulthood. Although label retention is simply a marker of the proliferative history of a cell, there is, nevertheless, evidence that it is a characteristic of epidermal stem cells. Here we review the early literature on LRC and then highlight two recent applications. We describe how LRC can be visualized by whole-mount labelling of the epidermis and how purified LRC can be used to screen for markers of the epidermal stem cell compartment.
The simplest definition of a stem cell in an adult tissue is that it is any cell with a high capacity for self-renewal that extends throughout adult life. In addition, stem cells are usually considered to have the potential to produce differentiated progeny (
). According to these criteria the epidermis has long been recognized as having a resident stem cell population. The tissue consists of a stratified squamous epithelium (interfollicular epidermis; IFE) with associated hair follicles and glandular structures (the sebaceous glands and sweat glands). The IFE undergoes continuous turnover and there is a constant requirement to replace the dead, terminally differentiated cells of the outermost cornified layers through proliferation of cells in the basal layer. Hair follicles undergo repeated cycles of growth, regression, and rest (
). These observations can be reconciled by proposing that there are separate stem cell populations within the hair, sebaceous gland and IFE. Each of these is capable of generating daughters that differentiate along any of the epidermal lineages. Under steady-state conditions, however, the stem cells normally give rise to a more restricted repertoire in response to signals from the local microenvironment (
). The fraction of labelled mitoses in the epidermis pointed to the existence of at least two distinct cell populations with different cell cycle times, together with a discrete post-mitotic compartment (
The position of the slowly cycling cells is not random but related to the architecture of the IFE. In the mouse, epidermis from many body sites is organized into precise columns of hexagonal cornified cells overlying groups of basal cells from which they are derived; these are known as epidermal proliferative units (EPU) (
). These observations, together with studies of the regeneration of the skin following radiation induced damage, led to the concept that the cell cycle characteristics of keratinocytes in the basal layer of the IFE are indicative of whether or not they are stem cells (
). The slow cycling cells were designated as stem cells and rapidly cycling cells as transit amplifying cells, cells destined to undergo terminal differentiation and leave the basal layer after a few rounds of division.
Having established that infrequent division could be a characteristic of stem cells there was a clear need for a detection method that obviated the need for complex mathematical analysis. One such method is to generate label-retaining cells (LRC) (
). Neonatal mice are given repeated injections of a nucleotide analogue such as 3H-thymidine or 5-bromo-2′-deoxyuridine (BrdU) at a stage when the tissue is hyperproliferative so that all dividing cells, whether or not they are stem cells, incorporate the label in newly synthesized DNA. The label is then chased for several weeks or months; only those cells that rarely divide retain the label into adulthood (LRC) (
). These data suggested that stem cells within the hair follicle reside in the permanent portion of the follicle, instead of the lower bulb.
LRC: Practical Considerations
There are a number of experimental variables in the generation of LRC that need to be considered. The first of these is the reagent used for labelling. In all of the early work LRC were detected by incorporation of 3H-thymidine. Two drawbacks of using 3H-thymidine are the length of time required for autoradiographic exposure of tissue sections and the technical difficulty of combining autoradiography with immunolabelling for potential stem cell markers. The advantage of using BrdU incorporation is that it can be detected by immunofluorescence with anti-BrdU antibodies.
The time when the DNA label is introduced into the mice is another consideration. Injection of pregnant females allows simultaneous labelling of multiple animals and can lead to incorporation of label into a high proportion of keratinocytes. Since BrdU is a teratogen (
) injection of neonatal mice may be preferable, however. It is obvious that a stem cell that was not synthesizing DNA during the labelling period will never be visualized as a LRC, and it is important not to assume that 100% labelling is achieved with any of the protocols that have been described (
The age of the mice at the time of euthanasia is another issue. In our experience the label-retaining population in tail epidermis stabilizes 24 d after BrdU injection and the pattern of LRC is remarkably constant until at least 140 d post-injection (
A final important consideration is how to define a LRC. It is necessary to have some type of operational definition of how many rounds of division a cell must undergo before it stops being a LRC. Label is progressively lost when LRC are induced to divide but is probably still detectable after two or three divisions, at which point the labelled cells may have become transit amplifying cells or committed progenitors (
). One potential concern about LRC is that they are incapable of dividing because they have sustained damage by virtue of 3H-thymidine or BrdU incorporation. This, however, turns out not to be the case. One way to recruit LRC to divide is to treat the epidermis with the phorbol ester 12-O-tetradecanoylphorbol-13-acetate (TPA) (
For obvious reasons LRC have not been generated by injecting human subjects. Instead clonal assays of proliferation of cultured human keratinocytes have been performed, with extensive self-renewal and generation of differentiated progeny being used as markers of the stem cell compartment (
). This indicates that in human epidermis the cells with highest self-renewal capacity in vitro are slow cycling in vivo.
In the mouse, direct comparisons of label retention in vivo and clonal growth in vitro have been hampered because of the low colony-forming efficiency of keratinocytes cultured from adult mice. Nevertheless an important study demonstrated that LRC isolated from mouse skin do have higher colony-forming potential in vitro than pulse-labelled cells from the same animal (
The functional relevance of the localization of LRC in the hair follicle bulge was confirmed by dissecting rat whisker and human hair follicles and assessing their clonogenic potential in vitro. In the case of the rat whisker follicles, 95% of keratinocyte colony-forming cells are located within the bulge region (
). More recently, chimeric follicles of rat whiskers were generated in which the bulge region expressed β-galactosidase. These studies confirmed that cells isolated from the bulge region of vibrissal follicles have low in vivo mitotic activity and high in vitro clonogenicity (
). Therefore, not only do cells with the highest label-retaining capacity reside within the bulge region of the hair follicle, but the majority of the clonogenic cells in the follicle are also derived from this region.
In the IFE the onset of terminal differentiation is linked to migration out of the basal layer because keratinocytes downregulate the function and expression of integrin extracellular matrix receptors as they undergo terminal differentiation (reviewed by
Any keratinocyte has the potential to acquire an oncogenic mutation, but most cells with such a mutation are lost through terminal differentiation. It takes more than one genetic lesion to found a tumor and multiple mutations are thought to be acquired only by long-term residents of the epidermis that have the capacity for clonal expansion. This has led to the concept that stem cells are largely responsible for epidermal tumorigenesis (reviewed by
). Exposure to a tumor promoter, such as TPA, induces proliferation of mutant cells within the epidermis, thereby increasing the target cell population that is susceptible to acquiring additional oncogenic mutations (
). In addition, tumor promoters appear to be involved in the selection of the specific target cell population that undergoes expansion, as evidenced by the outgrowth of tumors with dissimilar molecular characteristics following exposure to different types of tumor promoters (
A further aspect of stem cells is their ability to generate both stem cell daughters and daughters that undergo terminal differentiation. This can potentially be achieved through invariant asymmetric divisions or by populational asymmetry (
). There is no morphological evidence that epidermal stem cells undergo asymmetric divisions, since in postnatal IFE each division gives rise to two daughter cells in the basal layer, in contrast to the stratified squamous epithelium of the esophagus where vertically oriented mitoses occur (
). Asymmetric divisions could potentially involve an “immortal strand” of DNA: stem cells would arrange their sister chromatids at mitosis such that the same template DNA strands stay together through successive divisions (
). To test the immortal strand hypothesis template strands in the stem cells can be labelled during development or during tissue regeneration using 3H-thymidine. Labelling newly synthesized strands with a different marker (BrdU) allows segregation of the two markers to be studied (
). The technique has yet to be applied to epidermis; however, the fact that the BrdU label is completely lost when LRC are stimulated to divide tends to argue against the immortal strand hypothesis in the epidermis (
). We were interested to establish whether whole-mount labelling could also be applied to mouse epidermis. This technique facilitates visualization of rare cells, such as LRC, in large areas of the epidermis. In addition whole-mounts provide a way to examine changes in lineage specification in transgenic and knockout mice (
). The method is illustrated schematically in Figure 1.
We found tail epidermis easiest to work with, as we could generate whole-mounts at any stage of the hair cycle. In contrast dorsal epidermis is thinner and more fragile and we could only make good whole-mounts when the dorsal epidermis was in telogen. The hair follicles in the tail undergo growth cycles with similar timing to dorsal follicles, although the degree of synchronization is less tight. The tail hair follicles are arranged in groups of three and the central follicle tends to cycle with different kinetics to the other two (
). The main difference between the hair follicles of the tail and dorsal epidermis is that the tail follicles are larger. This is reflected by an increased number of LRC (range 10–79 for tail, compared with 0–15 for dorsum) (
We found scattered LRC throughout the IFE and at the periphery of the sebaceous glands and a concentration of LRC in the permanent portion of the follicles (Figure 2). There was no discernible pattern to the position of LRC in the IFE, but this is not surprising since the tail IFE has a different organization to the classical EPU (
; Figure 2). LRC were present in all follicles, irrespective of whether a distinct bulge was visible and their number was not significantly depleted by successive hair cycles (Figure 2). Whereas a physical protuberance of the ORS, when present, lies on one side of the follicle, the LRC are usually distributed symmetrically. Thus in the context of the localization of LRC, the “bulge” (
). It is also clear from examining whole-mounts at different stages of the hair cycle that the patterns of expression of markers such as the α6β4 integrin is very dynamic: high α6β4 expression is a good marker of LRC in telogen follicles but the expression pattern becomes less localized during anagen.
The recent development of multiple inducible and non-inducible promoters that can be used to target transgenes to different compartments in the skin has led to a dramatic increase in data regarding the molecules that regulate epidermal stem cell self-renewal and lineage commitment (
). These animal models can be used in conjunction with analysis of LRC to assess the changes that the transgenes induce within the epidermal stem cell compartment. This approach has been used to demonstrate that constitutive overexpression of cMyc from a keratin 14 promoter (K14.Myc2) leads to a 75% reduction of LRC in 3-mon-old mice, suggesting that deregulation of cMyc leads to epidermal stem cell depletion (
We have been using the whole-mounts in conjunction with LRC and lineage analysis to examine what happens to LRC when epidermal growth and differentiation are disturbed in genetically modified mice. Activation of cMyc (
). Increased sebaceous differentiation was not restricted to hair follicles, but also occurred in the IFE, suggesting that cMyc may re-program basal epidermal cells. In contrast, expression of N-terminally truncated Lef-1 to block β-catenin signalling (
). Thus by combining tail epidermal whole-mounts with label-retention studies, we have demonstrated that LRC are more sensitive to some proliferative stimuli than others and that changes in lineage can occur with or without recruitment of LRC into cycle.
Isolation of Slow-Cycling Cells by GFP-Expression
One long-standing limitation in epidermal research has been in the inability to isolate the slow-cycling population. Recently an elegant new technique has been published in which infrequently-cycling cells in the skin are labelled by crossing transgenic mice that contain histone H2B-GFP (
). Although the GFP-positive cells have not been compared directly with LRC, the advantage of this system is that the slow cycling population can be isolated despite the lack of cell surface markers due to their endogenous GFP fluorescence.
Using this approach, Tumbar and colleagues isolated a pure population of slow-cycling, GFP-positive cells from the epidermis. They were able to demonstrate that GFP-positive cells rarely divide within the bulge region of the hair follicle, but that the biochemical characteristics of the cells change once they are stimulated to leave the niche. For example, in response to wounding, GFP-positive cells exit the bulge to repopulate the IFE. During the anagen phase of the hair growth cycle the downgrowth of the follicle is primarily GFP-positive, suggesting that these new cells are derived from bulge cells. New populations of GFP-positive cells could be identified in the cycling portion of the follicle that displayed markers for outer root sheath, inner root sheath, hair and matrix.
The transcriptional profile of the GFP-positive cells was determined to define the characteristics of the stem cell niche. In total, greater than 100 messenger RNAs were preferentially expressed in GFP-positive stem cells located in the bulge region of the hair follicle (
. All of the reported bulge markers (including K15, K19, α6 integrin, β1 integrin, CD34, S100A4 and S100A6) were present, but the GFP-positive population was more restricted than the population defined by any of the individual markers. Possibly the most promising marker for epidermal stem cells in this study was CD34, a molecule that had been independently described as a putative epidermal stem cell marker (
). This study also identified a plethora of genes, in particular surface receptors and signalling proteins, which appear to be involved in maintaining the epidermal stem cell compartment.
Future studies will seek to characterize the roles that these candidate genes play in regulating the stem cell niche. In addition, the proliferative ability and multipotentiality of slow cycling, GFP-positive cells can now be assessed both in vitro and following transplantation in vivo. This approach will have general utility in isolating LRC from other self-renewing adult tissues where a slow-cycling stem cell population is present. Finally, comparison of the transcriptional profile of the slow-cycling cells of the bulge with the K15-positive population (
) will further define the characteristics of hair follicle stem cells.
Generation of LRC is an old technique that is finding new applications in the study of epidermal stem cells. Because label retention is simply a marker of the proliferative history of a cell it is important not to assume that LRC and stem cells are synonymous. We think it more likely that LRC represent a subset of epidermal stem cells. It will be interesting to examine the properties of cells that are not LRC yet reside in the bulge and express markers such as keratin 15 and CD34. Whole-mount labelling allows visualization of LRC and facilitates analysis of the effects of reprogramming epidermal lineages. GFP labelling of slow-cycling cells represents a great step forward, because for the first time LRC can be purified and characterized at a molecular level. Clearly, these are exciting times to be working on stem cells.
We are most grateful to Violeta Silva-Vargas for providing Figure 1. Kristin Braun is supported by an American Cancer Society Fellowship (PF-01-131-01-MGO). Both authors gratefully acknowledge the financial support of Cancer Research UK.
Changing patterns of localization of putative stem cells in developing human hair follicles.