Our central objective is to explore the mechanisms governing skin stem cells and their remarkable ability both to self-renew long term and to commit to proliferate and differentiate along a particular lineage. Skin is one of the few systems where adult stem cells can be maintained and propagated in the laboratory, a feature that facilitates our studies. We also use technologies for targeting foreign genes to epidermis and hairs of transgenic animals and for inactivating the expression of epidermal or hair genes either by removing the genes from the skin's chromosomes (conditional inducible knockout) or by our latest breakthrough in silencing their mRNAs specifically in skin (conditional knockdown by lentiviral-mediated in utero transduction of the embryo surface). By taking molecular approaches to skin biology and stem cell research, we aim to apply this knowledge to human genetics, cancers, and medicine.
During development, skin begins as a single layer of multipotent embryonic progenitors that are able to produce hair follicles (HFs) and epidermis. In response to specialized signals from underlying mesenchymal cells, skin progenitors develop into HFs; in the absence of these cues, the cells stratify to make epidermis. We have had a long-standing interest in how these extraordinarily distinct architectures are established. Over nearly three decades, we have uncovered many molecules that epidermal cells use to make adhesive connections and their associated mechanical frameworks (cytoskeletons). These investigations guided us to the genetic bases of blistering skin diseases as well as hyperproliferative disorders that resemble precancerous lesions. We are now gaining insights into how skin cells control and change the organization of their actin and microtubule cytoskeletons to achieve dynamic tissue organization. Our research has centered on actin-microtubule interactions with two types of junctions: intercellular junctions, known as adherens junctions, composed of E-cadherin and its associates α-catenin, β-catenin, and p120-catenin; and cell-substratum junctions, composed of α3β1 integrins and their regulatory "focal adhesion" kinase (FAK).
We have shown that by integrating cell junctions and cytoskeletal dynamics, cells within the epithelial tissue can coordinate contractions and movements. Actin-microtubule-cell junction connections also provide the potential to polarize microtubule-based cargo transport and orient the mitotic spindle. Our studies point to the view that such connections are important in orchestrating the orientation of cell divisions in the developing skin and in controlling cellular movements upon stem cell activation and during wound repair in adult skin. Moreover, without the proper contacts between neighboring cells and extracellular matrix, epidermis is unable to keep harmful microbes out and body fluids in, hairs are not produced, and wounds cannot be repaired. Genetic defects and changes in the expression of cell adhesion genes are prevalent in a variety of disorders, including human cancers, where tissue architecture is distorted. Our studies have shown that tumorigenesis in the skin is typically associated with a weakening of adherens junctions and an enhancement of integrin dynamics.
While exploring how actin and microtubule dynamics govern epidermal morphogenesis, we also discovered that expansion of epidermis into a stratified epithelium involves a reorientation of the mitotic spindle, causing basal cells to undergo cell divisions that are asymmetric relative to the underlying extracellular matrix to which integrin junctions adhere. Moreover, we have shown that integrins, adherens junctions, and the cortical actin cytoskeleton are essential for proper spindle orientation during embryonic skin development. In addition, our studies have revealed that the epidermis controls these asymmetric cell divisions by using a set of genes whose ancient origins go back to fruit flies and worms. By orienting the spindle in this way, the innermost daughter epidermal progenitor retains its ability to proliferate while the detached, uppermost daughter differentiates and continues its journey outward.
To form the HF, cytoskeletal-junctional associations must also be remodeled. In this case, invagination to make the hair bud occurs in response to a signal transduction pathway called Wnt. Wnt signaling results in stabilization of excess β-catenin that is not utilized in adherens junctions. If a member of the "Lef1/TCF" family of DNA-binding proteins is present, as it is in the developing hair bud, β-catenin can interact with it and influence the transcription of HF genes. Adherens junctions are also remodeled at this time, and our findings suggest that reductions in these junctions are integral to the invagination process. We have developed new technologies that will facilitate our continued efforts to unfold the molecular mechanisms underlying these complex communication networks in tissue morphogenesis.
While embryonic progenitors must form tissues, fully grown adult skin requires stem cells to replace tissue cells lost in wear and tear (homeostasis) and to repair wounds on injury. In adult epidermis, the proliferative innermost (basal) layer maintains a pool of long-lived stem cells responsible for generating the stratified, terminally differentiated layers that are continually shed from the skin surface. In addition, hairs must regrow periodically throughout life. The upper portion of every adult HF maintains a separate reservoir (niche) of stem cells that are used for this purpose. The initiation of a new cycle of hair growth is determined in part by interactions between the epithelial stem cells and specialized mesenchymal cells, referred to as dermal papillae (DP). Similar to the epithelial-mesenchymal interactions that specify HF formation in embryonic skin, signals from the DP to the epithelial stem cell niche are important to activate stem cells to fuel HF regeneration and produce a new hair. HF stem cells within this niche are also mobilized in response to injury, only in this case, stem cells migrate toward the wound, rather than downward. We are learning that Wnt signaling plays a role in both of these processes and is accompanied by changes in actin and microtubule dynamics and polarity.
To monitor the divisions and movements of adult skin stem cells, we devised methods to specifically label epithelial skin stem cells and DP with fluorescently tagged proteins. Using these methods, we learned that stem cells at the base of their niche are the first to be activated to re-form the new HF. Those stem cells remaining in this upper region of the HF divide only briefly to replenish stem cells expended in tissue regeneration. By contrast, the bulk of tissue growth is fueled by shorter-lived stem cell progeny, which divide rapidly and then progress to terminally differentiate into hair cells. Skin stem cells that are kept in a quiescent state are subjected less frequently to chromosomal replication, which might introduce genetic errors.
We have also discovered additional resident progenitors at the base of the sebaceous gland and, most recently, in the sweat glands. We have found that different stem cell populations within the skin participate in repairing different types of wounds. Epidermal stem cells can repair simple surface wounds, while stem cells in the HF and sweat glands are mobilized after deeper wounds. As we continue to unravel the mechanisms underlying wound repair, this information will be especially important in efforts to improve upon therapies for healing wounds.
We have also devised methods to purify and transcriptionally profile these various skin stem cells. By identifying genes that are preferentially expressed in each of these cell populations, we are beginning to decipher the unique properties of skin stem cells and how they relate to other stem cells of the body. We have also developed methods to isolate pure populations of DP cells to learn how these cells communicate with the epithelial stem cells to stimulate rounds of new hair growth. While we are still in the midst of these explorations, the ability to culture adult skin stem cells in the laboratory is a major advantage in evaluating how stem cells maintain a growth- and differentiation-inhibited environment in vivo, and how they become mobilized outside their niche.
We have thus far identified six different transcription factors and also chromatin (epigenetic) and post-transcriptional (microRNA) modifications in skin stem cells. Our loss-of-function mouse models suggest that these factors govern essential features of stem cell maintenance and integrity. For over a decade, we have been studying TCF3 and TCF4, which appear to inhibit HF stem cell activation and differentiation. Although the exact details remain to be elucidated, relief of TCF3/TCF4 repression is achieved by the TCF-interacting protein β-catenin, which is stabilized and becomes nuclear upon convergent receipt of activating Wnt signals and molecules that repress BMP signaling. Among the genes that are subsequently transactivated are those involved in cell proliferation and migration. When Wnt signaling is overactivated or when BMP signaling is irreversibly compromised, HF stem cells increase their susceptibility to tumorigenesis. Overall, we are learning that the very pathways involved in the activation of the skin stem cells may contribute to cancer when deregulated.
In the past few years, we have learned that squamous cell carcinomas (SCCs) can originate from mutations in HF stem cells. Moreover, we have now isolated and characterized the tumor-initiating cells from skin SCCs (cancer stem cells) and shown that at the single-cell level, these cells can initiate a skin tumor resembling the parental cancer. These cancer stem cells from SCCs have high levels of signaling through the pathway involving activated β1 integrin and its downstream kinase cascade composed of FAK, Src, and MAPK. Conversely, the cancer stem cells are repressed for both E-cadherin and α-catenin. They reside at the tumor-stroma interface rich in extracellular matrix, inflammatory cells, blood vessels, and fibroblasts. In addition, the numbers of cancer stem cells and the aggressiveness of the SCC appear to be predicated upon whether its stem cells can respond to transforming growth factor-β (TGFβ) made by the surrounding stroma. If the cancer stem cells lack the TGFβ receptor, they are refractory to this growth inhibitory "brake" signal, and their numbers soar.
Overall, our results suggest that the features of cancer stem cells can be highly influenced by their surrounding stroma or "tumor microenvironment," which in effect becomes a new niche for the cancer stem cell, and one that is very different from the normal stem cell niche. The combination of intrinsic differences in the cancer stem cell, due to mutations and/or epigenetic silencing, and extrinsic differences in the stem cell niche are perhaps best exemplified by the hundreds of changes in gene expression that we have identified when comparing cancer stem cells to normal skin stem cells. In the future, we will pursue the functional significance of this myriad of changes and elucidate how the transition from a controlled to a deregulated program of stem cell self-renewal comes about. The answers could hold promise for identifying novel targets for the diagnosis and treatments of SCCs, among the world's most prevalent and life-threatening of cancers.
Grants from the National Institutes of Health provided partial support for some of these projects.
As of March 19, 2013