sj rom the biblical tale of

Hair Today Untangling the biology of the hair follicle ELlA T. BEN-ARI sj rom the biblical tale of Samson and Delilah and L J the fairy tale of Rapunzel to the eponymous 1960s rock musical and the latest coiffures, hair has long been a source of fascination and obsession. Both too little hair and too much hair can cause psychological and social anguish. And the multibillion dollar industry devoted to products designed to alter hair growth and appearance attests to the importance that humans place on healthy and attractive hair. Aside from its importance in social interactions, hair serves a number of purposes for humans and other mammals. It provides insulation, acts as a sensory organ, and helps disperse perspiration and pheromones produced by glands in the skin. Specialized immune cells at the opening of the human hair follicle can detect pathogens at the skin's surface and activate an immune response. The follicle also contains epithelial stem cells that are essential for regenerating the follicle and are also involved in regenerating the epidermis, or upper layer of skin, when it is injured or abraded. For over a century, biologists have been working to understand the structure and function of the hair follicle and the factors that control this tiny but complex organ's growth and development. Recent progress in hair research, including advances made through molecular and genetic approaches, is revealing some of the intricacies of hair follicle biology and may eventually lead to the development of more effective treatments for hair growth disorders. The hair follicle is also attracting interest as a useful model system for studying a range of biological processes, including tissue development, epithelial cell differentiation, apoptosis (programmed cell death), and tumor formation. An introduction to hair Like fingernails, claws, feathers, and scales, hair is a specialized skin appendage. Each hair follicle is associated with a sebaceous (oil) gland that opens to the skin through the follicle. The mature hair follicle (see diagram page 305) reaches down from the epidermis into the lower layer of skin, or dermis, a dense connective tissue that is richly supplied with blood vessels and nerves. The hair shaft, or fiber, that is produced by the hair follicle consists of two or three concentric layers and is composed largely of proteins called keratins. These tough, fibrous proteins are produced by rapidly proliferating matrix cells in the hair bulb, the lowermost portion of the follicle. Melanocytes in the hair bulb produce the pigment, or melanin, that colors the hair shaft. In the mature follicle, three concentric layers of epithelial cells surround the hair shaft, forming the inner root sheath. The outer root sheath, which is continuous with the epidermis and the sebaceous gland, forms the outermost epithelial layer of the follicle. A specialized region of the outer root sheath, called the "bulge," which is located just below the sebaceous gland, contains epithelial stem cells. Attached near the bulge region is the arrector pili muscle, which makes the hair stand erect in response to cold and other stimuli. A thin dermal sheath encloses the epidermal components of the follicle and is connected to the dermal papilla, a small, pear-shaped cluster of cells that is encircled by the hair bulb and plays a central role in hair follicle growth and development. Humans are born with all of the hair follicles that they will ever haveapproximately 5 million-although the size of some follicles and hairs can change during a person's lifetime, for example, in response to hormonal changes at puberty. The downy hairs formed by the follicles after they first develop in the embryo, called lanugo hairs, are shed in utero or within weeks after birth. Each follicle then produces one of two general types of hair: short, fine, nonpigmented hairs, called vellus hairs, which are found over the entire body, or longer, thick, pigmented hairs, called terminal hairs, such as those on the scalp, underarms, pubic region, and beard. Only the palms and soles lack hair follicles. Where hair follicles come from In most mammalian embryos, hair follicle formation starts in one area of the skin and spreads across the body surface. The follicles are formed from the ectoderm, a layer of epithelial tissue that gives rise to the epidermis, and from the underlying mesoderm, a mesenchymal layer that generates the April 2000 / Vol. 50 No.4' BioScience 303

Features Glossary Dermis. The inner layer of the skin in vertebrates, beneath the epidermis. The dermis is made up of connective tissue and is richly supplied with blood vessels, nerves, and (in mammals) sweat glands. Ectoderm. The germ layer on the outside of an embryo that develops into the epidermis and its derivatives (for example, nails, hairs, and oil glands), nervous tissue, and the linings of the mouth, nasal cavity, and gastrointestinal tract. Epidermis. The outer layer of cells or outer tissue of an animal. In vertebrates, the epidermis consists of several layers of cells and forms the outer layer of skin. Epithelial tissue, or epithelium. A tissue made up of one or more layers of closely packed cells that forms the covering of most external and internal body surfaces and organs. Liposome. Bubble-like lipid membrane structure that can fuse with cell membranes and can be used to deliver DNA molecules, drugs, enzymes, and other substances to target cells or organs. Liposomes have been successfully used to deliver DNA to cells in the hair follicle. Melanocyte. A cell that is found in the lower part of the epidermis and produces the pigment melanin. Mesenchymal tissue, or mesenchyme. A form of embryonic connective tissue that usually underlies epithelial tissue. It is derived from the mesoderm and consists of a loose network of un specialized cells embedded in a gelatinous matrix. Mesoderm. The embryonic germ layer from which muscles, connective tissues, bone, and the circulatory and urogenital systems develop. Transcription factor. A protein that binds to one or more specific target genes, altering their expression by regulating the rate at which messenger RNA is synthesized. dermis. In the mature follicle, only the dermal papilla and dermal sheath are made up of mesenchymal cells; the other follicle structures are composed of epithelial cells. The final number and spacing of the hair follicles on the body is determined by genes that are turned on during the early stages of hair follicle development. The protein products of these genes trigger the formation of thickenings in the epidermis, called placodes, that grow downward to form the hair plugs. The expanding base of the hair follicle plug surrounds a group of dermal mesenchymal cells to form the dermal papilla, and the hair plug develops into the hair bulb matrix, whose rapidly dividing epithelial cells differentiate and move toward the surface of the skin to form the inner root sheath and hair shaft. Early tissue transplantation experiments in mouse embryo skin showed that a signal from the mesoderm to the ectoderm is critical for placode formation. A signal from the ectoderm to the mesoderm then leads to formation of the dermal papilla, and a second mesodermal signal triggers further follicle development in the epithelium. Similar mesenchymalepithelial interactions are required for the development of other organshence the appeal of the easily accessible hair follicle as a model system for studying cell differentiation and tissue development. What makes the hair cycle tick? Unlike other organs, hair follicles continually break down and regenerate. Each follicle cycles repeatedly through three distinct stages: anagen, catagen, and telogen (see illustration page 307). In anagen, the growth phase of the cycle, the hair follicle is regenerated and a new hair grows. The duration of anagen in a given follicle is genetically programmed and determines a hair's maximum length. Thus, human scalp hairs, whose follicles stay in anagen anywhere from 2 to 8 years, can grow much longer than eyebrow hairs, whose follicles are in anagen for only 2-3 months. By comparison, the growth phase for a typical mouse hair lasts only approximately 2 weeks, and sheep hair (wool) fibers grow continuously. In catagen, the next and shortest stage of the hair follicle cycle, the hair stops growing and the follicle regresses. Follicle regression is a controlled process that results from a burst of programmed cell death that destroys most of the epithelial cells in the follicle below the bulge region but leaves the dermal papilla intact. The follicle then enters telogen, the resting stage of the cycle, during which the now loosely anchored hair fiber is eventually shed. After approximately 2-3 months in telogen, the follicles enter anagen once more. The lower part of the follicle is then regenerated from epithelial stem cells in the bulge, and a new hair fiber begins to grow. Signaling between the dermal papilla and the overlying epidermal tissue is critical for perpetuating the hair follicle cycle. Near the end of catagen, after the lower part of the hair follicle has degenerated, the dermal papillawhich remains intact throughout the cycle-moves upward, reaching a position just beneath the bulge region of the follicle. At the onset of a new growth phase in the hair follicle, researchers believe, signals from the dermal papilla activate adjacent epithelial stem cells in the bulge, thereby initiating regeneration of the mature follicle. In most mammals, the cycle of hair growth is seasonal and is related to daylight length and hormonal activity, resulting in regular periods of shedding, or molting. By contrast, human hair follicles generally cycle independently of one another, with each follicle apparently having its own internal clock, the nature of which remains unknown. Clinical observations in humans and laboratory studies in mice have shown that a range of substances, such as growth factors, hormones, and drugs, can modulate the cycle. Identifying the factors that coordinate the transitions between stages of the hair follicle cycle and figuring out 304 BioScience April 2000 / Vol. 50 No.4

.-... -... ~... -.~.----~.-... ~.-,-.. -,.~ how these factors work are central problems of hair biology research. Because most hair growth disorders involve abnormalities in the cycle (see box page 306) rather than total destruction of the follicle and are thought to be reversible, at least in theory, researchers hope that understanding the signaling networks that regulate hair follicle cycling will provide the key to developing new and more effective treatments for various forms of alopecia (hair loss) and unwanted hair growth. Identifying hair-loss genes Many disorders of hair growth have an underlying genetic component, although the extent to which genetic factors are involved in such disorders is variable and, in many cases, still unknown. "If by genetic you mean that there is an identifiable abnormal gene or set of genes that either cause the disease or make some individuals more susceptible to getting the disease than others... then just about all [hair disorders 1 have some genetic component;' explains Alan Moshell, director of the Skin Diseases Branch at the National Institute of Arthritis and Musculoskeletal and Skin Diseases, part of the National Institutes of Health, in Bethesda, Maryland. But not all hair disorders are genetic in the sense of running in families, he adds. The most common hair disorder, androgenetic alopecia (commonly known as male pattern baldness), does run in families, but its pattern of inheritance is unclear, and the disorder is likely to involve multiple genes. Alopecia areata, another relatively common hair disorder, is thought to involve an autoimmune reaction, in which a person's immune system attacks the hair follicles. But genetic factors-perhaps determining susceptibility to the disorder-probably also playa role. To begin to identify genes involved in human hair loss, researchers have been studying rare hair disorders that have a simpler pattern of inheritance than the more common hair growth disorders. Angela Christiano and Wasim Ahmad, Permanent segment I,.r---- Outer root sheath j~~------innerrootsheath '-"<:...------- Matrix '---------- Dermal papilla In the mature (anagen) hair follicle, rapidly proliferating matrix cells in the lowermost portion, the bulb, produce the hair shaft, or fiber. The lower portion of the hair shaft is enclosed by the rigid inner root sheath, which acts much like a mold to shape the growing hair. Surrounding the inner root sheath is the outer root sheath, which is continuous with the epidermis and the sebaceous gland. A specialized region of the outer root sheath, called the "bulge," which is located just below the sebaceous gland, contains epithelial stem cells. The dermal sheath (not shown) encloses the epidermal components of the follicle and is connected to the dermal papilla, a small, pear-shaped cluster of cells at the base of the follicle that is encircled by the hair bulb and plays a critical role in hair follicle growth and development. Most of the lower part of the follicle (cycling segment) continually breaks down and regenerates during the hair follicle cycle. Illustration from U. Gat et al. 1998. Cell 95: 605-614. Cell Press. of Columbia University, and an international group of collaborators studied one such disorder in a large Pakistani family. In January 1998 they reported (Science 279: 720-724) that a rare form of hereditary human hair loss, congenital atrichia, is caused by a single mutation in the human version of a gene that is known to be mutated in the mouse strain known as "hairless." Studies of other large families with congenital atrichia by Christiano and her colleagues have since revealed that a variety of mutations in the human hairless gene are associated with the disorder. Mice of the hairless strain initially develop a normal coat of fur, but after approximately 2 weeks they begin to lose their hair. Hair loss progresses in a wave that starts at the snout and moves toward the tail, leaving a mouse completely bald within a week. Researchers had previously noted the similarity between hairless mice and people with congenital atrichia, who are born with normal hair but lose all of the hair on their head and body shortly after birth. In both hairless mice and humans with congenital atrichia, hair loss is inherited recessively-that is, both copies of the hairless gene must be mutated to cause the disorder. And, as in mice, the human hairless gene is expressed primarily in skin and brain tissue. The gene codes for a protein that, on the basis of its sequence, is believed to be a transcription factor. George Cotsarelis, a University of Pennsylvania dermatologist who stud- April 2000 I Vol. 50 No.4 BioScience 305

Features Some disorders of human hair growth Alopecia. General term for hair loss. Androgenetic alopecia, also known as male pattern baldness. The most common hair disorder, it involves progressive shortening of the anagen stage, miniaturization of the hair follicles, and replacement of terminal hairs with vellus hairs. Androgenetic alopecia affects women as well as men, although the patterns of hair loss differ between the sexes. Hirsutism and hypertrichosis. Excessive hair growth due to an extended anagen stage and abnormal enlargement of the follicles, with vellus hairs converted to terminal hairs. Alopecia areata. A relatively common form of hair loss in which inflammation of the follicle leads to termination of the anagen stage of the hair follicle cycle and forces follicles into catagen. Thought to be an autoimmune disorder, alopecia areata causes patchy hair loss on the scalp. It may progress to affect the entire scalp (alopecia totalis) or the entire body (alopecia universalis). Telogen effluvium. Transient hair loss that occurs when an increased number of follicles enter prematurely into the telogen stage of the hair cycle and shed their hair shafts. Causes include drugs, childbirth, fever, malnutrition, and endocrine abnormalities. Anagen effluvium. Loss of hairs in the anagen stage due to disruption of the rapidly proliferating bulb matrix cells. Caused by chemotherapy and radiation treatment, anagen effluvium is usually reversible, but high doses of radiation can cause permanent hair loss, most likely due to destruction of parts of the follicle that are critical for regeneration. ies hair biology, says that the discovery of the human homologue of the mouse hairless gene defect "beautifully shows that the hair follicle in the mouse is very similar to the human, [and) that's heartening because a lot of people are working on mouse and assuming that it's going to be relevant to humans." Indeed, Moshell says, "the relationship between mouse and human is close enough that there's a very good chance that you'll find a mouse model for almost any human hair disease if you look hard enough:' Detailed histological studies of the hair follicles in hairless mice by a group of researchers, including Ralf Paus, of the University of Hamburg, and Christiano and Andrei Pantaleyev, from Columbia University, showed that the absence of functional hairless protein causes follicles to enter catagen prematurely. Furthermore, the usually controlled degeneration of the hair follicle is aberrant in these mice, resulting in the destruction of the hair follicle's normal structure, cessation of hair follicle cycling, and shedding of the hair. On the basis of these findings, the researchers hypothesize that the normal hairless protein is a critical factor in the control of catagen. In April 1999, Christiano and a group of collaborators reported (Nature 398: 473-474) that they had identified a rare human syndrome that is homologous to the "nude" mouse phenotype, which results from mutations in the whn (winged-helix-nude) gene. Like the nude mice, people with this syndrome have a mutated wnh gene, are born without hair, and have severe immunological defects. Although it is not yet clear that determining the genetic basis of rare hair disorders in humans will yield information that is relevant to developing treatments for more common disorders, studying the function of genes such as hairless "will definitely tell us more about basic hair biology," Cotsarelis says. And, Moshell says, "it may tell [us) something about the normal cycling [or) the normal hair structure that may be related to other diseases:' Genes and hair development New information about basic hair biology and the genes involved in hair follicle growth and development is also emerging from developmental studies. "The developmental biologists have made really big strides in understanding the normal development and cycling of the hair follicle, and that sort of happened by accident," Cotsarelis says. A common scenario, he says, is that a researcher may find the mouse homologue of a gene that is important for development in the fruit fly Drosophila melanogaster, create "knockout" mice that do not express the gene, and then discover that hair development is abnormal in those mice. In studies of such knockout mice and of mice that overexpress specific genes, developmental biologists have identified a number of mammalian genes that are required for normal follicle development. These genes code for molecules that transmit growth and development signals both between and within cells. In one study (Cell 78: 1017-1025), developmental biologist Gail Martin and her colleagues at the University of California-San Francisco knocked out expression of the FGF5 gene in mice, hoping to learn about the growth factor's role in limb development. Instead, the researchers found evidence that FGF5 functions as an inhibitor of hair elongation and may participate in the transition from anagen to catagen. The FGF5 knockout mice have hair that is 50 percent longer than normal. In these mice-and in a previously described mouse strain known as "angora," which turns out to have a spontaneously occurring deletion in both copies of the FGF5 genethe absence of FGF5, which is normally expressed in the hair follicle just before the end of anagen, extends the growth stage of the hair follicle cycle. Another study, published in the 25 November 1998 issue of Cell (95: 605-614) by Uri Gat, Elaine Fuchs, and colleagues from the University of Chicago, showed that mice that are genetically engineered to overproduce 306 BioScience April 2000 I Vol. 50 No.4

Hair follicle development and cycling. A series of epithelial and mesenchymal signals in the mammalian embryo specify when and where the primordial hair follicles, which consist of an epithelial placode and dermal condensate (the precursor of the dermal papilla), will appear. Additional signals lead to formation of the mature hair follicle, which cycles repeatedly through three distinct stages postnatally: growth (anagen), regression (catagen), and quiescence (telogen). Illustration: Republished with permission of Journal of Clinical Investigation, from The Hedgehog and the hair follicle: a growing relationship. A Dlugosz. 104: 851-853. 1999; permission conveyed through Copyright Clearance Center, Inc. Ectoderm Placode ===\ ==== --+ ~,I ~... -+ ~? Dermal Condensate Inner Root Sheath Dermal Papilla Outer Root Sheath I Catagen I Sebaceous Gland an activated form of an intracellular signaling protein called ~-catenin in their epidermal cells form extra hair follicles in areas of skin between existing follicles. The extra follicles, which develop after the mice are born, are oriented abnormally in the skin, leading to incorrectly angled hairs. Mice that overexpress ~-catenin also develop two types of hair follicle tumors. Together, these findings suggest that ~ catenin, which is part of a signaling pathway that controls cell fates and cell proliferation in developing tissues in a wide range of organisms, may normally playa role both in determining the spacing and number of hairs and in regulating cell growth in the hair follicle. Similarities between hair follicle development and the regeneration of the mature follicle during subsequent follicle cycling-including the critical role of interactions between the dermal papilla and cells in the epidermis-have led some researchers to speculate that the two processes may involve some of the same molecular signals. The extent to which events controlling follicle development and regeneration are alike is not yet clear. But recent experimental evidence shows that some signaling proteins known to play a role in mammalian hair follicle development, including sonic hedgehog, which is involved in a wide range of developmental processes in fruit flies and other organisms, and Wnt, which triggers the activation of ~-catenin, can also modulate hair follicle cycling and hair growth. Hair follicle stem cells One likely target for signaling molecules that regulate hair follicle cycling is the hair follicle stem cell. Ten years ago, Cotsarelis and his colleagues showed that the bulge region of the follicle's outer root sheath contains cells with epithelial stem cell properties. Previously, Cotsarelis says, researchers had thought that the stem cells involved in regenerating the mature hair follicle were located in the hair follicle bulb and that "these cells would migrate up and down during the hair cycle. But we showed that they were stationary in this bulge area and [that 1 they're in a location that allows them to give rise to cells to regenerate the lower follicle and also to regenerate the epidermis." When activated, each stem cell is believed to divide and give rise to one cell that remains a stem cell and another cell that becomes a "transient amplifying cell." Transient amplifying cells "divide more frequently and more quickly [than the stem cells 1 and provide the basis for that new population of cells" that regenerates the base of the follicle, Moshell explains. Hair follicle epithelial stem cells also "play some part, maybe a major part, in re-epithelialization after wounding of the surface epidermis," he says, but whether there are other epithelial stem cells in the skin is still a matter of debate. Hair follicle stem cells may also give rise to some types of skin tumors. Formation of these tumors is believed to result from abnormal control of the hair follicle cycle. Because many types of human cancer originate from epithelial cells, understanding "what goes awry when you develop tumors that are related to the stem cell in the hair follicle would be a model system for any epithelial tumor system," Moshell says. Gone today, hair tomorrow Researchers who hope to develop gene therapy for hair and skin disorders are interested in hair follicle stem cells as a potential target, especially for in vivo April 2000 / Vol. 50 No.4' BioScience 307

Features Human hair transplants A recent article in Nature (402: 33-34), by Colin Jahoda and Amanda Reynolds, of Durham University, in Durham, UK, and several colleagues, showed that the mesenchymal portion of the hair follicle can induce new follicles to form in the skin of a human adult. The researchers transplanted dermal sheath tissue from the scalp hair follicles of a man (Jahoda) to the inner forearm of a genetically unrelated woman (Reynolds) and found that the transplanted tissue stimulated the formation of new follicles. The new follicles produced hairs that were larger and thicker than the normal vellus hairs on the arm, were mostly pigmented, and grew in variable directions. Jahoda and Reynolds had previously demonstrated that transplants of isolated rat hair follicle dermal cells could stimulate new follicle formation in adult rats, but analogous results had never before been achieved in humans. The experiments in humans need to be repeated by others, says University of Pennsylvania dermatologist George Cotsarelis, "but if it really turns out to be true, I suspect there'll be a lot of interest in developing techniques for transplanting mesenchymal cells to people's bald scalps and trying to induce big thick hair." The results of the human transplant experiments, he adds, "also bring up [the 1 fascinating idea that the follicle is immune privileged;' because hair follicles transplanted from one person to another, unrelated person were not rejected by the immune system. approaches such as topical application of a gene enclosed in a bubble of lipids (a tiposome) or direct injection of a gene into the skin. The entire hair follicle is attractive as a target for gene therapy because it is readily accessible, both because of its location on the surface of the body and because it is an anatomical break in the tough and generally impermeable outermost layer of the epidermis. And because hair follicle epithelial stem cells can give rise to epidermal skin cells, researchers hope that therapeutic genes introduced into these stem cells in vivo will be passed on to cells in the epidermis and thus be useful for gene therapy of some skin diseases. "It's not just going to be for treating baldness," Cotsarelis says of the potential for targeting gene therapy to stem cells in the follicle. "I think it will be important for treating lots of different skin conditions." As for the likelihood that gene therapy for hair disorderswith or without the use of stem cellswill become a reality, Cotsarelis says, it is impossible to predict. However, he says, "it's not crazy by any means... At some point in the future, I wouldn't be surprised if there were some type of gene therapy for hair loss." Moshell points out that "more typical drug-type therapy" with substances that block or stimulate one or more steps in the hair follicle cycle may also be useful for treating hair disorders. But before effective treatments can be developed, researchers will need to learn more about the steps involved in the control of hair follicle cycling and which steps in the process can be influenced by drugs or other treatments, preferably without causing unwanted effects in other parts of the body. Continued study of what longtime hair biology researcher Margaret Hardy, of the University of Guelph, in Ontario, Canada, has called "the secret life of the hair follicle" is bound to yield some hair-raising results. 0 308 BioScience April 2000 / Vol. 50 No.4