How Close Are We to Solving the Puzzle? Review of the Alopecia Areata Research Workshop
-David Norris, M.D. University of Colorado Health Sciences Center Denver, CO
Speakers at the Fourth Research Workshop on Alopecia Areata have clearly presented the state of current knowledge in three fundamental areas of research in alopecia areata : genetics, immunology, and targets, triggers and controls. The final speakers have reviewed what we know about the treatment of alopecia areata and what are the prospects for new treatment modalities in the future. It is clear that the research enterprise related to hair biology and alopecia areata in particular is now firmly established and expanding into important areas of fundamental biology including stem cell research and sophisticated studies of epidermal differentiation and morphogenesis. It is evident that the mechanism of hair follicle dysfunction in alopecia areata is immunological, controlled by activated T cells. Whether this response is truly autoimmune mains to be firmly established, but there are strong indications that self antigens, including melanocyte antigens are the key targets. Susceptibility to alopecia areata is likely to be genetically determined. Linkage to genes controlling the immune response is strongest in patients with severe AT and AU, and it is likely that AA is a polygenic complex trait, with distinct genetic linkages determining specific gene subsets. Our understanding of the mechanisms through which the cycling hair follicle is damaged and dysfunctional in AA have been thoroughly reviewed in this conference. “How close are we to solving the puzzle?” In spite of the progress reported here, we have a long way to completely understand the mechanisms of disease and to identify AA-specific targets for treatment. On the other hand, we may be very close to providing better treatment approaches for AA. Treatments proven to be effective in other autoimmune skin diseases can be used in AA in “proof of concept” trials which will both verify mechanisms of disease and provide practical approaches for treatment of this important disease.
Since alopecia areata is likely to be a complex genetic trait, understanding of the complexities of polygenic interactions is necessary to understand how genetic analysis of AA should proceed. Fred Nijhout from Duke University presented an elegant perspective on complex genetic interactions, using models developed in Drosophila to study the nature of robustness and genetic mechanisms. Complex traits don’t follow Mendel’s laws, the correlations are often not linear and that there are complex networks of interaction that can be mathematically modeled. But even within these networks there may be flat planes of response in which there is linearity. The effect of individual genes in such models depends entirely on context. The summation of properties law states that the sum of the squares of the effects of all genes in a pathway is one; identification of a gene with a major effect means that the effects of other genes in the pathway may be low. For example, if we identify an immune response gene that is necessary for the development of alopecia areata, then the other genes in the pathway may be rather minor in their importance and only some of them may be necessary for expression of disease. Certain genes may determine which subsets of disease such as AA or AU or AT may occur. The importance of context is clear in the hair follicle, where cross talk among different cell populations control the cycling of the follicle. This point was made repeatedly in many of the presentations here , especially in the elegant presentation by Elaine Fuchs on the control of epidermal and follicular differentiation.
Richard Kalish presented a summary of data that he and Amos Gilhar have produced over the past 10 years, demonstrating that AA is an autoimmune disease mediated by T lymphocytes in which autoantiogens are necessary to activate T cells that produce the disease. The model that they used is a SCID mouse (deficient in the ability to respond to human skin grafts) engrafted with hair bearing human skin. Leukocytes from AA patients are stimulated in vitro with hair follicle antigens and injected into these engrafted mice., producing immunologic hair loss similar to that seen in AA. Similar SCID mouse models have been effectively used to study autoimmune skin disease in psoriasis and photosensitive lupus, but the Gilhar/Kalish model is a particularly useful and robust model. Using this model they have shown that cells from AA patients, activated by hair follicle antigens, can transfer AA. In this Adoptive transfer model, both CD4 and CD8 T cells are necessary for maximal effect even though CD8 cells are sufficient to induce hair loss. They have more recently shown that the requirement for hair follicle antigens to stimulate T cells in this transfer model can be replaced by follicular melanocytes. This provides evidence that follicular melanocytes may be the targets for activated T cells in AA, as originally proposed by Tobin and Bystryn. It is interesting that control of certain melanocytic peptides are linked to immune response (MHC) loci, providing another possible point of control of AA by the MHC.
Jerry Niederkorn provided some new perspectives on how certain locations in the body, such as the eye, might have very unique immunologic environments, a concept developed by his mentor Wayne Streilein. The eye is a unique structure characterized by “immune privilege”, the resistance to induction of acquired immunity in that site. This resistance to immune response is based on a number of mechanisms: lack of lymphatic drainage, anterior chamber immune deviation, immunosuppressive effects of the aqueous humor itself , Low MHC expression, low expression of Langerhans cells or antigen presenting cells, and inhibition of natural killer cells. Dr. Niederkorn proposed the idea that immune privilege in the hair follicle is to also protect the key structures of the hair follicle such stem cells in the bulb from immune destruction.
Ralf Paus is the father of father of the concept of immune privilege in the hair follicle. MHC Class I molecules are not expressed in the hair matrix, in the dermal papilla or in the outer root closest to the bulb. Class I MHC molecules are necessary in initiate immune responses in which endogenous antigens are presented to activate CD8 T lymphocytes.
Interferon gamma generated by activated T lymphocytes and natural killer cells can up regulate Class I expression. In the hair follicle, such expression may be inhibited locally by growth factors such as a-MSH, TGF-B and IGF-1, acting in this instance as immune-modulating molecules. But it’s inhibited locally he proposes by a number of growth factors acting through the receptors, who in this sense act as immuno-modulating chemicals.
It may be that the purpose of immune privilege in the hair follicle is indeed to protect the stem cells in the hair follicle. I propose an alternative explanation: that the lack of MHC Class I expression in the lower regions of the hair follicle is intended to protect the hair follicle from induction of auto-immunity during regression of the catagen follicle. In regression of the catagen follicle, the epithelial cells and melanocytes of the matrix, and the epithelial cells of the root sheaths nearest the hair bulb the root sheaths, all undergo apoptosis. Apoptotic fragments of the cells are available for ingestion by cells locally. If those cells had Class I expression could internalize those antigens, present them on the surface in the context of Class I that would induce an immune response against self-antigens. The hair follicle is unique in that it repeatedly undergoes cycles of cell death and then regrowth, repeatedly leaving apoptotic materials within the dermis. The differentiating epidermis, on the other hand, disposes of apoptotic corneocytes by desquamation.
In this context, Ralf Paus proposes an “Immune privilege collapse model” to explain the development of autoimmunity in alopecia areata. In this model, infections, bacterial super antigens, or follicular damage trigger release of gamma interferon, which induces expression of Class I MHC on follicular cells leading to induction of CD8 positive cytotoxic cells, and also induction of Cass II MHC molecules leading to induction of CD4 help, and then to downstream autoimmune phenomenon and then downstream from all this there’s a spread of the immune response with antibody, macrophagus, expression of Fas ligand, apoptosis and damage to follicular cells. Is there a genetic predisposition to the breakdown of immune privilege in alopecia areata? I believe that is clearly one area of hair biology research that requires additional funding.
David Duggan provided a clear review of how gene arrays might be applied in general to the study of gene expression in cancer and inflammatory diseases. Several speakers presented their work on the application of micro array technology to the study of alopecia areata in human and animal models. I was impressed with Jerry Shapiro’s studies using micro array and proteomic analysis to look at gene and protein expression patterns in the rat alopecia areata model in response to therapy. Another important approach is that reported by John Sundberg, using induction of AA in the C3H/HEJ mice to study those genes that are turned on in the induction of AA. In both of these approaches, micro array changes are studied in the context of turning-on or turning off AA. Initial studies by Sinha of micro array analysis in human AA have shown predictable changes in keratins and in cytokines. David Duggan made a case for application of micro array technology to human AA, but we should be wary in the experimental approach employed. I would favor using micro arrays as they have been applied in animal models of AA, looking at experimental situations where AA is turned on or off, i.e. at AA in transition, not just in established disease.
C3H /HEJ mice develop spontaneous alopecia areata-like disease in about 20%, and will predictably develop multiple sites of alopecia if one grafts involved skin onto un-involved mice. This allows the development of large populations of alopecia mice with active progressing disease, which is an ideal model for gene and protein expression studies. Using micro array analysis, Sundberg found decrease in genes controlling pigment and hair proteins early on, which is just what one would expect. As the hair follicles are involuting, he found increases in gamma interferon and in antigen presenting cells, and later on induction of antibodies of multiple specificities. The changes in antigen-presenting cells may be an important determinant of the development of autoimmunity. This animal model, which has the capacity for induction of disease by grafting of alopecia skin grafts, may be the perfect model for the study of the breakdown of immune privilege in alopecia, and for the study of T cell responses in alopecia areata.
Epitope spreading is the expansion of the specificities of an immune response with time, characterized by the spreading of the immune response to recognize more epitopes on the same protein or to epitopes similar to the one that initiated the immune response. An epitope is a small polypeptide fragment recognized by antibody or T cell receptor. The C3H/HEJ animal model is the place to begin looking for epitope spreading, for T cell receptor repertoire, for specific receptors and what antigens they recognize, using the inducible aspect of the model to study early and later immune responses in this model. The genetic associations in this animal model show strong linkage to chromosome 17: DQB1, DRB1, TNF and LT and to chromosome 9: CD3 NCAM. These are clearly regions that control the immune response, and provide some idea of the genetic loci that will be identified in human alopecia areata.
Madeleine Duvic presented an update on the National Alopecia Areata Registry. This Registry will provide the means for collection of data on disease subsets and natural history of disease. In addition, we will focus on collection of multiplex families and sib pairs. These will be important in the study of genetic linkages in AA, AT and AU.
The power of genetic approaches was addressed by Angela Christiano. She reported on the results of study of 5 large multiplex families with AA, using some families that are not enrolled in the registry. Her initial results show significant linkages at four sites: Chromosome10, 6, 16, 18. So we’re starting down the road of finding the genes responsible for alopecia areata.
Elaine Fuchs discussed the WNT signaling pathway and how it controls cell proliferation, transcription and adhesion and attachment. WNT signaling is important in stem cell development, in hair follicle determination, and in committing hair matrix cells to formation of hair keratins. The details of how such signaling pathways link cell attachment, proliferation and differentiation will provide fundamental understanding of the molecular control of hair follicle development and cycling. In the context of alopecia areata, such knowledge may make it possible to maintain hair follicle growth even in the face of immune attacks in alopecia areata.
Another crucial signaling pathway controlling hair follicle development is the hedgehog signaling pathway, discussed by Andrzej Dlugoz. Sonic hedgehog is required for hair follicle growth but not differentiation, and mutations in this pathway produce basal cell carcinoma. Downstream mutations in this pathway stop hair follicle regression and induce tumors like BCC. We are developing fundamental understanding of control of the development and cycling of hair follicles involving multiple signaling elements and growth factors: WNT and hedgehog pathway, Keratinocyte Growth Factor, Fibroblast Growth Factor V, NOGGIN.
It is clear that neuropeptides are involved in controlling the immune response in psoriasis and atopic dermatitis. Vladimir Botchkarev has proposed that neurotropins may control CD8 lymphocyte populations in alopecia areata. His evidence is compelling that neurotropins are present in the hair follicles in AA lesions, and that neurotropins can kill peripheral blood CD8 positive T cells from AA patients. Do neurotropins eliminate CD8 suppressors intended to terminate the immune response? David Whiting has shown us that T cells are eliminated from lesions of long-established AA. Are neurotropins the mechanism for lymphocyte disappearance? We need new studies to determine what stimulates T lymphocytes in AA, and what makes them disappear in long-standing lesions.
Kevin McElwee presented additional work showing how valuable the C3H/HEJ mice model can be. Grafting involved skin from one animal to syngeneic mice without alopecia leads to the progressive development of AA. If these induced mice, there are complex immunologic changes: inflammatory infiltrates consisting of CD4> CD8 lymphocytes, macrophage and B cells; local increases in cytokines IL-12, IL-6> IL10, IL-4; all T cell subsets including T regulatory cells: and hair follicle changes requiring Fas and FasL. The specific role of T cell cytokines in inducing apoptosis in the follicle in AA will be an area of active investigation.
Desmond Tobin has continued and persisted in his quest to identify antibodies in different alopecia areata models – human, mouse, horse, chicken, and others. In human AA there are high titer antibodies to hair follicle antigen, sometimes as high as 1 to 10,000. In the rat model they precede hair loss. If you grow melanocytes from hair follicles these antibodies bind to them and the binding can be nuclear, cytoplasmic and membrane. In multiple animal models there are also antibodies to keratin or trichohyalin. Effective treatment of alopecia areata decrease the antibody levels and I think it’s further evidence for a complex immune response in alopecia areata. These antibodies can also be used as tools to determine immunologic specificity in AA: the antibodies can be easily sampled and one can measure antibody specificities more easily than with T cells, and I believe this is going to be a useful way to look at epitope spreading and maturation of the immune response in alopecia areata. It has been recently shown that immunotherpy in melanoma treatment can induce both regression of melanoma and vitiligo. The melanocyte destruction in these areas of alopecia appears to be mediated by autoantibodies, reawakening our interest in antibody destruction of melanocytes.
Jerry Shapiro showed that one can use modern gene expression techniques in animal models to determine the mechanism of action of some established treatment approaches: topical anthralin and nitrogen mustard. Anthralin induced hair regrowth in the rat alopecia model, along with decreases in gamma interferon, and in thymosin and JNK signaling. Topical nitrogen mustard also decreased gamma interferon, and increased IGF-1. These studies validate gamma interferon as an important mediate in AA, and suggest that other signaling pathways may be necessary for reemergence of new hair follicles.
Jonathan Vogel presented an interesting view of the future and how gene creams or solutions, or injecting DNA can be used to modify skin function and may directed to treat hair disorders such as AA. Even better, hair follicles may be removed, transduced ex vivo and then transplanted to alopecia areata patches. Will gene therapy make AA a surgical disease?
Amy McMichael and Vera Price presented view on the use of new modern immunomodulators developed in psoriasis and now available to test in alopecia areata.
These therapies fall into 4 major categories: killing activated T cells, inhibiting T cell activation, inducing immune deviation, and inhibiting key cytokines. The prototype drug for killing activated T cells is Ontak, a fusion protein containing a toxin and anti-IL-2R antibody. Drugs that block T cell activation interfere with macrophage/T cell interactions by binging the TCR or secondary adhesion and signaling pairs. Immune deviation drugs switch the T cell response from TH1 to TH2. Drugs that inhibit cytokines interfere with cytokines such as TNF-a, or their receptors (Enbrel and Remicaide). Another important approach in development is interference with leukocyte trafficking through binding important adhesion structures such as ICAM-1 and selectins.
The presentations at this Fourth International Alopecia Areata Workshop have taught us about the current state of knowledge in research on AA and hair biology, and on related topics in cutaneous biology. I think that it’s also important to ask what we can learn from others who have studied cutaneous autoimmune diseases and how we can apply some of the lessons from these diseases to better understand alopecia areata.
Lupus erythematosus is the prototypical autoimmune disease and there is much to learn from studies in lupus erythematosis. John Harley just presented at the Montagna Symposium on the Biology of Skin the results of 8 years of study of the genetics of lupus erythematosus. He has found is that many of the genetic associations in lupus could only be identified by study of carefully segregated subsets of disease based on organ involvement, autoantibody specificity, and clinical features [Harley 2002, Kelly 2002]. In implementation of the Alopecia Areata registry, we must take care to collect similar clinical data related to AA subsets.
Vicky Werth has studied the susceptibility to photosensitivity in cutaneous lupus and focusing on polymorphisms in genes associated with immune response and cell death, specifically TNF-a [Werth 2000]. This is a key mediator in inflammation in the skin, and in control of cell death. It is believed that these TNF polymorphisms may be important in the UV-induced cell death that leads to auto antibody production in photosensitive lupus. Polymorphisms in TNF, IL-1 and other MHC associated molecules will be important in studies of alopecia areata biology and genetics.
Landmark work done over the past six or seven years by Livia Casiola Rosen and Anthony Rosen at Hopkins has shown that many autoimmune diseases are produced by autoantibodies which are specific for fragments of self antigens produced during apoptosis, and presented to a defective immune system [Rosen 1999, Rosen 2001]. The prototype of this mechanism in photosensitive lupus, in which UVR-induced apoptosis induces fragments of Ro antigens [Casciola-Rosen 1996]. Apoptosis is the principle mechanism of cell death in the regressing catagen follicle and it may be that these apoptotic cells – hair matrix melanocytes, matrix keratinocytes, and root sheath keratinocytes- are being presented to the immune response and autoimmunity develops in alopecia areata because of genetically determined defects in the immune response. Is the immune response in alopecia areata directed against proteolytic fragments of auto antigens exposed during catagen? Is the principal genetic defect in AA related to how the cells of the hair follicle die during catagen?
Psoriasis provides many possible lessons for those interested in studying alopecia areata. It is a T cell dependent disease, often triggered by infection. What is most unique about psoriasis is that carefully planned clinical trials were used to understand the mechanisms of disease [Gottlieb 1995 , Krueger 2002 ]. The model was simple and direct: employ a highly specific treatment modality that targets a single point in the mechanism you’re interested in evaluating. Jim Krueger and Alice Gottlieb used highly selective anti T cell treatments to reverse intractable psoriasis. Those treatments which killed activated T (Ontak, UVR) cells lead to remissions of disease, while those that merely inhibited T cell activation (cyclosporin) produced clinical response but no remission when treatment was discontinued.
In alopecia areata, we must determine if killing T cells will also allow return of normal hair follicle function. If AA is triggered by an infection, will the immune response persist after the infection is gone? If there is epitope spreading, will the disease persist if the activated T cells are killed? Studies using specific anti-T cell treatments are essential in to answer these questions. There are a number of different models that we could use. Drugs that target pathogenic T cells, drugs that inhibit T cell activation, drugs that induce immune deviation towards TH2 cells, and drugs that inhibit cytokines like TNF-a. Killing pathogenic T cells is probably the best treatment because then you can see if you get a remission when the patient is taken off of treatment. Such drugs need to be used in “proof of concept” experiments.
Vitiligo is another autoimmune disease from which we can learn many lessons. Vitiligo is also a complex autoimmune disease characterized by both autoreactive T cells and autoantibodies. In vitiligo, repigmentation of affected skin is a separate process that involves more than just suppressing the immune response. Re-pigmentation of vitiligo is a distinct process different from the process of pigment cell destruction [Norris 1994]. It requires activation of a special population of melanocyte precursors in the outer root sheath [Horikawa 1996]. In alopecia areata, hair cycling returns when immunosupporessive are used, but not in every patient. In the patients that don’t respond to the blocking of the immune response do you have to add something else to directly stimulate the hair follicle? Better understanding of follicle cycling and control will provide us the tools to answer this question and perhaps to finds stimuli for anagen growth that overcome the inhibition imposed by the immune response in alopecia area.
This workshop has summarized the significant accomplishments in understanding hair biology and alopecia areata since the first NIAMS-sponsored workshop.
We’ve established that alopecia areata is a T cell mediated disease, and appears to be autoimmune. In alopecia areata, investigators have identified some of the antigens to which the immune response is directed, which has not yet been accomplished in psoriasis. We’ve had a tremendous growth in understanding the biology of the follicle and its role as an immune privileged site. We have initial genotyping of immune response associated genes, especially in AT and AU but we’re also identifying other genes and we’ll continue to do that the larger the patient populations become. We are collecting through the National Alopecia Areata Registry the patient data necessary to understand the natural history and epidemiology of this disease, and are beginning to stratify patients by disease subset, which is necessary for these studies and for the genetic analyses now underway. We have begun to correlate the different stages of pathology of the hair follicle, with disease subset, and with bio-chemical markers of immune disease and also hair follicle response, and with the complex genetic determinants of this alopecia areata and its subsets.
What needs to be done? We must continue to support basic research in hair follicle biology, as the foundation on which understanding of pathophysiology and treatment will be built. We need to support our Alopecia Areata Registry as the basis for fundamental mechanism and treatment trials. We need to strongly support genetic studies if we are to understand diseases mechanism and subsets. We need to fund opportunities to apply new research approaches like gene and protein expression through microarray analysis and proteomics. I strongly believe that we need to aggressively organize proof of concept clinical trials to verify that activated T cells are essential for development of AA and that anti-T cell drugs cause reversal of disease. We have to study the return of hair growth after blocking the immune response to see if additional stimuli. I think it’s time for us to take our animal models and develop the appropriate collaborations with fundamental T cell biologists to determine the T cell receptor repertoire in the lesions of AA. What are these cells recognizing? Is it the same, patient to patient? Is there epitope spreading? Does the T cell response cease late in disease? We are indeed close to making major breakthroughs in understanding of the mechanism and treatment of alopecia areata.
I must express sincere appreciation to NIAMS for their support of this entire field over the past 10 years, and I personally and specifically want to thank Alan Moshell, Head of the Skin Program at NIAMS who in so many ways has helped to support the investigators trying to understand alopecia areata and hair biology: through these meetings, through publication of these proceedings in the JID, and through good advice on funding approaches. Of course we also should thank Steve Katz, the Director of NIAMS for his support of the development of this area of research.
References:
- Werth VP, Zhang W, Dortzbach K, Sullivan K. Association of a promoter polymorphism of tumor necrosis factor-alpha with subacute cutaneous lupus erythematosus and distinct photoregulation of transcription. J Invest Dermatol. 2000 115:726-30.
- Rosen A, Casciola-Rosen L: Clearing the way to mechanisms of autoimmunity. Nature Medicine. 2001;7:664-5
- Rosen A, Casciola-Rosen L : Autoantigens as substrates for apoptotic proteases: implications for the pathogenesis of systemic autoimmune disease. [Review] Cell Death & Differentiation. 1999 6:6-12
- Norris DA, Horikawa T, Morelli JG: Melanocyte destruction and repopulation in vitiligo. Pigment Cell Res 7:193-203, 1994.
- Krueger JG. The immunologic basis for the treatment of psoriasis with new biologic agents. J Am Acad Dermatol. 2002 46:1-23. Review.
- Kelly JA, Moser KL, Harley JB. The genetics of systemic lupus erythematosus: putting the pieces together. Genes Immun. 2002 Suppl 1:S71-85.
- Horikawa T, Norris DA, Johnson TW, Zekman T, Dunscomb N, Bennion SD, Jackson RL, Morelli JG: DOPA-negative melanocytes in the outer root sheath of human hair follicles express premelanosomal antigens but not a melanosomal antigen or the melanosome-associated glycoproteins tyrosinase, TRP-1, and TRP-2. J Invest Dermatol 106:28-35, 1996.
- Harley JB. The genetic etiology of systemic lupus erythematosus: a short dispatch from the combat zone. Genes Immun. 2002 Suppl 1:S1-4
- Gottlieb SL, Gilleaudeau P, Johnson R, Estes L, Woodworth TG, Gottlieb AB, Krueger JG. Response of psoriasis to a lymphocyte-selective toxin (DAB389IL-2) suggests a primary immune, but not keratinocyte, pathogenic basis. Nat Med. 1995 1:442-7.
- Casciola-Rosen L,Rosen A, Petri M, Schlissel M: Surface blebs on apoptotic cells are sites of enhanced procoagulant activity: implications for coagulation events and antigenic spread in systemic lupus erythematosus. Proceedings of the National Academy of Sciences of the United States of America. 1996 93:1624-9.
