The hair follicle (HF) is the most prominent appendage of the skin and is a uniquely mammalian trait (1). Its secreted product, the hair fiber, is of considerable psycho-social importance for humans, rendering disorders of the HF highly impactful on our quality of life (2). Alopecia areata (AA), is a relatively common (lifetime risk of ~2%) immune-mediated disorder of HF growth, which can affect males and females of all ages and geographic ancestries (3–5). The diagnosis of AA is usually straightforward. It most commonly presents with patchy areas of hair loss on the scalp (so-called patchy AA) but can progress to loss of all scalp hair (A. totalis) and body hair (A. universalis).

AA is one of the most enigmatic of the common dermatoses, in large part due to its exceedingly unpredictable course, heterogeneous presentation, in the context of the unusual mosaic nature of human scalp hair growth where much autonomy is invested in each individual HF. Despite intimations to the contrary, the etiology of AA remains unknown. AA is associated, however, with several immune abnormalities, some non-selective while others appear to be more specific and point to an immune abnormality that targets a component(s) of the HF, especially when in the anagen phase of the hair growth cycle. Still, AA may also have features of systemic disease, as is characterized by frequent involvement of other tissues that so far include nails, eyes, and heart (6–9). AA is most readily seen on areas of the body with the highest proportion of HF in the growth phase of the hair cycle (called anagen) e.g., scalp and beard. However, it may appear on any hair-bearing sites (10, 11). AA has also been reported in several non-human mammalian species e.g., horses, dogs, inbred mice, rats, rhesus monkeys, cows, etc. (12–19). Treatment of AA is difficult; there is no FDA-approved drug, and there is currently no cure. Despite this, a broad range of treatment modalities can be used, including topical, systemic, and intralesional corticosteroids; topical immunotherapy; topical minoxidil; anthralin; and Psoralen plus Ultraviolet A (PUVA), and more recently JAK inhibitors (20, 21).

In this review we re-examine what is known about the antigenicity of HF proteins, potential HF autoantigens in AA, and whether post-translational modifications of the presumptive HF autoantigen trichohyalin, may render this crucial HF protein susceptible to triggering an inappropriate immune response to the HF in AA-susceptible patients. Specifically, we examine how two post-translational modifications (citrullination and deamidation) that can alter protein antigenicity, may be particularly relevant in the induction of AA.

Despite a very concerted research effort over the last 30 years, the etiology of AA remains elusive. What we call AA may even represent more than one disease entity. However, for simplicity, we can describe AA as multifactorial with a strong genetic predisposition in affected kinships (22, 23). There is a striking acquisition of ectopic immune-competence (most easily represented by upregulated MHC-I expression) in the cycling part of the growing HF (24), as well as other features of immunological dysregulation (25, 26). Dysregulated redox balance (27, 28), and microbial dysregulation (29, 30) may also feature in AA pathogenesis. While these and numerous other factors have been reported in AA over the decades (31), the classical features of active disease include the brisk infiltration of immune cells in a classic ‘Swarm of Bees’-like pattern (32), and the production of autoantibodies to HF-associated proteins (33–35). Recently, in an interesting transcriptomics study, Borcherding et al., observed clonal expansions of both CD4⁺ and CD8⁺ T cells, with shared clonotypes across varied transcriptional states in murine and human AA, suggesting autoantigen-dependent cellular autoimmune response in AA (36).

The curiously low immune-competence of the lower anagen HF is characterized by a range of devices, including the maintenance of low expression of classical MHC-I molecules in the HF epithelial components. These are described in considerable detail elsewhere (22, 24) and so will only be briefly alluded to here. This unusual low immune-competence of the epithelium in the growing HF is likely to be essential for the life-long cyclical (i.e., regenerative) nature of hair growth. Significant evolutionary selective pressure likely developed a range of devices to prevent the inadvertent and catastrophic immune-mediated damage to the mammal’s HF-dense coat in the wild (37). Thus, the growing HF, together with a small number of other sites (anterior chamber of the eye, placenta/fetus, testes, and central nervous system excluding brain) are able to ‘tolerate’ the production of antigens (self, non-self, modified, neoantigens, etc.) without eliciting an inflammatory immune response and so protect this vital skin appendage from direct or collateral damage.

Antigens in these appendages interact with T cells in an unusual way, including by inducing a form of tolerance to normally rejected stimuli (38). This is likely to be especially relevant in haired skin, where HF ostia provide for millions of potential ports-of-entry across our skin surface to ensure continuous connection between the skin organ and the outside world. However, to do this successfully, the skin must balance risks from the entry of microbes and other potential aggressins. Indeed, it is far from sterile under the epidermis of the skin, with the highest expression of prokaryotic 16S rRNA detectable at or close to the depth of the scalp anagen hair bulb (i.e. ~3 mm deep) (39).

AA is essentially a hair growth cycle disorder; as it presents with the greatest vulnerability to immune-mediated attack during early anagen, at a time coincident with the re-establishment of the HF fiber and melanin production systems (40). The HF is the only continually-cycling tissue throughout the entire lifespan of the mammal (41), and is an impressive example of a physiological deconstruction of a multicellular tissue (via programmed death called apoptosis), followed by its reconstruction (via massive cell proliferation, differentiation, and maturation) to facilitate a new generation of hair fiber growth. After initiation at the start of anagen (operationally divided into anagen I-V), active hair fiber production continues during anagen VI for a long time, typically 3-5 years on the human scalp, but in some people even longer, before coming to an end again with the involution of up to 70% of the growing hair follicle during a phase called catagen (42, 43). The catagen HF then transitions to a phase of relative dormancy (called telogen), around which time the hair fiber can be ejected from the human HF via a process called exogen (44). Then, the HF re-enters a new hair fiber-producing anagen phase, or in the aging mammal can remain empty as a so-called ‘kenogen’ HF for some further time (45). Thus, as the greatest vulnerability to immune-mediated attack in AA occurs only during early anagen, we must concentrate more attention on what is going on in the HF at this particular time. Moreover, if one attempts to search only for the footprints in the wake of key immune-mediated events in the AA immune attack, it is likely to be much too late.

In this game of dominoes, the immune cascade of successively falling tiles has already begun when HF tissue damage or immune cell infiltration has already become evident (i.e., the thief has been and gone). Moreover, to turn off this immune assault in early AA lesions or to stop/protect it from starting in the first place, one needs to identify the mover(s) of the first falling domino as well as identifying the nature of the first domino itself. In our view, this is likely to center on ‘flag-waving’ change(s) in the targeted HF i.e., the antigenic flags that first attract the attention of a misguided immune response. Given that this HF disorder commonly targets people with a (genetic) predisposition/susceptibility to inflammation and autoimmunity (e.g., atopy, Hashimoto’s thyroiditis, pernicious anemia, celiac disease, etc.), a preferred mode of the (auto)inflammatory response to the HF may already be established in these patients via an already altered immune-activation threshold (22, 46). Thus, we need to determine the cellular and tissue events that pre-date induction of ectopic immune-competence in the AA-susceptible lower HF. The latter is prevented under normal circumstances by an active or induced state of immune privilege or tolerance unique to key AA-targeted epithelial components of the HF (2, 47). Early work by Bystryn and others reported the presence of abnormal immune deposits in the lower HF in AA patients, which suggested this region of the HF to be a prime target of pathogenic autoinflammatory attention (40, 48, 49). This was further confirmed with the observation of circulating antibodies in the sera of AA patients that bound to proteins extracted from scalp anagen HF (33), and later to distinct compartments of the anagen HF itself (34). Patients with autoimmune polyendocrine syndrome type I (APS I) - a condition where up to 37% can have AA (50) - are also observed to have high titre autoantibodies directed against the anagen bulb matrix, hair fiber cuticle, and cortex keratinocytes and also against melanocyte nuclear antigens (51). Thus, it is evident from several studies that anti-HF autoantibodies, in addition to cellular drivers, are associated with AA pathogenesis, even if not directly as effector molecules. Still, most attention has been directed toward the cellular arm of the immune response to HF in AA, leaving humoral factors relatively ignored to date.

Histopathological observations suggest that the hair growth cycle is disrupted in AA, due to the premature involution of early anagen (anagen III-VI) HF into catagen (52, 53). This is key, as AA does not present with an inflammatory response to the overlying epidermis or even the upper non-cycling parts of the HF (infundibulum and isthmus) - despite that, cells of the same histological type reside there i.e., keratinocytes, melanocytes, etc. Thus, there must be something unique happening in the cycling part of the AA-susceptible HF during early/earliest anagen, which induces, exposes or in some other manner activates an auto-inflammatory process. As a result, the soon-to-be targeted anagen hair bulb is forced to leave its protective immune silence and move above the radar with often catastrophic consequences.

While the involvement of cellular (intra-/perifollicular immune cell infiltrates) and humoral (anti-HF autoantibodies) immunity is well-established in AA, much less progress has been made to identify the HF-associated autoantigen(s) that induces pivotal stages in the pathogenesis of this disorder (25, 31, 49). Recently, some have even proposed that the phenotype of AA may be derived from alternatively autoimmune and non-autoimmune pathomechanisms (54), with the latter mode being largely independent of autoantigens. Still, a small group of AA researchers has persisted against all (including funding) odds with their attempts to discover relevant autoantigens in AA. They do this, not only to determine whether AA can be definitively classified as truly ‘auto-immune’, but also to explore whether tolerization to a self-HF specific antigen(s) could shut-off the immune-mediated attack on the anagen HF and so treat (or even cure) this frustrating skin condition.

To date, a few potential autoantigens in AA have been suggested ( Table 1 ) and these are mostly derived from HF keratinocytes and HF melanocytes. The most prominent of these, Trichohyalin (TCHH) is located mostly in the HF inner root sheath (IRS) and importantly is only synthesized during the anagen stage of hair growth. Support for the involvement of TCHH comes from a variety of sources. Early work has shown the common expression of circulating IgG autoantibodies to TCHH in human (55) and animal (16, 56) patients with AA hair loss, and later work showed T cell reactivities to this protein in AA patients (57) (see below for a detailed discussion of TCHH as a presumptive autoantigen in AA). However, from these data, it is not yet clear which autoantigen(s)/epitope(s) is the first target of the inappropriate attention of the AA immune-system. We need to distinguish between targets of the initial stage of AA onset versus the antigens/epitopes that emerge during antigenic drift and epitope spread as AA disease progresses and/or in disease of long-standing (49, 55, 57, 60, 63, 64). Thus, discovery of the key HF-associated autoantigens that are truly responsible for triggering the initial immune-mediated cascade against the anagen HF in AA may help us to better understand disease onset and progression. Furthermore, identifying key autoantigens could possibly pave the way for developing novel therapeutic approaches.

Previously, we observed that hair bulb keratinocytes and hair bulb melanocytes were primarily affected in acute AA (34, 53, 58), suggesting some involvement of (precursor) keratinocyte- or melanocyte-associated antigens, which may become visible to the immune system as a result of either upregulation of immune-competence in targeted HF (e.g., via ectopic MHC-I/-II molecule expression) with or without the help of local antigen-presenting cells.

Autoimmune responses represent dysregulation of effector and regulatory immune mechanisms, which generally develop through phases of initiation and propagation, and often show periods of remission and relapse (83). Post-translational modification (PTM) is one of several normal devices for expanding or diversifying the functional protein repertoire, which can produce an extraordinarily complex ‘proteoform’ (84). Also, any one single protein may undergo significant modification during cellular differentiation and maturation. This plasticity of peptide epitope form is likely to influence the creation of potential (self)-antigens. PTMs can occur in self-proteins during the normal maturation/differentiation processes of cells, but may be further altered in the presence of additional cell stressors, such as oxidative stress, inflammation, aging/senescence, etc. How these (self)-antigens are perceived by an individual’s immune system will depend on a range of host factors, not least their susceptibility to autoreactivity. Prominent examples of the latter include associations with particular MHC-I and -II alleles (22), and polymorphisms including in PTPN22 (encoding a lymphoid-specific tyrosine phosphatase that is a master regulator of the immune response), and AIRE (encoding a transcription factor expressed in the thymic medulla and involved in the elimination of self-reactive T cells) (85)-. However, alone these may be insufficient for disease development. Previous studies have shown that negative selection of T cells specific for the PTM-variant of the same antigen requires transport of the peripheral antigen (e.g., TCHH here) to the thymus by professional APCs (86). However, those studies suggest that peripheral tolerance mechanisms and/or transport of peripheral self-antigens to the thymus cannot compensate for the lack of autochthonous (i.e., indigenous) thymic-induced central tolerance.

APCs enter the affected tissue, engulf cellular damage and potentially transport these proteins to the lymph node for presentation in the context of MHC-I/-II molecules to lymphocytes and subsequent amplification of the antigenic repertoire. Consequently, this leads to infiltration of autoreactive T cells and B cells into the host tissue where that may drive an auto-inflammatory or autoimmune reaction (87). Several autoinflammatory disorders exhibit striking post-translational modifications in tissue-restricted proteins that trigger autoimmune responses ( Table 2 ). In that way, genes that encode candidate antigens do not necessarily need to be reflected in patient GWAS studies. Still, Petukhova et al., reported previously that patients with AA exhibited several overlapping genetic susceptibility loci with other autoimmune diseases, including most commonly rheumatoid arthritis (RA), type I diabetes (T1D), and celiac disease (CD) (22), suggesting that these disorders may provide insights into the etiopathogenesis of AA. However, in terms of AA autoantigen discovery, our knowledge lags far behind that of autoantigens in these more definitive autoimmune conditions.

Here, we propose that some PTMs, such as citrullination, deamidation, and oxidation, may be particularly relevant with respect to autoantigen generation in AA.

The HF presents particular challenges for the identification of auto-reactive targets in disorders like AA. This is because self-antigens that are expressed in the lower HF during the growth phase only (even PTM-variant forms) are not usually easily recognized by the immune system, due to the (very) low expression of MHC-I antigens in the HF epithelium. This is not the case for most other autoimmune disorders that affect normally immune-competent tissues (e.g., RA, T1D, etc.). Thus, we do not yet know whether the auto-inflammatory response to HF in AA results from a) an ectopic immune response to native (i.e., unmodified) self-antigens expressed by the healthy HF, b) a normal immune response against modified self-antigens (or neoantigens), or c) a normal immune response against self-antigens (modified/non-modified) that were not preciously visible to the immune system (because they may be conformationally-hidden or sequestered) but now become exposed and presentable in an MHC-I/II molecule-restricted manner.

Several triggers have been proposed to explain the loss of HF ‘immune privilege’ in AA, including micro-trauma, viral infection, bacterial superantigens, psycho-emotional stress, mast cell degranulation, and other immuno-genetic factors (24, 25, 123, 124). Induction of the resultant ectopic HF immune-competence, in the anagen hair bulb, may be aided by other (secondary) factors, not least peri- and/or intrafollicular secretion of IFN-γ. This potent cytokine can upregulate MHC-I expression in the epithelium of the HF anagen bulb. In that context, HF self-antigens, which are normally undetected, may now be presented to the immune system as either modified-, non-modified or neo-antigens. However, immune privilege is not ‘absolute’ in the healthy growing anagen HF. While MHC-I expression may be low in the healthy hair bulb epithelium it is not fully absent, and there are also low numbers of various immunocytes (e.g., CD4⁺ and CD8⁺ T cells, macrophages, Langerhans, mast cells) both in the HF mesenchyme (dermal sheath, dermal papilla) and even within the hair bulb itself (75). Thus, there is already some machinery that may drive or facilitate immune dysregulation in this part of the HF, which may be recruited to the earliest AA-associated changes within susceptible HFs.

A key question in the etiopathogenesis of all auto-inflammatory disorders, including AA, is what is the first strike and what is the identity of the first domino to fall in the subsequent immune-mediated cascade. Some suggestions include loss or absence of central tolerance to ectopically MHC-I-presented epitopes on HF antigens, as well as an upregulation of NKG2D ligands (e.g., MICA and ULBP) expressed in the hair bulb and associated mesenchyme (22). These initial events likely precede the mass infiltration of auto-reactive cytotoxic CD8⁺-NKG2D⁺ lymphocytes that wreak such destruction of the anagen hair bulb in AA.

While promiscuous expression of tissue-restricted antigens, by medullary thymic epithelial cells, is required to establish effective central tolerance (125), and loss of immune tolerance is one of the hallmarks of autoimmune diseases (126), it is not known whether intra-thymic expression of peripheral tissue-restricted antigens also extends to those proteins with PTMs. In a PTM-dependent mouse model of autoimmunity, involving the tissue-restricted self-antigen collagen type-II, T cells specific for the non-modified, i.e., native, antigen were seen to undergo efficient central tolerance. However, the PTM-variant antigen-reactive T cells escaped thymic selection, despite that the PTM-variant constituted the dominant form of the antigen in the periphery (86). This implies, at least for this case, that the PTM protein is not present in the thymus or at least is unable to induce negative selection of developing thymocytes. Thus, PTM-variant antigens may exhibit a lower level of tolerance induction than their non-PTM i.e., native parent peptide. As most self-antigens are in fact post-translationally modified, this may appear a rather curious situation. Still, it does raise the possibility that central tolerance is regulated differently for non-modified versus PTM-variants of the same self-antigen (e.g., TCHH here). Thus, T cells specific for self-antigens naturally subjected to PTM, like TCHH, may also escape effective central tolerance. Regardless, T cell reactivity to PTM-variant self-antigens is a known initiating/perpetuating factor in the progression of autoimmune diseases, as such modifications significantly affect antigen binding to MHC molecules, and consequently affect T cell activation (97, 112).

Given that the focus of the cellular and humoral immune responses to HF in AA is concentrated in/around the anagen hair bulb, it is highly likely that relevant autoantigen expression is associated with cellular proliferation and differentiation dynamics restricted to this region of the HF. Thus, likely autoantigen candidates are expressed in corresponding hair bulb cells, including melanocytes and/or proteins of early keratinocyte differentiation. Trichohyalin (TCHH) is a prominent example of the latter category, as it is one of the very earliest differentiation proteins to be detected in early anagen. It is of note that the TCHH gene is located within the so-called Epidermal Differentiation Complex (EDC), containing over 50 individual genes involved in keratinocyte differentiation (127). TCHH protein is located within granules of precursor and differentiated cells of the IRS, where it confers mechanical strength to the HF, and in precursor and differentiated keratinocytes of the hair shaft medulla (when a medulla is present). TCHH-expressing cells are first detectable in the most peripheral and proximal anagen bulb ( Figure 1 ). Once these cells move distally, as they mature and differentiate, they reside in their final ‘inner’ root sheath position within the HF, bounded by keratinocytes of the ORS as these cells fully emerge close to the supra-bulbar HF. Thus, it is important to emphasize that proteins of earliest IRS differentiation are amongst the most ‘exposed’ in the HF to potential antigen presentation when cellular and humoral factors induce abnormal immune-competence in this tissue.

TCHH undergoes two key post-translational modifications (PTMs) during its maturation process, including citrullination by PADI enzymes and deamidation by TGases (81). Both PTMs are immunologically very important, as discussed above. Upon careful examination of co-immunostaining of anagen scalp HFs with AA patient sera and a monoclonal antibody to TCHH, we observed binding colocalization predominantly in the most exposed region of the IRS called the Henle’s layer ( Figure 2 ). The latter is the first IRS sublayer to harden/cornify (128). There are a few key points worth noting from our immunohistochemical findings, which may reflect variations in the antigenicity of TCHH protein in the IRS. First, there is significant variation in the level of TCHH protein expression in the three different constituent IRS sub-layers. Second, the TCHH epitopes targeted by AA serum IgG autoantibodies appear spatially-restricted within the IRS. Third, AA patient serum IgG antibodies contain specificities that can target the earliest TCHH-positive precursor IRS keratinocytes, long before there is any discernible morphologic evidence of this HF cell layer. These TCHH-positive cells, importantly, are located most externally in the anagen hair bulb, at the interface of the proliferatively-active matrix epithelium and the mesenchymal connective tissue sheath. Thus, these precursor IRS keratinocytes can readily interact with rare surveilling immunocytes, either resident normally in this tissue or transiting from the peri-follicular vasculature. Fourth, intense AA patient serum IgG binding can also be detected circumferentially in keratinocytes located in the supra-bulbar area of the anagen HF, at a level reminiscent of IRS cells within the redox-sensitive ‘ring of fire’ (78). The latter observation suggests the potential additional involvement of oxidatively-modified autoantigens/neoantigens in AA. Thus, we propose that citrullination, deamidation, and oxidation are key PTMs involved in autoantigen generation in AA.

The extant AA literature is voluminous, it is dominated by studies reporting the involvement of various immunocytes (including cytotoxic CD8⁺ T cells, classical natural killer (NK) cells, invariant NK T cells, γδ T cells, type-I innate lymphoid cells (ILCs) etc. (25). Remarkably, little research has focused on autoantigen discovery in AA or even on the initiating events in AA pathology. Thus, a much deeper understanding of the antigen targets of both autoantibodies and TCRs in AA will contribute towards more effective patient stratification, characterization, and management, as well as aiding the development of more targeted immunotherapies. These include the potential for induction of tolerance to the immunodominant autoantigen(s) in AA, that could provide for an ultimate cure. An important note of caution; most studies reporting on antigenic epitopes in autoimmune diseases rely on synthetic peptides or recombinant proteins (129); approaches that may ignore relevant PTMs that confer, potentially, the all-important antigenicity to these self-peptides. Thus, greater research focus needs to be directed to work on the affected tissue (e.g., scalp anagen HF) of AA patients at difference stages of their disease. Exciting recent discoveries of PTM-variant antigenic peptides in other autoimmune diseases, including RA and TID, demonstrate the very significant potential for this approach in AA.

SDJ and DJT were involved in the conceptualization, literature review, writing, and final editing of the original draft of the manuscript. Both authors have read and agreed to the published version of the manuscript.

This work was supported by a start-up grant to DJT from University College Dublin, Ireland, and a grant to SDJ & DJT from Alopecia UK, United Kingdom [AUK2021_003].

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.