An intense psychic shock may also exert pronounced effects on the skin, e.g. graying and generalized loss of hair (Hans Selye, 1950)1
Even today, it is still uncertain whether the above claim is based on fact or fiction. And yet, classical as well as modern literature, caricature, even fairy tales are full of anecdotes connecting hair loss and ‘nerves’. Also, in clinical practice one often meets patients convinced, that their hair loss has something to do with their perceived ‘weak nerves’ or is ‘stress’-related (1–3). For hair diseases such as alopecia areata, a ‘trophoneurotic’ or ‘psychogenic’ etiology has been suggested, although only accepted by few. However, even investigators, who consider alopecia areata an autoimmune disease, acknowledge that psychoemotional stress can indeed trigger or aggravate this hair-loss disorder, at least in a subgroup of patients (3–7) [Psychological stress can cause ‘overnight-graying’, possibly a fulminate attack of diffuse alopecia areata affecting pigmented hair follicles (HF) only, although a direct connection could only be proven occasionally (1)] (Table 1).
|1. Clinical observations of ‘stress’-induced alopecia (2,3)|
|2. Denervated anagen HF survive in culture and continue to produce hair shafts (32)|
|3. Hair growth occurs in denervated transplanted skin (129)|
|4. HFs are densely innervated, especially around the stem cell containing bulge region (32,33,35)|
|5. HF innervation is high during early anagen, during which period hair growth promotion by neurogenic factors occurs (32,33,35)|
|6. HF innervation is low during telogen and late anagen, during which period HF regression is promoted by excess neurogenic factors and neurogenic |
|7. Cutaneous neurotrophin overexpression is associated with both premature HF regression and hyperinnervation of HFs (60,61,130,131)|
|8. Hair growth inhibition by capsaicin-induced release of sensory neuropeptides from peripheral nerve endings in late anagen skin (73)|
|9. Arrest in telogen upon foot shock stress (76)|
|10. Proliferation stop and apoptosis associated with neurogenic inflammation upon sonic stress and SP (10–12)|
|11. Premature HF regression upon sonic stress, SP and untimely high SP-innervation in late anagen (12,13)|
|12. NK-1 or NGF blockade and mast cell deficiency reverse negative effects of stress, SP or NGF on hair growth (12,13,132,133)|
|1. Stressed individuals are not necessarily bald individuals|
|2. Denser hair growth in hyperinnervated skin, e.g. after wounding (111,122–124)|
|3. Hair growth promotion in early anagen by high doses of SP (32)|
|4. HF innervation is essential for hair growth, e.g. denervation through chemical neuralektomie or diabetes mellitus causes HF atrophy (125–128)|
|5. Cutaneous inflammatory processes and stress-exposure increase nerve fiber numbers in skin but are clinically not associated with hypertrichosis (50,134).|
If psychoemotional stress can exert any hair growth-inhibitory effects at all, these must be mediated via definable neurobiological, neurorendocrine, and/or neuroimmunological signaling pathways (8). Luckily, candidate pathways that could serve as a link between the central nervous system and the HF have increasingly become defined over the past decade. It is both logical and timely to examine stress, nerves, immune cells, and hair growth control from a truly interdisciplinary point of view, that integrates relevant progress in basic stress research, skin neuroimmunology, and skin neuroendocrinology (Table 1).
We begin this review by summarizing recent progress made on the most intriguing topic of ‘neural mechanisms of hair growth control’, because we first had covered this topic (9). This will provide us with the necessary background for any exploration of the role of psychoemotional stress in this clinical, biological, and psychological highly relevant context. Drawing upon our recent findings in an established murine stress model (10–13), we present evidence from the murine system in support of the ancient concept that ‘stress’ can indeed lead to hair growth inhibition (Table 2). In doing so, we focus on the concept of a ‘brain-HF axis’ (10) along which we visit the classical stress axis, the hypothalamic- pituitary-adrenal (HPA) axis. However, the potential pathways that underlie stress-induced hair growth inhibition elucidated so far focus on sensory neuropeptides and neurotrophins. We argue, that dissection of these pathways in the murine system will serve as a particularly instructive model to better characterize the complex effects of psychoemotional stress on skin biology in general, and on skin neuroimmunology and skin neuroendocrinology in particular.
|HF and their innervation develop parallel to each other (20)|
|HF and HF innervation development depend on each other (20)|
|First HF regression is associated with a decrease in peptidergic peri- and interfollicular innervation (20)|
|Cycling – physiological conditions|
|Innervation around and between HFs fluctuates hair cycle dependently|
|Low numbers of PGP 9.5+ (pan-neuronal marker), Gap-43+ (marker for neuronal plasticity or SP+ (peptidergic innervation) nerve fibers in telogen, late anagen, and catagen|
|High numbers of these nerve fibers in early anagen (32,33)|
|Mast cell contacts with SP+ nerve fibers are high in early anagen (24)|
|Hair growth control|
|SP promotes hair growth in early anagen skin organ culture (32)|
|Stress – pathological conditions|
| SP+ and Gap-43+ nerve fiber numbers increase in skin normally containing low numbers (telogen and late anagen) |
upon stress-exposure (12,50)
|Mast cell contacts with SP+ nerve fibers, mast cell degranulation and endothelial gapping increase upon stress-exposure (12,50)|
|Deleterious to the HF|
|HFs and interfollicular dermis show increased numbers of apoptotic cells and infiltration upon stress-exposure (12,50)|
|Stress is associated with arrest in telogen or premature catagen development (12,76)|
|Stress effects can be abrogated by NK-1-antagonist, NGF-neutralizing antibody, and surprisingly by minoxidil (10–13)|
The unusually dense and intricate innervation of the HF suggests roles beyond classical sensory and autonomic functions
HF innervation is surprisingly dense and complex [Sensory nerve fibers containing for example SP (substance P) terminate freely in the epidermis or dermis or on merkel cells, HFs, blood vessels, lymphatic vessels, or sweat glands (14,15,16,17,18,19,20,21,22). They are also frequently located in close physical contact to mast cells (23,24,25,26), endothelial cells (20,27), dendritic cells (28,29,30), and macrophages (31). On the murine pellage HF, we find SP+ nerve fibers in the subepidermal dermis, circulating the HF ostium; in single-nerve fibers in the interfollicular dermis and innervating subcutaneous blood vessels (32). Prominent SP+ nerve fibers can also be found in nerve fiber bundles at the border between dermis and subcutis, typically in close anatomical neighborhood to the insertion point of the arrector pili muscle into the HF bulge region, where they give rise to fibers innervating the arrector pili muscle (32,33)], and an elaborate nerve fiber network is present on each and every tylotrich and non-tylotrich pelage HF in the murine back skin (9,20,34) (Fig. 1). This observation has all along suggested a function beyond the sensory or autonomic (9,24,32,35). Given the multitude of neuronal signaling molecules contained in the peri- and intrafollicular neural plexus (5,9,14,17,20,36), it is tempting to speculate that secreted neurotransmitters, neuropeptides, and neurotrophins can exert hair growth-modulatory effects, e.g. by targeting epithelial stem cells in the HF bulge region (9,12,32,37). In fact, one of the first biological functions discovered of the prototypic sensory neuropeptide SP was its trophic role for the maintenance of epithelial appendages more than two decades ago (38), followed by the observation that SP is a growth factor for fibroblast and endothelial cells (39).
Today SP is widely acknowledged as a potent immunmodulator (40,41) and stress-mediator (42) and it does therefore not surprise that it can also affect hair growth indirectly via modulation of the perifollicular immune response. Most if not all peripheral axons transport vesicles containing neuropeptides such as SP, which are produced in the respective cell bodies in the dorsal root ganglia (22,42,43). Upon depolarisation, they are released from unmyelinated A-δ- [high threshold, rapid adapting pain, cold, and touch receptors with medium conduction velocity (44,45)] or C-fibers [low threshold, rapid adapting pain, temperature, and mechanoreceptor or sleeping nociceptors with slow conduction velocity (46)] to produce for example the classical features of neurogenic inflammation: mast cell degranulation, vasodilatation, and plasma extravasation (22,41,47,48). Intriguingly, this cascade can be initiated by stress-exposure in murine skin under the control of SP and NGF (nerve growth factor) (42,49,50).
Besides the growth and immunomodulatory response of cutaneous target cells such as keratinocytes and mast cells to neuronal stimulation (40,51–56), neuronal signaling molecules can also be produced by and released from these cells (57–62), for example after stress-exposure (13,49). Thus, stress-triggered release of neuropeptides from peripheral nerve fibers may set in motion a complex and multilayered interactive machinery (63,64), which ultimately leads to trophic effects in the pilo-neural complex (8–10,12). To better understand what may control this machinery, it is crucial to first understand the ‘hardware’ of HF innervation and pilo–neural interaction throughout the hair cycle.
HF cycling is associated with changes in HF innervation
Analysis of HF innervation in the back skin of adolescent mice has challenged the long-held dogma that peripheral innervation only changes during development, and under pathological conditions such as inflammation or injury. During the synchronized, cyclic transformations of mature HFs between stages of rapid growth (anagen), apoptosis-driven regression (catagen), and relative quiescence (telogen) (65), peri- and interfollicular innervation and nerve fiber-mast cell contacts fluctuate. They are low during telogen, rise dramatically in early anagen and then drop back to telogen levels in late anagen and catagen (24,33). Progression through the hair cycle thus was the first physiologically occurring tissue- remodeling process shown to be associated with fluctuations in nerve fiber density and neuroimmune interaction.
During this process, cutaneous SP content and nerve fiber number – among other neurotransmitters and neuropeptides – peak in early anagen (32,35,66,67). An increase, that reflects neuronal plasticity (33) and may be caused by a rise in the skin content of the neurotrophin NGF in early anagen just prior to the rise in SP (68,69).
When low levels of SP are released from subcutaneous implants, hair growth is induced in skin containing only HFs in the resting phase of the hair cycle (telogen) (66). Also, SP given to early anagen HFs in culture (Fig. 2) promotes hair cycle progression (32). This indicates that SP is a positive growth factor for murine keratinocytes and HFs in telogen and early anagen, which is in good keeping with a number of publications reporting on keratinocyte proliferation under SP treatment (52,70–72).
In contrast, SP-induced mast cell degranulation in vivo leads to decreased epithelial proliferation in rats (55) and induces premature, dystrophic catagen-development in mice in high anagen (73). These observations indicate a SP-triggered deleterious effect of neurogenic inflammation on hair growth. This is in line with the observation that SP does not act as a growth factor, e.g. for keratinocytes under culture and in vivo conditions that resemble the cytokine milieu of inflamed skin, a cytokine milieu also observed in catagen skin (70,72,74). In this context, it is interesting to note that SP content and nerve fiber number decrease in late anagen and that this coincides with a rise in neutral endopeptidase (NEP) activity in the skin (66). Thus, under physiological conditions, in late anagen, SP effects can be terminated on time by NEP to protect the HF and surrounding tissue from continuous inflammation.
Stress-exposure exerts profound hair growth-inhibitory effects on murine HFs in vivo
In the above mentioned mouse model for perceived psychoemotional stress, telogen HFs exhibited increased numbers of apoptotic cells, a finding which was associated with signs of perifollicular inflammation (10) (Table 2). This model utilizes a 24-h exposure to sound emitted by a rodent repellent device that had originally been established to investigate stress-triggered abortion (75). Under these study conditions, interfollicular SP-immunoreactive nerve fiber numbers increased as did Gap-43-immunoreactive nerve fiber numbers, indicative for axonal outgrowth and plasticity in skin upon stress-exposure (50) (Table 2). Likewise neuroimmune interaction was altered with increased nerve fiber-mast cell contacts and mast cell degranulation (50) (Table 2). These findings were suggestive for a deleterious effect of stress-triggered neurogenic inflammation on hair growth, which could subsequently be proved in mice with all their HFs in the late anagen phase of the hair cycle. Here, sonic stress induced neurogenic inflammation and led to premature HF regression (12,13) (Table 2). In a model employing a more physical stressor, electric foot shock, it was also shown that stress freezes HFs in the telogen stage (76).
Neurokinin-1 receptor blockade antagonizes stress-induced deleterious perifollicular inflammation and premature HF regression
These findings prompted the hypothesis of a ‘brain-HF axis’ to exist (10). With the above observations (10,12,50) in mind, we were thus not surprised to discover that systemically applied SP acts as a deleterious stress mediator to the HF epithelium (10–12). We found that SP is as potent as the physiological stress response to sound, in the induction of apoptosis and perifollicular neurogenic inflammation. This response involves mast cells and macrophages (10–13), and ultimately leads to premature HF regression. Most intriguingly, in mice, this stress response can be antagonized by the blockade of the SP receptor neurokinin-1 (NK-1) (10,12) (Fig. 3) and is lacking in NK-1–/– mice (Arck et al., in preparation).
Mast cells serve as important switchboards in stress-induced pilo–neural interactions
Close contacts between mast cells and nerve fibers have been described repeatedly in normal and inflamed skin by light- and electronmicroscopy (23,31,33,77). Mast cells can modulate the proliferative activity of HFs and thereby hair growth in more than one way (73,78). The functional significance of these contacts may be bi-directional (22,26,79–81). SP can degranulate mast cells and induce cytokine expression (22,52,82,83). Mast cells in turn produce the neurotrophin NGF and thereby promote survival and outgrowth of sensory nerve fibers (84,85). In addition, selective release of proteases upon stimulation with SP (22,86) can activate protease-activated receptors (PAR) on the surface of sensory nerve fibers, contributing to further activation and neuropeptide release (22). Intriguingly, in mice unable to respond to stem cell factor (SCF) stimulation due to severe reduction of peripheral mast cell numbers (KitW/KitW–v), stress fails to induce premature catagen. At the same time, the presence of SP+ nerve fibers in the skin of these mice is significantly lower than in wildtype mice.
Peptidases such as NEP or dipeptidyl peptidase IV (DP IV, CD26) time the effects of neuropeptides released by nerve fibers and other cells. Accordingly, NEP terminates SP-induced cutaneous inflammation (87). As mentioned above, neuropeptide activity in skin experiences rapid and effective termination by proteolytic enzymes such as trypsin from mast cells or NEP on the cell surface of Schwann cells and keratinocytes (87–90).
Up- and downstream of SP NGF regulates stress-induced pilo–neural interactions
Upstream of SP, NGF can promote outgrowth of SP+ nerve fibers, e.g. toward mast cells, while it can degranulate mast cells downstream of SP. When data were published that demonstrated systemic increase of NGF upon social stress exposure in mice and after parachute jumping in humans, NGF was added to the long list of stress mediators (cf. 91). We were therefore interested to know whether NGF was involved in the stress-induced neurogenic inflammatory response observed in our studies. We found that the stress response following sonic stress exposure is accompanied by an increase in NGF expression in skin and can be antagonized by the application of NGF-neutralizing antibodies (10–13) (Fig. 3).
Bi-directionality in pilo-neuroimmune communication
The hypothalamic hormone and neuropeptide corticotropin-releasing hormone (CRH) serve as a key control peptide upstream of neuroendocrine systemic stress responses (92). CRH can also induce mast cell degranulation and thereby local neurogenic inflammation of the skin (42,49). In transgenic mice overexpressing CRH, which are exploited as a murine model of chronic stress, hair loss has been reported (93), and foot shock stress upregulates CRH immunoreactivity in murine pelage HFs (76). Furthermore, this prototypic hypothalamic stress hormone (94) is also prominently detectable in the epithelium of murine and human HFs (95–97), and at least in mice, CRH immunoreactivity of the HF epithelium is hair cycle dependent. Together, these observations suggest that CRH may be involved not only in the local coordination of intracutaneous and intrafollicular stress responses but also in the modulation of HF pigmentation, growth, and/or immunology (92,96,98). However, the capacity of NK-1 receptor blockade to abrogate the pilo-neural stress response suggests a leading role for SP in the deleterious stress response of murine late anagen HFs.
Most recently, we could show that normal human scalp HFs even display a fully functional peripheral equivalent of the HPA axis. CRH is capable of inducing intrafollicular adreno corticotropic hormone (ACTH) and cortisol synthesis in organ-cultured human scalp HFs and regulatory feedback loops, similar to those found in the central HPA axis, are established in this miniorgan (99). Therefore, it is particularly intriguing to define to which extent neurally secreted products, that are released upon exposure to psychoemotional and other environmental stressors such as oxidative stress into the perifollicular environment (e.g. CRH, ACTH) change the locally established, CRH-driven, intrafollicular stress response system (Fig. 3).
Downstream of neuropeptides and neurotrophins, the detrimental effects of stress on the HF may also be brought about by cytokines such as tumor necrosis factor-α (TNF-α), acting as bi-directional mediators in the chemical crosstalk between the nervous and immune systems (79,100,101). Cytokines are important partners in this crosstalk (102) because stress perception is known to profoundly alter cytokine responses in the context of complex neuroendocrine–immune interactions (63,64,103–106). In addition to cytokine production by dermal mast cells and macrophages, γδ T-cell-receptor+ lymphocytes may also serve as stress-effector cells that generate, e.g. TNF-α (75,105,107,108). To identify ‘friends or foes’ of hair growth, γδ T cells clearly deserve special attention as they represent the vast majority of lymphocytes within the murine HF epithelium and form a key element of the HF immune system (109).
The notion, that increased expression of inflammatory cytokines may be detrimental to HF maintenance, is further supported by increased signs of apoptosis in the bulge region in stressed mice with all HFs in telogen or catagen (10,12). This region of the HF contains one putative stem-cell population of the HF (37), and it appears likely that in response to stress such a delicate area is one key target of immune cells. In the bulge region, inflammatory cells have the potential to damage the HF epithelium by upregulation of apoptosis through the secretion of cytokines, i.e. TNF-α, interleukin 1β, and interferon-γ(110–113).
The solid in vivo-evidence for the existence of a defined ‘brain-HF axis’ (10), which we have reviewed in this article, provides leads and lessons for future research both experimentally and clinically. The ultimate aim will be to develop a better understanding of neuroimmune interactions in skin. Wherever hair growth disorders are under investigation, modulation by chronic psycho-emotional stress must be taken into consideration. In addition, these findings provide an instructive model for nerve–skin interactions relevant to other skin appendages, such as the sebaceous glands, sweat glands, blood vessels, etc. Last but not least, nerve-immune–tissue interactions are also crucially important in inflammatory diseases and injured skin with a neurogenic component, such as alopecia areata, atopic dermatitis, or wound healing, diseases which are also frequently mentioned in the context of stress (114).
Surprisingly, the hair growth therapeuticum minoxidil (10–13) (Fig. 3) is able to abrogate stress effects in the HF. This observation may explain why minoxidil is only effective in a subgroup of patients receiving treatment for androgenetic alopecia or telogen effluvium. These patients may benefit from the reduction of stress effects on the HF, a hypothesis that deserves further exploration in clinical studies and in other inflammatory skin diseases. By way of analogy, this also invites the exploration of stress-reducing therapeutica and interventions such as progressive muscle relaxation in their capacity to alter neuroimmune interaction in skin and downregulate inflammatory processes.
Whether stress regulates peptidase activity in skin, and thereby contributes to the increased SP activity upon stress also remains to be elucidated, as does the response of the local HPA axis equivalent to the systemic stress response. Future hair research will exploit the many mouse mutants with defects in neuropeptide- or neurotrophin-signaling, PAR receptors, and HPA axis components. Enzymes represent another group of potentially fruitful targets to modulate neurogenic signaling. Important enzymatic pathways will include pathways that terminate the neurogenic stress response. Further exploration of NEP for example can clarify the hierarchy of HF damaging stress response in the search for optimal intervention strategies (40,41,62,115–117)2. Other enzymatic systems include the prohormone convertases and furins, important in neurotrophin and ACTH processing (92,97,118)3.
The analysis of second messenger systems up- and downstream of neuronal signaling will be another important step in the future analysis of cutaneous and HF stress responses. This line of research will enable us to specifically suppress the HF stress response with minimal side effects, for example by the suppression of SP-induced mast cell degranulation. New molecular targets for pharmacological intervention will be determined with great accuracy, thus reducing our dependency on serendipitous clinical observations, similar to those that have occurred under minoxidil treatment.
With regard to other skin appendages, it is quite intriguing that NGF is expressed by sebocytes (119) although sebaceous glands are not innervated. During inflammatory processes involving the sebaceous gland such as acne lesions, the NGF expression is increased, and SP+ nerve fibers can be observed close to the sebaceous gland (119,120). This aberrant innervation of the sebaceous gland in acne lesions by SP+ nerve fibers (121) suggests a role for stress-induced neurogenic inflammation for example in acne lesions, which certainly invites further exploration.
In collecting the evidence accumulated so far on the modern stress mediators SP and NGF (12,13,21,24,32,50), we of course ignored the many indicators that hint at an important role for the autonomic cutaneous innervation in the control of a local stress response. The sympathetic stress response is well characterized, and we already know that cholinergic and noradrenergic innervation in the skin is modulated during hair morphogenesis and cycle (20,24,35). Likewise, the contacts between cholinergic and adrenergic nerve fibers and mast cells fluctuate during the hair cycle (24), and many clinical observations suggest their role in hair growth control (1–4).
Finally, neuroimmune interaction, as it takes place in the control of hair growth, may serve as an indicator of malfunctioning neuroimmune communication in general. The HF provides us with highly complex, well-defined and easily accessible material for immunohistological and molecularbiological investigation. Stress-induced hair growth inhibition can therefore serve as a highly instructive model for exploring the ‘brain-skin connection’. From this point of view, it offers a unique model system for dissecting general principles of skin neuroendocrinology and skin neuroimmunology well beyond the HF.