There is a rapidly growing interest in the characterisation of the human intestinal microbiome, the composition of which remains largely uncharacterised (1). Associated with this has risen the knowledge and conviction that, in addition to local intestinal effects, there is a continually functioning interaction between the gut, its microbiome and the brain (‘gut–brain axis’) (2,3). For example, visual cues of food as well as food intake interact with this axis at the level of the hypothalamus, while gut peptides, in turn, are involved in the regulation of central processes such as satiety and feeding behaviour. Gut peptides also regulate numerous other processes including glucose regulation, insulin secretion and insulin sensitivity (2,3). In addition, gut peptides also impact on distant tissue functions, including possibly bone metabolism (3,4). Metabolic abnormalities and gastrointestinal infection impact on this axis, as shown in patients with type 2 diabetes (5) or in mice with chronic Helicobacter pylori infection (6).
Thus, the ‘gut–brain axis’ is now thought to contribute to the regulation of a multitude of metabolic, immune, endocrine and even nervous system processes, whose full scope is only slowly emerging (2,3,5,7). Also, the ‘gut–brain axis’ has increasingly become a target for therapeutic intervention in gastrointestinal diseases, obesity and metabolic syndrome (8,9). Furthermore, immense interest has recently been raised by the possible use of ingested commensal organisms for treating inflammatory and allergic conditions (10,11). While the potential clinical benefit of such probiotic strategies is generally acknowledged, its actual efficacy in clinical practice, the profound strain-specific effects of, e.g. different Lactobacilli in ameliorating inflammation and pain (12), and the underlying mechanisms of action through which these effects may be exerted in vivo all remain hotly debated (13,14).
Although clinicians have long felt this to be evident, it has only relatively recently become clear that, besides the ‘gut–brain axis’, there is also a ‘brain–skin axis’, whose molecular key players are increasingly understood (15–18). Interestingly, a number of gastrointestinal peptides (e.g., calcitonin gene-related peptide (CGRP), vasoactive intestinal peptide) have also been detected in intra-cutaneous nerve fibres (19).
A great deal of understanding mediators involved in the brain–skin axis came from studies in mice, where perceived stress-induced major neurogenic inflammation in mammalian skin (20,21) and even alters epidermal barrier function (17). For example, sound stress causes substantial substance P-, nerve growth factor- and mast cell-dependent neurogenic inflammation in mouse skin and even impacts on neuropeptide production by dorsal root ganglia sensory neurons (19,21,22). Interestingly, formation of functional associations between nerve fibres and mast cells appear to be very similar in the skin and gut (23,24). In the skin, perceived stress, at least in part via neurogenic inflammation, subsequently inhibits hair growth by premature induction of hair follicle (HF) regression (catagen), reduction in HF keratinocyte proliferation and stimulation of HF keratinocyte apoptosis, e.g. via activation of mast cells, macrophages and dendritic cells (20,23).
Here, we introduce our viewpoint that the ‘gut–brain axis’ and this more recently documented ‘gut–skin axis’ are intimately linked with each other. The large number of shared signals and cellular protagonists, the intricate innervation of both skin and gut, and the prominence of neurogenic inflammation in several gastrointestinal and dermatological diseases all make this an intuitively persuasive hypothesis (25,26).
However, the evident challenge is to provide initial experimental evidence in support of such a ‘gut–skin axis’ so as to encourage further exploration of this concept.
As a first step towards testing the hypothesis of a ‘gut–skin axis’, we have now probed whether probiotic strategies can affect neurogenic skin inflammation and its negative hair growth sequelae in the skin of stressed mice. This approach was encouraged not only by the rapidly growing literature which reports potent anti-inflammatory effects of probiotic treatment (10–14) but also by the recent report that probiotic pretreatment of rats exposed to repeated water avoidance stress completely abrogated stress-induced bacterial adhesion to the intestinal mucosa and translocation of bacteria to mesenteric lymph nodes (27).
On this background, the hair cycle was synchronised, and active hair growth (anagen) was induced in the back skin of 8- to 10-week-old female C57BL/6 mice by depilation, and mice were fed 109 colony-forming units of Lactobacillus reuteri by gavage (28) for 9 days, starting on day 6 after depilation [i.e. when all hair follicles (HFs) had fully entered anagen]. Control mice were gavaged with broth only. On day 13, half of the mice of the respective groups were exposed to sound stress for 24 h (280 Htz, 75 dB, emitted four times per minute as a 1-s-lasting tone from a device placed in the cage). This stress challenge likely includes an interference with the opportunity to sleep, as an increased food intake could be observed in mice under stress challenge. Skin was harvested for histomorphometric and immunohistological analyses on day 16 after depilation (20,23).
As shown in Fig. 1a, treatment with Lactobacillus significantly counteracted the stress-induced premature termination of active hair growth and HF regression (catagen). Strikingly, the percentage of catagen HF in probiotic-treated and stress-challenged mice [i.e. mice with a high hair cycle score (HCS)] was significantly reduced. This corresponded well to reduced levels of stress-induced apoptosis in the highly proliferative hair matrix epithelium of Lactobacillus-treated, stressed mice, compared to stressed mice without probiotic treatment (Fig. 1b). Stress-induced neurogenic skin inflammation also was greatly attenuated by probiotic treatment, reaching the levels seen in sham-stressed control mice. The number of HF that showed a dense perifollicular cluster of MHC class II+ inflammatory cells, which largely represent activated macrophages and dendritic cells (22), was significantly reduced by feeding Lactobacillus (Fig. 1c).
More than a century ago, the later Nobel laureate Metchnikoff proposed that longevity results from a good balance of the intestinal microbiota (29). But not until recently, the link between microbiota composition and health or disease has undergone systematic analysis. Five major microbiota sites, the nasal and oral cavities, the gastrointestinal and urogenital tracts as well as the skin, have undergone scientific scrutiny, whereby the majority of research focuses on the gut (Fig. 2). Nowadays, it is increasingly recognised that an in-depth understanding of the microbiota holds the potential to tailor the sites where microbiota reside to bias towards maintenance of allostasis (11,13). Hereby, excessive inflammation and the onset of diseases could be prevented through tempering the susceptibilities to infections, maintaining immune tolerance, facilitating bioavailability of nutrients and protecting tissue barrier functions. To date, the deployment of antibiotics, probiotics, and prebiotics has been shown to affect the composition and functions of microbiota (2,11,13). The intervention strategies utilising probiotics largely yielded insights on the ability to modify the gastrointestinal microbiota, e.g. by substituting absent microbial components with advantageous potential for the host, hereby ameliorating diseases of the gastrointestinal tract such as colitis, inflammatory bowel disease or irritable bowel syndrome. However, it should be noted that the effect of probiotics may not necessarily be beneficial and can also be ineffective in suppressing inflammation, e.g. in the context of atopic dermatitis (30).
In the present study, we report that the application of probiotics ameliorated stress-induced neurogenic skin inflammation and hair growth retardation in a mouse model. This supports the hypothesis that certain probiotics can modulate the immune system not only at local levels, the gut, but also systemically. In turn, immune defense mechanisms are improved, and inappropriate immune reactions are ameliorated, e.g. in the skin. This viewpoint is strengthened by recent observations revealing that the probiotic bacterial strain Lactobacillus johnsonii accelerates the recovery of skin immune homoeostasis after UV-induced immunosuppression (31). Further, human clinical studies suggested that a probiotic lysate was beneficial for the skin not only through the gastrointestinal route but also when applied topically. This non-replicating bacteria (Bifidobacterium longum sp. extract) improved sensitive skin and increased skin resistance against physical and chemical aggression when applied to the skin as a topical cream containing the active extract (32). This clinical finding could be supported by observations arising from the use of an ex vivo human skin explant model. Here, the probiotic lysate significantly inhibited capsaicin-induced CGRP release by neurones and improved signs of inflammation, such as a decrease in vasodilation, oedema, mast cell degranulation and tumor necrosis factor (TNF) α release. (32). Recent elegant findings uncovering a topographical and temporal diversity of the human skin microbiome gave a new twist to the effect of topical probiotic lysate administration, as the skin microbiome was thoroughly characterised at certain skin sites. Interestingly, physiologically comparable sites have been shown to harbour similar bacterial communities, and the site determines the complexity and stability of the respective microbial community. Future studies should venture the role of these bacterial communities in maintaining healthy skin (33).
We now report an amelioration of the skin homoeostasis by oral application of probiotics in a model of stress-induced skin inflammation. This tempted us to speculate the existence of a gut–brain–skin axis through which the alteration of the gastrointestinal microbiota would work towards skin homoeostasis. This concept is supported by some evidence published some decades ago revealing that peptide-containing cells in skin, brain and gut are linked by a common embryonic origin (34). Also, the comorbidity of inflammatory bowel and skin diseases strengthens the link between the gut and skin (35,36). Further, epidemiological studies have provided evidence for a link between altered intestinal microbiota to skin diseases, such as atopic eczema (37). Besides such, possibly embryonically wired, connection between the gut and skin, another scenario through which probiotocis may be operational in maintaining skin allostasis in the context of stress challenge can be envisioned: modifications of microbial components, which is referred to as dysbiosis, may be triggered by stress challenge, subsequently resulting in a generalised inflammation. Here, increased epithelial permeability may be an important cue in the development of such inflammation, as it may trigger T-cell activation and break tolerance mediated by immunosuppressive cytokines and T regulatory (Treg) cells. Pro-inflammatory cytokines then further perpetuate epithelial permeability, setting up a vicious cycle of chronic systemic inflammation, which also affects skin allostasis.
The deployment of probiotics may temper such systemic inflammation (38–40). Such probiotic-mediated dampening in inflammation has largely been described for gut-related inflammation (41,42). However, it can be postulated that the induction of Treg cells by probiotic microorganisms, which are decisive in limiting inflammation and disease, may not restricted to the gut but also affects the skin and other organs, such as the lung (28).
We here propose that the ‘brain–skin’ paradigm may be amended and rephrased as the ‘gut–brain–skin axis’, and future research should aim at identifying if and how epithelial permeability is affected by stress exposure and how stress-induced systemic inflammation and down-regulation of inflammation-dampening pathways, i.e. Treg cells, is communicated to the skin.
In conclusion, we are introducing a viewpoint, encouraged by our pilot observations and resting upon published prior evidence from the literature, which amalgamates previously proposed partial concepts into a new, unifying model, i.e. the gut–brain–skin axis. This concept has never been explicitly proposed before and proposes that modulation of the microbiota by deployment of probiotics can not only greatly reduce stress-induced neurogenic skin inflammation but even affect a very complex cutaneous phenomenon of (mini-) organ transformation, i.e. HF cycling. These observations raise the fascinating prospect that feeding of just the right kind of bacteria can exert profound beneficial effects on skin homoeostasis, skin inflammation, hair growth and peripheral tissue responses to perceived stress, at least if you are a mouse.