Baby’s skin bacteria: first impressions are long-lasting

Abstract

Early life is a dynamic period for skin microbial colonization and immune development. We postulate that microbial exposures in this period durably alter the skin immune trajectory and later disease susceptibility. Bacteria contribute to infant skin immune imprinting via interactions with microbes as well as with cutaneous epithelial and immune cells. Excellent research is underway at the skin microbiome–immune interface, both in deciphering basic mechanisms and implementing their therapeutic applications. As emphasized herein, focusing on the unique opportunities and challenges presented by microbial immune modulation in early life will be important. In our view, only through dedicated study of skin–microbe crosstalk in this developmental window can we elucidate the molecular underpinnings of pivotal events that contribute to sustained host–microbe symbiosis.

Host–microbe interactions in neonatal skin: a unique opportunity for discovery awaits

Our understanding of the mammalian skin microbiome (see Glossary), the cutaneous immune system, and the interactions between these two entities is based largely on studies of adult humans and animals. However, the composition and function of skin microbes and immune cells differ substantially in neonates versus adults (Box 1). In addition, the neonatal period has been shown to be a particularly formative window wherein host–microbe interactions can have life-long impact. We argue that only through separate, dedicated study of skin–microbe crosstalk during early development can we understand the biology of pivotal events that sustain cutaneous host–microbe symbiosis and durably shape skin immune function. Furthermore, by understanding factors that disrupt early-life microbial tuning of skin immune function and their consequences, we may identify ways to better support and restore optimal crosstalk so as to proactively correct the trajectory and maintain or restore skin immune health.

Box 1. Age-dependent changes in the skin microbiome and immune system.

Skin microbiome

The composition of the skin microbiota is distinctly different in human infants and children compared to adults. Birth mode influences the bacterial genera found on the skin in the first hours to days of life. Compared to cesarean delivery, vaginal birth is associated with a greater predominance of maternal vaginal versus skin-derived strains [84,85]. These early differences are largely undetectable after several weeks, by which time the human infant skin microbiota demonstrates a relative predominance of Firmicutes, including Staphylococci and Streptococci [86–88]. During late puberty, increasing skin sebum production and other hormonal changes facilitate another shift towards a more Actinobacteria-dominated skin microbial community [89]. Few metagenomic studies have been performed on infant skin, limiting our insight into the functional capacity of pioneer species and whether or how this differs from those that dominate cutaneous niches in later life.

Skin immune system

Early life is an equally dynamic period for the skin immune system. However, our current understanding of how the composition and function of skin immune cells evolves during infancy and childhood relies heavily on murine studies. Direct extrapolation from mice to humans is fundamentally flawed given the presence of species-specific cell types [90,91] and profound differences in timing of thymic [92–95], skin barrier [96,97], and hair follicle [98,99] development. However, two general principles derived from mice appear to be applicable to humans. First, there is a layered transition during which key tissue immune functions initially performed by innate-type cells are eventually superseded by an expanding population of tissue-resident memory αβ T cells [100]. Populations enriched early in life include commensal-responsive cell types such as mucosal invariant T (MAIT) cells [74] and other PLZF⁺ lymphocytes [101]. Although the timing of this transition and the specific cell types involved are better defined in mice [71], analogous principles are thought to apply in humans, although many remain to be fully demonstrated [19,79,80,102]. Second, studies on human tissues have consistently revealed an increased early-life propensity for immune regulation and immune tolerance [103,104]. This may extend to human skin, based on the observations that Tregs appear in skin during the second trimester of human fetal life [78] and are enriched in pediatric versus adult skin [105], as well as the finding that skin dendritic cells from human fetal versus adult skin have an increased capacity to promote Tregs while limiting effector T cell proliferation [81].

Early-life interactions with our commensal microbes influence the development of the immune system and have life-long implications for skin health and disease [1,2]. The ability of microbes to tune immune function is not merely transient. They can have enduring effects that are often referred to as immune imprinting (Figure 1). Although interactions with bacteria in adult mouse skin have demonstrated lasting changes in immune composition and function [3–5], early-life interactions can result in unique immune outputs [6]. An often-cited example of this phenomenon is the ‘hygiene hypothesis’, wherein children that grow up exposed to antigens abundant in farmlands with animals are less susceptible in the long-term to atopic conditions such as asthma versus those from urban areas [7–9]. Immune imprinting encompasses concepts such as tissue memory, which can refer to epigenetic changes in epithelial or stromal cells that alter their behavior [10], or the idea that cumulative antigen exposure establishes a population of tissue-resident memory lymphocytes [11] that will change the coordinated tissue immune response to subsequent antigen encounter. We use immune imprinting here as a term that can also extend to other mechanisms such as alterations in bacterial community composition or cytokine-mediated recruitment of polyclonal or innate-type lymphocyte populations that, when occurring in a time-limited neonatal window, have the potential for unique and enduring effects.

Figure 1. Model of microbial immune imprinting.

(A) Microbial exposure via barrier tissues such as the skin, gut, and lung in the early developmental window can imprint healthy immune function through to adulthood [6]. (B) Microbial perturbations such as antibiotic exposure in the early developmental window have been associated with chronic inflammatory conditions that affect barrier tissues, such as food allergies, asthma, and psoriasis [14–17].

Growing evidence in humans and mice suggests that unique molecular interactions between commensal microbes and cells of the neonatal immune system have the potential to tip the balance between health and disease in adult life. There are several such examples where bacteria in the neonatal gut shape immune function in a way that regulates later susceptibility to inflammatory disease. For instance, a recent study in human newborns used longitudinal systems-immunology analyses and metagenomic profiling of the infant gut microbiome to uncover a negative correlation between the prevalence of specific gut bacteria, Bifidobacteria, and markers of systemic or intestinal inflammation, for example increased blood neutrophils and concentrations of tumor necrosis factor α, interleukin (IL)-17A, IL-1α, and IL-13 [12]. Bifidobacteria are known to metabolize human breast milk oligosaccharides to generate organic compounds with immunomodulatory properties. In a small cohort of infants, Bifidobacterium infantis supplementation was sufficient to reduce systemic inflammatory markers. Moreover, fecal water from these infants skewed the in vitro differentiation of helper T cells towards a type 1 T helper (Th1) cell rather than a type 2 (Th2) or type 17 (Th17) cellular fate [12]; this in turn showed the potential impact of these bacteria on influencing future infant susceptibility to allergic or other immune-mediated diseases. Similarly, bacteria in the developing lung and intestine of mice have been shown to limit the accumulation of invariant natural killer T (iNKT) cells, thereby reducing subsequent susceptibility to oxazolone-induced colitis or ovalbumin-driven asthma [13]. Other studies have reported that early, antibiotic-mediated perturbation of the developing microbial diversity can be associated with a significantly altered later-life risk of developing inflammatory conditions such as food allergies [14,15], asthma [16], or psoriasis [17]. Although there are emerging examples of how in utero exposure to microbes or their products might tune the function of developing human intestinal immune cells [18,19], we focus here on the potential role of postnatal microbial exposures.

Inclusive of hair follicles, skin represents arguably the largest body surface that interfaces with microbes [20]. Our understanding of early microbial imprinting at this large external body site still lags behind that for the gut. However, it is a field ripe with opportunity to generate meaningful insights into the bacterial or host molecules and pathways that could be targeted for early correction of the microbe–skin–immune trajectory, saving years of disease and morbidity down the road. We are likely to see accelerated work in this area in coming years owing to renewed interest in neonatal immune development and tissue immunology [21,22], as well as the increasing availability of high-throughput system-level tools to study immune cells and bacterial communities [23,24]. Recent work has already begun to delineate the mechanistic categories by which skin bacteria can shape the composition and function of the developing cutaneous immune system. These include: bacteria–bacteria interactions, bacteria–epithelial cell interactions, and bacteria–immune cell interactions (Figure 2). We address each in turn, summarizing recent studies and highlighting areas for continued mechanistic investigation.

Figure 2. Potential mechanisms of bacterial immune imprinting.

Although much additional work will be necessary to elucidate how skin bacteria can mediate immune imprinting, the current literature would suggest the possibility of at least three mechanistic categories: (i) bacteria–bacteria interactions, (ii) bacteria–epithelial interactions, and (iii) bacteria–immune cell interactions. As discussed herein, bacteria–bacteria interactions could influence the composition and function of the neonatal skin microbiome, for example via bacteria-derived antimicrobial peptides (AMPs) [32,33] or inter-species inhibition of quorum-sensing pathways, as has been shown in vitro and in mice [35]. Bacterial interactions with human or murine keratinocytes could alter epithelial production of cytokines, thereby influencing the recruitment of skin immune cells, as seen in mice [57,61]. This might also occur by conferring epigenetically mediated inflammatory memory in murine or human epithelial stem cells (EpSCs) that could influence downstream wound healing [10]. Bacterial molecules could, as demonstrated in murine systems, interact with microbe-associated molecular patterns (MAMPs) on myeloid cells or lymphocytes [76,77,111], thereby changing their behavior or directly expanding bacteria-specific CD8⁺ T cells [3,4], mucosa-associated invariant T cells (MAIT) [71,75], and regulatory T cells (Tregs) [58].

Bacteria–bacteria interactions

Independently of their interaction with host cells, bacteria impact the skin environment via mechanisms that favor the outgrowth and niche establishment of particular bacteria over others. This has been well described in the context of the neonatal gut microbiome in which the early presence of so-called founder species helps to cultivate conditions that enable new strains to enter the niche and establish long-term residence [25,26]. This microbial succession is guided via various bacterially mediated processes, including the generation of metabolic byproducts, alteration of the pH or oxygen availability, and the production of small molecules that can influence the growth or behavior of other bacterial species. Intra-individual longitudinal tracking of skin bacteria has been studied in healthy human adults [27–30] and preliminarily in some disease states such as atopic dermatitis [31]. However, many early findings still require robust validation, and a more granular understanding will be necessary to determine which are the key founder species on infant skin and how they might support the evolution of the cutaneous microbiome over the first years of life.

Interactions among skin bacterial species have been studied mostly in the context of atopic dermatitis, in which particular strains can limit the growth or pathogenicity of the atopic dermatitis-associated pathobiont, Staphylococcus aureus. Coagulase-negative Staphylococci in particular, as well as other genera, have been shown to produce antimicrobial peptides such as lantibiotics and oleic acids that limit the growth of S. aureus [32,33]. Many of the same strains can produce small molecules such as short-chain fatty acids that inhibit pathogenicity by limiting biofilm-forming capacity [34] or by autoinducing peptides that limit the production of quorum-sensing-dependent toxins [35]. Although the field is already in the process of exploring the use of particular commensal strains as potential topical therapeutics to treat atopic dermatitis flares by reducing the cutaneous burden of S. aureus [36], it is intriguing to consider how similar interactions between bacterial species might play out even earlier – namely in neonates – before atopic dermatitis and its accompanying skin inflammation appear. Both sequencing- and cultivation-based studies of the skin microbiome in infants at risk for atopic dermatitis have documented an altered skin microbiome composition before disease onset, characterized by reduced Staphylococcus epidermidis prevalence or heightened S. aureus prevalence – especially S. aureus strains with a functional Agr quorum-sensing system [37–39]. Presumably, inherited host factors such as compromised barrier function and exposure of epidermal molecules for bacterial adhesion can augment the propensity for S. aureus skin colonization [40–42], but failure to establish a ‘protective’ microbiome might also be an important factor that allows this pathobiont to establish a skin niche.

Ultimately, the successful potential use of live bacteria or bacterially derived therapies to promote skin health or to treat particular skin diseases would rely on being able to safely introduce key species as stable members of the skin microbial community. Thus, it is important to seek a holistic understanding of the direct and indirect effects of such candidate therapies on skin microbial community composition. Preliminary work and clinical trials have shown promise for the use of these strains in the treatment of atopic dermatitis [36,43], and this is garnering excitement for further validation. Continuing this push in parallel with studies that probe interaction networks among skin commensals – especially for those bacteria that occupy the niche in infants – can help to clarify inter-species bacteria interactions that contribute to shaping the development and long-term maintenance of the skin microbial community composition. Such efforts can lay the foundation for opportunities to intervene in the host–microbe relationship to not only treat but perhaps even prevent disease.

Bacteria–epithelial cell interactions

Although bacteria and their byproducts can be found below the human skin surface [44], their density is greatest in the epidermis, both between and within hair follicles [45]. This results in a myriad of opportunities for bacteria–keratinocyte interactions that help to modulate the core physical and antimicrobial barrier functions of the epidermis. Studies using murine models and primary human keratinocytes have demonstrated that commensal activation of the aryl hydrocarbon receptor pathway in keratinocytes augments their expression of key skin differentiation genes to limit transepidermal water loss in the skin tape-stripping mouse model [46]. Conversely, proteases such as EcpA and staphopains expressed by S. epidermidis and S. aureus species, respectively, can contribute to epidermal barrier breakdown in murine and human skin under some pathogenic circumstances [47]. In addition to these effects on the physical properties of the epidermal barrier, skin bacteria can also augment the human and murine skin antimicrobial barrier and protection against pathogen assault by bolstering keratinocyte expression of antimicrobial peptides [48–50]. Collectively, these studies demonstrate the ability of commensal microbes to interact with epithelial cells and directly influence skin barrier health.

Keratinocytes, however, are also instrumental in coordinating cutaneous immunity. Under homeostatic conditions these cells – especially those in the hair follicle – produce chemokines and cytokines, such as IL-7, IL-15, CCL2, CCL8, or CCL20, that influence the recruitment and behavior of mouse innate lymphoid cells, dendritic cells (DCs), and T cells [51–54]. Additional murine studies have illustrated that skin bacterial dysbiosis associated with epidermal deletion of metalloproteinases ADAM10 or ADAM17 can modulate keratinocyte cytokine production, augment immune cell recruitment, and hasten skin pathology [55,56]. These interactions remain comparatively understudied in the neonatal period, but work in mice suggests that commensal-augmented production of the chemokine CCL20 in developing hair follicle keratinocytes can facilitate the recruitment of regulatory T cells (Tregs) into skin [57]. These Tregs support long-lasting immune tolerance to bacteria and other antigens, and this is a pivotal step in promoting healthy skin development [58]. Nevertheless, the extent to which keratinocyte-derived signals might coordinate the early recruitment or function of other cutaneous immune cell types remains an outstanding question with potential implications for neonatal immune imprinting.

From another angle, skin bacteria have the potential to influence essential skin functions via their effects on epithelial stem cells (EpSCs). EpSCs and the cytokines IL-1α and IL-1β are crucially involved in skin wound healing and hair growth [59]. S. aureus and coagulase-negative Staphylococci can elicit production of IL-1α and IL-1β [3,60], and have been shown to accelerate wound healing and post-wounding hair regrowth in mice via activation of the IL-1 receptor (IL-1R) on keratinocytes [61]. To further demonstrate the importance of bacterial signals in wound healing, the same investigators performed a small human cohort study that demonstrated delayed closure of skin biopsy wound sites treated with topical antibiotic ointment versus placebo [61]. Separate studies in adult mice showed that epidermal inflammation resulting from topical exposure to imiquimod, calcipotriol, physical abrasion, or Candida albicans led to persistent chromatin changes in loci within EpSCs, subsequently augmenting keratinocyte migration and wound closure in a model of biopsy-induced skin wounding [10]. This putative EpSC ‘inflammatory memory’ might represent an epigenetic [62] form of immune imprinting, as shown by recent work looking at the intestinal epithelium. Specifically, in a mouse model of maternal Yersinia sp. infection, increased concentrations of IL-6 induced epigenetic changes in fetal intestinal EpSCs, as determined from assays of transposase-accessible chromatin with high-throughput sequencing (ATAC-seq), relative to uninfected mice [63]. This study demonstrated persistent heritable differences in the transcriptional and epigenetic profiles of intestinal EpSCs in the mouse offspring, as well as heightened Th17-associated inflammation in the intestinal lamina propria of the mice relative to controls [63]. The data raise the intriguing possibility that microbially mediated activation of innate cytokines in neonatal skin might induce persistent epigenetic changes in EpSCs, with long-term implications for tissue immune tone. However, robust future investigations are warranted to examine if such hypotheses are true.

Bacteria–immune cell interactions

As discussed above, there are many ways that skin bacteria can shape skin immune function without interacting directly with immune cells. However, there is also significant evidence to suggest that early bacteria–immune cell communication might affect imprinting of cutaneous immune function.

Myeloid cells

Antigen-presenting cells (APCs), especially DCs, have a sentinel function in sensing and integrating signals from skin bacteria. These cells are densely decorated with an array of pattern recognition receptors (PRRs) that respond to microbe-associated molecular patterns (MAMPs) [64,65] and can be influenced by microbially derived metabolites [66]. Skin APCs actively phagocytose bacteria, even extending their dendrites across the intact skin epithelium, as has been shown for Langerhans cells [67]. Although much remains to be untangled regarding how specific skin APC subsets help to coordinate responses to skin bacteria, Langerhans cells, as well as type 1 and type 2 conventional DCs (cDC1, cDC2), have risen to the forefront as central quarterbacks in detecting and responding to cutaneous microbes [3,60,68]. However, homeostatic interactions between skin bacteria and non-DC subsets such as monocytes and macrophages remain largely understudied. Nevertheless, recent work has reported that early-life microbial exposure can influence the numbers of intestinal macrophages, thereby modulating colonic iNKT seeding [69]; indeed, this suggests another area of research focus for further understanding the microbial influences on cutaneous myeloid cell functions.

Lymphocytes

Lymphocytes, especially T cells, are central players in skin health and disease [70]. Lymphocytes that recognize and respond to commensal bacteria may be particularly important for skin homeostasis; For instance, these T cells have been demonstrated to exhibit enhanced Gata3 expression – a presumed reparative transcriptional signature – and to accelerate wound healing in mice [4,5,71]. There are now several examples of how skin bacteria-derived molecules can directly influence skin lymphocyte numbers and functions. This has been studied for skin CD8⁺ T cells in adult mice and non-human primates; in these models, colonization by specific strains of S. epidermidis, which produce N-formyl methionine-containing (fMet) peptides, expand fMet-specific, IL-17A-producing CD8⁺ T cells via non-classical antigen presentation [4]. Likewise, mycolic acids produced by Corynebacterium spp. have been shown to expand the numbers of murine Vγ4⁺ dermal γδ T cells and to augment their IL-17A production – an effect that was attenuated in IL-23-deficient (Il23r^−/−) mice or upon colonization with a Corynebacterium accolens mutant lacking the gene for mycolic acid synthetase [72]. These studies highlight how molecules produced by commensal microbes can substantially influence the composition and function of the cutaneous lymphocyte compartment.

The introduction of skin commensal bacteria in neonatal life, however, can elicit distinct immune effects that cannot be recapitulated by adult colonization. Two key examples include the preferential early-life expansion of cutaneous mucosa-associated invariant T (MAIT) cells and antigen-specific Tregs by S. epidermidis, as shown in mice (Figure 3). MAIT cells are innate lymphocytes that are enriched in barrier tissues, such as the gut and skin, and express semi-invariant αβ T cell receptors that recognize microbial products [73], including riboflavin metabolites produced by S. epidermidis and other bacteria [74]. Colonization of neonatal but not adult mice with riboflavin-producing bacteria can induce thymic accumulation of this metabolite and increase generation of MAIT cells, which then populate peripheral tissues, including the skin [75]. Subsequent S. epidermidis re-exposure in adult mice can further activate and expand these cutaneous MAIT cells [71]. Similarly, colonizing murine skin with S. epidermidis in the first 2 weeks of life leads to substantial enrichment of Tregs among the antigen-specific CD4⁺ T cell population. By contrast, analogous colonization of adult mice primarily generates antigen-specific proinflammatory effector CD4⁺ T cells [58]. Furthermore, the enduring effects of ‘missing’ this microbial imprinting in early life has been reported in relation to both MAITs and Tregs. For example, delaying the exposure of murine skin to MAIT-promoting S. epidermidis bacteria until adulthood resulted in a failure to generate MAIT cells and a consequent loss of their reparative benefits, as evidenced by delayed closure of biopsy-inducing skin wounds [71]. Similarly, in a tape-stripping model of epidermal injury, adult murine skin exposed to S. epidermidis – without undergoing prior neonatal S. epidermidis colonization to induce antigen-specific Tregs – displayed significantly increased tissue pathology and skin neutrophil numbers relative to the skin of mice exposed to this bacterium as neonates [58]. In both of these examples, beneficial interactions with cutaneous bacteria occurred preferentially during a time-limited window of early life in mice, emphasizing the importance of efforts to identify and dissect analogous or additional mechanisms in human infant skin.

Figure 3. Bacterial colonization of neonatal skin can shape the composition and function of the cutaneous immune system.

(Left) Colonization of neonatal murine skin by the commensal Staphylococcus epidermidis is detected by cutaneous dendritic cells (DCs). These DCs migrate to the draining lymph nodes and present S. epidermidis antigens to T cells. Antigen-specific regulatory T cells (Tregs) then accumulate in neonatal murine skin where they drive tolerance to commensal bacteria and establish healthy skin homeostasis [58,60]. (Center) Neonatal colonization of mice by the pathogenic strain, Staphylococcus aureus, generates a robust effector T cell (Teff) and a weak Treg response, primarily due to the release of the bacterial α-toxin which drives IL-1β release. This mechanism facilitates an immunological distinction of pathogen from commensal, resulting in failure to establish tolerance to pathogens and leading to robust skin inflammation upon subsequent re-exposure [60]. (Right) Exposure of neonatal mice to riboflavin-synthesizing bacteria such as S. epidermidis drives the accumulation of this metabolite in the thymus, leading to increased output of mucosa-associated invariant T (MAIT) cells and their subsequent accumulation in the skin. MAIT cells contribute to homeostatic functions in murine skin, including augmented wound closure after biopsy-induced wounding [71,75].

Molecular tuning of commensal-specific responses and considerations for early life

Although a significant amount of work will be necessary to decipher the molecular language by which bacteria tune skin cells and the ensuing immune responses, recent work has begun to identify bacterial molecules and receptors that may play a role. For example, there is preliminary evidence that S. epidermidis-mediated expansion of IL-17A-producing CD8⁺ T cells in adult mice [4] is supported by the recognition of S. epidermidis cell-envelope components by Toll-like receptor (TLR) 2 and the C-type lectin receptor, Dectin-1, on epithelial cells, and potentially on myeloid cells [76]. Although TLR2 signaling has also been implicated in expansion of colonic Tregs following exposure to intestinal bacteria [77], it remains to be seen whether this same pathway plays a role in mediating the expansion of S. epidermidis-specific Tregs following neonatal skin colonization [58]. Of note, myeloid cells, in particular cDC2s, have been implicated in early-life generation of S. epidermidis-specific Tregs based on the fact that they are a dominant subset of DCs that capture bacteria in neonatal murine skin [60]. Whether S. epidermidis or other commensal bacteria provide specific tolerizing signals to these cDC2s to promote their Treg generation is unknown, but interactions with α-toxin-producing S. aureus strains in neonatal mice can induce cDC2 production of IL-1β, thereby limiting S. aureus-specific Tregs, unlike S. epidermidis Tregs [60]. In addition to bacteria-directed signals, several host factors should be considered in future efforts to mechanistically dissect the basis of differential responses to colonization of neonate versus adult skin. These include the distinct composition and function of immune cells in neonatal skin (Box 1), for example; the predominance of different lymphocyte subsets [78–80] and potentially the higher tolerogenic capacity of neonatal skin APCs, as observed in human fetal skin [81]. Similarly, age-dependent differences in the expression of PRR receptors [82] on epidermal or immune cells, or the propensity for cytokines to lead to epigenetic modifications in immune and/or skin cells [83] – together with a heritable altered function of these cell types – might support neonatally restricted imprinting of skin immune function.

Concluding remarks

It is an exciting time for the investigation of skin microbiome–immune interactions, both in terms of deciphering basic mechanisms and identifying potential therapeutic applications. As emphasized herein, it will be important to focus on the unique opportunities and challenges afforded by an understanding of early-life immune imprinting by skin microbes. We posit that this area holds great potential for learning how early-life interventions could possibly be used to prevent or mitigate the severity of inflammatory skin diseases such as atopic dermatitis. Although microbially directed therapy holds great potential for treating established skin diseases [36], we should simultaneously explore its application in the context of early ‘course-correction’ in at-risk infants [37]. This strategy has merit in that targeting the skin–microbe–immune axis might ideally occur in infancy during a ‘window of opportunity’, presumably before the stable establishment of microbial communities and immune cell populations. These approaches, however, will need to consider the possibility that inherited host factors that increase the risk of developing a particular skin disease might themselves alter early immune responses to skin bacteria. Thus, principles gleaned from the study of healthy infants and neonatal animals may not directly apply, and will need special consideration in each disease context.

Exciting work is expanding in the academic and biotech spheres, and identification of microbially directed therapeutics or prophylactics will benefit from a better understanding of the molecular underpinnings and the immune and microbial sequelae of potential interventions. This is especially important for the pediatric population, where well-intended interventions may have inadvertent and undesirable long-term consequences. Given the complexity of the unresolved issues (see Outstanding questions), a combined approach that incorporates complementary studies in both mice and humans will be important to move the field forward. Ultimately, the future is bright for baby’s skin bacteria – so enjoy bath time, but go easy on the soap.

Outstanding questions.

How does the number and function of human skin myeloid and lymphoid cells evolve between birth and adolescence?

What are the key bacterial founder species on infant skin, how do they support the establishment of the cutaneous microbiome over the first years of life, and how do they differ functionally from strains that supplant them in adult life?

To what extent do keratinocyte-derived signals coordinate the early recruitment or function of cutaneous immune cell types in developing human skin?

Do neonatal skin microbes modulate epithelial stem cell ‘inflammatory memory’ via epigenetic modifications or other mechanisms in a way that alters later skin reparative capacity?

What are the key myeloid cell populations in neonatal skin that sample microbial antigens and integrate their signals to support the development of tolerance and homeostasis in the presence of commensals?

Which innate-type and/or classical T cell populations in developing human skin are particularly dependent on microbial signals for their abundance and/or functions?

What are the key host receptors and microbial molecules that govern bacterially mediated imprinting of skin immune function in early life?

How do genetic host factors that mediate susceptibility to pediatric inflammatory diseases, for example filaggin mutations in atopic dermatitis, alter the quantity and quality of early-life interactions between skin microbes and immune cells?

Highlights.

The composition and function of skin bacteria and immune cells differ substantially in mammalian neonates versus adults.

Although much remains to be understood about early-life skin immunity in humans, this period is demarcated by an increased capacity for tolerogenic function and a layered transition from innate-type lymphocytes to classical memory populations.

The ability of microbes to durably tune immune composition or function is often referred to as ‘immune imprinting’ – this phenomenon is preferentially active early in life and, if missed, some immune education events cannot be reproduced later on.

Growing evidence suggests that unique molecular interactions between commensal microbes and cells of the neonatal immune system have the potential to tip the balance between skin health and disease in adult life.

Mechanistically, early life bacterial exposure can influence skin immune function via bacteria–bacteria interactions, bacteria–epithelial cell interactions, and/or bacteria–immune cell interactions.

Only through separate, dedicated study of skin–bacteria crosstalk in the early developmental window can we understand the biology of pivotal events that impact on host–microbe symbiosis and skin immune function.

The safe and successful use of live bacteria or bacteria-derived therapies to promote skin health or treat skin disease relies upon the ability to stably introduce key species and understand their direct and indirect effects on the composition of the skin microbial community.

The strategy of targeting skin microbe–immune interactions in early life has the advantage of ‘correcting course’ before the stable establishment of microbial communities and/or immune cell populations, thereby not only treating but perhaps preventing skin disease.

Clinician’s corner.

Neonatal skin conditions, such as erythema toxicum neonatorum, acne neonatorum, and neonatal cephalic pustulosis, may reflect early-life sequelae of normal skin immune–microbe interactions [86,106,107].

Seborrheic dermatitis and infantile acne in young children likely represent more pathologic interactions with skin bacteria and fungi [108,109].

Atopic dermatitis is the poster child disease for disrupted early life skin–microbe crosstalk [110]. Emerging evidence indicates that this breakdown occurs well before profound flares of the disease are apparent [37–39], suggesting that early interventions to ‘correct course’ might help to mitigate disease severity or even prevent its onset, although this remains to be rigorously tested.

Even in the absence of any visible skin inflammation, early subclinical interactions between skin microbes and the developing cutaneous immune system may help set the stage for lifelong skin health and disease susceptibility [58,60].

Glossary

Actinobacteria

bacteria that are commonly found on skin; this group includes genera such as Corynebacterium and Cutibacterium. The prevalence of these bacteria on sebaceous sites such as the face, chest, and back increases during late puberty

Atopic dermatitis

a common inflammatory, allergic skin disorder with onset often in early life and that is characterized by flares typified by itching and redness, often associated with increases in S. aureus

Biofilm

a community of microbes that grow on surfaces encased in a tough matrix which is not easily removable from the surface

Epidermis

the outermost layer of the skin that comprises several layers of keratinocytes which terminally differentiate to form the stratum corneum

Epigenetic

reversible changes in genetic information that are not encoded in the DNA sequence but are instead mediated by DNA methylation or histone modifications. These types of changes can occur in response to various environmental factors including microbes and alter the function of the cells in which they occur

Epithelial stem cells

these reside in the basal layer of the epidermis and the hair follicle bulge region; they help to maintain skin homeostasis and hair regeneration, and participate in repair of the epidermis after injury

Firmicutes

bacteria that are commonly found on the skin. They are especially dominant in early life and include genera such as Staphylococcus and Streptococcus

Founder species

species of microorganisms that establish a niche, early in life, at a given anatomical site in the body, which then influences the future composition of the microbiome at that site

Immune imprinting

the concept that specific immune interactions (e.g., with microbes) can durably alter the composition and/or function of the immune system in a way that alters subsequent immune responses. This has classically been intertwined with the molecular mechanism of epigenetic modification, but there are examples of imprinting that do not rely on this mechanism

Keratinocytes

epithelial cells of the skin that form the epidermis, including that portion which lines the hair follicles

Microbe-associated molecular patterns (MAMPs)

molecules that are conserved and expressed across several species of microbes and that are recognized by pattern recognition receptors

Microbiome

the collective microbial composition, including but not restricted to, bacteria, viruses, and fungi, in any given anatomical niche

Mucosa-associated invariant T (MAIT) cells

these are resident mainly in mucosal tissues, have a semi-invariant αβ T cell receptor, and are characterized by their innate-like properties

Pattern recognition receptors

proteins that recognize and bind to conserved molecular patterns typically expressed by microbes. These proteins generally initiate signal transduction pathways upon binding their cognate molecules

Quorum sensing

a method of communication between bacteria via secreted molecules that in turn govern gene expression and foster community behavior. In Staphylococcus aureus, the accessory gene regulator (Agr) system is a key regulator of quorum-sensing behavior and related toxin production

Regulatory T cells (T regs)

a subset of CD4⁺ T cells defined by the transcription factor FoxP3; they play a key role in suppressing inflammation and preventing autoimmunity

Staphylococcus aureus

Gram-positive skin bacteria of the Firmicutes phylum; these can asymptomatically colonize nares and skin, but are also a major cause of skin bacterial infections and are associated with flares of atopic dermatitis

Staphylococcus epidermidis

one of several coagulase-negative Staphylococcus species that are commonly found on human skin

Type 1 T helper (Th1) cells

a subset of CD4⁺ effector T cells that are traditionally studied for their role in mounting immune responses to fight infection by intracellular pathogens and viruses via the production of IFN-γ

Type 2 T helper (Th2) cells

a subset of CD4⁺ effector T cells that are traditionally studied for their role in mounting immune responses to fight infection by extracellular parasites such as helminths; they are known producers of IL-4, IL-5, and IL-13. Overt production of these cytokines has been associated with allergic immune responses

Type 17 T helper (Th17) cells

a subset of CD4⁺ effector T cells that are deemed to have both protective and pathogenic roles; they are known producers of IL-17A and have a protective role against cutaneous bacterial infections (e.g., S. aureus). Th17-mediated inflammation has also been linked to disease pathology in skin conditions such as psoriasis