Targeting Skin Barrier Function in Atopic Dermatitis

Abstract

Atopic dermatitis (AD) is the most common chronic inflammatory skin disease in the general population. Skin barrier dysfunction is the central abnormality leading to AD. The cause of skin barrier dysfunction is complex and rooted in genetic mutations, interactions between the immune pathway activation and epithelial cells, altered host defense mechanisms, as well as environmental influences that cause epithelial cell activation and release of alarmins (such as thymic stromal lymphopoietin) that can activate the type 2 immune pathway, including generation of interleukins 4 and 13, which induces defects in the skin barrier and increased allergic inflammation. These inflammatory pathways are further influenced by environmental factors including the microbiome (especially Staphylococcus aureus), air pollution, stress, and other factors. As such, AD is a syndrome involving multiple phenotypes, all of which have in common skin barrier dysfunction as a key contributing factor. Understanding mechanisms leading to skin barrier dysfunction in AD is pointing to the development of new topical and systemic treatments in AD that helps keep skin borders secure and effectively treat the disease.

BACKGROUND AND AIMS

Atopic dermatitis (AD) is the most common inflammatory skin disease in the general population. Atopic dermatitis is genetically transmitted and is often the first step in the so-called “atopic march,” which leads to AD-associated food allergy (FA), eosinophilic esophagitis, asthma, and allergic rhinitis.^1,2 Recent studies have demonstrated that skin barrier dysfunction in AD skin facilitates penetration of foods from the environment followed by immunoglobulin E food sensitization to the allergen and gut-skin communication that results in mast cell hyperplasia in the gut and predisposition to food-induced anaphylaxis.³

During the past few years, new techniques have been developed that measure skin barrier function.⁴ These include documentation of transepidermal water loss (TEWL) that may account for the characteristic dry skin. Other devices have been found to measure vascular changes and thickening of the skin, as well as skin tape sampling that allows us to do microscopy of the skin and transcriptome profiling and analysis of lipid and protein composition. These techniques allow us to make measurements at an unparalleled level and provide hope to introduce precision medicine in skin barrier repair of atopic skin to replace missing components in AD skin. In this review, we aim to describe the abnormalities in skin barrier function and how to best determine them for the monitoring of disease and the effect size of therapeutic interventions. With this increased understanding of these mechanisms and methodologies, we strive to improve the treatment and future prevention of AD.

Understanding skin barrier function in patients with AD

Keratinocytes: builders of the skin barrier.

The origin of skin barrier abnormalities in AD are multifaceted, fueled by epidermis-intrinsic defects, extrinsic triggers, or caused by inflammatory mediators secreted by skin resident and infiltrating immune cells.⁵ Either way, the resulting skin barrier defects should be alleviated for disease management. Restoring the normal turnover and subsequent terminal differentiation of keratinocytes would, in principle, be an effective intervention or serve as a much-needed add-on therapy for acquiring long-lasting remission and prevention of disease flares and could omit the need for lifelong emollient use by AD patients. To identify and explore therapeutic strategies for effective keratinocyte-focused AD treatment, we first explore normal keratinocyte biology and epidermal barrier function.

Keratinocytes are the main cell population of the epidermis and build the distinct epidermal layers (strata) in a tightly controlled stratification process. Originating in the stratum basale, while still connected to the basement membrane, interfollicular stem cells give rise to transit amplifying cells and fuel epidermal renewal and regeneration. After the birth of a new keratinocyte, its journey of differentiation and movement toward the skin surface begins. Cell fate decisions are orchestrated by transcription factors (eg, p63), ligand-receptor (eg, Notch, WNT), and mechanosensitive signaling networks (eg, YAP/TEAD) that control proliferation rate and the delamination process allowing basal cells to exit the cell cycle and detach from the basement membrane (Figure 1, A). Entering the stratum spinosum, keratinocytes switch the expression of their keratin (KRT) protein profile from KRT5 and KRT14 to a new subset of KRTs, 1 and 10, providing increased cellular stiffness and mechanical strength. The final stages of keratinocyte terminal differentiation encompass the controlled cell death, enucleation, and cornification taking place in the stratum granulosum. The microscopically visible granules in this layer provide the key components for the skin barrier function as herein pro-filaggrin accumulates (Figure 1, B). Pro-filaggrin is the backbone protein from which filaggrin (FLG) monomers are released in a multiple-step process that involves dephosphorylation and proteolytic cleavage and are driven by phase-separation dynamics and regulated through environmental cues.⁶ The latter finding is of particular interest given the association of AD disease activity with seasonal changes and environmental humidity.⁷ Free FLG is bound to KRT filaments forming strong macrofibrils that anchor late differentiation proteins into the cornified envelope by various cross-linking processes. Filaggrin belongs to the S100-like fused-type proteins, a group that also comprises paralogs, FLG-2, hornerin, repetin, and trichohyalin. During the final stages of keratinocytes terminal differentiation, FLG monomers are further degraded and deaminated into hygroscopic free amino acids forming the so-called “natural moisturizing factor” (NMF) that is deposited in the stratum corneum.⁸ The NMF captures water molecules in the stratum corneum favoring skin hydration and its levels inversely correlate to TEWL and skin pH (Figure 1, C).

FIGURE 1.

The multifaceted skin barrier. (A) Epidermal barrier maintenance and repair starts skin deep in the stratum basale as basal keratinocytes proliferate under cues of master regulating transcription factors (p63, NOTCH) and signal transduction pathways (WNT, YAP) and exit the cell cycle, becoming early differentiated keratinocytes in the stratum spinosum. (B) After transitioning from the stratum spinosum to the granulosum, the characteristic keratohyalin granules (KGs) become apparent, holding the pro-filaggrin (FLG) protein that will be proteolytically processed into FLG monomers and later further deaminated into free amino acids that form the NMF. (C) Production of NMF from FLG protein is essential for stratum corneum hydration, and NMF levels correlate to barrier function parameters like TEWL. (D) The formation of the acid mantle is part of the chemical barrier next to the secretion of AMPs and terminal differentiation proteinederived cationic intrinsically disordered antimicrobial peptides (CIDAMPs), including FLG, hornerin (HRNR), and late cornified envelope (LCE) proteins. (E) Upon terminal differentiation in the last living cell layer, cell organelles are degraded, leaving only tightly crosslinked cornified envelopes (CEs) embedded in lipid matrix that form the stratum corneum and provide the physical barrier that withholds passive permeation of allergens and pathogens. (F) The commensal microbiota of the skin microbiome provides an important microbial barrier against pathogens by competition for skin surface, nutrients, and production of bactericidins. The commensal microbial metabolites feed back to the host and tune skin barrier formation by activation of host receptors, like the AHR. The AHR is involved in almost all processes that relate to skin barrier formation and function.

Filaggrin in AD: a gene-host-microbe interactor.

The importance of FLG levels to maintain a healthy skin became apparent in 2006 when the McLean group⁹ reported on the homozygous loss-of-function mutations in the FLG gene being causative of ichthyosis vulgaris and heterozygous mutations posing the major genetic risk factor for AD. Both diseases are characterized by extremely dry skin and poor skin barrier function. In following years, skin barrier malfunction and stratum corneum malformation (eg, increased TEWL, stratum corneum dehydration, increased stratum corneum permeation, altered corneocyte topography, and fragility) have been linked to the levels of FLG in skin. Notwithstanding the central role for FLG in AD pathogenesis, other key epidermal skin barrier proteins (eg, hornerin, involucrin, loricrin) are also downregulated in AD skin, possibly related to genetic polymorphisms or driven by the T-helper (Th) cell cytokine milieu in AD skin that disturbs transcription of major epidermal barrier genes (and constituents of the cornified envelope).^10,11

Of particular interest for the development of therapeutic modalities to fortify the skin barrier is the recent discovery that multiple, previously considered structural components of the cornified envelope exert potent antimicrobial activity against a variety of skin commensal and pathogenic bacteria and fungi. Proteins from the late cornified envelope family and the S100 family, including FLG and hornerin, are reclassified as cationic intrinsically disordered antimicrobial peptides (CIDAMPs), a term first posed by Jens-Michaël Schröder.¹² Remarkably, these CIDAMPs are even more potent in standard in vitro antimicrobial assays than beta defensins (hBD2),¹³ and may thus confer significant strength to the antimicrobial chemical barrier of the skin (Figure 1, D). Typically, (CID)AMPs are cationic peptides excreting broad structure-independent killing of microbes through membrane disruption, among others. Their oxidized or reduction states appear to critically affect the antimicrobial potency, and many antimicrobial peptides (AMPs; e.g. S100s, LCEs) have high cysteine contents at the protein C-terminus increasing their stability against proteolysis, high temperature, and pH changes,¹⁴ which are important features for an AMP in the stratum corneum. The clustering of cysteine-residues, with pairs of histidine and aspartic acid residues allows for the binding of transition metals that are essential to various microorganisms.¹⁵ Next to these microbe-directed actions, most AMPs also target host receptors to orchestrate danger signals and amplify the host defense response.

Downregulation of CIDAMP expression in AD, either through genetic predisposition or the local inflammatory milieu, could be a hitherto-overlooked aspect that confers to the lack in skin microbiota diversity and colonization and infection of AD skin by the disease-modifying pathogen, Staphylococcus aureus (S. aureus). The frequent S. aureus colonization of lesional AD skin and bacterial infestation of AD skin was putatively assigned to the lack of conventional AMPs, like defensins, because Th2 cytokines interleukin (IL) 4 and IL-13, and also thymic stromal lymphopoietin (TSLP) are known to dampen keratinocyte-derived AMP expression in vitro.¹⁶ In FLG-deprived skin, the colonization by S. aureus may further be substantiated by a lower abundance of the commensal gram-positive anaerobic cocci that utilizes NMF in the stratum corneum as an energy source and induces a rapid antimicrobial host defense response in keratinocytes, potentially controlling the defense against S. aureus.¹⁷ In addition, altered corneocyte topography and corneodesmosin patterns upon FLG deficiency or low NMF levels appear to favor S. aureus binding to the stratum corneum.¹⁸ To make matters increasingly complex, lipidomics analysis recently identified AD-related ceramide changes to correlate to FLG disease genotype and the colonization of skin by S. aureus.¹⁹ The lipid composition (including sphingolipids, cholesterol, and ceramides) and organization in the stratum corneum of AD patients is strongly affected,²⁰ leading to increased stratum corneum permeability and deprived skin barrier function. Therefore, the abnormalities in the physical barrier function (Figure 1, E) appear strongly intertwined with the chemical (Figure 1, D) and commensal microbial barrier (Figure 1, F) through complex gene-host-microbe interactions that we have only just began to explore and understand.

Investigative functional studies aiming to discover druggable pathways to fortify the skin barrier should, therefore, also consider the potential gene-environment interactions in AD. Integration of multi-omics data with well documented patient records may help to identify patient phenotypes and disease endotypes (as aimed for in the pan-European, IMI-funded consortium BIOMAP^21,22) that can be functionally validated in relevant experimental diseases models to complete the journey from association to causality and foster drug target identification and screening for precision medicine.

The aryl hydrocarbon receptor as a therapeutic target in AD and beyond.

For targeted and multilevel skin barrier repair in AD, we should find ways to reinstate FLG functions by (1) targeting transcription factors that regulate FLG expression, (2) replenishing FLG degradation products, or (3) discovering epidermal proteins that can overtake or compensate FLG (loss of) functions. Intriguingly, such treatment may just have been around for ages, namely, coal tar therapy. The discovery of the molecular mechanism of action, namely, the activation of the aryl hydrocarbon receptor (AHR) in keratinocytes leading to increased terminal differentiation¹¹ has sparked investigative studies into this promiscuous ligand-receptor activated pathway. A new class of nonsteroidal anti-inflammatory drugs has now emerged with the first single molecule therapeutic AHR-modulating agent being approved by the U. S. Food and Drug Administration for psoriasis treatment (Tapinarof) with ongoing trials in AD.²³ The need for alternatives to coal tar therapy is eminent, being considered an obsolete therapy, patient unfriendly (unpleasant smell/odor, staining of clothes, phototoxicity, potential mutagenicity), and having a stigma on potential, but never proven, carcinogenic effects.²⁴

Investigative studies into the function of the AHR in skin and in the dampening of skin inflammation have revealed a complex multifaceted mode of action that underpins the effectivity data of Tapinarof in psoriasis and AD clinical trials, with numerous epidermal processes being (co-)regulated through AHR signaling (Figure 1). Next to the strong induction of keratinocyte terminal differentiation by either transactivation of other transcription factors (eg, OVOL1, NRF2) or metabolic reprogramming in keratinocytes,²⁵ AHR activation upregulates antimicrobial protein and CIDAMPs expression²⁶ and normalizes skin microbiome dysbiosis in AD patients. Additional innate immune responses in keratinocytes, like cytokine and chemokine production, can also be dampened by AHR ligands, albeit this effect is highly dependent on the ligand source, the duration of receptor activation, and the local inflammatory cytokine milieu.²⁷ Interestingly, reported in other tissues than skin, AHR activation controls sphingolipid synthesis by controlling serine palmitoyltransferase,²⁸ the critical enzyme for ceramide biosynthesis. Deficiency of this enzyme in keratinocytes leads to barrier defects and skin inflammation in mice.²⁹ Knowing that the skin microbiome is a potent source of AHR ligands through the catabolism of tryptophan and excretion of indole metabolites, and that skin barrier function and its repair is dependent on the presence of AHR-modulating commensal skin microbes,³⁰ the loss of microbiome diversity that is characteristic for AD may directly drive skin barrier defects through deprived epidermal AHR signaling. Microbiome-mediated therapeutics that tune AHR signaling may thus have future potential in the treatment or even prevention of defects in the physical, chemical, and microbial barrier of an AD patient’s skin. In the meantime, old King Coal may just be a good alternative when other treatment options fail or leave patients ineligible for treatment.

Mapping of the skin barrier constituents in AD Skin proteome findings in AD.

Analysis of the human skin proteome is critical to our understanding of molecular mechanisms for the healthy skin homeostasis or pathogenic processes leading to the development of skin diseases. For minimal invasive sampling of skin proteomes, the use of self-adhesive skin tape strips (STS) has been successfully applied, with protein extracts then resolved by liquid chromatography mass spectrometry. Proteomic approaches to skin analysis allow the examination of small sample amounts and, thus, can be used for minimally invasive skin analysis in different age groups,^4,31 including infants,^32–34 can be applied for the longitudinal sampling, which is critical in clinical trials to monitor the therapeutic drug responses,³⁵ and can also provide opportunities to perform spatially lateral resolved proteome analyses from different depths of the skin through the analysis of consecutive strips.³⁶

An STS proteomic analysis in AD has been proven useful for the analysis of the cornified envelope formation and for the examination of the inflammatory cytokine production in the skin.³⁷ Many groups now have utilized protein extracts from STS for the profiling of inflammatory mediators in the skin of AD patients and healthy subjects.^{4,31,34,35,38,39} Significantly increased production of mediators associated with the type 2 inflammatory reponse, including CCL22, CCL17, and TSLP, was established in AD skin.³⁸ A decrease in IL-1a levels in AD skin has been shown, attributable to a compromised skin barrier in AD.³⁸

Together with lipidomic analysis, proteomic assessment has the potential to examine the expression of skin inflammatory mediators, as well as multiple structural proteins and enzymes that are involved in formation and re-enforcement of the cornified envelope and protein cross-linking to skin ceramides, and to evaluate whether any of these events are abnormal in early life skin samples of AD infants. The period between birth and 2 months of age is the critical period for the transition and adaptation to the dry and cooler environment from the wet, warm, vernix-laden environment in the womb. During this time, the regulations of skin desquamation, antimicrobial defense, moisturization, and acidification are established, which are critical components that control epidermal barrier development.⁴⁰ Although limited knowledge is available to date about skin barrier development during the first year of life, a recent study found that TEWL measurements at 3 months of age were associated with increased risk of allergen sensitization at 6 months, providing a concept that an altered epidermal barrier may potentiate allergic sensitization.⁴¹ In a Korean birth cohort study, increased epidermal TSLP in STS was identified in 2-month-old infants nearly 1 year before the onset of clinical AD, suggesting that early changes in epidermal barrier composition may support the onset of type 2 inflammation in the skin.³² Recently, nested in a prospective birth cohort study that examined the occurrence of physician-diagnosed AD in 300 children, STS samples were analyzed from 44 random children with the onset of AD in the first year of life matched on sex and season of birth with 44 children who did not develop AD. The study established increased levels of CCL17 in the skin of children at 2 months prior to AD disease onset.³⁴

In a recent prospective clinical study using STS global proteomics, it was established that the nonlesional skin of children with AD, and more so the nonlesional skin of children with atopic dermatitis and food allergy (AD+FA+), compared with healthy nonatopic controls is associated with an immature KRT profile consistent with keratinocyte hyperproliferation in AD skin.⁴

An unbiased, unsupervised analysis of the STS proteomics data identified a group of 45 proteins that were differentially expressed across the 3 study groups. This group of proteins had the highest expression in the skin of AD+FA+ children and were strongly associated with TEWL, immunoglobulin E, and food skin tests. Functional analysis of these proteins subgrouped them into the 3 major functional groups, namely: (1) KRT intermediate filaments that reflect altered expression of keratins in AD+FA+ skin; (2) proteins associated with inflammatory response (S100 proteins, alarmins, protease inhibitors); and (3) glycolysis and oxidative stress response proteins (glycolytic enzymes, oxidative stress response enzymes).³¹

Importantly, skin expression of these proteins distinguished nonlesional skin samples between healthy nonatopic, AD+FA-, and AD+FA+ groups, whereas TEWL alone was not a sufficient parameter to differentiate nonlesional skin of these patients. SERPINB3 and KRT77 were identified to have the greatest impact on skin TEWL. These findings were confirmed and validated in an independent cohort of AD adults with a history of clinical reactions to peanut. These data suggest that AD+FA+ patients are a unique endotype that persists into adulthood.³¹

The aforementioned established approaches for the STS proteomic analysis open up opportunities for the longitudinal proteomic studies of structural proteins and inflammatory mediators in AD skin and, thus, will shed light into the events in the skin that preceed AD development.

Skin lipid biochemistry under attack of AD pathological processes.

Skin barrier function is ensured by a multilayered hydrophobic proteo-lipid structure formed during keratinocyte terminal differentiation. Lipid biochemistry is dramatically changing in terminally differentiating keratinocytes that result in the disappearance of most common lipids (such as polar and neutral glycero- and sphingolipids with fatty acid chain length between 14 and 24 carbon). At the very surface of the skin, in the stratum corneum, unique, skin-specific ceramides that biophysically are highly rigid and hydrophobic appear. Furthermore, terminally differentiating keratinocytes launch the biosynthesis of ultra long-chain fatty acids (up to the chain length of 34 carbons) that replace shorter-chain length fatty acids in ceramides and make resulting ceramides even more rigid and hydrophobic. The actual hallmark of keratinocyte lipid biochemistry is the biosynthesis of unique skin-specific esterified omega-hydroxy fatty acid containing ceramides (EOS-CER) that are not only highly hydrophobic owing to their extreme total carbon number (up to 74 carbons) but also because of their unique role as an immediate precursor of protein-bound ceramides. After a multistep oxidation and removal of linoleic acid in EOS-CER, the resulting omega-hydroxylated ceramides are linked to such proteins as loricrin and involucrin in corneodesmosome by transglutaminase 1^42–45 and form a scaffold to lay upon free ceramides, cholesterol, and free fatty acids that comprise the majority of free lipids in the stratum corneum and are ultimately responsible for skin barrier properties.⁴⁶ What is not yet fully appreciated in the uniqueness of stratum corneum lipid biochemistry is the fact that stratum corneum ceramides have an extended chain length of not only fatty acids but also sphingoid bases.^47–49 In fact, ceramides with C20- and C22-sphingosine are at least as abundant as ceramides with the ordinary C18-sphingosine, and ceramides with C28-C32 omega-hydroxy fatty acids comprise the backbone of the scaffold-forming protein-bound ceramides.^47,49

Biophysical modelling suggests that the presence of any charged molecules, like sphingomyelin, in this highly organized ceramide-cholesterol-free fatty acid assembly will disrupt tight packing of lipids and decrease overall lamellae barrier properties.^46,50 Therefore, any use of emollients that contain lipids with charged polar head groups, like phosphatidylcholine, sphingomyelin, and phosphatydylethanolamine, is potentially detrimental if not for healthy skin, then for already damaged skin. Moreover, many hygiene products, like shampoos, contain natural lysophospholipids or synthetic lipids that possess strong detergent-like properties that will absolutely disrupt lamellaes and impair skin barrier function if allowed to be present on the skin for a prolonged time. Taking into account that lipid-like molecules, when applied to the skin, quickly penetrate deep into the skin,^51,52 a prolonged use of oil-based emollients that contain natural polar lipids and oils should be avoided unless specifically required.

In AD, it is this organized unique stratum corneum ultrastructure that is “targeted and destroyed” by type 2 cytokines IL-4 and IL-13. Recently, we have shown that IL-4 and IL-13 inhibit the expression of fatty acid elongases (ELOVL) 3 and ELOVL6 that increase the chain length of C16-C20 fatty acids.⁴⁴ Therefore, regardless of a concomitant increase in the expression of elongases ELOVL1 and ELOVL4 that actually form very long- and ultra long-chain fatty acids, a global generation of C24-C32 fatty acids and ceramides containing those fatty acids is inhibited.⁴⁴ Simultaneously, keratinocytes become abundant in a short-chain palmitic acid (16:0-FA) that now not only excessively incorporates into ceramides as fatty acid but also pushes the biosynthesis of ceramides with C18 sphingoid bases by serine palmitoyl transferase over ceramides with C20-C22 sphingoid bases.⁴⁹ This lack of ultra long-chain fatty acids in AD skin also leads to insufficient biosynthesis of EOS-CER.⁴⁹ As a result, keratinocytes form shorter-chain length ceramide molecules with insufficient proportion of critical EOS-CER that globally affects skin barrier property just from the total molecule hydrophobicity perspective.

Global shortening of stratum corneum ceramides is not the only way type 2 cytokines disrupt lamellae formation in AD skin. Our group was one of the first to show that IL-4 and IL-13 are able to drastically inhibit the expression of the FLG protein as well as the major cross-linkers for omega-hydroxylated ceramides loricrin and involucrin in keratinocytes in vitro.^45,53 Because FLG functions as a core protein that assembles and organizes other proteins in corneodesmosome, its insufficiency, together with the lack of key lipid acceptors as a result of IL-4/IL-13–induced signaling, leads to a formation of ultrastructurally disorganized lamellae^4,5 and the loss of skin barrier integrity. Together with genetic factors that affect FLG expression,⁵⁴ IL-4 and IL-13 provide a multitargeted insult to keratinocyte terminal differentiation that ends up in the loss of skin barrier function, development of eczema, and the onset of atopic march. What causes the activation of type 2 immunity in the skin originally is still a topic of major debates, but many factors point at environmental causes that provide an original insult.

The understanding of the major role played by IL-4 and IL-13 in the onset and development of AD allowed the development of therapeutic strategies based on the disruption of IL-4/IL-13–induced signaling by biologics. The first longitudinal study aimed to understand the link between clinical improvement during the course of treating AD patients with Dupixent (dupilumab), and the changes in stratum corneum lipids has not only confirmed a global decline in EOS-CERs and shift in favor of a shorter fatty acids in ceramide molecules in AD described by previously⁴⁴ but also demonstrated quick then steady improvement in all lipid parameters over the course of the treatment, revealing a distinct pattern in changes in ceramides with longer C18- and C20-sphingosine–containing ceramides versus ceramides with C18-sphingosine.⁴⁹ This work further pointed out the importance of stratum corneum lipids to maintain skin barrier function.

Skin barrier normalization: a critical factor for disease modification

Barrier function tests as a guidance in the management of AD.

Whereas skin barrier function, as demonstrated by TEWL analyses, closely associates with AD severity, the most interesting differences are observed in clinically clear skin. For example, Montero-Vilchez and colleagues⁵⁵ recently reported mean TEWL values of 11.60 g/m²/h for healthy skin compared with 13.15 g/m²/h and 28.68 g/m²/h, respectively, in the unaffected and affected skin of AD patients. These values highlight the marked effect of inflammation and the predisposition for suboptimal function in AD patients at unaffected sites. This calls for 2 distinct but complementary approaches to AD management: (1) the normalization of skin barrier function in nonlesional skin and (2) the suppression of inflammation once overt AD develops. On the surface, this fits with the well-established treatment paradigm of using emollients as baseline therapy on which topical and systemic treatments are added.⁵⁶ The misconception, however, is that emollients treat the skin barrier. By definition, an emollient simply softens and soothes the skin, and although they all do this well, they have very different effects on the function of the skin as a barrier.⁵⁷ For example, the regular use of the aqueous cream emollients reduces skin barrier function⁵⁸ and was associated with a high rate of adverse cutaneous reactions in children.⁵⁹ In stark contrast to this, the regular use of a complex emollient containing 5% urea was shown to improve skin barrier function⁶⁰ and prolong the period of remission compared with no treatment⁶¹ and the same regimen with a simple emollient cream.⁶²

This highlights that a greater focus on how interventions for AD affect the skin barrier is needed to improve the long-term control of the condition. Whereas TEWL provides water permeability barrier function, irritant and allergen challenge tests provide a useful marker of skin permeability to these environmental agents.⁶³ Notably pretreating the skin with emollients has different effects in these tests, with some increasing and some decreasing cutaneous irritation and inflammatory responses to the challenge.^64–67 The function of the skin as an antimicrobial barrier is also important to minimize potentially pathogenic S. aureus colonization and maintain the normally diverse microbial flora of the skin.⁶⁸ Early studies have demonstrated that achieving positive microbiome changes with emollients is dependent on their effect on the skin barrier.⁶⁹ Vaseline treatment of lesional AD skin also in part shifts the skin microbiome composition toward a healthy state, although targeting the AHR with coal tar resulted in a stronger shift, potentially due to the concomitant induction of antimicrobial protein expression, as introduced previously.⁷⁰

What lies beneath the surface: sampling skin proteins, metabolites, and lipids.

The quantification of functional skin changes is of utmost importance in our understanding of the actual disease processes, disease endotype, and severity and can be surprisingly easily obtained from samples collected on adhesive tapes. For example, urea-based emollients with positive effects on the skin barrier appear to increase NMF levels^71,72 consistent with the finding that urea promotes FLG expression in the skin.⁷³ Transepidermal water loss measurements taken in conjunction with STS to experimentally disrupt the skin have been used to assess the structural integrity of the stratum corneum. Leung and colleagues⁴ found that, whereas TEWL levels were similar between AD patients with and without FA, the stratum corneum integrity was significantly lower in the former group. Carriage of FLG loss-of-function mutations was also associated with decreased integrity even when TEWL appeared normal.⁷⁴ In both cases, the reduction in integrity was associated with significant changes in the composition of lipids, especially ceramides, and NMF components.⁷⁵ Alterations in the composition of the lipid lamellae have a marked effect on the way the lipids align together and compromise water barrier function.⁷⁶ Using attenuated total reflectance–Fourier transform infrared (ATR-FTIR) spectroscopy, Danby and colleagues⁷⁷ demonstrated that improvements in stratum corneum lipid structure following the use of an emollient cream containing skin lipids contributed to improved integrity and reduced responsiveness to irritants. Spectroscopic techniques like this have also been developed for the direct in vivo quantification of relative NMF, water, and lipid levels in the stratum corneum and show promise as at-the-bedside diagnostic and prognostic tools.^72,78,79 Gathering minimally invasive superficial stratum corneum samples on adhesive discs is an alternative technique allowing individual components to be quantified.

The importance of skin barrier function in the prevention of AD development and relapses.

The ability to predict AD and the promise of skin barrier normalization using emollient therapy have opened up the possibility of preventing (or at least delaying) the onset of AD from birth. Initial pilot studies suggested that, by regularly applying an emollient from birth, the relative risk of AD developing (by 12 mo of age) could be cut by up to 50%.^80,81 Subsequent larger studies did not confirm this and hinted at an increased risk of FA, contact dermatitis, and skin infections.^82–84 The recent finding that more frequent emollient use during the first 3 months of life is associated with higher rates of FA is a clear warning that we need to evaluate the longer-term consequences of emollient therapy as an intervention for AD.⁸⁵ Worryingly, a number of the top 20 most used emollients identified in this study have known detrimental effects on stratum corneum integrity.^58,71,86,87 Encouragingly, more recent primary prevention studies using more sophisticated emollients have reported protective effects of intervening shortly after birth.⁸⁸ The very stark differences in the effects of emollients on skin barrier function, which influences how the skin responds to key AD triggers, provide a compelling mechanism to explain the varied outcomes of the primary prevention studies conducted so far.

Immunosuppressive therapies also have pronounced effects on the skin barrier, which are defined by their specificity for type 2 inflammation. This is because the mediators of type 2 inflammation, like IL-4 and IL-13, downregulate the expression of a plethora of genes encoding structural proteins and enzymes required for the proper formation of the skin barrier.⁸⁹ Traditional immunosuppressants like corticosteroids, cyclosporin A, and methotrexate display very broad anti-inflammatory action but have well-established off-target effects that compromise their long-term safety.^90,91 At the other end of the spectrum, newly developed biologic therapies specifically targeting IL-4 and IL-13 signaling display improved long-term safety.⁹²

To limit the risk of adverse effects, corticosteroids are commonly delivered topically, where they effectively inhibit localized skin inflammation.⁹³ The problem, however, is that corticosteroids have broad effects, and their continued use in clear-appearing skin reduces skin barrier function and thins the epidermis compared with healthy controls.^94,95 This is partly explained by the inhibition of genes encoding structural proteins and lipid processing enzymes leading to reduced levels of NMF and abnormal formation of the extracellular stratum corneum lipid matrices.^96,97 A number of adverse effects are associated with the long-term inappropriate use of topical corticosteroids (TCS), including striae and telangiectasia.⁹⁸ Long-term steroid use and the associated rebound flare phenomenon upon discontinuation of steroid use are explained by the counteracting effects of TCS; while being applied, they prevent inflammation from developing, but upon discontinuation, the damage caused to the skin makes it prone to relapse.^99,100 Although not quite as efficacious as TCS, topical calcineurin inhibitors do not appear to exert these negative effects on the skin barrier and are better suited to delicate skin sites.^101,102

Topical therapies are commonly used reactively when the skin erupts and until it visibly clears. However, we know that inflammation persists in the clear-appearing skin of AD patients.¹⁰³ Maintenance regimens involve reduced frequencies of dosing (eg, twice weekly or only on a weekend) and significantly prolong the period of remission, highlighting the effect of subclinical inflammation on the propensity for future flares.^104,105 This means a fine balance must be struck between completely clearing the drivers of inflammation and avoiding skin damage when using TCS. Optimizing this balance requires new approaches to visualizing the skin’s subclinical condition. Optical coherence tomography is one such approach for obtaining virtual biopsies of the skin noninvasively from which sensitive metrics of subclinical inflammation and epidermal atrophy can be derived (Figure 2).¹⁰⁶

FIGURE 2.

Example of optical coherence tomography (OCT) images taken from the forearm of healthy skin and skin affected to different degrees by atopic dermatitis (AD) using the Vivosight OCT machine (Michelson Diagnostic, Maidstone, UK). (Top row) The raw images. (Bottom row) The same images have been segmented to highlight the changing structure of the skin. (Images provided by Dr. Robert Byers, Sheffield Dermatology Research, The University of Sheffield, Sheffield, UK.)

The (in)direct effects of anti-inflammatory therapies on skin barrier function.

Few studies have directly compared the effects of different classes of anti-inflammatory therapy on the skin barrier. In 1 example, TCS, cyclosporin A, and dupilumab regimens were all found to improve skin barrier function over a 16-week treatment period.¹⁰⁷ The improvement was more pronounced and resulted in a lower TEWL in the dupilumab group than in the other groups at both lesional and nonlesional areas. Treatment with dupilumab markedly reduces epidermal hyperplasia, the number of inflammatory cells, and the levels of proinflammatory cytokines in the lesional skin of patients.^108,109 The skin tissue alarmins TSLP, IL-15, and IL-25, which play a role in driving type 2 inflammation, are also inhibited. Moreover, the gene expression and levels of structural proteins and enzymes required for healthy skin barrier function, like FLG, LEKTI, and HBD-3, are normalized. Upregulation of genes involved in lipid metabolism has also been observed, including that of ELOVL3 required for the elongation of skin lipids. Lipidomic analysis of stratum corneum samples collected every 2 weeks revealed a progressive normalization of lipid levels, characterized by an increase in EOS-CER and a decrease in short-chain length ceramides with non-hydroxy fatty acids and sphingoid bases.^49,110 Functionally, TEWL rapidly normalizes over the first 2 weeks of dupilumab treatment, but stratum corneum integrity follows a slower trajectory, closely matching the changes in lipid composition and suggesting that further improvements could be made with continued treatment beyond 16 weeks. Commensurate with the improvement in skin barrier condition is a significant drop in S. aureus skin colonization and a concerted increase in microbiome diversity.⁶⁸

These findings demonstrate that the very specific targeting of the type 2 immune pathway effectively suppresses skin inflammation and normalizes the condition and function of the skin barrier. They also show that, whereas treatments like tralokinumab and dupilumab can bring about rapid changes in the function of the skin within just 2 weeks, longer treatment regimens are required to fully normalize the condition of the skin barrier. Less targeted therapies for AD can also adversely affect this normalization, particularly as the skin clears of inflammation. The negative effects of simple emollients are likely to confound this process and may even leave the skin more susceptible to environmental triggers, which is further discussed in the next section.

Optimal skin care and emollient use in AD

Skin care products affecting skin barrier function.

Although we here focus on skin barrier malfunction in established AD, we must not forget the healthy population who is potentially at risk of developing AD or dermatitis-like conditions later in life. For this lucky normal majority, the Hippocratic Oath applies to skin care specialists—by all means, “do no harm!” Fortunately, it is difficult to visibly insult normal skin, no matter which unguent you choose to apply. Oily emollients can be applied even to facial skin if one is not at risk for developing acne. Hence, most moisturizers can be considered safe in these fortunate individuals. Yet, despite having normal skin, these individuals typically use plenty of topical skin care products. The reason for this lies deep within the epidermal neuroendocrine system.¹¹¹ Unguent applications inevitably include stroking and even caressing of the skin surface, processes that almost certainly stimulate the production of endorphins, as well as the feel-good neurohormone, oxytocin. Small wonder, then, that so many users become addicted to various forms of repetitive skin care. Serious problems can occur, however, in those born with inherited forms of dry, sensitive skin, such as AD, as noted previously, and in the xerotic skin that accompanies chronological aging. Aging skin recovers very slowly, even after minor insults, requiring up to 1 week to normalize, as shown by persistent elevations in TEWL (Figure 3). These patients often are tempted to apply moisturizers, whose impact is typically short-lived, resulting in a vicious circle that demands ever more frequent applications. In fact, a large majority of these over-the-counter moisturizers further compromise the skin barrier.¹¹³

FIGURE 3.

Nine (22 to 62 year old; seven females and two males) nonatopic human volunteers were enrolled in this study. Creams (n = 3 to 4 subjects each) were applied to 3×3 cm² areas (of previously untreated skin sites) on the flexural surface of the forearm twice daily for 4 days. Two sites, 12–14 cm apart, were selected on each forearm. Untreated sites on contralateral forearms served as normal controls. During the study period, no detergents or skin care products were applied to the forearm flexors. On day 5, basal pH, hydration, and transepidermal water loss were measured with a Tewameter (Evalulab, Montreal, Canada). Barrier recovery rates were assessed 3 hours following barrier disruption by repeated tape stripping until transepidermal water loss levels ≥5 mg/cm²/h. Data are expressed as mean ± standard error of the mean. SB, styrene butadiene. Reprinted with permission from Elias et al.¹¹²)

Key determinants for effective emollients to foster skin barrier function.

Lipid deficiency is a key attribute to AD and a lengthy series of studies showed that the stratum corneum contains 3 key lipid species (cholesterol, nonessential free acids, ceramides) that are abnormal in AD (as described previously). Whereas each of these lipids are required individually for the permeability barrier, each one alone cannot correct the defective barrier in AD. To achieve normal recovery, they must be applied as an equimolar ratio of 3:1:1¹¹⁴ and applied at a high total concentration to replenish the missing 5% of total lipids in the stratum corneum. Very few products fulfill these criteria, and even fewer contain a mixture of the 3, required physiological lipid species. Two such novel products, developed in Korea, but available in the United States, are Atopalm cream (NeoPharma USA) and a Dr. Raymond Laboratories formulation (which also contains all 3 key lipid species), CURECODE Intensive1:1 Care Ointment with Neuromide. Man and coworkers¹¹⁴ developed a 3:1:1 optimal repair mixture, which is marketed as a prescription formulation, EpiCeram (Primus Pharma, Scottsdale, AZ). In 1 controlled trial of moderate-to-severe pediatric AD, EpiCeram proved comparable in outcomes to a midpotency, fluorinated steroid.¹¹⁵ Whether this formulation will prove to be efficacious in other forms of eczematous dermatitis is not yet known. Notably, EpiCeram is not yet available for sale outside of the United States.

Another important factor in skin care products or emollient development is the formulation’s pH. An acidic pH reflects the pH of the normal skin surface and is critical for both the permeability barrier and antimicrobial defense (Figure 1). Yet, at present, many moisturizers and cleansers are formulated at what is erroneously considered a normal, neutral pH (7 or higher). Such a neutral-to-alkaline pH may activates serine proteases, such as kallikrein 5 and 7, which are present in abundance in the stratum corneum.¹¹⁶ Not only are they directly destructive, but they also increase Th2-mediated skin inflammation by upregulating epidermal production of the cytokine instigator of allergic inflammation, TSLP.¹¹⁷ Fortunately, the consumer has a choice of several skin care products, formulated at an acidic pH, such as the Optical coherence tomography product, Lacticare, also known as Amlactin (Sandoz). Notably, the lipid replacement product, EpiCeram, described previously, is also formulated at an acidic pH (=5.0).¹

Suggested toolbox for skin barrier maintenance in daily practice.

The easiest at-home treatment is to refrain from prolonged hot baths and showers, which deplete the skin of its “greasy” lipids and place additional stress on the permeability barrier. Application of any water-in-oil moisturizer, such as petroleum, Aquaphor, or Eucerin, immediately after bathing will help retain the stratum corneum’s critical hydration levels but, in AD patients, will only provide temporary relief (up to 4 h). Dry skin follows shortly unless the moisturizer is applied repeatedly. The application of an acidic formulated, physiological lipid-based formulation, such as Atopalm or EpiCeram would be preferred based on the scientific evidence on skin biology and AD pathophysiology as reviewed herein. However, given our evolving knowledge of the heterogeneity of skin barriers in various endotypes of AD and the lack of uniform availability of creams optimized for skin barrier repair, this aspect of AD skin care still has room for improvement.

DISCUSSION AND FUTURE WORK

The stratum corneum plays a critical role in protecting human hosts against microbial invasion and allergen/irritation penetration. The stratum corneum or cornified envelope results from terminal differentiation of keratinocytes, which are susceptible to dysregulation by cytokines (eg, I-L4/IL-13, IL-22), microbes such as S. aureus, pollutants and detergents, or other irritants. The strength of the cornified envelope is dependent on the ideal mixture of proteins and lipids found in normal skin in a free form and further cross-linked to proteins by transglutaminase. Filaggrin deficiency, due to either loss-of-function mutations or increased IL-4/IL-13, has been well studied and found to cause skin barrier dysfunction. Recent studies have highlighted the importance of other epidermal proteins and, more importantly, the relative concentrations of ceramide (2), cholesterol (1), and free fatty acids (1) at an optimal molar ratio found in normal skin.

With this in mind, it is important that we take a holistic approach to evaluating new and existing treatment strategies for their effects on the skin barrier and, ultimately, the long-term control of AD. We now have a variety of techniques at our disposal for this purpose and to establish disease modification. On the one hand, multi-omics (transcriptomic, lipidomic, metabolomic, and proteomic) analysis of skin tissue enables very specific appraisal of treatment effects on individual dysregulated pathways, albeit at limited time points. On the other hand, biophysical techniques like optical coherence tomography, molecular spectroscopy, and TEWL enable noninvasive, frequent, and direct quantification of key biomarkers so that the dynamic effects of interventions can be monitored. Future efforts to combine these techniques will undoubtedly lead to improved and highly sensitive methods of monitoring this complex condition and improved therapeutics that will benefit the skin barrier function and overall people’s health and well-being.

In biophysical reviews, it has been concluded that any lipid composition that differs from natural composition of lipids in the stratum corneum will impair ultrastructural organization of lipids in lamellae because lipids will partition into natural lipid structure. Therefore, what is applied to the skin should be as close as possible to the natural lipid composition.^46,50 Aside from optimizing lipid composition of skin emollient creams, it is important to optimize timing of cream application to the infant skin because the transition from pregnancy to infancy results in a drastic change in humidity. Damage due to the perinatal skin exposure has not been addressed in any study examining optimal time of emollient application, number of emollient applications, and location of skin cream application. Most importantly, consistency in skin applications must be monitored and when eczema-prone infants need transition from simple skin hydration and barrier cream to anti-inflammatory therapies must be determined. Such studies will be more precise when noninvasive skin tape sampling can be done to analyze the skin pathobiology in response to various therapeutic approaches.

Abbreviations used

AD

Atopic dermatitis

AD+FA+

Children with atopic dermatitis and food allergies

AHR

Aryl hydrocarbon receptor

AMP

antimicrobial peptide

CIDAMPs

Cationic intrinsically disordered antimicrobial peptides

ELOVL

Fatty acid elongases

EOS-CER

Esterified omega-hydroxy fatty acid containing ceramides

FA

Food allergy

FLG

Filaggrin

IL

Interleukin

IV

Ichthyosis vulgaris

KRT

Keratin

NMF

Natural moisturizing factor

OTC

Over-the-counter

pro-FLG

Pro-filaggrin

STS

Skin tape strips

Th

T-helper cell

TCS

Topical corticosteroids

TEWL

Transepidermal water loss

TSLP

Thymic stromal lymphopoietin