Regulation of Skin Barrier Function via Competition between AHR Axis versus IL-13/IL-4‒JAK‒STAT6/STAT3 Axis: Pathogenic and Therapeutic Implications in Atopic Dermatitis

Atopic dermatitis (AD) is characterized by skin inflammation, barrier dysfunction, and chronic pruritus. As the anti-interleukin-4 (IL-4) receptor α antibody dupilumab improves all three cardinal features of AD, the type 2 cytokines IL-4 and especially IL-13 have been indicated to have pathogenic significance in AD. Accumulating evidence has shown that the skin barrier function is regulated via competition between the aryl hydrocarbon receptor (AHR) axis (up-regulation of barrier) and the IL-13/IL-4‒JAK‒STAT6/STAT3 axis (down-regulation of barrier). This latter axis also induces oxidative stress, which exacerbates inflammation. Conventional and recently developed agents for treating AD such as steroid, calcineurin inhibitors, cyclosporine, dupilumab, and JAK inhibitors inhibit the IL-13/IL-4‒JAK‒STAT6/STAT3 axis, while older remedies such as coal tar and glyteer are antioxidative AHR agonists. In this article, I summarize the pathogenic and therapeutic implications of the IL-13/IL-4‒JAK‒STAT6/STAT3 axis and the AHR axis in AD.


Introduction
Atopic dermatitis (AD) is a common inflammatory skin disease that accounts for almost 10% of dermatologic patients in Japan [1]. Its prevalence is 10% to 16.5% in children under 5 years old, which showed a tendency to increase from the 1980s to the early 2000s globally [2,3]. The patients with AD show heterogeneous clinical presentation and at least six subtypes are defined by their natural course with the early-onset-early-resolving type being the most frequent one [4,5]. However, subsequent recurrence is also frequently seen and the recurrent symptoms are usually more indolent than the initial ones [4,5]. AD is characterized by skin inflammation, skin barrier dysfunction, and chronic pruritus, which reduce quality of life and treatment satisfaction among afflicted patients [6][7][8]. Their adherence to treatment is also markedly deteriorated [9][10][11]. Skin barrier dysfunction causes frequent colonization of Staphylococcus aureus, which can lead to further deterioration of the barrier function [12,13]. Although not as conspicuously as in psoriasis, AD is known to be associated with systemic inflammatory diseases such as cardiovascular diseases and metabolic syndrome [14].
In addition to IL-4Rα/IL-13Rα1, keratinocytes also express IL-13Rα2 ( Figure 1). IL-13Rα2 is a high-affinity receptor for IL-13, but it does not signal. Therefore, IL-13Rα2 serves as a decoy receptor and decreases the concentration of IL-13 in the microenvironment [51,52]. Notably, IL-13 does not alter the expression levels of IL-4R and IL-13Rα1, but it up-regulates the expression of IL-13Rα2 in keratinocytes [51]. Scratch injury has also been found to enhance the expression of IL-13Rα, but not IL-4Rα and IL-13Rα1, in keratinocytes [51]. Thus, IL-13 promotes the development of atopic inflammation via IL-4Rα/IL-13Rα1, but simultaneously triggers a negative feedback signal via the decoy receptor IL-13Rα2, which lowers excess amounts of extracellular IL-13 [51]. Pruritus-mediated scratching appears to exacerbate dermatitis [53]. However, scratch injury on keratinocytes also results in the up-regulation of IL-13Rα2 expression and forms another negative feedback circuit to inhibit excess IL-13 activity [51,52].

Role of IL-31 and IL-13/IL-4 in Atopic Pruritus
Pruritus is the major subjective symptom in AD [54]. Histamine is the essential pruritogen of urticaria and immediate-type allergic response, for which the first-line treatment is H1 receptor antagonists [55][56][57]. However, the anti-pruritic efficacy of H1 antagonists is quite limited in AD [58].
IL-31 is a type 2 cytokine produced from Th2 cells [59,60]. Its expression is elevated in the lesional skin of AD [22]. IL-31 induces elongation and branching of sensory nerve fibers [61,62]. Notably, the administration of IL-31 evokes scratching behavior in mouse, dog, and monkey, as well as pruritus in human [59]. The transcription factor hypoxia-inducible factor-2α (HIF-2α), also called endothelial Per-Arnt-Sim (PAS) domain protein 1 (EPAS1), is required for the production of IL-31 [60]. Binding of IL-31 to IL-31-sensitive dorsal root ganglion cells induces the production of neurokinin B and activates its neurokinin 3 receptor, which then triggers the production of an itch mediator, gastrin-releasing peptide [63]. A recent study also revealed a significant role of P2X3 receptor-positive nerves in chronic pruritus [64]. In addition, chronic pruritus is known to be associated with STAT3-dependent astrogliosis in the spinal dorsal horn [65].
Injection of the anti-IL-31 receptor A (IL-31RA) antibody nemolizumab significantly reduces the pruritus of patients with AD [66,67]. The monthly administration of nemolizumab for 52 weeks was also found to continuously inhibit the pruritus of AD [68]. Moreover, cotreatment with nemolizumab and topical steroid was reported to significantly augment the anti-pruritic effects of topical steroid monotherapy [69]. The anti-IL-31 antibody lokivetmab is now commercially available for the treatment of canine AD [70]. This highlights the pivotal role of IL-31 in atopic pruritus.
Colonization of Staphylococcus aureus up-regulates the expression of antimicrobial peptides such as human β-defensin 2 and -3 and RNAse7; however, IL-31 inhibits their expression and may accelerate the Staphylococcal infection [73]. Staphylococcal superantigen may also induce glucocorticoid insensitivity by inhibiting the nuclear translocation of glucocorticoid receptor α [74].

Regulation of Skin Barrier Function by Competition between AHR Axis and IL-13/IL-4-JAK-STAT6/STAT3 Axis
Under physiological conditions, homeostasis of the skin barrier function is regulated by the coordinated expression of barrier-related proteins, intercellular lipids, and corneodesmosomes in the granular and cornified layers [75,76]. The genes encoding many barrier-related proteins such as FLG, LOR, and involucrin (IVL) are encoded in the epidermal differentiation complex (EDC) region located on chromosome 1q21.3 [76]. The EDC region includes members of the cornified envelope precursor gene family such as LOR and IVL, the S100A protein gene family such as S100A7 and S100A8, and the fused gene family such as FLG and FLG2, with this latter family being derived from fusion of the cornified envelope precursor gene and the S100A protein gene [76] (Figure 2).
Among various transcription factors, aryl hydrocarbon receptor (AHR) is essential to the coordinated up-regulation of EDC genes [25,[77][78][79]. Several endogenous and exogenous ligands up-regulate the expression of FLG, LOR, and IVL [5,[77][78][79] (Figure 3). For example, photoproducts produced by ultraviolet rays [80][81][82][83], bioproducts from commensal cutaneous microbiomes such as Malassezia and Staphylococcus epidermidis [84,85], and bioproducts from intestinal microbiomes [86] activate AHR and up-regulate the expression of FLG. Phytochemicals used in folk medicine [87][88][89][90][91] and some cosmetic ingredients [30] contain AHR ligands and up-regulate FLG expression. Medicinal coal tar and soybean tar glyteer are AHR ligands that increase FLG expression [31,92] (Figure 3). granular and cornified layers [75,76]. The genes encoding many barrier-related proteins such as FLG, LOR, and involucrin (IVL) are encoded in the epidermal differentiation complex (EDC) region located on chromosome 1q21.3 [76]. The EDC region includes members of the cornified envelope precursor gene family such as LOR and IVL, the S100A protein gene family such as S100A7 and S100A8, and the fused gene family such as FLG and FLG2, with this latter family being derived from fusion of the cornified envelope precursor gene and the S100A protein gene [76] (Figure 2). The epidermal differentiation complex includes cornified envelope precursor gene family members such as LOR and IVL, S100A protein gene family members such as S100A7, and The epidermal differentiation complex includes cornified envelope precursor gene family members such as LOR and IVL, S100A protein gene family members such as S100A7, and fused gene family members such as FLG and FLG2. The fused gene family was generated by fusion of the cornified envelope precursor gene and the S100A protein gene.
Ligation of AHR induces its cytoplasmic-to-nuclear translocation and the expression of genes encoding FLG, LOR, IVL, and other barrier-related proteins in the EDC loci [25,[77][78][79] (Figure 3). In addition to the AHR-mediated direct up-regulation of barrier-related proteins, the activation of AHR up-regulates the gene and protein expression of OVO-like 1 (OVOL1) transcription factor [81,90]. Cytoplasmic OVOL1 then translocates into the nucleus and further up-regulates the gene and protein expression of FLG and LOR [81,90]. However, OVOL1 is not involved in IVL up-regulation [90]. Activation of AHR simultaneously induces the expression of xenobiotic-metabolizing enzymes such as cytochrome P450 1A1 (CYP1A1), which can potentially metabolize AHR ligands [80]. Therefore, the activation of AHR by ligands is self-limiting, ensuring constitutive homeostatic regulation [80].
In the lesional skin of AD, the expression of FLG is down-regulated compared with that in healthy control skin [81,92]. Transepidermal water loss (TEWL) is significantly increased and the skin hydration is decreased in AD compared with the levels in healthy individuals [93]. Serum CCL17 levels are positively correlated with TEWL and negatively correlated with skin hydration in AD [93]. Topical steroids improve the skin inflammation of AD and normalize the increased levels of TEWL and decreased expression of FLG and LOR [41].
Genome-wide association studies (GWASs) have identified more than 30 genes conferring susceptibility to AD [94][95][96]. A meta-analysis of GWASs for those of European, Chinese, or Japanese ancestry revealed that the top three highly significant susceptibility genes are FLG, OVOL1, and IL13 [96]. Among these, loss-of-function mutation of FLG is most significantly associated with the development of AD [75,96]. However, the majority of patients with AD do not exhibit loss-of-function mutation of FLG [97]. In addition, the loss-of-function mutation of FLG is less common in Southern Europe than Northern Europe [97], and is not found in some parts of Africa [98,99]. On subtropical Ishigaki Island, the prevalence of pediatric AD is lower than that in mainland Japan [100] and the loss-of-function mutation of FLG is not associated with the development of AD on this island [101].
up-regulates the gene and protein expression of OVO-like 1 (OVOL1) transcription factor [81,90]. Cytoplasmic OVOL1 then translocates into the nucleus and further up-regulates the gene and protein expression of FLG and LOR [81,90]. However, OVOL1 is not involved in IVL up-regulation [90]. Activation of AHR simultaneously induces the expression of xenobiotic-metabolizing enzymes such as cytochrome P450 1A1 (CYP1A1), which can potentially metabolize AHR ligands [80]. Therefore, the activation of AHR by ligands is self-limiting, ensuring constitutive homeostatic regulation [80].  . Up-regulation of FLG, LOR, and IVL by the AHR axis. AHR is activated by various ligands such as ultraviolet ray (UV) photoproducts, bioproducts of Malassezia and Staphylococcal epidermidis, bioproducts of intestinal microbiomes, phytochemicals, cosmetic products, and medicinal coal tar and glyteer. Once activated, cytoplasmic AHR translocates into the nucleus, binds EDC, and up-regulates the expression of FLG, LOR, and IVL. AHR also up-regulates the expression of OVOL1. Cytoplasmic OVOL1 translocates into the nucleus, binds EDC, and up-regulates FLG and LOR. AHR also up-regulates the expression of xenobiotic-metabolizing enzymes such as cytochrome p450 1A1 (CYP1A1), which rapidly metabolize the AHR ligands. Thus, homeostatic regulation of barrier-related proteins is operated by the AHR axis.
Down-regulation of barrier-related proteins by IL-13/IL-4 is probably more significantly involved in the development of AD than loss-of-function mutation of FLG [31,81,89,90,92,102]. A recent report from Croatia revealed that loss-of-function mutation of FLG was detected in only 4 of 91 AD patients and none of 47 non-AD controls; it also showed that elevated TEWL is associated with skin inflammation but not with FLG mutation [103].
These results indicate that the IL-13/IL-4-JAK-STAT6/STAT3 axis affects several signaling pathways to inhibit the expression of barrier-related proteins. In parallel with this, upon IL-4 treatment, the permeability barrier of the cultured keratinocytic sheet is disrupted [106] and the distribution of cell surface E-cadherin is altered [107]. Given that the expression levels of barrier-related proteins are regulated via competition between the AHR axis and the IL-13/IL-4-JAK-STAT6/STAT3 axis [25,31,81,92], STAT6-deficient conditions may enhance the AHR axis and potentiate the skin barrier function. Findings have suggested that this is indeed the case. The skin barrier function in STAT6-deficient mice is significantly up-regulated compared with that in wild-type mice, as demonstrated by decreased TEWL, increased water content, decreased pH, decreased permeability of Evans blue, and increased LOR and FLG expression [108]. Given that the expression levels of barrier-related proteins are regulated via competition between the AHR axis and the IL-13/IL-4-JAK-STAT6/STAT3 axis [25,31,81,92], STAT6-deficient conditions may enhance the AHR axis and potentiate the skin barrier function. Findings have suggested that this is indeed the case. The skin barrier function in STAT6-deficient mice is significantly up-regulated compared with that in wild-type mice, as demonstrated by decreased TEWL, increased water content, decreased pH, decreased permeability of Evans blue, and increased LOR and FLG expression [108].

Skin Barrier Dysfunction Stimulates Keratinocytes to Produce TSLP, IL-25, and IL-33 and Promotes Type 2 Immune Deviation
Excessive activation of the IL-13/IL-4-JAK-STAT6/STAT3 axis and subsequent down-regulation of barrier-related proteins are likely to cause skin barrier dysfunction and the development of AD. On the other hand, it is known that epidermal keratinocytes in barrier-disrupted skin promote type 2 immune deviation. Epicutaneous application of hapten and mite antigens to barrier-disrupted skin dysfunction. An IL-13/IL-4-rich milieu up-regulates the production of CCL17 and CCL22, which induce the preferential recruitment of Th2 cells. An IL-13/IL-4-dominant microenvironment also up-regulates the production of IL-5 and CCL26, which attract eosinophils. IL-13 is responsible for the pro-fibrotic process and is also involved in lichenification and tissue remodeling.

Skin Barrier Dysfunction Stimulates Keratinocytes to Produce TSLP, IL-25, and IL-33 and Promotes Type 2 Immune Deviation
Excessive activation of the IL-13/IL-4-JAK-STAT6/STAT3 axis and subsequent down-regulation of barrier-related proteins are likely to cause skin barrier dysfunction and the development of AD. On the other hand, it is known that epidermal keratinocytes in barrier-disrupted skin promote type 2 immune deviation. Epicutaneous application of hapten and mite antigens to barrier-disrupted skin was found to induce increased production of IL-4 and IgE in the regional lymph nodes compared with the levels in control mice with these antigens applied on barrier-intact skin [114]. Keratinocytes obtained from tape-stripped epidermis were also shown to produce larger amounts of CCL17, CCL22, and CCL5 and chemoattract IL-4-producing immune cells and eosinophils compared with the findings in their counterparts with intact epidermis [115]. Moreover, mice kept in dry air conditions show dry skin, with increases in the number of dermal mast cells and histamine concentration [116]. The application of moisturizer was also found to improve the dry skin and normalize the dermal histamine concentration [116]. In normal healthy individuals, decrease of skin water content significantly elevates serum CCL17 levels [93]. These results suggest that dry skin or skin barrier dysfunction can potentiate type 2 immune deviation in mammals.
IL-33 is a member of the IL-1 family that is produced from peripheral tissues and induces type 2-dominant immune deviation [130]. It is overexpressed in keratinocytes derived from tape-stripped, barrier-disrupted epidermis [131]. The expression of IL-33 is also up-regulated in keratinocytes with herpes virus infection [132]. Therefore, the rapid exacerbation of AD in Kaposi's varicelliform eruption may be attributable to IL-33 overexpression [132]. House dust mites also increase the production of IL-25 and IL-33 by keratinocytes via toll-like receptor 6 activation [133]. Many human allergens, such as those of house dust mites, fungi, and pollen, exhibit protease activity [130,134,135]. IL-33 is susceptible to the protease activity of these allergens, generating short-chain IL-33. Notably, this short-chain IL-33 shows much stronger biological activity than its original long-chain counterpart [130]. IL-33 also stimulates ILC2s to express OX40L and up-regulates the Th2 differentiation [125] (Figure 5). IL-25 can also enhance OX40L expression in ILC2s, but its potency is lower than that of IL-33 [125]. On the other hand, TSLP up-regulates the OX40L expression in dendritic cells, but not in ILC2s [122][123][124][125]. IL-33 also stimulates DCs to initiate type 2-prone immune deviation [136].
In a murine Schistosoma mansoni infection model, individual knockout treatment of each of Tslp, Il25, and Il33 was shown not to interfere with IL-13/IL-4 production and subsequent fibrosis, whereas simultaneous knockdown of Tslp, Il25, and Il33 was found to block the IL-13/IL-4 production and subsequent fibrosis [137]. These results suggest mutually overlapping or redundant bioactivity among TSLP, IL-25, and IL-33 [137]. Recent clinical trials have revealed that the anti-TSLP antibody tezepelumab is not efficacious for AD [138]. However, a single injection of the anti-IL-33 antibody ANB020 attenuates the skin symptoms in all 12 patients with AD [139]. These results may further stress the pivotal role of the IL-33-ILC2 axis in the pathogenesis of AD. Figure 5 is a simplified scheme on the pathogenesis of AD. AD is actually quite a heterogeneous skin inflammation, so this scheme may not represent all of its aspects. However, it is now clear that the excessive production of IL-13, but probably not IL-4, in the skin activates the IL-13/IL-4-JAK-STAT6/STAT3 axis, down-regulates the expression of barrier-related proteins, induces barrier dysfunction, up-regulates the production of TSLP/IL-25/IL-33, and further accelerates type 2 immune deviation. The pruritogenic type 2 cytokines IL-31 and IL13/IL-4 stimulate sensory nerves and induce chronic pruritus. Pruritus evokes mechanical scratch, which further disrupts the barrier function and drives a vicious cycle toward type 2 immune deviation.

Th17/Th22 Cells and Chronicity in AD
Excessive IL13/IL-4 signaling is the critical driver in the pathogenesis of AD, while IL-17A-producing Th17 cells are present in the lesional skin of AD [36,[140][141][142]. IL-22 produced from Th17 cells and Th22 cells is also detected at high levels in lesional skin compared with that in non-lesional skin and in chronic rather than acute lesions in AD [17,18,37]. The involvement of IL-17A has also been reported to be conspicuous in AD in Asians [141,142]. Moreover, serum IL-22 levels are significantly associated with the serum levels of CCL17 [143]. It has also been shown that the number of IL-22-producing CLA-positive Th cells is more increased in adult patients with AD than in pediatric ones [37]. The expression of IL-22 likely overwhelms IL-17A expression in the lesional skin of AD [36].
IL-22 increases the proliferative activity of keratinocytes and has been proposed to be involved in the chronicity of AD [144]. The anti-IL-22 antibody fezakinumab is more efficacious in severe than in mild AD [145] and in patients with high rather than low serum IL-22 levels [146]. Moreover, the decreases of IL-13 and IL-22 expression by topical steroid and tacrolimus are associated with the improvement of AD lesions [41,147,148]. These results suggest the role of IL-22 in AD chronicity [145]. In addition, some patients actually manifest psoriasiform eruption with IL-23/IL-17A overexpression during dupilumab therapy, suggesting that IL-23/IL-17A axis is potentially active and meaningful in these particular AD patients [149]. Multipolarity of these cytokines in AD may explain the heterogeneity and age/race differences of AD [141]. On the other hand, blockade of the IL-13/IL-4 signal by dupilumab alone was shown to normalize the increased IL-17A/IL-22 signals including S100A proteins, elafin and IL-23p19 [22], which may support the notion that excessive IL-13/IL-4 signal also causes the elevation of IL-17A/IL-22.
Th17/Th22 cells express CCR6 [150]. CCL20 is the only chemokine to recruit CCR6-positive immune cells [150]. Although epidermal keratinocytes are an abundant source of CCL20, its production is not influenced by IL-13/IL-4 signal [151]. However, the production of CCL20 is rapidly and significantly induced by epidermal scratch injury [152,153], suggesting that itch-scratch behavior itself may trigger CCL20 production to recruit Th17/Th22 cells [151,154,155]. The recruited Th17/Th22 cells may enhance their accumulation because IL-17A further stimulates keratinocytes to produce CCL20 [153]. It is possible that the blockade of itch-scratch behavior by dupilumab may attenuate the infiltration of Th17/Th22 cells.
Th1 signatures such as interferon-γ (IFN-γ) are also detected more often in chronic than in acute lesional skin of AD [16][17][18]. IFN-γ binds a heterodimeric receptor, IFN-γ receptor I (IFNGR1) and IFN-γ receptor II (IFNGR2), and activates downstream JAK1/JAK2 and STAT1 [166]. Conflicting results have been reported on the effects of IFN-γ on the regulation of FLG expression: down-regulation [165] or up-regulation [167]. In our study, IFN-γ was shown to up-regulate the epidermal permeability barrier [106]. It is well known that IFN-γ inhibits Th2 cell differentiation and reduces the production of IL-14 and IL-13 [168,169]. The Th1 cell infiltration may be a compensatory reaction to attenuate excessive type 2 deviation in AD.
Another potential explanation for Th1 and Th17/Th22 cell infiltration in chronic AD is endothelin 1. Keratinocytes are an abundant source of endothelin 1 [170]. Endothelin 1 is abundantly expressed in basal keratinocytes under physiological conditions [119,171], but it is variably overexpressed in inflamed epidermis [119,172]. Endothelin 1 is one of the pruritogenic cytokines and has been confirmed to induce pruritus in mouse and human [173]. As mentioned above, dendritic cells treated with TSLP, IL-25, or IL-33 shift the immune response toward type 2 dominance [122][123][124][125]136]. In sharp contrast to this, DCs treated with endothelin 1 prompt T cells to differentiate toward Th1, Th17, and Th22 cells [172]. Thus, endothelin 1 is possibly involved in the chronic (namely complexed or intermingled) immune response in inflammatory skin diseases [172,174]. In keeping with this notion, topical application of endothelin receptor antagonist was found to attenuate mite-induced dermatitis [62] as well as imiquimod-induced psoriasiform skin inflammation [175]. Notably, there is a mutual feedforward regulatory circuit between IL-25 and endothelin 1. IL-25 up-regulates the production of endothelin 1, while endothelin 1 also up-regulates the production of IL-25 in keratinocytes [119].

IL-13/IL-4-JAK-STAT6/STAT3 Axis and Oxidative Stress
Oxidative stress is one of the most important cellular reactions in dermatitis [176]. The release of IL-1β from dendritic cells is essential for the efficient induction of hapten-specific T cells [177]. Hapten activates Syk and induces the production of pro-IL-1β [177]. Pro-IL-1β has to be cleaved by caspase 1 to produce mature IL-1β before its release. The hapten-mediated generation of reactive oxygen species (ROS) is indispensable for this caspase 1 activation [177]. In vivo, it has been confirmed that ROS are generated in dermatitis [178]. This in vivo generation of ROS can be visualized by observing the rate of reduction of redox compound tempol signal using dynamic nuclear polarization magnetic resonance imaging (DNP-MRI) [178]. It has been confirmed that mite antigen-induced dermatitis generates a large amount of ROS in local inflamed skin [178].
IL-13 is also known as a pro-fibrotic cytokine and is believed to be involved in lichenification in AD or tracheal remodeling in asthma [185,190,191] (Figure 5). IL-13 is more pathogenic than IL-4 regarding fibrosis in atopic disorders [192]. IL-13-mediated periostin production is also involved in atopic fibrosis and the periostin production is dependent on IL-13-induced ROS generation [185]. Antioxidative phytochemicals such as cinnamaldehyde actually activate NRF2 and reduce the IL-13-mediated ROS generation and periostin production [185].
nuclear polarization magnetic resonance imaging (DNP-MRI) [178]. It has been confirmed that mite antigen-induced dermatitis generates a large amount of ROS in local inflamed skin [178].

Mechanisms of Action of Pharmaceutical Agents in AD
In the lesional skin of AD, excessive activation of the IL-13/IL-4-JAK-STAT6/STAT3 axis inhibits the production of barrier-related proteins, disrupts the barrier function, and induces oxidative stress by ROS generation. Pharmaceutical agents used in standard therapy in AD can inhibit the IL-13/IL-4-JAK-STAT6/STAT3 axis. Corticosteroid [193][194][195], tacrolimus [194][195][196][197], and cyclosporine [193,195,197,198] can inhibit the production of IL-13/IL-4 from immune cells. The reduction of IL-13/IL-4 may attenuate the activity of the IL-13/IL-4-JAK-STAT6/STAT3 axis and decrease ROS generation. It has been shown that 0.05% topical betamethasone dipropionate or 0.05% clobetasol propionate reduces the lesional levels of IL-13, which correlates well with the significant improvement of clinical symptoms in AD [41]. The decrease of IL-13 levels is also significantly associated with the improvement of TEWL, decreased expression of CCL17 and CCL22, and restoration of FLG and LOR expression [41]. Notably, in asthmatic patients, tracheal ILC2s are more tolerant of steroid treatment than circulating ILC2s [199,200].
Dupilumab inhibits the IL-13/IL-4-JAK-STAT6/STAT3 axis by interfering with the binding of IL-13/IL-4 and IL-4Rα. Therefore, dupilumab inhibits the elevated production of type 2 chemokines such as CCL17, CCL22, and CCL26 and restores the decreased levels of FLG and LOR [22]. Blockade of the IL-13/IL-4-JAK-STAT6/STAT3 axis using either steroid or dupilumab simultaneously decreases the elevated levels of IL-13, IL-31, IL-17A, and IL-22 [22,41]. These results stress that excessive activation of the IL-13/IL-4-JAK-STAT6/STAT3 axis is the fundamental abnormality in AD and that other immunological responses may be secondary to it.
Coal tar and glyteer are historical remedies for treating dermatitis. They activate the AHR axis, up-regulate the expression of FLG, and restore the skin barrier function [31,92]. In general, mere activation of AHR can only restore the FLG expression, but cannot inhibit ROS production. However, medicinal coal tar and glyteer are antioxidative AHR agonists and can activate AHR as well as NRF2 [31,92] (Figure 6). These AHR and NRF2 dual activators can up-regulate various antioxidative enzymes and neutralize the IL-13/IL-4-mediated generation of ROS [92,[179][180][181]. A reduction of ROS may also down-regulate STAT3 activation [206,207]. Moreover, the activation of NRF2 can potentially inhibit the function of ILC2 [208].
In addition to their effects on keratinocytes, antioxidative AHR ligands can inhibit the phosphorylation of STAT6 and subsequent production of CCL17 and CCL22 in dendritic cells [26]. Antioxidative AHR ligands can also block the enhancing effects of IL-4 on the IL-31-mediated stimulatory function of dendritic cells [214]. Therefore, competitive regulation between the IL-13/IL-4-JAK-STAT6/STAT3 axis and the AHR axis is possibly involved in other biological responses. As their mechanistic features are different, combined treatment with AHR/NRF2 dual activators and IL-13/IL-4-JAK-STAT6/STAT3 axis inhibitors may become a more suitable therapeutic approach for AD. In addition, many other biologic and small molecular drugs including anti-IL-13 antibodies and JAK inhibitors are in clinical trials as reviewed elsewhere [215]. Considering the favorable efficacy of some emerging agents, future therapeutic strategy for AD seems quite promising [215].

Conclusions
In this review article, I mainly describe the regulatory mechanisms of skin barrier-related proteins, focusing on the IL-13/IL-4-JAK-STAT6/STAT3 axis and the AHR axis, which are mutually competing systems. Although AD patients are rather heterogeneous, excessive activation of the IL-13/IL-4-JAK-STAT6/STAT3 axis and subsequent barrier dysfunction appear to be common and cardinal features in all AD patients. AD appears to be characterized by skin inflammation, barrier dysfunction, and chronic pruritus, which are all attributable to excessive activation of the IL-13/IL-4-JAK-STAT6/STAT3 axis and its associated immune reaction. However, it remains unclear whether AHR axis affects pruritus or not. Accumulating evidence underscores the notion that AD can be re-defined as a form of skin inflammation attributable to excessive IL-13/(IL-4) signaling in individuals with atopic diathesis. Targeting the IL-13/IL-4-JAK-STAT6/STAT3 axis as well as the AHR axis is a promising strategy to develop new drugs for AD. The combined use of remedies for these two axes may bring further benefits.
Funding: This work was supported by a grant from the Ministry of Health, Labour and Welfare, Japan (H30-Shokuhin-Shitei-005), and Grant-in-Aid for Scientific Research (C, FAG0K08692) from the Ministry of Education, Culture, Sports, Science and Technology, Japan in 2020.