“Structural imprinting” of the cutaneous immune effector function

ABSTRACT

Keratinization provides tolerance to desiccation and mechanical durability. Loricrin, which is an epidermal thiol-rich protein, efficiently stabilizes terminally differentiated keratinocytes and maintains redox homeostasis. The discovery of the largely asymptomatic loricrin knockout (LKO) phenotype decades ago was rather unpredicted. Nevertheless, when including redox-driven, NF-E2-related factor 2-mediated backup responses, LKO mice provide opportunities for the observation of altered or “quasi-normal” homeostasis. Specifically, given that the tissue structure, as well as the local metabolism, transmits immunological signals, we sought to dissect the consequence of truncated epidermal differentiation program from immunological perspectives. Through a review of the aggregated evidence, we have attempted to generate an integrated view of the regulation of the peripheral immune system, which possibly occurs within the squamous epithelial tissue with truncated differentiation. This synthesis might not only provide insights into keratinization but also lead to the identification of factors intrinsic to the epidermis that imprint the immune effector function.

KEYWORDS:

• Squamous epithelium
• epidermis
• stratum corneum
• metabolism
• KEAP1/NRF2 signaling
• keratinization
• thiol
• disulfide
• loricrin
• dendritic cells
• retinoid

Introduction

Facing the external environment, the epithelial tissue serves as an insulation barrier. The regulation of the paracellular permeability through cell-cell adhesion is a universal strategy for insulation. The prime example could be the intercellular adhesion complex tight junction (TJ), whose components and functions are evolutionary ancient.¹ Transmembrane proteins, such as claudins or occludins, are thought to represent the diffusion barriers, and decreased expression of claudin 1 is associated with ichthyosis¹ and atopic dermatitis (AD) in humans.² Another strategy could be active secretion of protective substances, such as mucus or lipids, to the surface area. Lamellar granules (LGs), which function as lipid storage and secretory organelles, are conserved across many cell types.³ For instance, LG-derived lipid film on the cell surface optimizes lung gas exchange and protects against aggressive gastric juice.³ To cope with the harsh terrestrial environment, the stratified squamous epithelium epidermis is armed with a specialized barrier, the stratum corneum (SC). Densely packed keratin intermediate filaments provide the epidermis with mechanical durability, and LG-derived, covalently bound lipids confer desiccation tolerance, both of which are compared to “the brick” and “the mortar,” respectively.⁴ The epidermal differentiation program promotes the assembly of the macromolecular complex, the cornified cell envelopes (CE).⁵ For the establishment of “the mortar,” the earlier CE precursors envoplakin/periplakin/involucrin participates in the formation of the lipid-based outer coat corneocyte lipid envelope (CLE),⁶ thus serving as a scaffold for the lipid-based paracellular permeability barrier.⁷ For the reinforcement of “the brick,” the later CE precursor loricrin (LOR) facilitates corneocyte maturation through promoting the cross-linking of cytoskeletal proteins at the cell periphery.⁴ Following the fundamental principle that desiccation tolerance the priority in terrestrial life, recent studies on medical genetics have revealed the pathogenic importance of the “broken mortar” situation. This leads to various degrees of epidermal hyperkeratosis.⁸ Given that keratinization is considered an example of adaptive tissue responses, the ichthyosiform hyperkeratosis may be an epitome of tissue maladaptation compensating for structural defects or metabolic disorders.⁹

The “thiol-based sensor–effector apparatus” the Kelch-like erythroid cell-derived protein with the cap-n-collar homology-associated protein 1 (KEAP1)/NF-E2-related factor 2 (NRF2) signaling pathway mechanistically revealed its own importance as a regulator of keratinization.^10,11 Several mouse models have also shown that aberrant activation of the stress signaling pathway not only lead to autosomal recessive congenital ichthyosis (ARCI)-like hyperkeratosis,^12,13 but also contributes to the pathogenesis of the most common skin disorders, such as atopic dermatitis (AD),¹⁴ psoriasis,¹⁵ or acne.¹⁶ In accordance with the “brick and mortar” model,⁴ the major later CE precursor LOR was dispensable for the lipid-based SC permeability barrier per se.^4,17 Although somewhat anticipated from the structural basis,⁴ the discovery of the very mild LOR knockout (LKO) phenotype was counterintuitive and rather surprising.^17,18 Nonetheless, LKO mice provided us opportunities to uncover the backup responses that maintain the quasi-homeostasis, both structurally and functionally.^17,19–21

Because LOR is nonessential for the formation of the lipid-based SC permeability barrier,^17,19–23 there is a need to revisit the fundamental principle of keratinization (i.e. sulfur metabolism^24,25); it could be hypothesized that the LKO epidermal microenvironment, including altered metabolism of sulfur,²⁰ serves as an intrinsic microenvironmental cue that affects even the immune effector function. To be specific, as in AD,²⁶ the attenuated cell-mediated immune response in LKO mice^27,28 possibly suggests the immunoanatomical principle of epidermal differentiation: structural maturation possibly establishes immunological competence, or conversely, desiccation tolerance can provide reinforcement of immunological tolerance. In broad terms, this may be an answer to the conundrum, “Why is dry-surfaced skin highly immunogenic as compared with the wet-surfaced inner mucosa?”^29,30 This theoretical and hypothetical article, in an attempt to synthesize the structure and function of the epidermis from multiple points of view, focuses on the principle of epidermal differentiation with special attention paid to the cross talk between structure and metabolism.

Beyond gene hunting

The past several decades were an era of advances in medical genetics, which revealed previously uncharacterized pathomechanisms of inherited diseases. The major pillars of such discoveries were loss-of-function (LOF) mutations in transglutaminase 1 (TGM1),³¹ ATP-binding cassette subfamily A member 12 (ABCA12),^32,33 and filaggrin (FLG),³⁴ resulting in ARCI1 (lamellar ichthyosis, OMIM #190195), ARCI4B (harlequin ichthyosis, OMIM #2425500), and ichthyosis vulgaris (IV; a common dry skin condition, OMIM #146700), respectively. TGM1 catalyzes the ε-(γ-glutamyl) lysine cross-linkages between protein–protein or protein–lipid in the CLE, whereas ABCA12 transports ω-hydroxyceramides to the CLE.⁴ FLG undergoes proteolytic cascades and ultimately serves as a humectant in corneocytes.³⁵ Because a major threat to fully terrestrial lifestyles is desiccation, higher levels of transepidermal water loss can evoke protective responses and manifest as pathologic keratinization.^9,33 Thus, it is likely that clinical phenotypes depend on the degree of the breach of the epidermal paracellular barrier, ranging from mild, “flaky” retention keratosis in IV^34,36 to severe, platelike hyperkeratosis in ARCI4B.^8,32 It is important to note here again that the phenotypes are derived solely from the host’s tissue response¹⁰ rather than from the defects in structure or function. In particular, this could be the case with multigenic disorders because there is often a distance between disease-susceptibility genes and disease phenotypes. For instance, to manifest as AD, the LOF mutation in the FLG gene (ATOD2, OMIM #605803) requires another predisposing factor of internal (gene variants)^37,38 or external origin (environmental factors). The endogenous factor not only depletes the gene expression product, but it can sometimes also evoke backup responses,^9,17 often rendering the disease phenotypes invisible.¹⁷ The regulation of gene expression is profoundly affected by an exogenous factor, thus evoking adaptive tissue responses.¹⁴ Therefore, regardless of the consequences of gene variants (homeostasis, disease, or death), the phenotype consists of an integrated sequence of biological reactions that involve many forms of interactions between the surrounding environment and genes.³⁹

Cross talk between structure and epidermal metabolism: ARCI as an example

The epidermis is considered as a regenerative tissue, and impaired metabolism affects the permeability barrier of the SC.⁸ To date, autosomal recessive conditions that are relatively rare have been a clear demonstration of the genotype–phenotype correlation, and we use ARCI as an example. As shown in Table 1, ARCI1 is caused by an LOF mutation in TGM,³¹ which catalyzes the structural basis of CLE, and other forms of ARCI result from gene mutations due to impaired lipid metabolism (ARCI2–ARCI10 and ARCI13–ARCI14)^{40,41,42,43,44,45,46,47} or proteolytic processing of FLG (ARCI11–ARCI12).^48,49,50 The accumulated evidence suggests that either impairment in the formation of CLE (the “mortar”) or production of corneocyte humectant is the cause of the ARCI phenotypes, most of which are phenocopied by gene deletion in mouse. Reactive hyperkeratosis and accompanying inflammation constitute the universal ARCI phenotype, and these features are caused by the host’s adaptive response but not the gene mutations themselves.¹⁰

Table 1. ARCI and causative mutated genes

ARCI #	Chromosome	Gene	Function	Reference #
1	14q12	TGM1	CLE cosslinking	21
2	17p13.1	ALOX12B	Linoleic acid metabolism	33
3	17p13.1	ALOXE3	Linoleic acid metabolism	33
4	2q35	ABCA12	Ceramide linoleic ester formation	22, 23, 34
5	19p13.12	CYP4F22	Catalyzing ultra-long-chain fatty acids	35
6	5q33.3	NIPAL4 (Ichthyin)	Mg^2+ transport	38
7	12p11.2-q13.1	N/D	N/A	N/A
8	10q23.31	LIPN	Lipase	32
9	15q26.3	CERS3	C24-ceramides synthesis	36
10	6q21.31	PNPLA1	Esterification of linoleic acid	31
11	11q24.3	ST14 (Matriptase)	Filaggrin processing	41
12	19p13.12	CASP14	Filaggrin processing	40
13	12q13.3	SDRC9C7	Ceramide linoleic ester formation	37
14	19q13.33	SULT2B1	Sulfoconjugation of neutral steroids and sterols	39

Abbreveations: TGM1, Transglutaminase; ALOX12B, Arachidonate 12-lipoxygenase, ALOXE3, Arachidonate lipoxygenase 3; ABCA12, ATP-binding cassette sub-family A member 12; CYP4F22, Cytochrome P450, family 4, subfamily F, polypeptide 22; NIPAL4, NIPA-like domain-containing 4; LIPN, Lipase family member N, CERS3, Ceramide synthase 3; PNPLA1, patatin-like phospholipase domain-containing 1; ST14, Suppressor of tumorigenicity 14 protein; CASP14, Caspase 14; SDRC9C7, Short Chain Dehydrogenase /Reductase Family 9 C Member 7; SULT2B1, Sulfotransferase family 2B member 1.

Keratinization as sulfur metabolism

It is well known that epidermal differentiation involves the metabolism of sulfur, and thiol (–SH) groups of the proliferative basal layer are converted to covalent disulfide (–S–S–) bridges of keratin molecules.²⁵ In support of this principle, radioactive tracer experiments have revealed that the keratinocytes (KCs) in the uppermost living layer (stratum granulosum) uptake radiolabeled cystine (disulfide cross-linked cysteine) nearly twice as much as that in the basal or spinous layer.²⁴ These observations imply that KCs in the stratum granulosum undergo robust sulfur metabolism to finalize epidermal differentiation.⁵¹ This notion could be supported by the clinical evidence indicating that the topical application of a cysteine antioxidant N-acetyl cysteine (NAC) improves the hyperkeratotic phenotype in ARCI1.⁵² This suggests that NAC serves as a free cysteine donor (nucleophile), reduces disulfide (–S–S–) bridges, and attenuates the hyperkeratosis.

The KEAP1/NRF2 system in keratinization

The following question is raised: What causes the hyperkeratotic tissue response? Vertebrates are armed with the KEAP1/NRF2 signaling pathway, a thiol-based sensor-effector apparatus.¹¹ The phylogenetically conserved interaction between the thiol-rich cytoskeletal protein KEAP1 and transcription factor NRF2 regulates myriad stress responses.¹¹ KEAP1 forms a complex with cullin 3-based ubiquitin ligase in the steady state and subjects NRF2 to proteasome-mediated degradation in the cytosol, thus repressing the activity of NRF2.^10,11 The oxidative stress that causes redox imbalance, such as reactive oxygen species or electrophiles, modifies KEAP1’s cysteine residues and results in conformational changes.^10,11 This redox exchange unleashes NRF2’s transcriptional activity and upregulates the detoxification/antioxidant enzyme, ATP-binding cassette transporters, and other biomolecules that are associated with stress responses.⁵³ However, witnessing the consequence of the genetic deletion of Keap1 in mice would have been surprising. Although the Keap1-null mutation systemically enhanced the detoxifying responses, the mice suffered from postnatal malnutrition because of keratinized materials that filled the lumina of the esophagus and forestomach.¹³ The concomitant deletion of NRF2 rescued the severe phenotype, which suggests that the KEAP1/NRF2 system is an important keratinization regulator.¹³ In accordance with the inner squamous epithelium phenotype, the epidermis exhibited orthohyperkeratosis, which somewhat resembled ARCI.¹³ In another setting, the ARCI-like hyperkeratosis was reproducible, in which mutant NRF2 incapable of interacting with KEAP1 is overexpressed in the epidermis.¹² Taken together, the KEAP1/NRF2 system serves as a central regulator of sulfur metabolism and thus keratinization in the epidermis.¹⁰

The consequence of the absent epidermal thiol-rich protein

Despite the fact that the keratin intermediate filament constitutes the structural backbone of the epidermis, the amino acid composition of the differentiated keratins K1/K10 cannot explain the unusually high sulfur content in the differentiated epidermal layer.⁵¹ Therefore, researchers hypothesized the existence of a sulfur-rich CE precursor, and it was not until the 1990s that Mehrel and Roop et al. cloned and characterized the sulfur-rich CE precursor LOR.⁵¹ Subsequently, LKO mice were generated, but because CE maturation does not take place in submerged KC cultures,^5,54 the mouse model taught us the consequence of the loss of the major thiol-rich epidermal structural protein. The most important finding from LKO mice is that LOR is nonessential for the lipid-based SC permeability barrier¹⁷ but is mandatory for corneocyte maturation.^23,55 We attribute the rather unpredicted outcome at that time to the following two causes. First, the LG system is ubiquitous and is even observed in nonepithelial tissue.³ Second, the CLE forms outside corneocytes subsequent to scaffolding, and the following reinforcement process precipitates disulfide (–S–S–) at the inner periphery, but not the outside, of corneocytes.⁴ An additional important finding is the induction of a compensatory response, which induces a massive amount of the cross-bridging CE precursor, the small proline-rich protein (SPRR)^17,20 or the late cornified envelope (LCE).¹⁹ These alternative reinforcing CE components are somewhat structurally similar to LOR but less efficient executors of ε-(γ-glutamyl) lysine or disulfide (–S–S) cross-linkages.¹⁹ Consequently, the LKO epidermis suffers from structural fragility, rather than an impaired SC permeability barrier.¹⁷ Nonetheless, as was the case with ARCI, which can be caused by defects in either the structure or the metabolism,⁸ LKO affects the epidermal sulfur metabolism; KEAP1 senses the cytosolic redox imbalance, and NRF2 serves as an important effector that massively induces the SPRR²⁰ or the LCE.¹⁹ Because both the SPRR^56,57 and LCE^58,59 are induced in response to primary injury, such as ultraviolet irradiation or tape stripping, it may be likely that the LKO phenocopies an adaptive tissue response, and the superior nucleophilic property of SPRR⁵⁶ would prevent further oxidative stress-mediated tissue injury, alongside enhanced LG biogenesis (unpublished observation).

Metabolic control of the adaptive tissue response

Now that we know that LKO affects the epidermal metabolism rather than the lipid-based SC permeability barrier, let us take (1) the electrophilic counterattack response⁶⁰ and (2) contact hypersensitivity (CHS)⁶¹ as redox-mediated adaptive tissue responses.

7,12-dimethylbenz(a)anthracene (DMBA) is a prototypic environmental carcinogen. DMBA is converted into the ultimate mutagenic carcinogen DMBA-3,4-diol-1,2-epoxide (DMBADE) through the cytochrome P450-mediated metabolic conversion of the host.⁶² DMBADE possesses a strong electrophilic property, forms adducts with genomic DNA, and exerts genotoxicity.¹¹ The host’s antioxidative response is raised by a weak electrophilic stress, which could protect against higher doses of the same or other electrophiles, including the carcinogen. Because the KEAP1/NRF2 signaling pathway plays a pivotal role in phase II detoxification and is activated in LKO mice, we questioned whether the structural protein LOR, which promotes corneocyte maturation,^23,55 is simply chemopreventive or whether the NRF2-mediated compensatory response in LKO mice could be equivalent to the electrophilic counterattack response.⁶⁰ In line with the susceptibility against ultraviolet B,⁵⁵ LOR protected against the chemical carcinogen DMBA.²⁸

CHS is a cell-mediated adaptive tissue response in which a strong, antigen-specific cytotoxic immune response is evoked at a place distant from the primary exposure site.⁶³ Most reactive haptens are electrophilic and potentially activate the KEAP1/NRF2 signaling pathway.¹⁴ Nevertheless, if the concentration is low enough or devoid of irritancy (such as the lipophilicity),⁶⁴ CHS does not ensue, and this phenomenon is known as low zone tolerance.⁶⁵ Similar to the defects in the lipid-based SC permeability barrier^37,38 or FLG deficiency,³⁴ reduced tolerance to desiccation might increase sensitivity. However, it was found that LKO mice tolerated desiccation well, LG biosynthesis was increased (unpublished observations), and corneocytes were abundant with unlinked FLG.^23,55 Accordingly, we recently discovered that LKO mice exhibited attenuated CHS in the effector phase.²⁸

Immunoanatomical bases of the squamous epithelium

Both CHS and tumor eradication involve cytotoxic immune responses that are strong, self-destructive, and potentially harmful.⁶¹ Regardless of the antigens involved, the epidermis is more immunogenic than the mucosal epithelium.⁶⁶ For instance, one of the most common sites of graft-versus-host reaction and adverse drug reaction is the skin.⁶⁷ Likewise, percutaneous entry of allergens favors type 2 immunological memory (atopy), whereas permucosal routes lead to tolerance (regulatory T-cell memory).²⁹ It is presumed that the epidermal tissue renders immunogenicity, which might make it a more amenable target organ for topical immunotherapy or immune checkpoint inhibition⁶⁸ as compared with nonkeratinizing squamous mucosa.^69,70 We compared, from an immunoanatomical perspective, the epidermis with nonkeratinizing squamous epithelia.

The cornea covers the eye’s anterior chamber surface and allows efficient penetration of visible light. Because clear eyesight requires corneal transparency, the eye needs to tolerate foreign materials, such as microbes or viruses.^66,71 This strategic hyporesponsiveness (attenuated cytotoxicity) – or more precisely, tailor-made immune effector function – has been simply referred to as the immune privilege or the anterior chamber-associated immune deviation.⁷¹ This immunoanatomical peculiarity renders corneal transplantation one of the most successful solid organ graft procedures, with success rates as high as 90%.⁷¹ In cases of severe herpetic viral infection, the stromal Fas ligand prevents the destruction of the immune-mediated tissue via the induction of apoptotic death of infiltrating effector cells.⁷² By contrast, antigen-presenting cells deliver tissue-associated signals and maintain immune homeostasis in the steady state,⁷³ but autoimmune disorders can result from the absence of dendritic cells (DCs).^74,75 Squamous epithelia, including the cornea, are home to Langerhans cells (LCs), which sample tissue antigens and migrate to a secondary lymphoid organ.^76,77 By definition, LCs constitutively express major histocompatibility complex (MHC) class II antigen and C-type lectin receptor CD207/Langerin,^76,77 corneal LCs lack MHC class II⁷⁸ but express CD207 and CD11c (the conventional DC marker).⁷⁹ Because of the importance of CD207 in maintaining tolerance in the squamous epithelium⁷⁶ as well as the intestinal epithelium,⁸⁰ it is conceivable that the corneal epithelial LCs may endow the nonkeratinizing epithelium with a protective barrier⁸¹ via tolerizing exogenous agents that invade through the paracellular route, such as bacteria or viruses.

In the same vein, permucosal exposure of foreign antigens, particularly through the intestinal epithelium, favors tolerance (regulatory T-cell memory), whereas exposure through the broken SC permeability barrier breaks tolerance.⁶⁶ It is worth noting, however, that the percutaneous route does not necessarily break tolerance (and thus atopy) because the induction of transcutaneous tolerance against exogenous antigens is feasible.⁸² Instead, it would be legitimate to consider that the nature (or the class) of the immune responses depends on the status of the tissue in which the biological response arises.⁶⁶

Metabolic imprinting of cutaneous immune effector functions

Although myriad epithelial tissue-derived cytokines that skew immune effector classes toward type 2 immunity, such as interleukin (IL)-33, IL-25, or thymic stromal lymphopoietin,⁸³ have been characterized over the past decades, the important example of the metabolic cues mediated by the lipophilic nutrients vitamin A (VA) and vitamin D (VD).⁸⁴ It appears that the VA and VD signaling pathways work somewhat antagonistically in the epidermis. The former promotes the hydrolysis of polar lipids and plays a role in the maturation of lipid-based SC permeability (the “mortar”),^85,86 whereas the latter promotes the assemblage of the cytoskeletal filament network (the “brick”).⁸⁷ Persistent activation of epidermal VD receptor signaling through the application of calcipotriol (a low-calcemic calcitriol analog)⁸⁸ or the genetic ablation of the retinoid X receptor⁸⁹ results in the production of AD-like phenotypes, whereas systemic retinoid treatment proves useful for the management of chronic hand eczema⁹⁰ as well as ARCI,⁹¹ both conditions that are closely associated with “broken mortar” situations.^8,9

When taking into consideration the regulation of immune effector functions in the periphery, lessons from VA deficiency (VAD) may be important.^92,93 With the aid of food-derived VA, small intestine (SI) epithelial cells produce a tissue-derived cue through the metabolic cascade, which converts retinols (VA) into retinoic acids (RAs)^92,94,95 (Figure 1). SI epithelial cells express retinaldehyde dehydrogenase 1 (RALDH1) and can convert cellular VA storage (retinyl esters) into RAs, which, in turn, activate RALDH2 expressed in DCs, further amplifying the tissue-derived cue.⁹²

Figure 1. Overview of the retinoid metabolism

Retinol is bound in the cytosol to a cellular retinoid binding protein (CRBP). The red text shows the catalytic enzymes, whereas the blue text presents the binding proteins. 1. Extracellular transport Retinol is secreted into the circulation by retinol-binding protein (RBP) bound with transthyretin (TTR). 2. Intracellular storage The storage of retinyl ester occurs in cytoplasmic lipid droplets. The esterification of retinol is regulated by lecithin:retinol acyltransferase (LRAT) and retinyl ester hydrolase (REH). 3. Nuclear receptor-mediated signaling Alcohol dehydrogenase (ADH) and retinaldehyde dehydrogenase (RDH) convert retinol into retinaldehyde, which, in turn, is metabolized by retinaldehyde dehydrogenase (RALDH), yielding bioactive retinoic acid (RA). The signaling pathways mediated by the transcription factors RA receptor (RAR) or retinoid X receptor (RXR) are activated by a cellular RA binding protein (RABP)-bound RA. RA is converted into polar metabolites by cytochrome P450 family (CYP) members.

Because this positive feedback loop mediated by the RA receptor (RAR) maintains immune effector function, VAD patients with infectious diseases suffer from persistent diarrhea, which is ameliorated by VA supplementation.⁹³ Other salient features of VAD are ectopic keratinization of nonkeratinizing squamous epithelia and squamous metaplasia,⁹⁶ the latter of which accompanies the appearance of CD207⁺ cells.^80,97 This might be analogous to the development of the epidermal LC network in mice. LC precursors that originate from the fetal liver or yolk sac are seeded into the epidermis around embryonic day 18 and expand rapidly during the neonatal period⁹⁸ depending on the relatively low levels of RAR signaling in the CD11c⁺ lineage (conventional DC).⁸⁰ Similar to SI epithelium,⁹² RALDH3-expressing epidermal KCs can augment RAR signaling,⁹⁹ and defective signal processing results in a dry, scaly VAD skin phenotype.^85,86 For this reason, it is tempting to speculate that this epidermal metabolic cue profoundly affects cutaneous immune effector function and that LCs (or dermal DCs) in the vicinity of KCs deliver and/or amplify the signaling, downstream of which are cutaneous immune effector functions.⁸⁰

Conclusion

Because deviated cutaneous immune effector function, but not defective permeability, is unpredictable in the absence of the thiol-rich protein LOR, our interest in the immunological consequence of the absent epidermal thiol-rich protein has grown.^17,18 However, after reviewing the evidence on this topic that has accumulated over the past decades, we acknowledge here that LOR primarily functions as the “brick” protein that establishes the epidermal sulfur metabolism, which is the la raison d’être of the epidermis, other than the lipid-based SC permeability barrier.¹⁰⁰ This important epidermal structural protein not only reinforces structural integrity⁴ but also supports the robustness of immune effector functions, which could manifest as a strong cytotoxic immune response to CHS.²⁸ Although these preliminary findings are intriguing, the following questions persist: (1) What are the metabolic cues located downstream of LKO, other than the KEAP1/NRF2 signaling pathway? Can we define VA- or VD-mediated signaling pathways? (2) Is the LKO-mediated signaling received by LCs (or dermal DCs) merely enhanced migration or do factors intrinsic to the epidermis matter more? (3) Is DC migration merely the consequence of relative epidermal instability or any other cue (i.e. cytokines or lipid-mediated factors)? (4) Are the detuned immune effector functions analogous to anterior chamber-associated immune deviation, and if so, are deviations in the subset of LCs or other innate/adaptive immune cells induced by LKO? (5) What is the significance of having LOR in the stratified squamous epithelium, functionally, phylogenetically, and so forth?

Disclosure statement

The authors state no conflict of interest.

Additional information

Funding

This research was supported in part by a JSPS KAKENHI grant, Grant-in-Aid for Research Activity Start-up (16H06663 to YI), and Early-Career Scientists (18K16018 to YI).