In addition to providing a protective shield against external insults and dehydration, the epidermis is a major source of vitamin D for the body. Although much of the provitamin D in our bodies is activated by hydroxylation in the liver and kidney, the epidermis is also capable of producing active vitamin D3, which may play a role locally. Under the influence of ultraviolet exposure, the photolysis of 7-dehydrocholesterol results in the endogenous production of vitamin D. In addition, keratinocytes also possess the necessary enzymes (e.g., CYP27A1, CYP27B1, 25-hydroxylase, and 1-hydroxylase) to further metabolize vitamin D to its active form , 1,25-dihydroxyvitamin D (1,25(OH)₂D) (Bikle 2011, 2012).

The most striking feature of VDR-knockout mice is the development of alopecia, which is also found in many patients with mutations in the VDR, a condition referred to as hereditary vitamin D resistance (Malloy et al. 1999; Li et al. 1997; Bikle et al. 2006). These mutant mice also exhibit a defect in epidermal differentiation, as exhibited by reduced levels of involucrin and loricrin, as well as a loss of keratohyalin granules (Xie et al. 2002). They also exhibit a reduction in the lipid content of the lamellar bodies concomitant with a reduction in glucosylceramide production and transport to the lamellar bodies, which leads to a defective permeability barrier. The activation of VDR by 1,25(OH)₂D exerts an antiproliferative effect by increasing cyclin-dependent kinase inhibitor activity, reducing proliferation in the basal layer of the epidermis, and promoting the sequential differentiation of keratinocytes to form the upper layers of the epidermis. 1,25(OH)₂D increases the expression of involucrin, transglutaminase, loricrin, filaggrin, phospholipase C (PLC), and the calcium receptor at sub-nanomolar concentrations (Smith et al. 1986; Hosomi et al. 1983; Bikle and Pillai 1993; Pillai and Bikle 1991). The prodifferentiation effect of the VDR is thought to be partly due to its involvement in the regulation of intracellular calcium concentrations in addition to its effect on the promoter activity of numerous keratinocyte differentiation genes (Bikle et al. 2001, 2002). Increased expression of the calcium receptor renders the keratinocytes more sensitive to the prodifferentiating action of calcium. All of the PLC family members are induced by 1,25(OH)₂D and calcium, and blocking PLC-γ1 expression prevents both 1,25(OH)₂D-stimulated and calcium-stimulated differentiation. Calcium and 1,25(OH)₂D also functionally interact to induce involucrin and transgluataminase (Bikle et al. 2004; Bikle 2004). A possible mechanism for this synergy regarding the induction of the involucrin gene is that the calcium response element and vitamin D response element in the involucrin promoter are in close proximity. A recurring mechanistic action for VDR and other nuclear hormone receptors is the pivotal involvement of transcription factor AP-1. Mutations in the AP-1 site of the calcium response element block both calcium and 1,25(OH)₂D induction of the involucrin genes; however, mutations of the vitamin D response element only block its response to 1,25(OH)₂D. The specificity of VDR action within the skin is attributed, at least in part, to the involvement of different coregulators (Oda et al. 2004, 2009). For the proliferation of keratinocytes in the basal layer of the epidermis, the mediator complex Vitamin D receptor interacting protein (DRIP) is the dominant coregulator. In the more differentiated keratinocytes of the epidermis, the steroid receptor coactivator (SRC) complexes, which share similar nuclear receptor binding determinants with DRIP but constitute functionally distinct complexes, dominate the VDR functions. At a high-nanomolar range of 1,25(OH)₂D, the antiproliferative effect of VDR is accompanied by reduced c-myc and cyclin D1 expression and increases in the cell cycle inhibitors p15, p21, and p27. In addition, agonist-activated VDR regulates the processing of long-chain glycosylceramides, which are critical for the formation of the permeability barrier (Bikle 2011). Thus, owing to its clear prodifferentiation effects, 1,25(OH)₂D is used in the treatment of psoriasis (Samarasekera et al. 2013). Prescription cream or solution containing calcipotriene (Dovonex, Sorilux) is used to treat mild to moderate psoriasis, but this treatment cause irritation on sensitive area of the skin. Calcitriol (1,25(OH)₂D), marketed under various trade names including Rocaltrol (Roche), Calcijex (Abbott), Decostriol (Mibe, Jesalis), and Vectical (Galderma), is equally effective and possibly less irritating than calcipotriene.

The dermatological use of retinoids (vitamins A derivatives) preceded the description of their mechanism of action and the discovery of their receptors (Rees 1975). There are two major groups of retinoid-responsive nuclear hormone receptors: the RARs and the RXRs. Each of these groups of retinoid receptors has three subtypes (α, β, γ). RARs and RXRs are activated by 9-cis-retinoic acid, while all-trans-retinoic acid only activates RARs (Fisher et al. 1995; Fisher and Voorhees 1996). The functional retinoid receptor is a heterodimer of one RAR and one RXR molecule. RXR is also the obligate dimerization partner of other nuclear hormone receptors, such as PPAR, VDR, and LXR; thus, the effects of retinoids are expected to be wide-ranging. In the epidermis, retinoid signaling is carried out by RARα, RARγ, RXRα, and RXRβ. The protein expression levels of RXRs far exceed those of RARs, suggesting that RXRs can also partner with other nuclear hormone receptors present in the same cell. RARγ and RXRα are the most abundantly expressed retinoid-responsive nuclear hormone receptors in both human (Fisher et al. 1994) and mouse epidermis (Darwiche et al. 1995).

When topically applied to adult skin, retinoic acid induces epidermal hyperplasia that results from hyperproliferation of basal keratinocytes, which leads to thickening of the differentiated spinous and granular layers (Fisher and Voorhees 1996). Retinoic acid treatment also decreases the integrity of the stratum corneum, impairing the adult skin permeability barrier and increasing trans-epidermal water loss (Elias et al. 1981). Despite the well-known effect of retinoic acid in induced epidermal hyperplasia, the skin of RARα- and RARβ-null mice appeared normal, while the skin of RARγ-null mice displayed minor defects of granular keratinocyte differentiation (Chapellier et al. 2002; Lohnes et al. 1993; Lufkin et al. 1993; Ghyselinck et al. 1997). The skin of the mouse devoid of all three RARs exhibited an epidermal phenotype similar to that of the RARγ-null mouse, suggesting little functional redundancy between RARα and RARγ for the proliferation and differentiation of adult keratinocytes. Importantly, it also indicated that RARs in basal keratinocytes do not perturb homeostatic epidermal self-renewal, suggesting that RAR-dependent signaling is dispensable for homeostasis. As expected, retinoic acid treatment of control mice led to marked epidermal thickening associated with increased proliferation. However, this hyperproliferation was not observed in retinoic acid-treated skin that was devoid of RARγ specifically only in the suprabasal layer, or in RARγ-null mice. Thus, RARγ is required in suprabasal keratinocytes for RA-induced epidermal hyperplasia, partly through increased paracrine signaling involving heparin-binding epidermal growth factor (EGF)-like growth factor (HB-EGF) (Li et al. 2000; Chapellier et al. 2002). Although RXRα and RXRβ were detected in epidermal keratinocytes, genetic evidence indicated that RXRα has a clearly dominant role. Indeed, the skin of adult RXRβ-null mice appeared normal. Interestingly, the mice with an epidermal-specific deletion of RXRα developed progressive alopecia with typical features of degenerated hair follicles in addition to utriculi and dermal cysts, which can all be attributed to defects in hair cycles (Li et al. 2000). With respect to hair follicle development, the epidermal-specific RXRα knockout mice phenocopied the VDR-null mice (Li et al. 2000, 2001). In contrast to VDR-null mice, RXRα mutant mice also exhibited interfollicular keratinocyte hyperproliferation, as well as abnormal terminal differentiation and increased dermal cellularity associated with a skin inflammatory reaction (Li et al. 1997, 2000; Yoshizawa et al. 1997). These features most likely reflect the involvement of other heterodimeric partners of RXRα, such as the RARs, thyroid hormone receptors, and the PPARs.

The in vitro and in vivo effects of retinoic acid on keratinocytes differ (Törmä 2011). For example, retinoic acid treatment induced keratinocyte growth arrest in in vitro monolayer keratinocyte culture, while it stimulated keratinocyte proliferation in reconstructed three-dimensional (3D) human skin, similar to when topically applied to the skin in vivo. Furthermore, retinoids also induced different keratin expression profiles in cultured monolayer keratinocytes compared with 3D skin equivalents (Törmä 2011). Keratin gene expression is affected by retinoids in many complex ways, and it is obvious that classical activation of retinoic acid response elements (RAREs) by RAR/RXR heterodimers is not always involved. For example, a number of keratin genes (e.g., genes encoding keratins 5, 6, 14, and 17) reportedly carry positive or negative RAREs, while no RAREs have been identified in keratin 2 or keratin 4 genes (Kerns et al. 2010; Ma et al. 1997). Thus, it is likely that these latter two retinoid-regulated keratins are among the many genes that are indirectly influenced by retinoids through mechanisms that do not require RAREs (Lu et al. 1994; Balmer and Blomhoff 2002; Radoja et al. 1997). Among the indirect mechanisms through which retinoic acid and its receptors regulate the differentiation and proliferation of epidermal keratinocytes is the antagonistic activity of AP-1 and NF-κB (Törmä 2011).

PPARs are among the most recently identified nuclear hormone receptors. They have attracted attention due to their crucial role in lipid homeostasis and their potential to regulate cell differentiation (Desvergne and Wahli 1999; Icre et al. 2006; Michalik and Wahli 2007). PPARs form permissive heterodimers with RXR, thus, allowing both the ligands of PPAR and RXR to regulate the transcription of PPAR target genes (Tan et al. 2005). None of the three PPARs (α, γ, and β/δ) are detectable in the adult mouse interfollicular epidermis (Michalik et al. 2001). The expression of PPARα and PPARβ/δ were transiently upregulated at wound edges after injury. Epidermal PPARβ/δ expression is also stimulated by phorbol esters, hair plucking, and epidermal inflammation (Tan et al. 2001, 2003, 2004b; Michalik et al. 2001). Although all three PPARs are detectable in the human epidermis, PPARβ/δ is the most abundant subtype, followed by PPARα and PPARγ.

Macroscopic examination of the skin of adult PPARα-knockout mice showed normal skin architecture (Lee et al. 1995). However, a detailed microscopic analysis of these mutant mice revealed a thin stratum granulosum with focal parakeratosis, suggesting impaired keratinocyte differentiation. Fetal epidermal development in PPARα-knockout mice was delayed, with defects in the formation of the stratum corneum. The overexpression of PPARα in the epidermis of transgenic mice resulted in impaired embryonic epidermal development associated with a thinner epidermis and fewer hair follicles than normal mice; however, no abnormality was detected in the adult epidermis (Yang et al. 2006). These observations suggest that PPARα may be important for embryonic epidermal development, but dispensable for adult epidermal homeostasis. Consistent with the very low level of PPARγ expression in the epidermis, no defect in epidermal maturation was observed in PPARγ-heterozygous or PPARγ-knockout mice derived from placental rescue, demonstrating that PPARγ is not required for epidermal differentiation (Rosen et al. 1999; Barak et al. 1999). Histological analysis of PPARβ/δ-knockout mice did not reveal any defect in skin epidermal architecture during fetal development or adulthood (Michalik et al. 2001; Peters et al. 2000). However, PPARβ/δ-knockout mice exhibited retardation of postnatal hair follicle morphogenesis, particularly at the hair peg stage (Di-Poï et al. 2005). This finding was attributed to the PPARβ/δ-mediated temporal activation of the antiapoptotic Akt1 pathway in vivo, which protects keratinocytes in hair pegs from apoptosis and is required for normal hair follicle development (Di-Poï et al. 2005).

PPARα, PPARβ/δ, and PPARγ are ligand-activated nuclear hormone receptors, and their agonists have been shown to lead to the stimulation of the expression of numerous genes necessary for the formation of the cornified envelope and lamellar membrane barrier. Such genes include involucrin, filaggrin, loricrin, transglutaminase 1, aquaporin 3 (AQP3), angiopoetin-like 4 protein (ANGPTL4), and the ATP-binding cassette subfamily G members 1 and 12 (ABCG1, ABCG12) (Pal et al. 2011; Hanley et al. 1997; Mao-Qiang et al. 2004; Jiang et al. 2008, 2010, 2011). AQP3, a member of the aquaglyceroporin family, which transports water and glycerol, is robustly expressed in the epidermis and plays important roles in stratum corneum hydration, permeability barrier function, and wound healing (Jiang et al. 2008). ABCG1 is expressed in cultured human keratinocytes and murine epidermis and is induced during keratinocyte differentiation, with increased levels observed in the outer epidermis (Jiang et al. 2010). ABCG1-null mice display abnormal lamellar body contents and secretion leading to impaired lamellar bilayer formation, indicating a potential role for ABCG1 in normal lamellar body formation and secretion (Jiang et al. 2010). Using 3D skin substitutes, ligand-activated PPARβ/δ indirectly stimulated keratinocyte differentiation, as is required for de novo gene transcription and protein translation (Pal et al. 2011). PPARβ/δ stimulates the expression of ANGPTL4; deficiency of ANGPTL4 in human cultured keratinocytes and the skin of mice results in diminished expression of various protein kinase C isotypes and phosphorylated AP-1, whose roles in keratinocyte differentiation are well established (Pal et al. 2011).

Interestingly, many effects of PPAR agonists on epidermal differentiation and lipid synthesis have also been observed with respect to agonists of another nuclear hormone receptor, LXR. Such findings suggest that both PPAR and LXR nuclear hormone receptors likely mediate the expression of differentiation-associated genes through a common pathway (Feingold and Jiang 2011; Demerjian et al. 2009; Man et al. 2006; Feingold 2007; Schmuth et al. 2008). The promoter regions of many genes whose expression levels increased during differentiation, including involucrin, loricin, and transglutaminase 1, have AP-1 binding sites. ANGPTL4, encoded by a PPAR target gene, mediates the activation and binding of JUNB and c-JUN to the promoter region of human involucrin and transglutaminase type 1 genes, respectively (Pal et al. 2011). Various studies have shown that the deletion or mutation of such AP-1 sites abolishes the stimulation of those genes by PPARs and LXR activators (Kömüves et al. 2000; Hanley et al. 2000).

Wound healing occurs as a high-priority survival response to skin injuries. The expression of PPARβ/δ is upregulated in adult epidermis by inflammatory stimuli during skin injury, which also provokes keratinocyte activation (Michalik et al. 2001; Tan et al. 2001, 2007). The activated PPARβ/δ has an antiapoptotic effect on the keratinocytes. Hence, it is protecting them from cytokine-induced apoptosis during the inflammatory phase of wound repair, which maintains a sufficient number of viable migratory keratinocytes for the reepithelialization phase of the healing process (Di-Poi et al. 2002). PPARβ/δ in wound fibroblasts plays a regulatory role in controlling keratinocytes proliferation (Chong et al. 2009). Thus, our studies and that of others have provided compelling evidence for a role of PPARβ/δ as a valuable pharmacologic wound healing target, impacting numerous events essential for wound healing (Tan et al. 2003, 2004a, b; Michalik and Wahli 2006, 2007).