Peroxisome proliferator-activated receptors (PPARs) are nuclear hormone receptors expressed in the skin. Three PPAR isotypes, α (NRC1C1), β or δ (NRC1C2) and γ (NRC1C3), have been identified. After activation through ligand binding, PPARs heterodimerize with the 9-cis-retinoic acid receptor (RXR), another nuclear hormone receptor, to bind to specific PPAR-responsive elements in regulatory regions of target genes mainly involved in organogenesis, cell proliferation, cell differentiation, inflammation and metabolism of lipids or carbohydrates. Endogenous PPAR ligands are fatty acids and fatty acid metabolites. In past years, much emphasis has been given to PPARα and γ in skin diseases. PPARβ/δ is the least studied PPAR family member in the skin despite its key role in several important pathways regulating inflammation, keratinocyte proliferation and differentiation, metabolism and the oxidative stress response.
1. PPARdelta: The Least Studied PPAR Isoform
Peroxisome proliferator-activated receptors (PPARs) are transcription factors belonging to nuclear hormone receptor superfamily. Three PPAR isotypes, α (NRC1C1), β or δ (NRC1C2) and γ (NRC1C3), have been identified in mammals (henceforth, we refer to the β/δ isoform simply as PPARδ). After activation through ligand binding, PPARs heterodimerize with the 9-cis-retinoic acid receptor (RXR), another nuclear hormone receptor, to bind to specific PPAR-responsive elements in regulatory regions of target genes, mainly involved in organogenesis, cell proliferation, cell differentiation, inflammation and metabolism of lipids or carbohydrates. Endogenous PPAR ligands are fatty acids and fatty acid metabolites.
PPARδ is ubiquitously expressed in murine tissues with highest expression in liver, muscle, adipose tissue, placenta, small intestine and skin. PPARδ is expressed twofold, 10-fold and 30-fold more in mouse keratinocytes (KCs) compared to mouse liver, quadriceps muscle and thymus, respectively. In most tissues, PPARδ localizes to the nuclear fraction of cells and is hardly detectable in the cytoplasm
[1]. In humans, PPARδ mRNA and protein are highly abundant in the thyroid gland and placenta whereas high amounts of mRNA and moderate amounts of protein are detected in the cerebral cortex, skin and esophagus. Of note, inconsistency between protein and RNA levels of PPARδ has been observed in many human tissues and cell types (
https://www.proteinatlas.org/ENSG00000112033-PPARD/tissue, accessed on 7 July 2021). There are five human and mouse PPARδ isoforms generated by alternative splicing, which is a mechanism potentially involved in PPARδ regulation, as some PPARδ splice isoforms exhibit reduced translation efficiency
[2][3].
The ligand-binding pockets of PPARs have a distinct three-armed T shape, which allows not only straight fatty acids to bind them, but also ligands with multiple branches such as phospholipids and synthetic fibrates. The ligand-binding pocket of PPARδ is smaller than that of PPARγ or PPARα, which limits the binding of large ligands when compared to the other two PPAR isoforms
[4]. PPARδ is activated by several endogenous ligands including certain long chain fatty acids (regardless of saturation status), dihomo-γ-linolenic acid, eicosapentaenoic acid, 15(S)-hydroxyeicosatetraenoic acid (HETE), and arachidonic acid, with affinities in the low micromolar range (
Table 1). Supraphysiological doses of 8(S)-, 12(S)-, 12(R)-, and 15(S)-HETE efficiently activate PPARδ. 13(S)-hydroxyoctadecadienoic acid (HODE) is considered as weak PPARδ activator
[5][6]. Controversial results have been found for prostacyclin (PGI2) and all-trans retinoic acid
[7][8]. It has also been reported that 4-hydroxynonenal (4-HNE) and 4-hydroxydodecadienal (4-HDDE), the peroxidation products of polyunsaturated fatty acids, can activate PPARδ, although the mechanism remains unknown
[9][10]. Synthetic PPARδ ligands include GW501516, GW0742 and L165041, which preferentially activate PPARδ as compared to PPARα or PPARγ
[6]. Recently, 27 new synthetic PPARδ agonists (13 with low nanomolar EC
50 values) have been discovered
[11]. However, it is important to stress that preferential ligand does not mean exclusive ligand and that supraphysiological doses of any of the PPARδ ligands will activate other PPAR isoforms, and the same is true for all PPAR isoforms. For example, bezafibrate, which is known as a PPARα ligand, activates all three PPARs at concentrations ranging from 55 to 110 μM
[12]. In the absence of ligand binding, the heterodimer PPARδ-RXR is associated with corepressors and histone deacetylases (HDACs), which inhibit its transcriptional activity. After ligand binding, PPARδ undergoes conformational changes that induce the release of the corepressors and allow it to bind coactivators
[7].
Table 1. PPARδ potential endogenous ligands.
Compounds |
Weak Ligands |
Ligands |
ω3-PUFA |
α-Linolenic acid C18:3 |
EPA C20:5 |
γ-Linolenic acid C18:3 |
Dihomo-γ-linolenic acid |
DHA C22:6 |
ω6-PUFA |
Linoleic acid C18:2 |
|
Arachidonic acid C20:4 |
ω9-MUFA |
Palmitoleic acid C16:1 |
Oleic acid C18:1 |
Elaidic acid C18:1 |
Erucic acid C22:1 |
Nervonic acid C24:1 |
Saturated fatty acids |
Myristic acid C14:0 |
Arachidic acid C20:0 |
Palmitic acid C16:0 |
Stearic acid C18:0 |
Behenic acid C22:0 |
Eicosanoids |
5-HpETE |
5(S)-HETE |
8(S)-HETE |
15(R)HpETE |
15(S)HpETE |
15(R)-HETE |
15(S)-HETE |
12-HETE |
12-HpETE |
LTB4 |
LTA4 |
LTC4 |
9(R)-HODE |
9(S)-HODE |
12-HpODE |
5,6-diHETE |
13(S)-HODE |
|
5,15-di-HpETE |
|
Prostaglandins |
PGA2 |
PGF1α |
PGB1 |
PGB2 |
PGD1 |
PGD2 |
PGD3 |
PGF2α |
PGF3α |
PGI2 |
Lipoxins |
|
LXA4 |
4-Hydroxyalkenals |
4-HDDE |
|
The transcriptional activity of PPARδ is modulated by several factors, which are not well characterized but include post-translational modifications such as phosphorylation. Epidermal growth factor receptor (EGFR) has been recently shown to induce PPARδ phosphorylation at Y108 in response to epidermal growth factor (EGF)
[13]. Although PPARδ contains several putative phosphorylation sites (Y108, T252, T253, T256), (
https://www.phosphosite.org/proteinAction.action?id=24004&showAllSites=true (accessed on 9 May 2021))
[14], little is known about phosphoregulation of PPARδ, in contrast to PPARα and PPARγ. Both cyclic adenosine monophosphate (cAMP) and protein kinase A (PKA) activators increase the ligand-activated and basal activity of PPARδ and could be upstream signals that commit PPARδ to the regulation of glucose and lipid metabolism
[14]. In contrast, PPARδ can also be sumoylated at K104, which inhibits its activity
[14]. Desumoylation of PPARδ by small ubiquitin-like modifier (SUMO)-specific protease 2 (SENP2) promotes the transcriptional activity of PPARδ, which, in turn, upregulates fatty acid oxidation by enhancing the expression of long-chain-fatty-acid–CoA ligase 1 (ACSL1), carnitine palmitoyltransferase Ib (CPT1b) and mitochondrial uncoupling protein 3 (UCP3) in muscles of mice fed a high fat diet
[15]. Moreover, PPARδ contains several ubiquitylation sites, which suggests a potential role of ubiquitin–proteosome degradation in the regulation of its cellular turnover (
https://www.phosphosite.org/proteinAction.action?id=24004&showAllSites=true (accessed on 9 May 2021)). Degradation of PPARδ via the proteasome might prevent its accumulation in the nucleus and thereby moderate its cellular activity
[16]. In line with this, overexpression of PPARδ in fibroblasts leads to its polyubiquitylation and rapid degradation, a process partially prevented by exposure to the PPARδ synthetic ligand GW501516
[17].
PPARs can also engage in transrepression of other transcription factors. Although transrepression between nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB), activator protein 1 (AP-1), CCAAT-enhancer-binding protein (C/EBP), signal transducer and activator of transcription (STAT) and nuclear factor of activated T-cells (NF-AT) has been well characterized for PPARα and PPARγ, little is known about transrepression in the context of PPARδ
[18][19]. L-165041 is a PPARδ ligand that is less potent and selective than GW501516, yet it promotes the binding of PPARδ to the p65 subunit of NF-κB exerting anti-inflammatory effects
[5][20]. Moreover, in the absence of ligand, PPARδ binds directly to the transcription factor B-cell lymphoma 6 (BCL-6), leading to increased expression of proinflammatory cytokines. Indeed, BCL-6 is a transcription factor repressing the expression of various inflammatory genes via direct binding to their promoters or via inhibition of the transcription of nucleotide-binding oligomerization domain-like receptor (NOD)-like receptor family pyrin domain containing 3 (NLRP3)
[21][22]. Binding of PPARδ to an agonist disrupts the PPARδ-BCL-6 complex, thus reversing the transcriptional repression of inflammatory genes
[23]. Thus, ligand binding to PPARδ alleviates inflammation by enhancing its binding to NF-kB, hence neutralizing the transcriptional activity of NF-kB and/or the release of the anti-inflammatory transcription factor BCL-6. However, PPARδ has also been shown to bind to the N-terminal part of p65 in the absence of exogenous ligand
[5]. Therefore, the pro- vs. anti-inflammatory role of PPARδ might be context- and ligand-dependent. Moreover, conformational changes experienced by PPARδ after ligand binding might potentially strengthen or weaken the affinity of PPARδ to p65; however, this has not been studied to date.
32. PPARdelta in Psoriasis and Atopic Dermatitis
Atopic dermatitis and psoriasis are two chronic and pruritic inflammatory skin diseases exhibiting pathophysiological commonalities, including impaired epidermal barrier function, immune hyper-responsiveness, and local and systemic symptoms modulated by environmental factors such as the skin microbiome and stress. Moreover, both diseases are associated with a major genetic risk factor, i.e., Filaggrin (
FLG) loss-of-function mutations in atopic dermatitis and the HLA-Cw0602 allele in psoriasis vulgaris
[24][25]. Furthermore, in both atopic dermatitis and psoriasis patients, nonlesional and lesional skin coexists, but the mechanism of transition from the non-affected to the affected condition remains unclear. Atopic dermatitis is one of the most common inflammatory skin diseases worldwide and characterized by skin features such as erythematous and papulovesicular eruptions with oozing, crusting and pruritus as well as associated systemic signs such as food allergies, allergic asthma and rhinitis, anxiety and sleep disorders. At the cellular level, atopic dermatitis is characterized by (a) the complex interplay between impaired epidermal barrier function owing to altered lipid composition of the stratum corneum lipid matrix i.e., a reduction in the chain length of structural lipids (fatty acids and ceramides), (b) a complex Th2-driven inflammation, (c) skin infiltration by eosinophils, basophils and inflammatory dendritic cells, and (d) an altered skin microbiota
[24][26][27][28][29][30][31][32][33]. In psoriasis vulgaris, genetic risk factors predominantly affect innate immunity, and to some extent adaptive immunity (IL12p/IL-23R axis, Th1, Th17 cells). Similarly to atopic dermatitis, skin immunological abnormalities in psoriasis are complex and associated with comorbidities (e.g., arthritis and cardiovascular manifestations), pointing to a systemic immune hyper-responsiveness
[25][31][34][35][36][37].
PPARδ is expressed in all skin cell types, including KCs, fibroblasts, sebocytes, hair follicle cells, melanocytes and Langerhans cells
[19][38][39][40]. PPARδ is the predominant isoform in human KCs and is expressed throughout all epidermal layers
[41][42]. Activation of PPARδ with synthetic ligands promotes the expression of human KC differentiation markers such as involucrin (
INV) and transglutaminase 1 (
TGM1)
[42]. Although there is consensus on the pro-differentiative effects of PPARδ ligands and PPARδ activation in KCs, the effects on KC proliferation are more controversial, with studies showing reduced
[42] or enhanced
[43] KC proliferation after treatment with the PPARδ ligand L-165041 or GW-501516. Treatment of human KCs with L-165041 gave opposite outcomes in two distinct studies
[43][42]. Yet, the use of different treatment regimens of L-165041, i.e., 0.05 μM for 3 days
[42] and 1 μM for 7 days
[43], might have been responsible for these divergent results, for example by inducing the recruitment of different cofactors and thus engaging PPARδ in different metabolic pathways. Moreover, the direct effects of ligands should not be underestimated because the use of PPARδ siRNA to test the requirement for PPARδ in the cellular response was not carried out in either studies
[43][42]. In line with this, L-165041 can activate other PPAR isoforms, i.e., PPARα, PPARγ1 and PPARγ2 at doses as low as 0.05 μM
[42]. This underscores that PPAR ligands can exert receptor-independent effects, that metabolic effects might vary with ligand concentrations (e.g., U- or bell-curves), and that the relative contribution of other PPAR isoforms after treatment with ligands might significantly influence experimental results, hence stressing the need for cautious interpretation of data
[27]. Human KCs infected with a lentivirus containing an RNAi sequence directed toward PPARδ displayed reduced proliferative capacity, suggesting that PPARδ promotes, rather than dampens, proliferation of human KCs
[43]. However, it is also possible that PPARδ exerts both proliferative and differentiative functions according to the cellular context, i.e., basal cells (early KCs, progenitor and stem cells) or suprabasal cells (differentiated cells). As in other cell types, PPARδ is likely a master regulator of fatty acid metabolism in KCs by increasing the uptake of long-chain fatty acids via upregulation of CD36 and fatty acid β-oxidation
[42] (
Table 2). However, the role of PPARδ in epidermal lipid and glucose metabolism remains under-investigated. Interestingly, the PPARδ target genes in KCs are not identical to those in other organs and cell types (
Table 2), suggesting PPARδ has specific cellular functions in the epidermis.
Table 2. PPARδ target genes and associated pathways in keratinocytes.
|
Upregulated |
Downregulated |
Fatty acid metabolism |
FABP5 |
LASS6 |
FABP7 |
GPD1L |
ACADVL |
PRKAB2 |
ACOX1 |
CHPT1 |
CD36 |
|
ALOX12B |
|
LDLR |
|
PLA2G3 |
|
ECHB |
|
OACT5 |
|
BDH1 |
|
GDPD3 |
|
CRABP2 |
|
GM2A |
|
Cholesterol metabolism |
HMGCS1 |
|
HMGCR |
MVD |
CYP51 |
SQLE |
FDPS |
LSS |
FDFT1 |
DHC7 |
KC proliferation |
HB-EGF |
EGFR |
EPS15 |
EPS8 |
MCC |
RBL2 |
CCNG1 |
DUSP3 |
PDGFRA |
PDGFC |
CDKN1C |
KC differentiation |
INV |
DCN |
TGM1 |
KRT15 |
TGM3 |
DUSP3 |
S100A8 |
|
S100A9 |
|
S100A16 |
|
KRT6B |
|
KRT16 |
|
KRT17 |
|
KRT75 |
|
SPRR1B |
|
CNFN |
|
EHF |
|
KC apoptosis |
CIDEA |
|
Inflammation |
MMP9 |
TGFBR2 |
IL1F9 |
TGFBR3 |
IL1F5 |
LIFR |
IL1B |
IL1R1 |
IL1F6 |
|
IL1F8 |
|
ILA |
|
IL1RA |
|
IL18 |
|
IL17 |
|
IL23A |
|
IL22 |
|
STAT3 |
|
Glucose metabolism |
PDK1 |
PDK4 |
Oxidative stress |
SOD2 |
|
CAT |
ABCC3 |
Other |
HAS3 |
RBL2 |
GGH |
AXL |
UCK2 |
RHOC |
ATP10B |
TTC3 |
CCNB1 |
LFNG |
MAPK13 |
FXR1 |
CCNB2 |
FBLN1 |
GSPT1 |
GAB2 |
XPC |
|
|
PIK3IP1 |
Unknown |
AKR1B1 |
SERINC1 |
ATP12A |
EID1 |
ACPP |
KLF6 |
MAP4K4 |
RAI14 |
MREG |
MTCP1 |
FGFBP1 |
REEP5 |
ARL8B |
NENF |
GAS7 |
|
CD81 |
|
CCDC50 |
|
TACC1 |
|
|
OSR2 |