INTRODUCTION

Collagen IV is a major anchoring protein found mostly in lamina densa of the dermal–epidermal junction of human skin. As a major constituent of the basement membrane, collagen IV provides a three-dimensional scaffold with other macromolecules such as laminins, fibronectin and heparan sulphate proteoglycans in promoting anastomosis between the epidermis and the underlaying dermis. It also plays a significant role in wound healing and embryogenesis and promotes cell adhesion and proliferation [1-3]. Six human genes code for collagen IV chains, that is COL4A1, COL4A2, COL4A3, COL4A4, COL4A5 and COL4A6 [4]. Type IV collagen anomalies are associated with several diseases, including Alport and Goodpasture's syndromes, several rheumatological and dermatological diseases. Major changes and degradation of collagen IV are linked to subepidermal blistering diseases, including bullous pemphigoid. Autoimmune subepidermal bullous diseases (AISBDs) lead to irregular collagen type IV immunohistochemical staining, which is a useful technique for diagnosis of AISBDs [5] and classification of cutaneous autoimmune subepidermal blistering disorders [6]. Blistering diseases are rare skin diseases that occur when the immune system attacks the skin and mucous membranes.

Skin ageing occurs due to both intrinsic and extrinsic factors. Intrinsic ageing is a physiological process that leads to thin, dry skin, fine wrinkles and gradual dermal atrophy. Air pollution, smoking, poor nutrition and sun exposure are extrinsic ageing factors that result in crepey skin, loss of elasticity, increase in sagging and rough-textured appearance [7, 8]. With advancing age, there is a decrease in the collagen IV expression of matured dermal fibroblasts [9]. Furthermore, a previous immunohistochemical study revealed that significant reduction in the levels of collagen IV in non-photodamaged skin of older individuals (mean age: 71.5 years) compared with the younger individuals (mean age: 23.4 years) [10].

Peroxisome proliferator-activated receptors (PPARs) are nuclear hormone receptors similar to retinoid receptors. They act as transcription factors that modulate the expression of target genes by binding to specific DNA sequences. This ligand-activated transcription has been shown to regulate activities across various skin cell types and metabolic activities. During its activation, PPAR-alpha heterodimerizes with retinoid X receptor beta (RXRβ), and together, these receptors activate the acyl-CoA oxidase gene promoter [11]. Acyl-CoA oxidase metabolizes very long-chain fatty acids via oxidative processes.

The aldo–keto reductase (AKR) superfamily metabolizes carbohydrates, steroids, prostaglandins, endogenous aldehydes, ketones and xenobiotic compounds. AKR1B10 mediates retinoic acid homeostasis by diminishing the cellular levels of retinoic acids and reducing the availability of retinaldehydes through the conversion of retinaldehydes into retinols [12, 13].

Olive leaf and olive leaf extract are known for beneficial health properties such as antioxidant, anti-inflammatory and anti-bacterial effects. The leaves contain a complex variety of bioactive substances, including phenolic compounds, triterpenic acids, iridoids and sugars, such as oleanolic and maslinic acids, oleuropein, oleuropeoside, ligustroside, hydroxytyrosol and mannitol [14]. Both oleanolic acid and oleuropein have been reported to activate the PPAR-alpha receptor [15, 16]. Furthermore, oleanolic acid also demonstrates dual agonist action on two isoforms of PPAR γ/α and a potent inhibitory action against AKR1B10 [17-19].

In other PPAR-related skin care applications, 10-hydroxystearic acid, a vegetable oil-derived synthetic fatty acid, has been reported to stimulate collagens I and III in cell culture models and improves facial age spots and conspicuous pores in in vivo [20]. However, there have been limited number of studies, which explore PPAR activation induced by complex mixtures containing more than one ligand [21-23].

Different therapeutic modalities have been proposed for blistering diseases; however, all of them have their own limitations and side effects on long-term use. A previous study observed that agonists of PPAR can act as antagonists for AKR1B10 [24]. Therefore, the identification of a mixture of compounds targeting PPAR activation with minimal or no side effects for management of ageing skin and blistering diseases could be of value for formulating both topical skin care and therapeutic products.

In skin care applications, the availability of ingredients that are known to induce synthesis collagen IV is limited. A compound extracted from Psoralea corylifolia called bakuchiol has been shown to upregulate types I and IV collagen in a DNA microarray study and also showed stimulation of type III collagen in a mature fibroblast model [25].

Keeping this in view, this study was conducted to examine whether a complex mixture of PPAR ligands (Linefade) may cooperatively bind in the PPAR-α cavity to activate gene expression and lead to collagen IV biosynthesis. To explore this possibility, an in silico technique was employed to model two of the main constituents (oleanolic acid and glyceryl monoricinoleate) in a co-binding arrangement. In vitro cell culture and transcriptome analysis were used to confirm PPAR activation and gene expression. Stimulation of collagen IV synthesis in cultured human dermal fibroblasts was assessed using an ELISA assay. Further corroboration of collagen IV synthesis was examined ex vivo in human skin explants using immunostaining and image analysis.

MATERIALS AND METHODS Materials

Dimethyl sulfoxide (DMSO), caprylic/capric triglyceride (CCT), magnesium ascorbyl phosphate (MAP) and GW590735 were obtained from various suppliers. Indigo Biosciences Reporter Assay System was used for luciferase reporter assay.

Preparation of linefade

Linefade was extracted from the powdered leaves of Olea europaea via an alcoholic extraction. The wet cake was dried into a powder, and oleanolic acid was assayed by HPLC method to not less than 80.0% w/w on dried basis. The remaining 20% contained a variety of alcohol-soluble compounds commonly found in olive leaves. Next, 2% w/w of the dried solid powder was resolubilized with a 1:1 solvent mixture of glyceryl monoricinoleate and dimethyl isosorbide. Heat and agitation were applied to achieve a clear solution without any precipitate. The resulting composition contains not less than 1.6% oleanolic acid and 0.4% of undetermined compounds derived from olive leaves.

In silico modelling technique and validation

The 3D structure of PPAR-α (PDB ID: 2P54), bound with SRC1 peptide and GW590735, was considered as the known PPAR-α activator to determine the active sites. The retrieved structure of PPAR-α from the RSCB database (PDB: 2P54) was optimized by the addition of hydrogens, removal of all water molecules and optimizing hydrogen bonds using Discovery Studio Visualizer (ver 2.5). The optimized structure was defined as target macromolecule in Autodock Vina Wizard of PyRx (ver 0.8). GW590735 and major constituents of Linefade were docked to the ligand-binding domain. The docking experiment was run in the PyRx software workspace on default parameters of the number of generations and energy evaluation for ten steps of the run. The predicted binding affinity was expressed as kcal/mol. While docking, different binding poses were generated among which the best and top-ranked binding pose was selected for visual inspection using Discovery Studio Visualizer (ver 2.5).

In vitro cell studies PPAR transcription

Linefade was stored at room temperature and was diluted in DMSO or distilled sterile water at 10 mg/ml. The stock solution was further diluted with compound screening medium as per kit manufacturer's instructions (Indigo Biosciences, State College, PA, cat. # IB00111-32, Technical Manual version 7.2). Three concentrations (4, 20 and 100 µg/ml) were tested in duplicates on CHO cells transfected with the PPAR-α receptor-controlled bioluminescent protein for ~20 h. GW590735 (0.143 µg/ml) was the positive control. The chemoluminescent signal, proportional to the PPAR-α-driven promoter activation, was quantified using Thermo Scientific Luminoskan Ascent Microplate Luminometer. This instrument has passed DLReady™ validation at Promega Corporation.

Cell viability and type IV collagen output in the adult human dermal fibroblasts Cell line and cell culture

Human dermal fibroblasts (HDFs) from adult skin (early passage, cat.# 106-05aCell Applications, San Diego, CA) were maintained in DMEM-based medium containing 10% FBS medium and 1% penicillin-streptomycin at 37°C in a humidified, 95% air/5% CO2. All colorimetric measurements were performed using Molecular Devices (San Jose, CA) microplate reader MAX190 and SoftMax3.1.2PRO software.

Measurement of cell viability

The effect of Linefade on cell viability was determined using the MTT assay, which measures the activity of mitochondrial dehydrogenases, such as succinate dehydrogenase, implicated in the respiratory electron transport chain in mitochondria [26]. Linefade was dissolved at 20 mg/ml in DMSO and further diluted using distilled water. Samples were added in triplicates to confluent adult human dermal fibroblasts. The positive control contained magnesium ascorbyl phosphate (MAP). After 72 h, the experiment was terminated, and the effect of Linefade on mitochondrial metabolism was measured.

Quantitative detection of collagen IV

Type IV collagen was quantified in the cell culture medium (soluble fraction) via sandwich ELISA assay after treatment with Linefade for 72 h, using unlabelled anti-type IV antibody (cat.#1340-01) for capture, followed by biotinylated anti-type IV collagen antibody (cat.# 1340-08), streptavidin-HRP and TMB reagents from SouthernBiotech (Birmingham, AL).

Assay on reconstituted skin substitutes

EpiDermFT tissues (MatTek; Ashland, MA) were equilibrated overnight; then, samples were added in duplicates at 3 mg/cm2 with a positive displacement pipette and were spread evenly on top of the tissues. Sterile distilled water was the negative control. After 24 h incubation, RNA was extracted and purified with RNeasy Mini Kit cat.# 74104 from Qiagen (Germantown, MD), using QiaCube Connect robotic station (Qiagen). Purified total RNA was assessed at 260 and 280 nm with NanoDrop Lite (Thermo Fisher Scientific, Waltham, MA).

Isolated RNA samples were sent on dry ice to Thermo Fisher Scientific Microarray Research Services Laboratory (Santa Clara, CA) for transcriptome profiling using Clariom S cartridge on GeneTitan™ Microarray System instrumentation. The resulting CHP files containing probe set analysis results generated with Affymetrix software were uploaded, and differential gene expression as well as functional interaction networks were analysed using the TAC software version 4.0.2.15 (Applied Biosystems by Thermo Fisher).

Ex vivo study

The ex vivo study was conducted as per the Good Laboratory Practices (Decree of August 10, 2004), as well as in compliance with the validated procedures and SOP of Laboratoire BIO-EC. Forty-two explants of an average diameter of 11 mm (±1 mm), including 15 delipidated explants, were prepared from an abdominoplasty coming from a 28-year-old woman (reference P2302-AN28) with a skin phototype V. These skin tissues were obtained from surgical residues and are in compliance with the Declaration of Helsinki and the article L.1243-4 of the French Public Health Code. The delipidated set was utilized for examining ceramide synthesis (to be reported elsewhere). Explants were placed in BEM culture medium (BIO-EC’s Explants Medium) at 37°C in a humid, 5% CO2 atmosphere. The Linefade complex was solubilized in caprylic/capric triglyceride (CCT) at 1% and 2.5% v/v concentrations. The study samples were divided into four groups: untreated control, excipient control, test 1 and test 2. The untreated control group was a blank. The excipient control group was treated with CCT alone. Test 1 and test 2 groups were treated with 1% and 2.5% Linefade, respectively. CCT with or without Linefade was topically applied to skin explants at days 0 and 2 on the basis of 2 μl per explant (2 mg/cm2) and spread using a small spatula. On day 2, half of the culture medium was replaced by fresh medium. On day 0, the untreated control explants were collected and cut into two parts: One part was fixed in buffered formalin solution for 24 h, and the other part was frozen at −80°C. On day 3, three explants from each batch were collected and processed in the same way as the untreated control.

Tissue morphology

The explants were fixed for 24 h in buffered formalin, dehydrated and then impregnated in paraffin using a Leica PEARL dehydration automat. The samples were embedded using a Leica EG 1160 embedding station. Next, 5-μm-thick sections were made using a Leica RM 2125 Minot-type microtome, and the sections were mounted on Superfrost® histological glass slides. The tissue morphology of the epidermal and dermal structures was assessed through microscopic observation of formalin-fixed paraffin-embedded (FFPE) skin sections after Masson's trichrome staining, Goldner variant. The degree of staining was assessed by microscopic observation.

Quantification of collagen type IV

The frozen explants were cut into 7-μm-thick sections using a Leica CM 3050 cryostat. The sections were then mounted on Superfrost® plus silanized glass slides. The microscopic observations were realized using a Leica DMLB, an Olympus BX43, or BX63 microscope. Pictures were digitized with a numeric DP72 or DP74 Olympus camera with cellSens storing software. Collagen IV immunostaining was performed on frozen sections with a monoclonal anti-collagen IV antibody (Dako, ref. M 0785, clone CI22) diluted at 1:50 in PBS-BSA 0.3%-Tween 20 and incubated for 1 h at room temperature and revealed by AlexaFluor488 (Life technologies, ref. A11008). The nuclei were counterstained with propidium iodide. The immunostaining was assessed by microscopic examination and image analysis. All the images were analysed using Cell^D software.

Statistical analysis

All data are expressed as Mean ± Standard Deviation (SD) or Mean ± Standard Error of Mean (SEM). The differences between control and treatment groups were evaluated by Student's double-tailed t-test. Statistically, significant variation was defined as ≥20% variation from the control with a p-value < 0.05.

RESULTS In silico modelling

PPAR-α (PDB ID: 2P54) contains 16 helices, 4 beta-strands, 11 bends and 16 turns (Sierra et al., 2007). GW590735 occupies the binding pocket formed by the helices, bends and beta-strands. The binding site of GW590735 in the ligand-binding pocket of PPAR-α is composed of the amino acid residues CYS-275, CYS-276, GLN-277, THR-279, ALA-333, MET-330, MET-355, SER-280, TYR-314, HIS-440 and TYR-464. GW590735 formed hydrogen bonds with SER-280, TYR-314, HIS-440 and TYR-464 [27]. The binding affinity of GW590735 and major constituents of Linefade with target receptor (PDB ID: 2P54) was observed to be in the range of −5.7 Kcal/mole to −10.4 Kcal/mole (Figure 1 and Table 1).

Major constituents of Linefade—oleanolic acid (OA) and glyceryl monoricinolate (GMR)—docked in the active site of PPAR-α (PDB ID: 2P54), highlighting cooperative interactions

TABLE 1. Binding affinity of major constituent of Linefade and GW590735 with PPAR-α and their hydrogen bond interactions at the active site. Bold-typed residues represent common interacting residues with GW590735 Compound name

Binding affinity

(−Kcal/mol)

Amino acid interacting residues at a distance criterion of 5Ȧ Number of common amino acid interacting residues with GW590735

GW590735

(synthetic PPAR-α agonist standard)

−10.4 SER-280, TYR-314, TYR-464, HIS-440, CYS-275, CYS-276, GLN-277, THR-279, ALA-333, MET-330, MET-355 _ Oleanolic acid −5.7 HIS-274, CYS-278, TYR-334, HIS-47, LEU-254 0 Glyceryl monoricinolate −6.1 THR-279, GLU-21, CYS-275, CYS-276, SER-280, ARG-271, ILE-272, PHE-273, HIS-274, THR-246, LEU-247, CYS-248, MET-249, ALA-250, GLU-251 4

Oleanolic acid

+

Glyceryl

monoricinolate

−6.7 CYS-276, SER-280, LEU-254, CYS-275, THR-279, ILE-272, MET-330, LEU-331, MET-355, GLN-277, THR-283, LEU-247, ALA-250, GLU-251 7 In vitro cell studies PPAR transcription

Linefade stimulated the activation of transcription of the luciferase reporter gene under the PPAR-α-controlled promoter by 511% (p-value = 0.000) at a 20 µg/ml dilution when applied from a DMSO stock solution. When applied from a water stock solution, Linefade stimulated the activation of transcription of the luciferase reporter gene under the PPAR-α-controlled promoter in a dose-dependent manner with an increase of 418% occurring at a 100 µg/ml dilution. DMSO aids the dispersion of lipid-soluble Linefade in this cell culture model but is not required to produce a significant activation. The positive control GW590735 (0.143 µg/ml) strongly upregulated the expression of the reporter gene (p-value = 0.001) indicating experimental reliability (Figures 2 and 3).

Degree of transcriptional activation of Linefade by PPAR-α depicted by the Reporter Assay System with DMSO-prepared stock solution of Linefade. Data are presented as Mean values ± SEM, n = 16; **p-value < 0.01 vs. water blank

Degree of transcriptional activation of Linefade by PPAR-α in the Reporter Assay System with water-prepared stock solution of Linefade. Data are presented as Mean values ± SEM, n = 16; **p-value < 0.01 vs. water blank

Cell viability and collagen IV of human dermal fibroblasts (HDFs)

Cell viability was examined to exclude cytotoxic concentration of Linefade. Linefade did not show any cytotoxic effect on HDFs at concentrations of up to 200 μg/ml compared with water-treated control. Linefade significantly stimulated type IV collagen output, being the most active at a concentration of 100 μg/ml. It was observed that treatment with 100 μg/ml Linefade led to a 319% increase in collagen IV level compared with the water-treated control (p-value = 0.000) (Figure 4).

Dose-dependent stimulatory effect of Linefade on the synthesis of collagen type IV expressed as percentage of water control, wherein 50 µg/ml MAP is a positive control. Data are presented as Mean values ± SEM, n = 6; **p-value < 0.01 vs. water blank

Transcriptome analysis

Differential expression comparing a water blank (five samples) and Linefade (two samples) was processed with TAC 4.0 software at a fold change of either >2 or <−2 and p-value <0.05. Total number of genes in the data set was 21 448. Of these, 280 (1.31%) genes passed the filter criteria, among which 76.79% genes were upregulated and 23.21% genes were downregulated. Activity in relevant pathways was assessed with WikiPathways by significance and count in Table 2.

TABLE 2. Gene expression after treatment with water vs. Linefade assessed using Transcriptome Analysis Console 4.0.2.15 under filter criterion of fold change >2 or <−2 and p-value <0.05 WikiPathways Path significance Up Down NRF2 1.81 CES4A, EGR1 HSPA1A, PGD, SLC39A2, SLC39A14 Nuclear receptor meta pathway 1.03 CES4A, EGR1 CDK4, SLC39A2, SLC39A14, ACOX1, PGD, HSPA1A PPAR-alpha pathway 0.51 - CKD4 PPAR signalling pathway 1.82 SORBS1 MMP1, ACOX1, MMP1 PPAR-gateway pathway 1.64 - PLIN2, PCK2 PPAR-gamma pathway 1.64 - PLIN2, PCK2 Retinoid metabolism and transport 0.38 - RDH11 Integrin-mediated cell adhesion 1.27 SORBS1, CAPN3 ITGAE, CAPN1 Matrix metalloproteinases 1.64 - MMP1, MMP3

Genes related to epidermal differentiation (keratinocyte cornification), retinoid metabolism and collagen degrading MMPs were identified as shown in Table 3.

TABLE 3. Gene expression after treatment with water vs. Linefade assessed using Transcriptome Analysis Console 4.0.2.15 under filter criterion of fold change >1.5 or <−1.5 and p-value <0.05 Fold Change p-value Gene symbol Description Comments Cornified envelope precursor genes 6.99 0.0263 LCE1D Late cornified envelope 1D Late cornified envelope (LCE) genes within the epidermal differentiation complex 4.52 0.0166 LCE1F Late cornified envelope 1F 4.26 0.0249 LCE1A Late cornified envelope 1A 4.14 0.0286 LCE1B Late cornified envelope 1B 3.07 0.0471 LCE1C Late cornified envelope 1C 2.51 0.0481 LCE2D Late cornified envelope 2D 2.33 0.0181 LCE5A Late cornified envelope 5A 2.8 0.0358 S100A7A S100 calcium binding protein A7A S100A7 (Psoriasin) interacts with epidermal fatty acid-binding protein and localizes in focal adhesion-like structures in cultured keratinocytes [28] 1.61 0.0127 S100A7 S100 calcium binding protein A7 2.26 0.0063 S100A12 S100 calcium binding protein A12 2.51 0.0193 SPRR3 Small proline-rich protein 3 The small proline-rich proteins constitute a multigene family of differentially regulated cornified cell envelope precursor proteins [29] 1.91 0.0311 SPRR2G Small proline-rich protein 2G −1.8 0.0486 CRNN Cornulin 3.07 0.0424 KPRP Keratinocyte proline-rich protein Keratinocyte proline-rich protein deficiency in atopic dermatitis leads to barrier disruption [30] Matrix metallopeptidases −2.38 0.0163 MMP3 Matrix metallopeptidase 3 Matrix metalloproteinase-3 is the key effector of TNF-α-induced collagen degradation in skin [31] −2.82 0.0437 MMP1 Matrix metallopeptidase 1 Matrix metalloproteinase-1 is the major collagenolytic enzyme responsible for collagen damage in UV-irradiated human skin [32] −1.78 0.021 MMP14 Matrix metallopeptidase 14 (membrane-inserted) Retinoid metabolism 1.71 0.0241 CRABP1 Cellular retinoic acid-binding protein 1 1.55 0.0316 RDH13 Retinol dehydrogenase 13 (all-trans/9-cis) Catalyses conversion of retinol to retinal −1.59 0.0466 DHRS3; MIR6730 Dehydrogenase/reductase (SDR family) member 3; microRNA 6730 −1.65 0.0248 ALDH1A3 Aldehyde dehydrogenase 1 family, member A3 −1.71 0.0328 RARB Retinoic acid receptor, beta −1.73 0.017 PALM Paralemmin −1.76 0.0037 APOE Apolipoprotein E −2.08 0.0034 RDH11 Retinoldehydrogenase 11 (all-trans/9-cis/11-cis) −1.76 0.0231 AKR1C1 Aldo–keto reductase family 1, member C1 Reduces retinaldehyde to retinol conversion, limiting availability of retinoic acid Human explants (Collagen IV)

An ex vivo study was performed using human skin explants to further investigate whether Linefade could increase collagen IV synthesis in the ex vivo model. On days 0 and 3, tissue morphology was observed to be good in the epidermis and papillary dermis in the untreated control explants. On day 3, the excipient-treated control (CCT) samples showed slight alteration in the epidermis. The samples treated with 1% Linefade showed slight alteration in the epidermis while treatment with 2.5% Linefade did not result in any modification when compared with the untreated control group. On day 3, the samples treated with 1% Linefade did not show any modification in the epidermis and papillary dermis, while treatment with 2.5% Linefade led to a slight improvement in tissue viability in the epidermis when compared to the excipient control group as shown in Table 4 and Figure 5.

TABLE 4. Tissue morphology microscopic assessments of the epidermal and dermal structures of all explant batches is shown here Explant batch Tissue morphology observations Epidermis Dermis Untreated Control Day 0 G G Untreated Control Day 3 FG G Excipient Control Day 3 SA G 1% Linefade Day 3 SA G 2.5% Linefade Day 3 FG G Note Morphology legend: G = good, FG = fairly good, VSA = very slightly altered, SA = slightly altered, MA = moderately altered, FCA = fairly clearly altered, CA = clearly altered, VCA = very clearly altered.

Tissue morphology of day 0 and day 3 batches

Immunostaining revealed a significant increase in collagen IV along the DEJ as compared to untreated control and excipient-treated control groups on day 3.

Image analysis of collagen IV was analysed in terms of percentage of collagen IV surface. It was observed that on day 3, 1% Linefade showed a significant increase in collagen IV along the DEJ (52%; p-value <0.01) as compared to untreated control group. Furthermore, treatment with 1% Linefade resulted in a significant increase in collagen along the DEJ (41%; p-value <0.05) as compared to excipient-treated control group. Treatment with 2.5% Linefade did not produce a statistically significant result (Figures 6 and 7).

Effects of Linefade on collagen IV synthesis in skin explants vs. untreated and excipient-treated controls. Collagen IV was detected by immunostaining using specific antibodies. Number of explants per condition = 3

Quantification of collagen expression levels in Figure 6 (n = 9 with 3 images per explant) Data are presented as mean ± standard deviation. Calculated proportion of the surface to collagen IV along the DEJ. Treated samples vs. untreated control (Day 0); ** for p-value p-value p-value  DISCUSSION

Skin ageing is a continuous process, accelerated by oxidative stress and governed by intrinsic age-related metabolic changes. PPAR-α is a key regulator of this process via prevention of damage and control of repair after exposure to endogenous or environmental stressors [33]. PPAR-α mediates the transcription of lipid metabolizing genes and regulates the genes involved in maintaining the redox homeostasis, energy metabolism and integrity of tissues [34].

Altered DEJ is a characteristic of skin ageing. Type IV collagen, a fundamental constituent of the DEJ, provides a scaffold for other macromolecules and plays a key part in maintaining mechanical steadiness and resilience [35]. Therefore, the identification of compounds that can increase the level of collagen IV in DEJ is advantageous for managing changes associated with skin ageing.

To elucidate the mechanism of action of Linefade on skin agei