Lipid deregulation in UV irradiated skin cells: Role of 25-hydroxycholesterol in keratinocyte differentiation during photoaging

Skin photoaging due to UV irradiation is a degenerative process that appears more and more as a growing concern. Lipids, including oxysterols, are involved in degenerative processes; as skin cells contain various lipids, the aim of our study was to evaluate first, changes in keratinocyte lipid levels induced by UV exposure and second, cellular effects of oxysterols in cell morphology and several hallmarks of keratinocyte differentiation. Our mass spectrometry results demonstrated that UV irradiation induces changes in lipid profile of cultured keratinocytes; in particular, ceramides and oxysterols, specifically 25-hydroxycholesterol (25-OH), were increased. Using holography and confocal microscopy analyses, we highlighted cell thickening and cytoskeletal disruption after incubation of keratinocytes with 25-OH. These alterations were associated with keratinocyte differentiation patterns: autophagy stimulation and intracellular calcium increase as measured by cytofluorometry, and increased involucrin level detected by immunocytochemistry. To conclude, oxysterol deregulation could be considered as a common marker of degenerative disorders. During photoaging, 25-OH seems to play a key role inducing morphological changes and keratinocyte differentiation.


Introduction
Population aging constitutes one of the most significant trends of the 21st century; indeed, nowadays one in eight people in the world are aged 60 or over and this proportion tends to increase dramatically [1]. As a result, skin aging appears as a growing concern. Skin aging is characterized in particular by wrinkles, sagging skin and elasticity loss. Physiologically, the most abundant cells of the skin are keratinocytes, which are located in the epidermis, the outermost layer of the skin. Epidermis continuously regenerates throughout the life: keratinocytes migrate into the upper layer of the epidermis during a process called differentiation. During this process, keratinocytes are gradually modified to become flat cells named corneocytes, which have lost their nucleus and cytoplasmic organelles [2]. Changes occurring through keratinocyte differentiation are characterized, amongst others, by remodeling of actin cytoskeleton [3], increased levels of involucrin and keratin [4], increased calcium level [5] and autophagy [6]. Keratinocyte differentiation can be abnormal or accelerated under certain stresses such as UV irradiation after sun exposure [7]. In this particular case, the abnormal and accelerated keratinocyte differentiation is part of photoaging.
In this in vitro study, our aim was to evaluate first, the modifications of keratinocyte lipid levels induced by UV irradiation, and second, the cellular effects of oxysterols on cell morphology and hallmarks of keratinocyte differentiation.

Cell culture
HaCaT cells, spontaneously transformed human keratinocytes, were obtained from Cell lines service (Cell lines service-CLS-Germany). The cells were cultured in Dulbecco's modified Eagle's medium (DMEM, Eurobio, Courtaboeuf, France) supplemented with 10% fetal calf serum, 2 mM of glutamine, 50 IU/ml of penicillin and 50 IU/ml of streptomycin (Eurobio). Cell cultures were maintained in controlled atmospheric conditions: CO 2 5%, humidity 95% and temperature 37 C. When the cells reached confluency, they were dispersed using trypsin and counted. Depending on cellular concentration, the cellular suspension was diluted and seeded in flasks or microplates.

UV irradiation
UV irradiation was performed with a solar light simulator Suntest CPS+ (Atlas, Mount Prospect, IL, USA). This simulator equipped with a xenon arc lamp and special glass filters restricting transmission of light below 290 nm, provides irradiance that approximates sunlight. HaCaT cells were seeded in flasks and irradiated at a dose of 1 or 2.5 J/cm 2 . The cells were subsequently rinsed and incubated for 24 h in culture medium. Non-irradiated cells were used as control.

Lipidomic analysis
Lipid composition of keratinocytes was analyzed by ultraperformance liquid chromatography coupled to high resolution mass spectrometry (UPLC-HRMS). After cell dispersion using trypsin, cell pellets were dissolved in 600 mL double-distilled water, vortexed for 30 s and sonicated for 5 min. Total lipids were extracted by the method of Bligh and Dyer [17]. Lipids extracts were resuspended in a 35:35:20:10 v/v/v/v acetonitrile/isopropanol/chloroform/water solution and analyzed using UPLC-HRMS on a Synapt TM G2 HDMS TM mass spectrometer (Waters MS Technologies, Manchester, UK). Data were analyzed using unsupervised principal component analysis (PCA) and supervised partial least squares discriminant analysis (PLS-DA). Moreover, an orthogonal partial least squares discriminant analysis model (OPLS-DA) was built on the PLS model using SIMCA-P+ software version 13.0.3 (Umetrics, Umeå, Sweden). Annotation of lipid species was performed using LIPID MAPS and METLIN online databases with a tolerance window for the mass accuracy of 5 ppm. Expected and actual retention times were compared for each lipid to confirm the previous annotation, a relative difference between these two retention times below 15% was accepted [18].

Sterols dosage
Quantitation of sterols in human keratinocytes was performed according to the method developed by Ayciriex et al. [19] using ultra-performance liquid chromatography-high resolution mass spectrometry analysis (UPLC-HRMS). After cell dispersion using trypsin, cell pellets were dissolved in 600 mL double-distilled water, vortexed for 30 s and sonicated for 5 min. Sterols were extracted with a hexane/methanol mixture (7:1, v/v) under agitation for 40 min and dried under reduce pressure. Sterols were derivatized into carbamate using a solution of 4-(dimethylamino)phenyl isocyanate in dichloromethane. Dichloromethane was evaporated under reduced pressure, derivatized sterols were resuspended in an acetonitrile/isopropanol mixture (1:1, v/v) and analyzed using UPLC-ESI-HRMS on a Synapt TM G2 HDMS TM mass spectrometer (Waters MS Technologies, Manchester, UK). Oxysterol levels were normalized to protein content measured by BCA method.

Keratinocytes incubation with 25-hydroxycholesterol
HaCaT cells were incubated with 25-hydroxycholesterol (25-OH, Sigma-Aldrich, Saint Louis, MO, USA). 25-OH was dissolved in absolute ethanol to obtain a 40 mM stock solution. Solutions were sonicated to solubilize oxysterol. 25-OH was diluted in culture medium to obtain targeted concentrations ranging from 5 to 40 mM.

Cell viability evaluation
Cell viability was evaluated through membrane integrity using the neutral red assay. HaCaT cells, cultured in 96-well microplates, were incubated with 25-OH for 48 h. After this incubation time, the cells were washed with PBS and incubated with a 50 mg/mL neutral red solution for 3 h at 37 C according to Borenfreund and Puerner validated protocol [20]. Then, the cells were rinsed with PBS and lysed with a solution of acetic acidethanol (ethanol 50.6%, water 48.4% and acetic acid 1%). After homogenization, the fluorescence signal was scanned (lexc = 540 nm, lem = 600 nm) using a cytofluorometre (Safire, Tecan, Männedorf, Switzerland).

Necrosis evaluation
Cell necrosis was assessed using the lactate dehydrogenase (LDH) release assay. In case of membrane damage, LDH, a cytoplasmic enzyme, is released in the extracellular compartment. The extracellular rate of lactate dehydrogenase is therefore correlated with cell death [21]. LDH mixture was prepared according to manufacturer's instructions (Sigma-Aldrich). 50 mL of cell supernatant were added to 50 mL of LDH mixture, and the microplate was agitated for 30 min at room temperature. Reaction was stopped with 10 mL HCl 1 N and absorbance was read at 490 nm (lref = 690 nm) using a cytometer (Safire, Tecan).

Cell morphology assessment
Cell morphology was studied on living cells with the label-free technique digital holographic microscopy using a HoloMonitor M3 (Phase Holographic Imaging PHI AB, Lund, Sweden). HaCaT cells, cultured in 6-well microplates, were incubated with 25-OH for 48 h. Cell morphological changes (thickness, roughness, volume, area) were analyzed by the software provided with the HoloMonitor (HoloStudio).

Cell cytoskeleton actin analysis
Cytoskeletal changes in actin stained with phalloidin were observed by confocal microscopy. HaCaT cells were cultured in Lab-Tek 1 (Nalgene Nunc International, Rochester, NY, USA). After a 48-h incubation with 25-OH, the keratinocytes were washed with PBS and fixed with paraformaldehyde 2% for 20 min. The cells were permeabilized with Triton TX-100 (0.2%) for 5 min, and then incubated with Alexa Fluor 1 488 Phalloidin (1:40 dilution, ThermoFisher Scientific, Illkirch, France) during 2 h. The cells were rinsed with PBS and TO-PRO-3 (1:500 dilution, Thermo-Fisher Scientific) was added for 10 min to stain nuclear DNA.

Statistical analysis
In all 25-OH experiments, no difference was observed between negative control (culture medium) and absolute ethanol solvent (data not shown).
Statistical analysis was performed on at least three independent experiments with GraphPadPrism 6 software. A one-way ANOVA followed by a Dunnett test with a risk a at 5% was used. Thresholds of significance were ***p < 0.001, **p < 0.01 and *p < 0.05 compared to culture medium.

Changes in keratinocyte lipid composition after UV irradiation
As degenerative diseases are associated with increases in ceramides and oxysterols, we explored lipid modifications after UV irradiation of HaCaT keratinocytes using mass spectrometry.
First, an untargeted approach was used to characterize the changes in keratinocyte lipid content after UV irradiation. A twocomponent PCA score plot of UPLC-HRMS data was used to visualize general variation of lipids between irradiated and nonirradiated cells. Undistinguished clusters of irradiated HaCaT cells at 1 J/cm 2 and non-irradiated cells were observed (Fig. 1A), whereas a clear separation between 2.5 J/cm 2 irradiated HaCaT cells and non-irradiated HaCaT cells was observed (Fig. 1B). A supervised analysis using OPLS-DA method was realized to enhance the identification of similarities or differences between non-irradiated cells and 2.5 J/cm 2 irradiated cells. The score plot of the OPLS-DA method showed an excellent separation between 2.5 J/cm 2 irradiated HaCaT cells and non-irradiated HaCaT cells (Fig. 1C). The lipid changes we observed after 2.5 J/cm 2 UV irradiation on keratinocytes were: increases in ceramides, phosphatidylethanolamine and phosphatidylcholine, and decreases in sphingomyelin and phosphatidylcholine-plasmalogen (Table 1). Details on chemical formulas and fold changes were listed on supplementary data (table S1).
Second, a targeted approach was used to investigate changes in cholesterol and oxidized metabolites of cholesterol called oxysterols in UV-irradiated cells. Among the studied oxysterols, 25-OH level was significantly increased after irradiation (Fig. 2). This increase was UV dose-dependent; indeed, 25-OH level was higher after a 2.5 J/cm 2 irradiation (753.10 À6 nmol/mg protein compared to 174.10 À6 nmol/mg protein in control cells, corresponding to a 4.3-fold increase in Fig. 2B) than after a 1 J/cm 2 irradiation (348.10 À6 nmol/mg protein compared to 185.10 À6 nmol/mg protein in control cells corresponding to a 2.1-fold increase in Fig. 2A). No significant change was observed in the level of either cholesterol or other studied oxysterols. These data suggest a link between UV irradiation and 25-OH level in keratinocytes. That is why we further investigated the effects of 25-OH on keratinocytes.

25-OH cytotoxicity on keratinocytes
Cytotoxicity was evaluated after a 48-h incubation time with 25-OH through two markers: cell viability using the neutral red assay, and necrosis using the LDH release assay. As we can observe in Fig. 3A, the viability percentage was equal or superior to 70 until 40 mM. Considering a 30% viability loss as an acceptable cytotoxicity is a standard practice in toxicology [26]. Consequently, for further experiments, we selected concentrations ranging from 5 mM to 40 mM. No change in LDH release was observed after 25-OH incubation from 5 mM to 40 mM (Fig. 3B).

Keratinocyte morphology and cytoskeleton after 25-OH incubation
Cell morphology was studied using digital holographic microscopy. After a 48-h incubation time with 25-OH, we observed a concentration-dependent increase in thickness average, meaning that HaCaT cells are thicker than in culture medium (Fig. 4). These morphological alterations could be linked to changes in cytoskeleton, so we further studied 25-OH effects on keratinocytes actin cytoskeleton.    Actin cytoskeleton was observed by immunofluorescence on confocal microscopy. We observed numerous organized actin filaments in control cells (culture medium), whereas they became fewer after a 48-h incubation time with 25-OH at 10 mM and absent at 40 mM (Fig. 5). As cells undergo remodeling of actin cytoskeleton during keratinocyte differentiation [3], we decided to study other hallmarks of keratinocytes differentiation.

Autophagy after 25-OH incubation
As autophagy occurs during keratinocyte differentiation [6], we evaluated autophagy using the monodansylcadaverine (MDC) assay. Microscopic observations of MDC show increase in MDC fluorescence signal after 25-OH 10, 25 and 40 mM compared to culture medium (Fig. 6A). To confirm our observations, we quantified MDC fluorescence signal using microplate cytometry. A statistically significant increase in autophagy was observed at 25-OH 10, 25 and 40 mM (x1.23, 1.40 and 1.23, respectively compared to negative control, Fig. 6B). This increase reached a maximum at 25-OH 25 mM.

Involucrin level after 25-OH incubation
Involucrin, a specific marker of keratinocyte differentiation [4], was studied with immunofluorescence using confocal microscopy and flow cytometry. As seen in confocal microscopy pictures, involucrin level seemed higher after a 48-h incubation time with 25-OH than in culture medium (Fig. 7A). This increase is statistically significant for 25-OH 10 mM and 25-OH 40 mM (Fig. 7B).

Intracellular calcium quantitation after 25-OH incubation
Calcium plays a key role in keratinocyte differentiation process [5]. Intracellular calcium was quantified using the Fluo-4 Direct assay (Fig. 8). A significant increase in intracellular calcium was observed after a 48-h incubation time with 25-OH 10, 20 and 40 mM (x1.21, 1.26 and 1.32, respectively compared to negative control).

Discussion
The photoaging of skin is a degenerative process induced by sun exposure. UVA or UVB spectra are mostly used to reproduce sun exposure in labs; and yet, skin is exposed to a combination of UVA and UVB. Consequently, we chose an irradiator that approximates sun UV irradiation to expose keratinocytes to both UVA and UVB. As many degenerative diseases are correlated with lipid deregulation [8,9], we explored changes in lipid composition after UV irradiation of HaCaT human keratinocytes. HaCaT cell line constitutes a suitable model to study this parameter as these cells have the same total lipid content and the same distribution of major lipid species as native keratinocytes [27]. No change in keratinocytes lipid composition was observed after 1 J/cm 2 UV dose, whereas significant modifications were observed after 2.5 J/ cm 2 UV dose. Marionnet et al. showed that UV daylight dose is  comprised between 11 and 70 J/cm 2 in April around the world [28]. Our irradiation model, consisting of a single exposure at a dose close to UV daylight dose, is sufficient to induce dramatic changes in keratinocyte lipid profile.
In our study, UV irradiation exposure led to an increase in ceramides content of human keratinocytes. Our results complete those obtained by Wefers who showed that UV irradiation increased ceramides abundance in human stratum corneum [29]. The increase in ceramides observed in our study was concomitant with a decrease in sphingomyelin and an increase in phosphatidylcholine. As sphingomyelin can be metabolized by sphingomyelinase to form ceramides and phosphocholine, an intermediate of phosphatidylcholine, it would be interesting to study the activity of sphingomyelinase after keratinocyte UV irradiation in further experiments.
To us, oxysterols appear of most interest in UV irradiation studies because oxysterols are oxidized derivatives of cholesterol and UV irradiation is known to induce oxidative stress [30]. For the first time, we observed an increase in oxysterol levels, mainly 25-OH, after UV irradiation on keratinocytes. Previous studies demonstrated the involvement of ceramides and oxysterols in degenerative diseases [11][12][13][14]. Taken together, our observations and the literature data tend to highlight that ceramide and oxysterol deregulation appears to be a hallmark of degenerative disorders.
Oxysterols were the lipids with the greatest increase after UV irradiation. Therefore, we focused on oxysterols, especially 25-OH, for the following experiments.
In our model, 25-OH increased keratinocyte thickness and disrupted cytoskeleton. Previous studies underlined morphological changes in photoaging models after UV irradiation [31,32] and actin cytoskeletal network disruption in aging skin [33][34][35]. This actin cytoskeleton remodeling can trigger keratinocyte differentiation [3]. During early steps of keratinocyte differentiation, autophagy pathway is activated to recycle cellular wastes [6]. We observed that 25-OH acts not only on autophagy but also on other hallmarks of keratinocyte differentiation such as involucrin and intracellular calcium. Involucrin is a protein rich in glutamine and lysine, which are essential for crosslinking by transglutaminase to build the cornified envelope during the maturation of keratinocytes in corneocytes [36]. The onset of involucrin synthesis marks an early stage in terminal keratinocyte differentiation [37]: a layer of involucrin serves as a scaffold for the subsequent attachment of other reinforcement proteins such as loricrin, small proline rich proteins and filaggrin, and of the lipid envelope [38]. Therefore, involucrin expression is widely used in keratinocyte differentiation studies. Especially, the expression of involucrin is strictly calcium dependent. Indeed, increase in intracellular calcium stimulates the expression of markers of cell differentiation such as involucrin [39,40]. Keratinocyte differentiation is observed during photoaging process [41] and in our study, 25-OH, generated after UV irradiation, induced typical hallmarks of skin aging or photoaging.
To conclude, ceramide and oxysterol deregulation could be considered as a marker of degenerative disorders, including photoaging. During this process, 25-OH seems to play a key role inducing several patterns of keratinocyte differentiation. Fold increase in intracellular calcium Fig. 8. Intracellular calcium evaluation using Fluo-4 Direct assay (n = 12). **p < 0.01 and ***p < 0.001 compared to control cells.