Fucoxanthin and Rosmarinic Acid Combination Has Anti-Inflammatory Effects through Regulation of NLRP3 Inflammasome in UVB-Exposed HaCaT Keratinocytes

Excessive exposure to ultraviolet (UV) radiation is the main risk factor to develop skin pathologies or cancer because it encourages oxidative condition and skin inflammation. In this sense, strategies for its prevention are currently being evaluated. Natural products such as carotenoids or polyphenols, which are abundant in the marine environment, have been used in the prevention of oxidative stress due to their demonstrated antioxidant activities. Nevertheless, the anti-inflammatory activity and its implication in photo-prevention have not been extensively studied. Thus, we aimed to evaluate the combination of fucoxanthin (FX) and rosmarinic acid (RA) on cell viability, apoptosis induction, inflammasome regulation, and anti-oxidative response activation in UVB-irradiated HaCaT keratinocytes. We demonstrated for the first time that the combination of FX and RA (5 µM RA plus 5 μM FX, designated as M2) improved antioxidant and anti-inflammatory profiles in comparison to compounds assayed individually, by reducing UVB-induced apoptosis and the consequent ROS production. Furthermore, the M2 combination modulated the inflammatory response through down-regulation of inflammasome components such as NLRP3, ASC, and Caspase-1, and the interleukin (IL)-1β production. In addition, Nrf2 and HO-1 antioxidant genes expression increased in UVB-exposed HaCaT cells pre-treated with M2. These results suggest that this combination of natural products exerts photo-protective effects by down-regulating NRLP3-inflammasome and increasing Nrf2 signalling pathway.


Introduction
Skin is considered as the outmost protective barrier in the body, protecting from detrimental substances, mechanical damage, pathological invasion and radiation that could cause perturbations to the skin structure. In this sense, skin is a well-known essential piece of the immune system. Several factors can contribute to the initiation and development of cutaneous alterations. In this line, the excessive exposure to UV radiation remains the main risk factor for the skin cancer [1]. Solar UV radiation consists of three broad ranges of wavelength: UVC (100-280 nm), UVB (280-315 nm) and UVA (315-400 nm). UVB causes dermal changes, affects epidermal function and is the main factor

Effects of RA and FX on Cell Viability
The UVB intensity was selected in a preliminary study in which three different intensities (100, 150 and 225 mJ/cm 2 ) were tested. To study the cell viability was used MTT assay (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide). As expected, a UVB dose-dependent Mar. Drugs 2019, 17, 451 3 of 14 effect on cell death and morphology was observed. The intensity of 100 mJ/cm 2 was chosen for further assays since it allowed over 50% of cell survival and showed a significant difference in comparison with non-irradiated control group ( Figure 1A). In addition, MTT results showed that the treatment with RA or FX at 5 µM, M1 (2.5 µM RA + 5 µM FX) and M2 (5 µM RA + 5 µM FX) significantly increased cell viability in UVB-exposed cells, being more effective M1 and M2 treatments ( Figure 1B). RA + 5 µM FX) and M2 (5 µM RA + 5 µM FX) significantly increased cell viability in UVB-exposed cells, being more effective M1 and M2 treatments ( Figure 1B).

Effect of RA and FX on UVB-induced ROS Production.
It is well known that intracellular ROS levels can be increased by UVB exposure [24]. In this sense, HaCaT cells were treated with the compounds for 1 h and then were irradiated with UVB (100 mJ/cm 2 ). Cells exposed to UVB showed a significant increase in ROS production in comparison with non-exposed control cells (p < 0.001). The pre-treatment with either RA at the concentration of 2.5 or 5 µM or FX at 5 µM notably reduced intracellular ROS levels by 12%, 12% and 10%, respectively (p < 0.01). Nevertheless, concomitant administration of RA and FX (M1 or M2) showed a more marked decrease in ROS production, by 21% and 22%, respectively (p < 0.001) than the single treatments ( Figure 1C). (C) Intracellular ROS generation was evaluated 30 min after UVB irradiation by relative fluorescence intensity using 2',7'-dichlorodihydrofluorescein diacetate (DCF-DA) assay. For experiments B and C, cells were pre-treated with RA (2.5 and 5 µM), FX (5 µM) and two combinations M1 (2.5 RA plus 5 µM FX) and M2 (5 µM RA plus 5 µM FX), for 1 h. Then, cells were irradiated with selected UVB dose (100 mJ/cm 2 ) and incubated for the required times. Results are expressed as percentage respect to untreated control cells (B) or UVB-exposed control cells (C), and bars represents mean ± SEM of at least six independent experiments (n = 6) in duplicate. The mean was significantly different compared to control cells (*** p < 0.001; Student t test). The mean was significantly different compared to UVBirradiated cells ( + p < 0.05, ++ p < 0.01 and +++ p < 0.001; one-way ANOVA followed by Bonferroni's multiple comparison test).

Effect of RA and FX on Apoptosis
It is well reported that apoptosis is the most important type of programmed cell death in response to cell damage induced by UVB exposure [25]. In this sense, Annexin V (annexin family of  (C) Intracellular ROS generation was evaluated 30 min after UVB irradiation by relative fluorescence intensity using 2',7'-dichlorodihydrofluorescein diacetate (DCF-DA) assay. For experiments B and C, cells were pre-treated with RA (2.5 and 5 µM), FX (5 µM) and two combinations M1 (2.5 RA plus 5 µM FX) and M2 (5 µM RA plus 5 µM FX), for 1 h. Then, cells were irradiated with selected UVB dose (100 mJ/cm 2 ) and incubated for the required times. Results are expressed as percentage respect to untreated control cells (B) or UVB-exposed control cells (C), and bars represents mean ± SEM of at least six independent experiments (n = 6) in duplicate. The mean was significantly different compared to control cells (*** p < 0.001; Student t test). The mean was significantly different compared to UVB-irradiated cells ( + p < 0.05, ++ p < 0.01 and +++ p < 0.001; one-way ANOVA followed by Bonferroni's multiple comparison test).

Effect of RA and FX on UVB-induced ROS Production.
It is well known that intracellular ROS levels can be increased by UVB exposure [24]. In this sense, HaCaT cells were treated with the compounds for 1 h and then were irradiated with UVB (100 mJ/cm 2 ). Cells exposed to UVB showed a significant increase in ROS production in comparison with non-exposed control cells (p < 0.001). The pre-treatment with either RA at the concentration of 2.5 or 5 µM or FX at 5 µM notably reduced intracellular ROS levels by 12%, 12% and 10%, respectively (p < 0.01). Nevertheless, concomitant administration of RA and FX (M1 or M2) showed a more marked decrease in ROS production, by 21% and 22%, respectively (p < 0.001) than the single treatments ( Figure 1C).

Effect of RA and FX on Apoptosis
It is well reported that apoptosis is the most important type of programmed cell death in response to cell damage induced by UVB exposure [25]. In this sense, Annexin V (annexin family of intracellular protein)-FITC assay allows to differentiate cell subpopulations in different stages as necrotic (R1), viable (R2), early apoptosis (R3) or late apoptosis (R4). For this assay, only the adherent cells were considered, which means that the dead cells in suspension after UVB irradiation were not analysed and thus significant changes were not observed in this phase [26]. Our results showed that UVB exposure, in comparison with non-irradiated cells, significantly decreased the percentage of viable cells from 88% to 20%, increased the number of cells in early apoptosis from 1% to 5% and strongly augmented the percentage of cells in late apoptosis from 8% to 72% (Figure 2A). Regarding to pre-treatments, neither RA nor FX exhibited percentages changes in the different subpopulations, respect to UVB-exposed cells. Nevertheless, the treatment with M1 or M2 allowed a significant reduction of the percentage of late apoptotic cells and a marked increase of viable cells in comparison with irradiated control cells ( Figure 2A). strongly augmented the percentage of cells in late apoptosis from 8% to 72% (Figure 2A). Regarding to pre-treatments, neither RA nor FX exhibited percentages changes in the different subpopulations, respect to UVB-exposed cells. Nevertheless, the treatment with M1 or M2 allowed a significant reduction of the percentage of late apoptotic cells and a marked increase of viable cells in comparison with irradiated control cells (Figure 2A).

Effect of RA and FX on Cell Cycle
It is well known that UV light exposure promotes skin photo-damage such as cell death and DNA damage and, consequently, cell cycle arrest [27]. Our results indicated that UVB-exposure highly affects the percentage of cells at G0/G1 phase, showing a reduction in this percentage from 68% to 50%, accompanied by an increase in apoptotic sub-G1 subpopulations from 3% to 14%, compared with non-exposed cells ( Figure 2B). Moreover, UVB-exposed control group showed a significant increase in the percentage of cells in S phase, from 16% to 24%, and a marked decrease of cells in G2/M phase from 13% to 3% in comparison with non-irradiated control group. Pre-treatment with RA or FX at 5 µM, as well as concomitant administration of RA and FX at 5 µM (M2) significantly augmented the number of cells at G0/G1 phase in comparison with UVB-exposed cells. However, the percentage of cells in S phase was only reduced by M2, reaching similar values that those in control group ( Figure 2B).  Results are expressed as percentage and bars represents mean ± SEM of four independent experiments (n = 4) in duplicate. The mean value was significantly different compared with control group (*** p < 0.001, ** p < 0.01, * p < 0.05 vs. untreated cells; Mann-Whitney U-test). The mean value was significantly different compared with UVB-irradiated cells + p < 0.05, ++ p < 0.01 vs. UVB-irradiated cells; Kruskal-Wallis test followed by Dunn's multiple comparison test).

Effect of RA and FX on Cell Cycle
It is well known that UV light exposure promotes skin photo-damage such as cell death and DNA damage and, consequently, cell cycle arrest [27]. Our results indicated that UVB-exposure highly affects the percentage of cells at G0/G1 phase, showing a reduction in this percentage from 68% to 50%, accompanied by an increase in apoptotic sub-G1 subpopulations from 3% to 14%, compared with Mar. Drugs 2019, 17, 451 5 of 14 non-exposed cells ( Figure 2B). Moreover, UVB-exposed control group showed a significant increase in the percentage of cells in S phase, from 16% to 24%, and a marked decrease of cells in G2/M phase from 13% to 3% in comparison with non-irradiated control group. Pre-treatment with RA or FX at 5 µM, as well as concomitant administration of RA and FX at 5 µM (M2) significantly augmented the number of cells at G0/G1 phase in comparison with UVB-exposed cells. However, the percentage of cells in S phase was only reduced by M2, reaching similar values that those in control group ( Figure 2B).

Effect of RA and FX on Inflammasome Regulation
It has been well documented that inflammasome can be activated by several factors, among them UVB exposure [8]. In the present study, we demonstrated that UVB irradiation led to up-regulation of NLRP3, ASC and caspase-1 expression and, consequently, promoted a significant increase in IL-1β production in comparison with non-exposed cells ( Figure 3A-D). Pre-treatment with either RA or FX at 5 µM did not significantly modify the expression levels of inflammasome-related proteins. Nevertheless, combination of RA and FX at 5 µM (M2) significantly down-regulated NLRP3 and ASC expression levels (p < 0.01) as well as caspase-1 levels (p < 0.05) in comparison with UVB-exposed cells. Furthermore, pre-treatment with M2 reduced IL-1β production in a significant manner (p < 0.01) ( Figure 3D).

Effect of RA and FX on Inflammasome Regulation
It has been well documented that inflammasome can be activated by several factors, among them UVB exposure [8]. In the present study, we demonstrated that UVB irradiation led to up-regulation of NLRP3, ASC and caspase-1 expression and, consequently, promoted a significant increase in IL-1β production in comparison with non-exposed cells ( Figure 3A-D). Pre-treatment with either RA or FX at 5 µM did not significantly modify the expression levels of inflammasome-related proteins. Nevertheless, combination of RA and FX at 5 µM (M2) significantly down-regulated NLRP3 and ASC expression levels (p < 0.01) as well as caspase-1 levels (p < 0.05) in comparison with UVB-exposed cells. Furthermore, pre-treatment with M2 reduced IL-1β production in a significant manner (p < 0.01) ( Figure 3D). on inflammasome components expression and IL-1β levels in UVB-exposed HaCaT keratinocytes. Cells were pre-treated with RA (5 µM), FX (5 µM) and M2 (5 µM RA plus 5 µM FX) for 1h, then were UVB-irradiated (100 mJ/cm 2 ) and incubated for 24 h. Densitometry analysis of (A) nucleotide-binding domain, leucine-rich-repeat-containing family, pyrin domain-containing 3 (NLRP3), (B) inflammasome adaptor protein (ASC) and (C) caspase-1 positivity were performed following normalization to the control (β-actin housekeeping gene). (D) IL-1β production was evaluated by ELISA assay. Results are expressed as percentage respect to UVB-irradiated control and bars represents mean ± SEM of four independent experiments (n = 4). The mean value was significantly different compared with control group (* p < 0.05, ** p < 0.01 and *** p < 0.001; Mann-Whitney U-test). The mean value was significantly different compared with UVB-irradiated cells ( + p < 0.05 and ++ p < 0.01; Kruskal-Wallis test followed by Dunn's multiple comparison test). Cells were pre-treated with RA (5 µM), FX (5 µM) and M2 (5 µM RA plus 5 µM FX) for 1h, then were UVB-irradiated (100 mJ/cm 2 ) and incubated for 24 h. Densitometry analysis of (A) nucleotide-binding domain, leucine-rich-repeat-containing family, pyrin domain-containing 3 (NLRP3), (B) inflammasome adaptor protein (ASC) and (C) caspase-1 positivity were performed following normalization to the control (β-actin housekeeping gene). (D) IL-1β production was evaluated by ELISA assay. Results are expressed as percentage respect to UVB-irradiated control and bars represents mean ± SEM of four independent experiments (n = 4). The mean value was significantly different compared with control group (* p < 0.05, ** p < 0.01 and *** p < 0.001; Mann-Whitney U-test). The mean value was significantly different compared with UVB-irradiated cells ( + p < 0.05 and ++ p < 0.01; Kruskal-Wallis test followed by Dunn's multiple comparison test).

Effect of RA and FX on Nrf-2 Antioxidant Signaling Pathway
To further elucidate the photo-protective mechanism of the combination of RA and FX, we evaluated the capacity to activate antioxidant pathways as Nrf2, which promotes the transcription of antioxidant genes and detoxification of enzymes such as heme-oxygenase 1 (HO-1) to protect against an oxidative damage as UVB exposure [28]. Our results showed that UVB irradiation prevented the increase of Nrf2 and HO-1 expression ( Figure 4A-B). Non-irradiated control cells (sham) did not exhibit the highest expression levels of these antioxidant proteins because these cells were not exposed to any stress. Nevertheless, UVB-exposed HaCaT cells and pre-treated with FX at 5 µM or the combination of RA and FX at 5 µM (M2) significantly up-regulated Nrf2 expression levels (p < 0.05 and p < 0.01, respectively), reaching higher values that those in control cells. Regarding HO-1 expression, only M2 was able to significantly enhance this protein levels (p < 0.05).

Effect of RA and FX on Nrf-2 Antioxidant Signaling Pathway
To further elucidate the photo-protective mechanism of the combination of RA and FX, we evaluated the capacity to activate antioxidant pathways as Nrf2, which promotes the transcription of antioxidant genes and detoxification of enzymes such as heme-oxygenase 1 (HO-1) to protect against an oxidative damage as UVB exposure [28]. Our results showed that UVB irradiation prevented the increase of Nrf2 and HO-1 expression ( Figure 4A-B). Non-irradiated control cells (sham) did not exhibit the highest expression levels of these antioxidant proteins because these cells were not exposed to any stress. Nevertheless, UVB-exposed HaCaT cells and pre-treated with FX at 5 µM or the combination of RA and FX at 5 µM (M2) significantly up-regulated Nrf2 expression levels (p < 0.05 and p < 0.01, respectively), reaching higher values that those in control cells. Regarding HO-1 expression, only M2 was able to significantly enhance this protein levels (p < 0.05). . Data shown are expressed as percentage respect to UVB-irradiated control and bars represents mean ± SEM of four independent experiments (n = 4). The mean value was significantly different compared with control group (* p < 0.05; Mann-Whitney U-test). The mean value was significantly different compared with UVBirradiated cells ( + p < 0.05 and ++ p < 0.01; Kruskal-Wallis test followed by Dunn's multiple comparison test.

Discussion
Skin represents the first barrier that protects us from the deleterious effects of solar UV radiation, which is the main cause for skin cancer. Nevertheless, human skin is normally exposed to UV irradiation in a non-controlled way. Whereas a short-term UV exposure might suppress immune function or trigger an inflammatory response, chronic UV exposition conduces to photo-aging, DNA damage, or carcinoma [29]. Skin has an antioxidant system, composed by enzymatic and nonenzymatic antioxidants, to prevent against toxic exogenous/endogenous ROS levels. Nowadays, it is well known the importance of natural substances that support the endogenous antioxidant systems of the skin via diet or dermatological preparations [30]. To develop new skin photo-protective agents it is necessary to understand the molecular mechanism of UV-induced cellular responses and determine how these products may act in the skin cells. In this sense, many treatments with a unique compound have demonstrated successful results against photo-damage, but a more hopeful impact was obtained when a combination of several compounds was used [31]. The efficacy of combinations could be due to the synergistic effect of their components. In this regard, natural products may be good candidates for skin photo-protection due to their low toxicity and high antioxidant capacity . Data shown are expressed as percentage respect to UVB-irradiated control and bars represents mean ± SEM of four independent experiments (n = 4). The mean value was significantly different compared with control group (* p < 0.05; Mann-Whitney U-test). The mean value was significantly different compared with UVB-irradiated cells ( + p < 0.05 and ++ p < 0.01; Kruskal-Wallis test followed by Dunn's multiple comparison test.

Discussion
Skin represents the first barrier that protects us from the deleterious effects of solar UV radiation, which is the main cause for skin cancer. Nevertheless, human skin is normally exposed to UV irradiation in a non-controlled way. Whereas a short-term UV exposure might suppress immune function or trigger an inflammatory response, chronic UV exposition conduces to photo-aging, DNA damage, or carcinoma [29]. Skin has an antioxidant system, composed by enzymatic and non-enzymatic antioxidants, to prevent against toxic exogenous/endogenous ROS levels. Nowadays, it is well known the importance of natural substances that support the endogenous antioxidant systems of the skin via diet or dermatological preparations [30]. To develop new skin photo-protective agents it is necessary to understand the molecular mechanism of UV-induced cellular responses and determine how these products may act in the skin cells. In this sense, many treatments with a unique compound have demonstrated successful results against photo-damage, but a more hopeful impact was obtained when a combination of several compounds was used [31]. The efficacy of combinations could be due to the synergistic effect of their components. In this regard, natural products may be good candidates for skin photo-protection due to their low toxicity and high antioxidant capacity [32]. Thus, the present study aimed to evaluate the preventive effect of a combination of RA and FX in UVB-exposed human HaCaT keratinocytes.
It is well reported that UVB radiation increases ROS production and DNA damage as well as promotes a strong inflammatory response characterized by pro-inflammatory mediator production [33]. In this condition, apoptosis activation can eliminate the irreversibly damaged cells that could harbour oncogenic mutation [25]. In this line, FX has demonstrated antioxidant activity offering protection against DNA damage and preventing cellular apoptosis in HaCaT human keratinocytes [34,35]. In a similar way, several studies have reported that RA reduces ROS production, protects against DNA damage and regulates apoptotic markers in vitro [31,36]. According to these findings, our results showed that the pre-treatment with FX or RA at 5 µM significantly attenuated UVB-induced damage by increasing cell viability and inhibiting ROS production in human HaCaT keratinocytes, but did not modify the number of apoptotic cells after UVB exposure. Nevertheless, the selected combination of RA and FX at 5 µM at equal dose (M2) not only improved cell viability and ROS levels in comparison with FX and RA administered separately, but also enhanced the cellular protection, reducing the number of cells in apoptosis after UVB irradiation. In addition to cellular responses previously detailed, UVB exposure also affected cell cycle dynamics. Previous studies have reported that UV exposure activates cell cycle arrest in both G1 and G2 phases [37]. At this respect, our results reported that UVB irradiation significantly reduced cells at G0/G1 phase when compared with non-exposed cells. Furthermore, a S-phase delay was observed in UVB-exposed cells although it did not show to be highly affected by the treatments. The pre-treatment with FX, RA and concomitant administration of RA and FX (M2) significantly increased the percentage of cells at G0/G1 and reduced the percentage of cells at sub-G1. In this sense, our observations are in accordance with the results obtained in apoptosis assay. M2 prevented cell damage, reducing the percentage of cells under apoptosis 24 h after UVB exposure. In consequence, this combination promoted a higher percentage of viable cells at 48 h in UVB-exposed cells.
In terms of inflammation, it has been reported that pro-IL-1β is cleaved and IL-1β is released from cells as a pro-inflammatory cytokine after NLRP3 inflammasome activation [38]. Thus, it could be concluded that IL-1β production in keratinocytes is inflammasome-dependent [6]. Moreover, NLRP3 inflammasome activation by UVB increases not only IL-1β but also other pro-inflammatory cytokines such as IL-1α, IL-6, TNF-α or the mediator PGE2 [7]. In this line, NLRP3 inflammasome plays an important role in UVB-induced skin inflammatory responses and has a critical function in skin pathologies initiation such as cancer [8]. In addition, the inflammasome adaptor protein ASC has shown anti-tumour activity by activation of tumour suppressor protein p53 in UVB-irradiated keratinocytes [39]. Due to the high connection between inflammasome, inflammation and cancer, the therapeutic strategies targeting inflammasome-related products are currently in development and suppose a pivotal tool for skin pathologies and cancer prevention [40]. In previous studies, FX has shown anti-inflammatory activity through NF-κB and MAPKs signalling pathways regulation [41], and consequently the down-regulation of IL-1β, TNF-α, inducible nitric oxide synthase, and cyclooxygenase-2 expression [42][43][44]. On the other hand, the polyphenol RA has been previously reported to inhibit pro-inflammatory cytokines production, including IL-1β, and to down-regulate NF-κB signalling pathway in skin cells [45]. In addition, this polyphenol has shown to modulate inflammasome activation, through down-regulation of caspase-1, NLRP3 and ASC expression in a poly(I:C)-induced inflammation in an infectious model by using epidermal keratinocytes [21]. Moreover, due to its anti-inflammatory properties, RA has been proposed to ameliorate allergic reactions in allergic rhinitis and rhinoconjunctivitis [46]. Furthermore, this compound has shown to prevent cisplatin-induced apoptosis in chemotherapy by down-regulating the caspase-1 expression [47]. However, the role of FX or RA on NLRP3 inflammasome regulation in human keratinocytes has not been studied yet. In this line, our findings reported for the first time that the combination of RA and FX notably reduced inflammasome-related proteins such as NLRP3, ASC and caspase-1 in UVB-irradiated HaCaT keratinocytes. Interestingly, this effect was not detected when both compounds were administered separately, suggesting the beneficial effect of concomitant treatment with RA and FX.
The antioxidant role of FX and RA has been well-described. FX has shown to promote AKT/Nrf2/GSH-dependent antioxidant response in keratinocytes [48,49], as well as reduce wrinkles formation and epidermal hypertrophy in mice [50] and supress melanogenesis and prostaglandin synthesis [51]. In this line, due to its antioxidant activity, this carotenoid has been proposed as a photo-protective compound by stimulating skin barrier restoration in UVA-induced sunburn [52]. Furthermore, a previous study by our group evidenced that FX increased Nrf2 and HO-1 expression in UVB-induced acute erythema in SKH-1 hairless mice [44]. These results are in line with other authors that suggest that this compound could be an interesting strategy to counteract UVB-induced oxidative damage in skin [53]. At this regard, similar properties have been demonstrated with RA applications, confirming the antioxidant activity of this polyphenol via activation of Nrf2-antioxidant system [54,55]. In accordance with these findings, the combination between FX and RA (M2) used in this study showed an up-regulation of Nrf2 transcriptional factor and its main target gene HO-1. These data indicate that the protective activity of M2 may be due to the reduction of oxidative stress through modulation of Nrf2 signalling pathway.

Compounds
RA and FX were obtained from Sigma-Aldrich (St. Louis, MO, USA). RA and FX stocks were prepared in DMSO at 10 mM and diluted to the desired concentration with culture medium. Controls were incubated with the corresponding quantity of DMSO, which was always below 1% and did not affect cell viability

Cell Line
Human immortalized keratinocytes HaCaT were obtained from the American Type Culture Collection (ATCC) and maintained in high glucose Dulbecco's modified Eagle's medium (DMEM, GIBCO, Grand Island, NY, USA) containing 2 mM l-glutamine. The culture media was supplemented with 10% heat-inactivated fetal bovine serum, 100 U/mL penicillin and 100 mg/mL streptomycin (PAA Laboratories, Pasching, Austria). Cultures were incubated in an atmosphere of 5% CO 2 at 37 • C.

UVB-Irradiation
The CL-1000M UV Crosslinker system (UVP, Upland, CA, USA), formed by 5 UVB tubes (8 W), was used to submit an energy spectrum of UVB radiation (peak intensity 302 nm, inside of the spectrum of UVB light 280-315 nm). To prevent UVB light absorption by the cell culture medium the medium was replaced by a thin layer of phosphate buffer solution (PBS) to cover the cells during irradiation. Three UVB intensities were evaluated (100, 150 and 225 mJ/cm 2 ) to determine the most suitable to evaluate cell cycle alterations and apoptosis [22]. In this sense, 100 mJ/cm 2 was selected for the study.

Cell Proliferation Assay
Cell viability was evaluated by the colorimetric 3-(4,5-dimethyl-2-thiazolyl)-2,5-diphenyl tetrazolium bromide (MTT) (Sigma-Aldrich, St. Louis, MO, USA) assay, which determines the formation of purple formazan in viable cells and allows estimate cellular viability [56]. MTT assay was performed according to protocol described with slight modifications. Briefly, HaCaT cells were seeded at a density of 10 4 cells/well in a 96-well plate for 24 h. Cells were pre-treated with different doses either RA (2.5, 5, 7.5 and 10 µM) or FX (5, 10, 15 and 20 µM) and the sixteen possible combinations between RA and FX, for 1 h. Then, the culture medium was replaced with a thin layer of PBS and exposed to a single dose of UVB radiation at 100 mJ/cm 2 . Cells were supplied with fresh culture medium and incubated for 24 h. Absorbance was measured at 570 nm using a Synergy HT Multi-mode Microplate Reader (BioTek Instruments, Winooski, VT, USA).

Apoptosis Determination
Apoptosis was evaluated by flow cytometry using Annexin V-FITC Apoptosis detection kit from BD Pharmingen (Franklin Lakes, NJ, USA), as previously reported [26]. Briefly, HaCaT cells were seeded at 5 × 10 5 cells/well in 6-well plates and incubated for complete adhesion. The day after seeding, cells were treated with RA (2.5 and 5 µM), FX (5 µM) and mixtures M1 (2.5 RA plus 5 µM FX) and M2 (5 µM RA plus 5 µM FX) for 1 h. After that, treatments were removed and cells were irradiated at 100 mJ/cm 2 , and then incubated with fresh medium for 24 h. Cells were harvested and suspended in binding buffer at 10 5

Cell Cycle Determination
For cell cycle determination, a similar procedure to apoptosis evaluation was carried out, with the difference of that after irradiation, cells were incubated with fresh medium for 48 h. After incubation, keratinocytes were harvested by trypsinization as previously detailed [26], fixed in 85% cold ethanol (5 × 10 5 cells/mL) and kept at −20 • C until further analysis. Ethanol was eliminated, cells were resuspended in PBS, and the cell suspension was filtered through a 35 µm nylon mesh to separate aggregated cells. Then, 50 µL RNase (1 mg/mL) and 50 µL PI (1 mg/mL) were added to each sample, which was then incubated for 20 min in darkness at room temperature until analysis. DNA content was determined on a Coulter Epics XL Flow Cytometer (Beckman Coulter, Hialeah, FL, USA). Data were acquired using the SYSTEM II (v. 2.5.) software examining 10 4 events. Percentages of cells in apoptotic-sub G1, G0/G1, S and G2/M were calculated using CXP software.

Determination of IL-1β Production
HaCaT cells were seeded at 5 × 10 5 cells/well in 6-well plates, incubated for 24 h and treated with either RA (5 µM), FX (5 µM) or M2 (5 µM RA plus 5 µM FX), for 1 h. Then, the cells were irradiated at 100 mJ/cm 2 , and incubated with fresh medium for 24 h. Supernatant fluids were collected and stored at −80 • C until use. Commercial enzyme-linked immunosorbent assay (ELISA) kits (Diaclone GEN-PROBE, Besançon cedex, France) was used to quantify IL-1β production according to the manufacturer's protocol. The absorbance at 450 nm was read by a microplate reader (Sinergy HT, Biotek ® , Bad Friedrichshall, Germany). To calculate the concentration of IL-1β (pg/mL), a standard curve was constructed using serial dilutions of cytokine standards provided with the kit.

Statistical Analysis
All values in the figures and text are expressed as arithmetic means ± standard error of the mean (S.E.M.). Data were evaluated with GraphPad Prism ® Version 5.00 software (San Diego, CA, USA). In all cases, the Shapiro-Wilk test was used to verify the normality of the data. The Mann-Whitney U-test was chosen for non-parametric values. The parametric values groups were analysed by one-way analysis of variance (ANOVA) followed by Bonferroni's multiple comparison test. p values < 0.05 were considered statistically significant.

Conclusions
In conclusion, we have demonstrated for the first time that the combination of RA and FX shows additional benefits in comparison with the compounds administered separately on UVB-irradiated keratinocytes. The studied combination (M2) increased cell viability by reducing UVB-induced apoptosis and decreased intracellular ROS production. Moreover, the concomitant administration of RA and FX showed anti-inflammatory activity through NLRP3 inflammasome modulation, as well as antioxidant properties via up-regulation of Nrf2/HO-1 antioxidant system. Overall, we propose the combination of RA and FX as a natural promising tool in prevention of UVB-induced skin alterations as photo-aging, skin inflammation and its derivation to pre-cancerous lesions and skin carcinomas. Funding: This work was supported by grants from the Consejería de Innovación, Ciencia y Empresa-Junta de Andalucía POLFANAT-P12-AGR-430 and from University of Seville "V Plan Propio US-PPI2015-1.5". Thanks are due to FCT/MCTES for the financial support to CESAM (UID/AMB/50017/2019), through national funds. FCT is also acknowledged for the research contract under Stimulus of Scientific Employment 2017 to H. Oliveira (CEECIND/04050/2017). We thank "Centro de Investigación, Tecnología e Innovación (CITIUS)" at University of Seville for providing technical assistance.

Conflicts of Interest:
The authors have no conflicts of interest to declare.