The effect of Fernblock ® in preventing blue‑light‑induced oxidative stress and cellular damage in retinal pigment epithelial cells is associated with NRF2 induction

Blue light exposure of the ocular apparatus is currently rising. This has motivated a growing concern about potential deleterious effects on different eye structures. To address this, ARPE-19 cells were used as a model of the retinal pigment epithelium and subjected to cumulative expositions of blue light. The most relevant cellular events previously associated with blue-light-induced damage were assessed, including alterations in cell morphology, viability, cell proliferation, oxidative stress, inflammation, and the induction of DNA repair cellular mechanisms. Consistent with previous reports, our results provide evidence of cellular alterations resulting from repeated exposure to blue light irradiation. In this context, we explored the potential protective properties of the vegetal extract from Polypodium leucotomos , Fernblock® (FB), using the widely known treatment with lutein as a reference for comparison. The only changes observed as a result of the sole treatment with either FB or lutein were a slight but significant increase in γH2AX + cells and the raise in the nuclear levels of NRF2. Overall, our findings indicate that the treatment with FB (similarly to lutein) prior to blue light irradiation can alleviate blue-light-induced deleterious effects in RPE cells, specifically preventing the drop in both cell viability and percentage of EdU + cells, as well as the increase in ROS generation, percentage of γH2AX + nuclei (more efficiently with FB), and TNF-α secretion (the latter restored only by FB to similar levels to those of the control). On the contrary, the induction in the P21 expression upon blue light irradiation was not prevented neither by FB nor by lutein. Notably, the nuclear translocation of NRF2 induced by blue light was similar to that observed in cells pre-treated with FB, while lutein pre-treatment resulted in nuclear NRF2 levels similar to control cells, suggesting key differences in the mechanism of cellular protection exerted by these compounds. These results may represent the foundation ground for the use of FB as a new ingredient in the development of alternative prophylactic strategies for blue-light-associated diseases, a currently rising medical interest.


Introduction
There is a general concern about the impact of blue light exposure on human health, particularly in relation to the induction of damage in eye structures [1].With the incremental time of indoor activities and the use of electronic devices in daily life (mobile phones, computer screens, etc.), there has been an important increase in the exposure to artificial light sources by the general population; being light emitting diodes (LEDs) the main source of artificial lighting [2].The light emitted by those sources appears as white, but it peaks at wavelengths between 400 and 500 nm, corresponding with the blue range of the visible spectrum [3].Even though the main source of blue light is still considered to be sunlight, the amount, intensity, and frequency of blue light exposition are expected to keep continuously increasing owing to the modern lifestyle.In this context, the risk of blue light-induced damage to ocular structures needs to be addressed [4].
Although there is no current evidence that a single exposure and normal use of LED devices represent a direct hazard to the human eye [5], increasing evidence is highlighting the appearance of temporary or even permanent damage to different eye structures [5,6].The retinal pigment epithelium (RPE) is key to maintain the retinal functionality [7,8].This outermost layer is particularly sensitive to blue light, and its deterioration triggers very common eye problems [3,9].Nowadays, important efforts in research are directed at understanding the molecular mechanisms behind blue light-associated damage in RPE [3,10].Among the specific alterations observed, it is worth mentioning increased cell death rates and apoptosis [11,12], changes in cell morphology [12], decrease in cellular division [13,14] associated with induced DNA damage and subsequent activation of DNA repair mechanisms [15][16][17], and the activation of inflammatory responses [18][19][20].All these events are associated with an increment in reactive oxygen species (ROS) levels, which is a central event in the blue light-triggered intracellular cascade of events [17,21].For this reason, beyond physical protection different antioxidant compounds are currently being explored as preventive and therapeutic approaches to combat blue light-derived damage [3].
In light of the aforementioned pathogenesis of blue light-induced eye damage, previous reports have demonstrated the protective effect of different compounds with antioxidant or anti-inflammatory effects such as lutein, curcumin, vitamin E, silibinin and Prunella vulgaris [3,19].In fact, the xanthophyll carotenoid lutein (LUT) is an antioxidant compound naturally present in the macular pigment of the retina [22,23].LUT is especially interesting in the context of blue light, as not only has it anti-oxidative and anti-inflammatory properties [24], but also can absorb light of wavelengths around 450 nm, acting as a blue light filter by itself [25].Even when it is present at high concentrations in the human retina, the human body is unable to synthesize lutein de novo, so it can only be obtained from the diet [26].Consequently, lutein food supplements have become the most popular non-physical prophylactic measure to prevent eye conditions specially when associated with blue light exposure [24].
In this scenario, the hydrophilic natural extract from Polypodium leucotomos, commercialized as Fernblock® (FB), emerges as a promising compound with improved photoprotective properties that have been thoroughly reported [27].Particularly, FB exerts skin photoprotection through mechanisms involving the inhibition of ROS production, prevention of DNA damage and lipid peroxidation, limiting inflammation and preventing immunosuppression [28].Due to its robustly safe toxicological profile [29][30][31][32], it is being used in photoprotective strategies both as food supplements and included in topical cosmeceutical products [27].In addition, we recently demonstrated the photoprotective effect of FB in skin cells exposed to blue light [33,34].Thus, FB is an ideal candidate as a natural compound for photoprotection against blue light-induced damage in eye tissues.
Bearing in mind the rising global concern about blue light-induced damage and the need of identifying new active compounds to be used in preventing eye degenerative diseases, we aimed to investigate the protective effects of FB.First, we analysed the significant morphological, physiological, molecular, and cellular changes induced by the exposure to artificial blue light in RPE cells grown in vitro.With this aim, we established an experimental setup which reasonably simulates a common eye exposure routine to blue light.Then, we characterized the protective effect of FB on cell morphology and viability using LUT as a protective compound of reference in all the assays.After identifying the concentrations of FB and LUT exerting the most efficient photoprotective effect, we proceeded with a deeper characterization of the underlying molecular mechanisms involved.Based on previous studies, we addressed the most relevant events of the intracellular cascade underlying the harmful effects of blue light in the retina.Among them, as stated before, antioxidant effects have previously been attributed to FB.Thus, we measured ROS production upon irradiation to investigate whether it could prevent the blue light-induced oxidative stress.Related to the cell cycle arrest induced by blue light, we evaluated the expression levels of P21, a well-known cyclin-dependent kinase inhibitor which mediates P53-induced G1/S cell cycle arrest [35].In addition, γH2AX is a well-established marker to evaluate the presence of double-strand DNA breaks [16] and TNF-α is an inflammatory cytokine produced during inflammation upon blue light exposure [3,36].Therefore, γH2AX and TNF-α were studied as readouts of DNA damage and inflammation, respectively.In all cases, we observed the detrimental consequences of blue light radiation and assessed the capacity of FB (in parallel to LUT) to alleviate them.Finally, as the NRF2 protein is considered a master coordinator of the cellular responses to stress related to redox unbalance [37] and in light of the antioxidant attributes of FB, we decided to study the NRF2 expression pattern.NRF2 is also a key element in the development and progression of different diseases, being appointed as the link between oxidative stress and inflammation [37,38].Interestingly, although both FB and LUT exerted protection against blue light, their effects on the NRF2 dynamics were opposed, suggesting that this pathway can be key to characterize the specific responses orchestrated by each one of these antioxidant compounds.
Therefore, in the context of a growing concern about a plausible damage of the retina induced by blue light, this work has addressed some of the most characteristic events underlying blue light damage in RPE cells from retina and characterized the protective effect of FB.Our results support the potential of this natural extract for the prevention of blue light-related eye diseases and open new roads for the development of prophylactic strategies.

Cell treatments
For cell treatments, cells between passage 3 and 9 were seeded at a density of 10,000 cells/ml using 24-well culture plates.Two compounds were used: Polypodium leucotomos extract, commercialized under the trademark name of Fern-block® (FB), and Lutein (LUT).FB (provided by Cantabria Labs, Madrid, Spain) was prepared from a stock solution of 229 mg/ml and diluted in culture medium with 10% FBS to the desired concentrations.LUT (Thermo Fisher Scientific Inc., Rockford, IL, USA) stock solution of 2 mg/ml was also diluted in culture medium with 10% FBS.The treatment with the compounds started 24 h before the first event of irradiation with blue light.The compounds were removed before the irradiation and freshly added to the medium after each irradiation event.The treatments were maintained in the culture medium until the time point of evaluation.

Blue light irradiation
ARPE-19 cells were irradiated at room temperature with artificial blue light with a narrow-band LED lamp as blue light source, which emits light of 450-465 nm wavelength and 42.05 mW/cm 2 power (Segainvex, Madrid, Spain).The experimental setup was designed to replicate the effect of the cumulative exposure to blue light.In this sense, the cells received a total of 5 events of irradiation (101.3J/cm 2 each, corresponding to a 40 min irradiation interval) over a period of 3 days (with a maximum of two cycles per day, spaced 5 h between them).The total irradiation doses applied in this study (cumulative irradiation dose) ranged from 101.3 (in a single irradiation event) to 506.6 J/cm 2 (when a total of 5 irradiation events were given, implicating a total irradiation time of 3 h and 20 min over a period of 3 days).When cells reached an approximately 60-70% confluence, they were irradiated in phenol red-free DMEM containing 1% FBS.To avoid any potential shielding effects exerted by the culture medium, a flat irradiator devise was used and the culture plates were placed on top of it, proceeding to irradiate the cells through the plastic.As the plates were closed at all times, the irradiation was performed outside the culture hood.Non-irradiated cells were moved out of the incubator in parallel and kept in the dark for the same period of time.Immediately after each irradiation, fresh medium with the corresponding treatment was added, and the cells were maintained in the incubator.

Cell viability assessment
Cell viability was evaluated 24 h after the last irradiation event using the MTT assay.For this purpose, a stock solution of 1 mg/ml 3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT, Sigma-Aldrich, St. Louis, MO, USA) in PBS was diluted in culture medium with 10% FBS to reach a final concentration of 50 μg/ml.The incubation with MTT solution was performed during 3 h under normal culture conditions (37 °C, 5% CO2).Then, MTT solution was removed and DMSO (Panreac, Barcelona, Spain) was added to dissolve the formed formazan crystals.The metabolic cell activity was inferred from the absorbance measured at 542 nm using a SpectraFluor Tecan (Zürich, Switzerland) plate reader.The results are presented as relative absorbance (percentage of the control).Data obtained from irradiated and non-treated (control) cells were used as a reference for normalization.

Reactive oxygen species determination
ROS production was determined by fluorescence microscopy using the 2,7-dihydrofluorescein diacetate (DHF-DA) probe (Abcam, Cambridge, UK).Cells were first incubated for 24 h with the selected concentrations of LUT or FB.Then, cells were incubated with 7.5 M DHF-DA for 30 min under normal culture conditions and they were subsequently irradiated with blue light.Untreated cells and cells treated with FB or LUT in the absence of blue light were used as controls.To assess the levels of ROS, cells were observed under the fluorescence microscope using blue excitation and the fluorescence intensity was quantified using ImageJ software (version 1.8.0) (NIH, Bethesda, MD, USA).

Indirect immunofluorescence
For immunofluorescence (IF), cells were grown on glass coverslips and fixed with 3.7% formaldehyde (Panreac, Barcelona, Spain) in PBS at 4 °C for 30 min.Then, cells were washed three times with PBS and permeabilized with 0.5% Triton X-100 in PBS at room temperature.
Coverslips were subsequently incubated with the primary antibodies against γH2AX and NRF2 (Cell Signaling Technology, Inc, Danvers, MA, USA) diluted in 0.5% bovine serum albumin (BSA, Sigma-Aldrich, St. Louis, MO, USA) in PBS for 1 h at 37 °C.After washing with PBS, cells were incubated with the secondary antibodies AF546 goat antirabbit IgG (for γH2AX) and AF488 (for NRF2) goat antimouse [Thermo Fisher Scientific Inc (Rockford, IL, USA)] for 45 min at 37 °C.Cells were then washed again, and nuclei were counterstained with Höechst-33258 (Sigma-Aldrich, St. Louis, MO, USA) for 5 min at 37 °C.Finally, the coverslips were washed in PBS and mounted with ProLong (Thermo Fisher Scientific Inc (Rockford, IL, USA)).The slides were observed using an epifluorescence microscope (Olympus BX61) equipped with a pE-300 LITE LED lamp and the filter sets for fluorescence microscopy corresponding to ultraviolet irradiation (360-370 nm, UG-1 filter) for nuclei visualization (Höechst-33258), blue light irradiation (450-490 nm, BP 490 filter) for NRF2 and green light irradiation (570-590 nm, DM 590 filter) for γH2AX.Positive nuclei counting and quantification was performed using the FIJI software (ImageJ, version 1.53 NIH, USA).

Measurement of TNF-α
To evaluate the production of Tumor Necrosis Factor Alpha (TNF-α), the Human TNF-α ELISA Kit from FineTest (Wuhan Fine Biotech Co., Ltd., Wuhan, China) was used.Cells were first seeded in 24-well plates and treated as indicated above.After irradiation, the medium was switched to phenol red-free DMEM (Thermo Scientific Hyclone) supplemented with 1% FBS and 1% antibiotics.After 24 h, supernatants were harvested, centrifuged at 480 g for 5 min and stored at −20 °C.In parallel, the cells remaining in the wells were counted using a TC20TM automated cell counter (BioRad, Hercules, CA, USA) to normalize the TNF-α.

5-Ethynyl-2-deoxyuridine (EdU) incorporation assay
The Click-iT EdU cell proliferation kit is commonly used for the detection of replicating cells based in the incorporation of the thymidine analog 5-ethynyl-2-deoxyuridine (EdU), which can be detected by the use of fluorescent azide dyes via a copper-mediated "click" reaction [39].Cells were grown in glass coverslips and subjected to the treatments indicated above.

Statistical analysis
All the experiments were repeated at least three times.The GraphPad Prism software (GraphPad Software Inc, San Diego, CA, USA, version 6.05) was used to perform the statistical analysis (one-way ANOVA or t test) and data representation.Statistical significance was set at p < 0.05.

Photoprotective effects of Fernblock® on cell viability and morphology after blue light irradiation
First, we evaluated the detrimental effect of blue light by addressing cell viability and morphology on ARPE-19 cells.Cell survival after blue light irradiation was assessed using the MTT assay.A decrease in cell survival was found to correlate with increasing cumulative doses of blue light, supporting the potential harmful nature of blue light irradiation to ARPE-19 cells (Fig. 1a).Accordingly, phase-contrast images showed the appearance of morphological alterations in a blue light dose-response fashion (Fig. 1b).Based on these results, we selected 202.6 J/cm 2 as the sub-lethal dose to be used in subsequent assays, since it represented the threshold of inducing moderate cellular changes while keeping the survival levels around 80% (thus corresponding to the LD20).
We next evaluated the capacity of FB to prevent the observed harmful effects of blue light irradiation in terms of cell survival and morphology, using the previously selected dose of blue light.For the sake of comparison, the protective effects of LUT as an eye-endogenous antioxidant of reference [23,[40][41][42] were tested in parallel.For both compounds, the concentration ranges used were set based on previous reports identifying relevant biological effects: 0.1-1 μg/ml of LUT [23] and 0.001-0.1 mg/ ml of FB [34,[43][44][45].Cells were incubated with LUT or FB during the 24 h period preceding the irradiation, the treatments were removed immediately before the irradiation, and fresh LUT or FB were added upon irradiation and maintained until the evaluation.
As expected, effective photoprotection against blue light was exerted by 0.1 μg/ml LUT (Fig. 2a).No alterations in cell morphology compared to the control were observed under this experimental condition.Thus, 0.1 μg/ml LUT was the concentration selected for further assays (Fig. 2b,  c).Regarding FB, both 0.05 and 0.1 mg/ml prevented the blue light-induced significant decrease in cell viability.Given that the morphological features were better preserved in the cells treated with 0.1 mg/ml FB prior to blue  2b, c), this was the concentration selected for subsequent treatments with FB.

Protective effect of Fernblock® against blue light-induced oxidative stress
One of the best characterized effects of blue light irradiation is the increase in the intracellular levels of ROS.In fact, these high levels of ROS account for the major detrimental biological consequences of blue light exposure [14,17,21].
Given that the previously reported protective mechanisms of both LUT and FB have been associated with their ability to improve ROS management by the cells, ROS levels were assessed in blue light-irradiated ARPE-19 cells.With this aim, ROS production was evaluated using the DHF-DA fluorescent probe in ARPE-19 cells.Prior to the incubation with DHF-DA, cells were treated with LUT or FB, followed by irradiation with 202.6 J/cm 2 blue light.Blue light irradiation induced a significant increment in ROS levels compared to non-irradiated (control) cells.Both LUT and FB pre-treatments effectively prevented the overproduction of ROS, hence suggesting that these compounds could minimize the oxidative stress-induced damage derived from blue light exposure.In the absence of blue-light irradiation, the sole treatment with FB or LUT had no significant effect on the ROS levels (Fig. 3a, b).

Fernblock® partially rescues the blue light-induced decrease in DNA synthesis
Blue light irradiation has been widely reported to induce cell cycle arrest, preventing cell proliferation [2].We next evaluated the effect of blue light and the impact of the pretreatments with FB and LUT on the cell cycle dynamics using the EdU Click-iT assay kit.
As shown in Fig. 4a, b, the incorporation of the EdU nucleoside in blue light-irradiated cells (in the absence of LUT or FB pre-treatments) was significantly lower than in non-irradiated (control) cells.Treatments with LUT or FB significantly rescued the values of nucleoside analogue incorporation in blue light irradiated cells.Nonetheless, nonirradiated cells treated with LUT or FB did not show changes in the number of positive nuclei compared with the nonirradiated (control) cells (Fig. 4a, b).These results reveal active DNA synthesis in ARPE-19 cells that are treated with LUT or FB prior to blue light irradiation.Accordingly, these compounds seem prevent the cell cycle arrest induced by blue light, potentially alleviating the impact of blue light on cell proliferation.
We focused on investigating whether active DNA synthesis promoted by LUT and FB upon exposure to blue light was related to the prevention of the cell cycle arrest, or rather revealing the induction of DNA repair mechanisms.Previous  The concentration of the compounds used for pre-treatment were 0.1 μg/ml LUT and 0.1 mg/ml FB.C refers to non-treated and nonirradiated cells works using different experimental systems identified a role of the P21 protein, a well-known inhibitor of the cell cycle progression, in mediating the cell cycle arrest induced by blue light radiation [17].Thus, to dig deeper into the intracellular cascades underlying the cell cycle arrest observed in our model system, we evaluated the expression of P21 by Western blot.The results revealed higher P21 expression in blue light-irradiated cells compared to non-irradiated (control) cells (Fig. 5).However, the treatment with LUT or FB preceding the irradiation did not prevent the blue-lightinduced overexpression of P21 (Fig. 5).In addition, LUT or FB in the absence of irradiation did not induce any changes in the level of P21.
Overall, these results suggest that FB, as well as LUT, are priming the cells to activate DNA repair mechanisms more efficiently in response to blue light irradiation.

Evaluation of the capacity of Fernblock® to reduce blue-light-induced DNA damage
H2AX is a histone variant belonging to the H2A family that undergoes phosphorylation at the Ser139 residue early upon the appearance of DNA double-strand breaks [46].This modification, known as γH2AX, is commonly used as a sensitive maker of DNA damage [47], and it is essential to coordinate DNA damage response, playing an important role in the assembly of the DNA repair complexes [48,49].
To address the DNA damage and the cellular repair mechanisms induced by blue light exposure, we evaluated by immunofluorescence the nuclear expression of γH2AX in response to the different treatments (Fig. 6a).First, we observed a striking, significant increase in the percentage of γH2AX positive nuclei in response to blue light irradiation.Notably, although not as large as the increase induced by blue light, both LUT and FB induced a significant increase in the percentage of γH2AX positive nuclei compared to control cells in the absence of blue-light exposure (Fig. 6b).Conversely, the addition of LUT or FB prior to blue light exposure significantly limited the increase of γH2AX positive nuclei (Fig. 6a, b), suggesting effective protection in response to both antioxidant compounds.These results are in accordance with previous reports using FB treatments in basal conditions or in response to other stimuli [50,51].Importantly, the percentage of γH2AX positive nuclei in the irradiated cultures treated with FB was significantly lower than that observed in response to LUT, supporting that FB exerts an improved protective action to limit blue-light induced DNA damage (Fig. 6b).

Effect of Fernblock® on blue-light-induced expression of inflammation-related markers
As part of its deleterious effects, blue light irradiation has been reported to induce the production of pro-inflammatory cytokines [3], mainly TNF-α [18,52,53].We used a specific ELISA kit to quantify the levels of TNF-α secreted to the culture medium by ARPE-19 cells, 24 h after irradiation in the different experimental conditions.The highest levels of TNF-α were detected in response to blue light irradiation (Fig. 7).While LUT was only able to partially prevent the blue light-induced increase in TNF-α secretion, the TNF-α levels detected in cells exposed to blue light after the treatment with FB were fully rescued, reaching similar values to those of the control (untreated) cells.In addition, LUT or FB in the absence of blue light induced no significant changes in TNF-α secretion (Fig. 7).

Effect of FB on NRF2 dynamics
NRF2 is a transcription factor capable of inducing the expression of antioxidant enzymes [37,38].The translocation of the NRF2 protein into the nucleus is one of the cellular responses to environmental challenges, especially oxidative stress.Thus, NRF2 expression levels and its nuclear translocation are considered key elements as part of the cellular endogenous mechanisms triggered to manage intracellular ROS.In this sense, we next evaluated the activation of NRF2 in response to blue light irradiation and the potential role of FB in affecting NRF2 dynamics.As expected in light of the strong induction of ROS after blue light exposure, a significant increment in the levels of nuclear NRF2 was observed in ARPE-19 cells in response to blue light irradiation (Fig. 8a, b).The treatment with FB or LUT in the absence of irradiation induced a significant increment of the NRF2 levels, evidencing the ability of both compounds to induce cell endogenous mechanisms involved in the control of ROS balance.In this scenario, while the treatment with LUT prevented the blue-light-induced increase in NRF2 expression, the ARPE-19 cells treated with FB and subsequently exposed to blue light significantly upregulated nuclear NRF2 to similar levels than blue-lightirradiated cells (Fig. 8a, b).These results highlight some similarities as well as differences in the protective mode of action of both antioxidant compounds.

Discussion
Human exposure to blue light has increased over the past few years mainly due to the incorporation of light-emitting diodes (LEDs) to everyday use technologies.This raises relevant concerns about the ability of blue light to induce certain damage.Ocular structures can be a target of particular interest, as they are directly exposed to irradiation, and links between blue light exposure and some eye disorders (such as dry eye and age-related macular degeneration) have been previously proposed [5,54].The present work was conceived in the context of simulating chronic exposure to blue light of a representative RPE human cell line, ARPE-19.In this sense, the workflow was designed to recreate consecutive short exposures (40 min) to blue light, up to a total exposure of 3 h and 20 min.LUT, a well-known xanthophyll carotenoid with antioxidant potential, was used in parallel to FB along the study for the sake of comparison [41,55].Additional value of our experimental setup is given by the fact that the irradiation was performed in the absence of the compounds (FB or LUT) to avoid the filter effect attributed to LUT in previous studies, and thus revealing in this case the direct effect of the compounds on the cells.Our results indicated that increasing doses of blue light induce a significant decrease in cell viability, accompanied by morphological changes compared to what is observed in physiological conditions.This matches previous results generated using both RPE and skin epithelial cells exposed to different blue light doses [14,56], validating our experimental setup and supporting its use for investigating both the downstream effects of blue light exposure and the protective effect of the hydrophilic botanical extract from Polypodium leucotomos, FB.FB and LUT doses capable of restoring cell viability of blue-light-irradiated cells to that observed in non-treated and non-irradiated cells were selected for this study.
The aim of this study was to evaluate the photoprotective capacity of FB against blue-light-induced damage.Evidence has been gathered by our group and others supporting FB as an interesting candidate compound to exert protection against the cellular alterations triggered by blue light.First of all, FB has strong antioxidant and anti-inflammatory properties [57].Second, using cellular models mainly derived from the skin, FB has been shown to prevent DNA damage induced by UV and visible light and to reduce inflammation in skin models [28,33,34].Third, the treatment with concentrations of FB similar to those used in this study prior to the exposure to blue light resulted in overall positive outcomes in skin cells, mainly by preventing photooxidation [33].Interestingly, FB is currently used in clinical practice bringing important benefits not just for general photoprotective strategies, but also as a particularly effective compound in specific pathological conditions [27,58].All this preceding work has paved the way to test the potential benefits of FB against the harmful effects of blue light in RPE cells.
Blue light can highly increase ROS production in soft tissue models, which has motivated its use as an anti-tumoral tool capable of promoting ROS-induced apoptosis in the context of cancer research [59].Related to detrimental oxidative stress in our experimental framework, we evaluated whether the pre-treatment with FB affects ROS production in response to blue light irradiation.Indeed, we found significantly higher ROS levels induced by blue light alone compared to the cells treated with FB prior to blue light exposure.This resembles the results also obtained using LUT, supporting that both compounds exert antioxidant effects partially preventing the strong blue-light-driven induction of ROS, but yet not leading to reach similar ROS levels of the non-irradiated cells.One of the main detrimental effects of a sharp increase of intracellular ROS is DNA damage.In this study, γH2AX has been used as a well-established marker to evaluate the induction of double-strand DNA breaks caused by blue light.According to previous results reported by Chen et al. [16], positive expression of γH2AX was detected in a significantly higher proportion of ARPE-19 cultured cells after exposure to blue light.Although this blue light-induced increase in γH2AX positive nuclei was prevented by the pre-treatment with both LUT or FB, the effect of FB was significantly more potent.In this scenario, as previous studies have pointed at the supplementation of LUT through the diet as a tool to restrain the induction of γH2AX in the outer nuclear layer of the retina [60], FB comes up as an ideal candidate to be further tested for this aim using in vivo models.However, the limited blue-light-mediated induction of γH2AX in ARPE19 cells pre-treated with FB could also be revealing an impaired deposition of this modified histone and thus leading to a delayed activation of repair mechanisms in the event of DNA damage.Importantly, this was ruled out in light of the significant induction of γH2AX detected in response to the sole treatment with the compounds (in non-irradiated cells), which suggested that FB and LUT could be priming the cells to be able to efficiently activate DNA repair mechanisms upon blue light irradiation.In agreement with this, and connecting the results related to cell viability, oxidative stress, and cellular systems put in motion to trigger DNA repair mechanisms, we observed a partial rescue of cell survival and ROS levels upon blue light irradiation in cells pre-treated with the compounds, supporting the action of the compounds to limit DNA damage.This led us to seek further evidence of cell cycle arrest mechanisms and to investigate the presence of newly synthetized DNA that may correspond to recently repaired regions.
With this in mind, as blue light also influences the cell cycle dynamics [13], we carried out EdU incorporation assays that allow the detection of this thymidine analogue with fluorescent azide dyes via a copper-mediated click reaction.Our results revealed a significant decrease of EdUpositive nuclei in ARPE-19 cultures 24 h after blue light exposure, which was slightly prevented by pre-treatments with LUT or FB, although the resulting levels did not reach those of the control.The cell cycle arrest induced by blue light was in accordance with the increased expression of the well-known cyclin-dependent kinase inhibitor P21 detected by Western blot.Particularly, P21 has been shown to mediate p53-induced G1/S cell cycle arrest [35].In line with previous data by Nishio et al. [17] the expression of P21 significantly increased after blue light exposure, while the pre-treatments with LUT or FB were both unable to prevent this effect.As the γH2AX-related results suggested only a partial prevention of DNA damage exerted by the pre-treatment with any of the compounds, the trigger of elevated P21 levels even in the presence of FB or LUT can be interpreted as a strategy to allocate the time needed for DNA repair.In fact, taken together, our findings reveal that the administration of FB or LUT prior to blue light irradiation not only was associated with increased levels of P21, but also enabled ARPE-19 cells to partially restore the rate of EdU incorporation compared to non-irradiated cells.In agreement with a reduction observed in the percentage of γH2AX positive nuclei, this could also be suggesting the active contribution of DNA repair mechanisms [61] that may be favoured by FB and LUT.Overall, this points at the capacity of FB to prime the cells to engage in effective DNA repair upon irradiation with blue light.
Mechanistically, among the events triggered inside the cell in response to oxidative stress, the induction of expression and nuclear translocation of the NRF2 transcription factor plays a central role for the regulation of cell survival and maintenance of redox homeostasis.ROS can specifically lead to the dissociation of the NRF2 protein from its cytoplasmic inhibitor, which is followed by the subsequent translocation of NRF2 into the nucleus [62].Our results using FB were aligned with previous work using keratinocytes [63], showing that FB induced a mild increase of nuclear NRF2 by itself but boosted it to much higher levels under stress conditions, here represented by blue light exposition.While higher concentrations of LUT have been reported to induce nuclear translocation when total NRF2 levels are boosted by a different stimulus such as hyperglycemia [64], our experimental setup revealed that the pre-treatment with LUT restored the NRF2 nuclear levels to values similar to those of the control.Interestingly, the pre-treatment with FB exerted opposed effects to LUT, permitting a comparable increase of nuclear NRF2 to that observed in response to blue light alone.Overall, our results suggest that the NRF2 axis is activated in ARPE-19 cells responding to blue light irradiation after pre-treatment with FB and may be a key pathway to mediate its protective effect.In this regard, future studies are needed to more deeply characterise the time course of the NRF2 induction and nuclear translocation in response to blue light following the pre-treatment with FB, as the precise dissection of this response may shed light on the distinctive protective action of FB.
Finally, the secretion of inflammatory cytokines is another hallmark of the cellular responses observed in several retinal diseases that are associated with elevated ROS levels [65].Related to this, blue light has been attributed a dual role, both as a pro-inflammatory stimulus and as a potential modulator of the secretion of inflammatory cytokines.In accordance to the pro-inflammatory role of blue light described by Yoo et al. [52], TNF-α secretion significantly increased upon blue light irradiation in our in vitro model system using ARPE-19 cells.Thus, our results not only confirmed the role of LUT to partially limit this increase in TNF-α secretion, which is in agreement with previous literature [55], but also show how notably FB was able to fully rescue the levels of secreted TNF-α to those observed in non-irradiated cells.It is known that low level chronic inflammation (known as inflammaging [66]) can trigger numerous disorders.Among them, specific degenerative pathologies affecting eye structures such as age-related macular degeneration have been previously associated with increased circulating levels of proinflammatory factors.Therefore, the distinctive ability of FB to control the production of inflammatory mediators by epithelial cells may bear a high translational value in different scenarios, particularly interesting to be investigated in the context of barrier tissues such as the skin and the intestinal epithelia.

Conclusions
Our results highlight the potential benefits of FB to prevent retinal damage by reducing the detrimental impact of increasing exposure to blue light from digital devices.FB specifically acts similarly to LUT by limiting the induction of both ROS production and signs of DNA damage; also preventing the drop in cell viability and in the rate of DNA synthesis, the latter being likely associated with DNA repair responses.Interestingly, the production of inflammatory mediators by RPE cells was much more efficiently controlled by FB than LUT.On the other hand, nuclear translocation of NRF2 induced by blue light is restricted by LUT but maintained at elevated levels upon pre-treatment with FB, suggesting that this may constitute a key feature contributing to orchestrate the protective response characteristic of FB.In sum, here we expand the previous knowledge about the beneficial effects of FB by providing direct evidence of its photoprotective role against cellular damage induced by blue light irradiation in the RPE.In light of these findings, FB emerges as a promising candidate to be clinically used for the design of novel prophylactic strategies that limit the impact of blue light on human health.

Fig. 1
Fig. 1 Effect of blue light irradiation on the viability and morphology of ARPE-19 cells.Cells were irradiated with increasing cumulative doses of blue light.a Results of the MTT assay relativized to the control (non-irradiated) cells.Error bars denote ± S.E.M. (n = 3, one-way ANOVA, *p < 0.05; **p < 0.01; ***p < 0.001).b Phase-contrast images illustrating the changes in cell morphology after the specified irradiation conditions.All images were taken at the same amplification.Scale bar = 40 μm

Fig. 2 Fig. 3
Fig. 2 Effect of LUT and FB treatments on cell viability and morphology of blue-light-irradiated ARPE-19 cells.Cells were treated with different concentrations of LUT or FB. a Results of MTT assay in response to blue light and increasing concentrations of LUT.b Results of MTT assay in response to blue light and increasing concentrations of FB. c Phase-contrast microscopy images illustrating the changes in cell morphology after the treatments.All images were taken at the same amplification.Scale bar = 50 μm.For (a) and (b), error bars denote ± S.E.M. (n = 3, one-way ANOVA, *p < 0.05; **p < 0.01; ***p < 0.001; and n = 3, oneway ANOVA #p < 0.05).C refers to non-treated and nonirradiated cells

Fig. 5
Fig. 5 Expression of P21 protein in ARPE-19 cells.Expression levels of P21 in the different experimental conditions tested.GAPDH was used as a loading control.All the results are normalized to nontreated and non-irradiated (control) cells.Bands were quantified and analysed using the Image Lab version 3.0.1 software.Error bars denote ± S.E.M. (n = 3, one-way ANOVA, *p < 0.05).The concentration of the compounds used for pre-treatment were 0.1 μg/ml LUT and 0.1 mg/ml FB.C refers to non-treated and non-irradiated cells

Fig. 8
Fig. 8 NRF2 expression and localization in ARPE-19 cells after treatments.a Fluorescence microscopy images of NRF2 in ARPE-19 cells.Counterstain with Hoechst-33258 was used as a control for nuclei visualization.All images were taken at the same amplification.Scale bar: 40 µm.b Quantification of nuclear NRF2 expression as the mean fluorescence intensity (MFI) using FIJI Software.Error bars