The UVR Filter Octinoxate Modulates Aryl Hydrocarbon Receptor Signaling in Keratinocytes via Inhibition of CYP1A1 and CYP1B1

Abstract

Ultraviolet radiation (UVR) is a consistent part of the environment that has both beneficial and harmful effects on human health. UVR filters in the form of commercial sunscreens have been widely used to reduce the negative health effects of UVR exposure. Despite their benefit, literature suggests that some filters can penetrate skin and have off-target biological effects. We noted that many organic filters are hydrophobic and contain aromatic rings, making them potential modulators of Aryl hydrocarbon Receptor (AhR) signaling. We hypothesized that some filters may be able to act as agonists or antagonists on the AhR. Using a luciferase reporter cell line, we observed that the UVR filter octinoxate potentiated the ability of the known AhR ligand, 6-formylindolo[3,2-b]carbazole (FICZ), to activate the AhR. Cotreatments of keratinocytes with octinoxate and FICZ lead to increased levels of cytochrome P4501A1 (CYP1A1) and P4501B1 (CYP1B1) mRNA transcripts, in an AhR-dependent fashion. Mechanistic studies revealed that octinoxate is an inhibitor of CYP1A1 and CYP1B1, with IC50 values at approximately 1 µM and 586 nM, respectively. In vivo topical application of octinoxate and FICZ also elevated CYP1A1 and CYP1B1 mRNA levels in mouse skin. Our results show that octinoxate is able to indirectly modulate AhR signaling by inhibiting CYP1A1 and CYP1B1 enzyme function, which may have important downstream consequences for the metabolism of various compounds and skin integrity. It is important to continue studying the off-target effects of octinoxate and other UVR filters, because they are used on skin on a daily basis world-wide.

Throughout life, humans are consistently exposed to varying levels of ultraviolet radiation (UVR), which includes 2 main types of UVR: UVA (320–400 nm) and UVB (280–320 nm). This exposure produces some beneficial effects; however, the connection between UVB exposure with an increased risk of skin cancer is well established (Leffell, 2000; Narayanan et al., 2010). DNA absorbs UVB, which can result in potentially mutagenic lesions that may promote the development of cancer (Svobodova et al., 2006). To reduce this risk, sunscreens are widely used to protect the skin from the spectrum of UVA and UVB that pass through the earth’s atmosphere (Diffey et al., 2000; Sambandan and Ratner, 2011). Commercially available sunscreens usually contain a mixture of organic and/or inorganic compounds that function as UVR filters by either absorbing or reflecting UVR away from the skin. UVR filters have been shown to be effective in reducing both sunburns and DNA damage from UVR exposure (Green et al., 2011; Sambandan and Ratner, 2011).

Despite this protection, literature suggests UVR filters do not stay exclusively on the skin surface. Several organic filters are able to penetrate through the outer layers of the epidermis and reach systemic circulation at measurable concentrations (Hayden et al., 2005; Janjua et al., 2004). Matta et al. (2019) very recently demonstrated that, under maximal application conditions, the plasma levels of selected UVR filters exceeded FDA thresholds for waiving certain toxicological tests. Even more concerning is the evidence suggesting that some UVR filters may have hormone disrupting properties, such as estrogenic and/or antiandrogenic effects (Schlumpf et al., 2004). Further research is necessary to determine if UVR filters may have other off-target effects that could negatively impact human health. This is particularly important for babies and young children because their skin differs substantially from adult skin. Their skin is thinner and contains lower of levels of melanin, and many commercial products specifically target this population (Paller et al., 2011; Quatrano and Dinulos, 2013).

Many UVR filters are hydrophobic and contain aromatic rings, making them potential candidates for interactions with the Aryl hydrocarbon Receptor (AhR). The AhR is a ligand-activated transcription factor that regulates the expression of various enzymes involved in the metabolism of both exogenous and endogenous compounds, and it has been shown to be important in immune function and skin integrity (Furue et al., 2014; Ikuta et al., 2009; van den Bogaard et al., 2013, 2014). When inactive, the AhR remains in the cytoplasm complexed with chaperone proteins. Upon ligand binding, a conformational change occurs resulting in the translocation of the complex to the nucleus, where it interacts with the aryl hydrocarbon receptor nuclear translocator protein. This complex is then able to bind response elements in various target genes, such as cytochrome P4501A1 (CYP1A1) and P4501B1 (CYP1B1), to promote gene transcription. Some of these target proteins are able to metabolize the ligand that initially activated the AhR, effectively turning off the signaling pathway activation (Chiaro et al., 2007; Schiering et al., 2017; Weiss et al., 1996). Studies show that the AhR has a flexible-binding site and is able to accommodate compositionally diverse ligands, many of which have hydrophobic and aromatic groups (Denison et al., 2002; Soshilov and Denison, 2014). Although there remains much to understand about the regulation of the AhR, literature suggests that the balance of AhR signaling may have important consequences for the skin (Di Meglio et al., 2014; Kiyomatsu-Oda et al., 2018; Tauchi et al., 2005).

One of the most well-known AhR ligands is 6-formylindolo[3,2-b]carbazole (FICZ), a tryptophan derivative that can be formed by light-dependent as well as light-independent pathways, which has been detected in cell culture media, in human UV-exposed keratinocytes, and in human skin (Fritsche et al., 2007; Öberg et al., 2005; Rannug et al., 1987; Schallreuter et al., 2012; Smirnova et al., 2016). FICZ is a potent AhR ligand, although its half-life is relatively short (Wei et al., 1998, 1999). Metabolites of FICZ have been detected in human urine, suggesting that it is likely important for AhR signaling in humans (Wincent et al., 2009). To date, there has been little work done exploring potential off-target interactions between organic UVR filters and the AhR. The goal of this study was to screen commonly used, FDA approved, organic UVR filters to identify if any are capable of acting as an AhR agonist or antagonist and to identify mechanisms behind any observed AhR modulation.

MATERIALS AND METHODS

Cell culture

H1L1.1c2 cells (Garrison et al., 1996), a gift from the Dr Paige Lawrence Lab (University of Rochester Medical Center) and originally developed by Dr Michael Denison (UC Davis), were used to screen UVR filters for AhR agonist or antagonist function. These cells are a mouse hepatoma cell line that has been stably transfected with a luciferase expression vector containing dioxin response elements that confer responsiveness to AhR ligands (Garrison et al., 1996). HaCaTs are a spontaneously transformed immortal human keratinocyte cell line that is commonly used in keratinocyte and skin studies (Boukamp et al., 1988). Both cell types were grown in Dulbecco’s Modified Eagle Medium (DMEM) (Gibco 11995-065) supplemented with 10% fetal bovine serum (Gibco 10082-147) and 2% penicillin/streptomycin (Gibco 15140-122) in a 37°C 5% CO₂ incubator. For UVR studies involving HaCaTs, unless otherwise specified, the media used in these experiments contained supplemental tryptophan (Sigma T8941) so the final concentration of tryptophan was 1 mM. Supplemental tryptophan was not necessary for any of the UVR experiments involving the H1L1.1c2 cell line.

Reagents

All of the UVR filters that were studied, with the exception of octinoxate (OT) (Spectrum Chemical, New Brunswick, New Jersey, catalog number: O1123), were purchased from Sigma-Aldrich (St Louis, Missouri, Oxybenzone catalog number: PHR1074, Avobenzone catalog number: PHR1073, Octisalate catalog number: PHR1081, Homosalate catalog number: PHR1085, Octocrylene catalog number: PHR1083). Stock solutions were made in dimethyl sulfoxide (DMSO) and the final concentration of DMSO applied to cells was less than or equal to 0.1%. For cells that were treated with both UVB and OT, OT treatment occurred immediately after UVR exposure.

H1L1.1c2 luciferase reporter cell assay

H1L1.1c2 cells were seeded in 12-well plates at approximately 0.3 × 10⁶ cells per well. The following day the cells were incubated with their respective treatments for the specified timepoints. At the end of the incubation, each well was washed with PBS and then 200 µl of Mammalian Protein Extraction Reagent (MPER ThermoFisher Scientific catalog number: 78501) was used to extract protein according to the manufacturer’s instructions. Samples were centrifuged for 10 min at 12 000 rpm, the supernatant was transferred to a clean tube. From each sample, 50 µl of supernatant was added to a 96-well white plate, and 50 µl of reconstituted Bright-Glo luciferase assay reagent (Promega catalog number: E2610) was added. The plate was immediately covered in foil and allowed to incubate at room temperature for 10 min. A Turner Biosystems Modulus Microplate Reader (9300-002) was used to read the luminescence intensity of each well.

In vitro UVR exposure

UVR exposures of cells were conducted using a Hand-Foot UVB system (National Biological Corporation 1100-A) that was calibrated to UVB output using an IL1700 Research Radiometer (International Light SED240). The UVB system contained FS20 Sunlamps (Westinghouse) that emit between 290 and 320 nm and have a peak emission of 311 nm. The radiometer was used to measure the average flux (J/s) within the exposure area. The exposure time required to achieve a particular dose was calculated as previously described (Mortensen et al., 2008, 2013). Cells were exposed to UVR through a UVC blocking filter (Schott WG 295 glass filter BES Optics) inX1 PBS. After exposure the PBS was removed and media was added.

RNA isolation and qPCR

After treatment, cells were washed with PBS and then incubated at room temperature for 5 min with 250 µl/well (12-well plate) of TRI reagent (ThermoFisher Scientific catalog number: AM9738). Total RNA was extracted using the E.Z.N.A Total RNA Kit I (Omega Bio-Tek catalog number: R6834-01) according to the manufacturer’s instructions. The RNA concentration of each sample was measured using a NanodropLite Spectrophotometer (ThermoScientific 3259). After RNA isolation, 500 ng of RNA was reverse transcribed to cDNA using a qScript cDNA Kit (Quantabio catalog number: 95047-100) and a SimpliAmp Thermal Cycler (Applied Biosystems by Life Technologies A24812). The resulting cDNA was used to determine relative mRNA expression of various target genes using the PerfeCTa SYBR Green SuperMix Reaction Mix (Quantabio catalog number: 95047-100) and a CFX Connect Real Time System (Biorad CFX Connect Optics Module). HaCaT mRNA levels were expressed as they related to the housekeeping gene HPRT levels. Primer sequences (Integrated DNA Technologies) used are as follows: for human CYP1A1 (Gene ID: 1543) forward 5′-TAGACACTGATCTGGCTGCAG-3′ and reverse 5′-GGGAAGGCTCCATCAGCATC-3′ (Fritsche et al., 2007); for human CYP1B1 (Gene ID: 1545) forward 5′-CATGCGCTTCTCCAGCTTTGT-3′ and reverse 5′-GGCCACTTCACTGGGTCATGA-3′ (Borland et al., 2014); for human AhR (Gene ID: 196) forward 5′-TGGTCTCCCCCAGACAGTAG-3′ and reverse 5′-TTCATTGCCAGAAAACCAGA-3′ (Frauenstein et al., 2013); for human HPRT (Gene ID: 3251) forward 5′-TGCTGAGGATTTGGAAAGGG-3′ and reverse 5′-ACAGAGGGCTACAATGTGATG-3′ (Hoang et al., 2017).

For RNA isolation from mouse skin, approximately 20 mg of ear tissue was placed in a bead homogenizer tube (VWR catalog number: 10158-612) with 500 µl of Tri Reagent. The tissue was homogenized for 30 s on medium speed using an Omni International Bead Ruptor 12 homogenizer. An additional 500 µl of Tri reagent was added, and RNA isolation and qPCR from this point on was conducted as described above. Primer sequences (Integrated DNA Technologies) used are as follows: for mouse CYP1A1 (Gene ID: 13076) forward 5′-CCTCATGTACCTGGTAACCA-3′ and reverse 5′-AAGGATGAATGCCGGAAGGT-3′; for mouse CYP1B1 (Gene ID: 1545) forward 5′-ACATCCCCAAGAATACGGTC-3′ and reverse 5′TAGACAGTTCCTCACCGATG-3′; for mouse GAPDH (Gene ID: 14433) forward 5′-TCTCCCTCACAATTTCCATCCCAG-3′ and reverse 5′-GGGTGCAGCGAACTTTATTGATGG-3′ (Siddens et al., 2015).

CellTiter-Glo assay

CellTiter-Glo Luminescent Cell Viability Assay (Promega catalog number: G7572) was performed using HaCaT cells according to the manufacturer’s protocol. Briefly, HaCaT cells were seeded into white opaque walled 96-well plates at a density of 1.5 × 10⁴ cells per well and incubated for 24 h with the indicated treatment. In the case of UVB treatment, the cells were exposed to UVB as previously described and the assay was performed 24 h after exposure. After incubation, 100 µl of reconstituted CellTiter-Glo reagent was then added to each well and the plates were gently mixed for 2 min on a rocker plate. Luminescence was recorded using a PerkinElmer EnSpire Multimode Plate Reader (2300).

AhR siRNA knockdown

Knockdown experiments were conducted using the HiPerFect Transfection Reagent and vendor protocol (Qiagen 301705). Hs_AHR_5 FlexiTube siRNA (Qiagen catalog number: SI02780148) was used for AhR knockdown, and the AllStars Negative Control siRNA (Qiagen catalog number: 1027280) was used for a negative control. HaCaT cells were seeded in 24-well plates at a density of 4 × 10⁴ cells per well and treated with the appropriate siRNA following the Fast Forward Protocol provided by the vendor. Forty-eight hours after transfection the cells were treated and incubated for 6 h. RNA was isolated and qPCR was performed as described above.

Cell-based EROD assay

7-Ethoxyresorufin-O-deethylase (EROD) assays were conducted using H1L1.1c2 cells according to the previously published protocols (Zhang et al., 2012). CYP450 enzymes are able to convert 7-ethoxyresorufin (7-ER) to resorufin, which is a highly fluorescent compound with an emission maximum at 587 nm. This assay traditionally has been performed using hepatocytes (Barhoumi et al., 2000; Lee et al., 1993). Here, H1L1.1c2 cells were incubated with the specified treatments for 24 h in 12-well cell culture plates and the plates were frozen to lyse the cells. For the assay, plates were rapidly thawed, placed on ice, and 250 µl of EROD reaction mix containing the substrate 7-ER (Cayman Chemical catalog number: 16122) and NADPH (Millipore Sigma catalog number: 10107824001) was added to each well. The reaction mix contained 168.75 µl 50 mM Tris pH 7.4, 50 µl of 5.32 mg/ml BSA in 50 mM Tris, 12.5 µl of 100 µM of 7-ER in MeOH/50 mM Tris 15%/85%, and 18.75 µl of 6.7 mM NADPH in 50 mM Tris. After addition of the reaction mix, the plate was covered in foil and transferred to a 37°C incubator for 20 min. The reaction was terminated using 300 µl of 2 M glycine. Well contents were removed and centrifuged at 12 000 rpm for 2 min, 100 µl of the supernatant was transferred to a 96-well plate, and fluorescence intensity at 580 nm was immediately measured.

Cell-free CYP1A1/1B1 inhibition assays

To test for CYP1A1 and CYP1B1 inhibition specifically, P450-Glo CYP1A1 and P450-Glo CYP1B1 Assay Systems (Promega catalog numbers: V8751, V8761) were purchased and their protocols for biochemical CYP inhibition assays were followed. Briefly, 12.5 µl of water (untreated) or varying concentrations of OT were added to a 96-well plate. The CYP1A1/1B1 reaction mixtures were prepared according to the vendor’s instructions and 12.5 µl were added to the test wells. The plate was preincubated at 37°C for 10 min, after which the reaction was initiated by adding 25 µl of x2 NADPH regeneration system (Promega catalog number: V9510). The plate was incubated for an additional 10 min at 37°C, and then 50 µl of reconstituted luciferin detection reagent was added. After mixing gently, the plate was incubated at room temperature for 20 min. Luminescence was recorded using a Turner Biosystems Modulus Microplate. Additional reagents for the assays were recombinant human CYP1A1 enzyme (Sigma-Aldrich catalog number: C3735-1VL), CYP1B1 enzyme (Sigma-Aldrich catalog number: C3860-1VL), and NADPH regeneration system (Promega catalog number: V9510).

Mouse model and exposure

We maintain a colony of SKH1 mice that were previously partially backcrossed onto a C57Bl/6 background (approximately 60% C57Bl/6 as determined by a Charles River SNP panel) (Jatana et al., 2017; Palmer et al., 2019a,b). These mice harbor a mutation in the Hairless (hr) gene on chromosome 14. They experience a normal initial hair growth cycle, and then progressively lose their hair until they reach weaning age at which they are completely hairless. The mice were kept on a 12-h light/dark cycle, and had access to food and water ad libitum. Four to five-month old female mice were treated with 20 µl of either a vehicle (1:1 acetone: DMSO), 10 µM FICZ, 7.5% OT, or both on their ears (10 µl per side). Nine hours after treatment, the mice were euthanized and the ears were removed for RNA isolation and qPCR analysis. For UVR experiments, mice were exposed to varying doses of UVR emitted from UVA-340 lamps (Q-Labs) calibrated along the UVB wavelength using an IL1700 Research Radiometer (International Light SED240). These lamps emit UVR in the wavelength range of 365 to 295 with a peak emission at 340 nm. The amount of time required to achieve the specified doses was calculated from the average flux of the lamps, as described above. All of the mouse experiments were approved by the University Committee on Animal Resources at the University of Rochester Medical Center (UCAR No. 2010-024E/100360), and all experiments were conducted in accordance with the appropriate regulations and guidelines.

Statistical analysis

All statistical analyses were performed using GraphPad Prism 5. One-way and 2-way ANOVAs were conducted as needed and the Bonferroni post hoc test was used to determine significance between different groups. The * symbol was primarily used to indicate a group is significant when compared with the specified control. In cases of figures making multiple comparisons, * indicates a group is significant when compared with untreated groups, and ^ indicate a group is significant when compared with the FICZ alone or UVR alone treated group. A difference was considered statistically significant if the p value was < .05. Error bars depict standard error mean.

RESULTS

Screening of Common UVR Filters Using H1L1.1c2 Cells

Many UVR filters are hydrophobic and contain aromatic rings, and we hypothesized that they may be able to modulate AhR signaling. To examine this, we first used the H1L1.1c2 AhR luciferase reporter cell line. This cell line has been stably transfected to express a luciferase reporter gene that is responsive to AhR activation, allowing a correlation of luminescence with AhR activation (Han et al., 2002). FICZ was shown to strongly induce AhR activation in this cell line in a dose-dependent manner (Figure 1A). We examined several FDA approved organic UVR filters that are commonly found in commercial sunscreens including oxybenzone (Oxy), avobenzone (Avo), octisalate (Octi), homosalate (H), octocrylene (Octo), and octinoxate (OT). The doses chosen were approximately 100 times lower than the calculated maximum human exposure in an area equivalent to a 12-well cell culture plate, based on FDA guidelines for formulations and application (Supplementary Table 1) (Ou-Yang et al., 2012). Supplementary Table 2 contains the structures for each of the filters and the wavelength range that each filter absorbs in Sambandan and Ratner (2011). All of the UVR filters were tested either by themselves, to determine if the compound is an AhR agonist, or simultaneously in solution with 10-nM FICZ, to determine if it is an AhR antagonist. Studies were done using 10 nM of FICZ because it sufficiently induced AhR activity in the H1L1.1c2 cell line, but the signal was not so strong that it might mask potential changes from cotreatments (Figure 1A). Of the compounds tested, none exhibited intrinsic AhR agonist or antagonist activity after a 6-h incubation at the specified concentrations. However, when H1L1.1c2 cells were treated with OT and FICZ simultaneously we observed a statistically significant increase in luminescence that was greater than what FICZ was able to elicit alone (Figure 1B). These results suggest that OT is not an AhR agonist, but it is able to potentiate FICZ’ ability to activate the AhR. Luciferase induction in the H1L1.1c2 cell line is transient, and it has been reported that maximal luciferase induction occurs 4- to 6-h postactivation (Han et al., 2002). We next examined if the potentiation effect was observable at later timepoints. We quantified the ratio of the luminescence intensity of H1L1.1c2 cells treated with both OT and FICZ to the luminescence intensity of cells treated with FICZ alone at various timepoints. We observed that the potentiation effect is statistically significant starting at 6-h posttreatment and remains elevated through 24 h (Figure 1C).

Figure 1.

Octinoxate (OT) elevates aryl hydrocarbon receptor (AhR) activation via 6-formylindolo[3,2-b]carbazole (FICZ) in H1L1.1c2 cell line. A, The H1L1.1c2 luciferase reporter cell line was treated with varying concentrations of FICZ to demonstrate that FICZ activates the AhR in these cells in a dose-dependent manner. Six hours after treatment, the cells were lysed and luminescence was quantified as a function of AhR activation. B, Selected ultraviolet radiation (UVR) filters were screened with and without 10 nM of FICZ to examine effects on AhR activation after a 6 h incubation period. The UVR filters and concentrations tested include 5 µM of avobenzone (Avo), 10 µM of oxybenzone (Oxy), 8 µM of octisalate (Octi), 10 µM of octocrylene (Octo), 15 µM of homosalate, and 10 µM of OT. Six hours after treatment cells were lysed the luminescence intensity was quantified as a function of AhR activation. * indicates significant when compared with untreated, ^ indicates significant when compared with cells treated with 10 nM FICZ using 1-way ANOVA with Bonferroni posttest (p < .0001). C, A time course of FICZ and/or OT incubations was conducted in H1L1.1c2 cells, and the fold change in luminescence compared with the untreated group and relative to FICZ was examined. ^ indicates significant when compared with FICZ alone at the respective timepoint using a 2-way ANOVA with Bonferroni posttest (p < .0001), n = 3 samples per group.

Many commercial sunscreen formulations contain multiple UVR filters to extend the coverage of protection across multiple wavelengths. To test whether or not OT’s effects are altered by the presence of additional UVR filters, we treated H1L1.1c2 cells with mixtures of the UVR filters we previously tested with or without FICZ. The concentrations for each of the individual UVR filters were the same concentrations used in our initial screening experiment (Figure 1B). Our results showed that, at the concentrations tested, the potentiation effect was still observable in cells treated with mixtures of 2, 4, or 6 UVR filters combined together with FICZ (Supplementary Figs. 1A–C).

It has been previously reported that UVR is able to activate the AhR both in vitro and in vivo, potentially via the formation of FICZ (Ma, 2011). To determine if coexposures of UVR and OT could potentiate AhR activation, we treated H1L1.1c2 cells with OT immediately following a low dose of UVR (40 mJ/cm² UVB) and examined luminescence 6 h after exposure. Our results confirm that OT is also able to potentiate the activation of the AhR by UVR in H1L1.1c2 cells (Supplementary Figure 2).

OT’s Effect on CYP1A1 and CYP1B1 in HaCaTs

Although our observations in the H1L1.1c2 liver cell line are interesting and clinically relevant given the observed skin penetration and systemic circulation of OT (Janjua et al., 2004; Mota Ade et al., 2013), it is important to discover if the AhR potentiation effect occurs in keratinocytes because they experience the greatest exposure to UVR and UVR filters. To examine this, we repeated the treatments on HaCaT keratinocyte cells and measured CYP1A1 and CYP1B1 mRNA expression 6-h posttreatments using qPCR. CYP1A1 and CYP1B1 are gene products of AhR transcription and are commonly used markers of AhR activation (Denison et al., 2002; Hao and Whitelaw, 2013). First, we performed a FICZ dose response on the HaCaT cells and examined the change in CYP1A1 and CYP1B1 mRNA levels. Our results showed that, whereas FICZ was able to increase mRNA levels of both genes, a greater amount of FICZ was required to produce statistically significant changes in CYP1B1 compared with CYP1A1 (Figs. 2A and 2B). This agrees with the published works of others who have shown that, in HaCaTs, FICZ does not produce as strong of an increase in CYP1B1 mRNA levels when compared with CYP1A1 mRNA levels (Nair et al., 2009). We then repeated the cotreatments of FICZ and OT using HaCaTs and we observed that the cotreatment potentiated the ability of FICZ to elevate CYP1A1 and CYP1B1 transcript levels (Figs. 2C and 2D). Overall, our data suggest that the potentiation effect we observed in the H1L1.1c2 cell line also occurs in keratinocytes.

Figure 2.

Cotreatments of 6-formylindolo[3,2-b]carbazole (FICZ) and octinoxate (OT) elevates CYP1A1 and CYP1B1 mRNA levels in HaCaT cells. A, HaCaT cells were exposed to varying concentrations of FICZ to demonstrate that FICZ elevates CYP1A1 and CYP1B1 mRNA levels relative to the housekeeping gene HPRT in a dose-dependent manner. Six hours after treatment the cells were lysed and relative mRNA expression was quantified using qPCR. B, HaCaT cells were treated with either 10 nM FICZ, 10 µM OT, or both compounds together and CYP1A1 and CYP1B1 mRNA levels were examined 6 h after treatment. * indicates significant when compared with untreated, ^ indicates significant when compared with FICZ alone using a 1-way ANOVA with a Bonferroni posttest (p < .0001), n = 5 samples per group.

As with the H1L1.1c2 cell line, we next examined if UVR, in the place of FICZ, could induce the AhR potentiation effect in HaCaT cells. It has been shown that in HaCaTs, the ability of UVR to activate the AhR depends to some extent on the presence of tryptophan in the cell culture media (Fritsche et al., 2007; Wei et al., 1999). The commercially available cell culture media DMEM (Gibco) contains tryptophan at a concentration of 78 µM. We suspected that this concentration might be too low to allow observable changes in CYP1A1 and CYP1B1 mRNA levels in UVR-exposed HaCaTs. Hence, we first compared the effects of various doses of UVR in either regular DMEM or DMEM supplemented with additional tryptophan (final concentration 1 mM) and examined CYP1A1 and CYP1B1 mRNA expression. At all of the UVB doses tested, except 10 mJ/cm² UVB, significant effects were observed in HaCaTs that received media with supplemental tryptophan, and not in those that received standard DMEM (Supplementary Figure 3A). In light of this finding, we treated HaCaTs with UVR and OT in tryptophan supplemented media for 6 h. UVR (40 mJ/cm² UVB) was able to increase the mRNA levels of both CYP1A1 and CYP1B1, but, at the timepoint tested the potentiation effect was only observed for CYP1B1 mRNA (Supplementary Figs. 3B and 3C). This dose of UVB is frequently used in HaCaT cell culture (Mera et al., 2010; Mortensen et al., 2015; Xu et al., 2006, 2009) and we confirmed that it does not significantly affect cell viability up to 24-h postirradiation (Supplementary Figure 4).

AhR Dependency of OT’s Potentiation Effect

Although CYP1A1 and CYP1B1 mRNA levels are commonly measured endpoints for AhR activation, there is literature to suggest that they are not solely controlled by the AhR (Delescluse et al., 2000). To determine if the potentiating effects of OT depend on the presence of functional AhR, we knocked down the AhR in HaCaT cells using siRNA and treated them with either FICZ, OT, or a combination of FICZ+OT. Our mRNA analysis of the AhR showed that the siRNA treatment effectively reduced AhR expression in treated HaCaTs (Figure 3A). This AhR knockdown completely inhibited the effects of FICZ and FICZ+OT to increase CYP1A1 and CYP1B1 mRNA levels, strongly suggesting that the AhR is required for the observed effects of OT on CYP1A1 and CYP1B1 mRNA levels (Figs. 3B and 3C).

Figure 3.

The effects of OT on mRNA levels in HaCaTs are aryl hydrocarbon receptor (AhR) dependent. HaCaTs were treated with either negative control siRNA or AhR-specific siRNA. Forty-eight hours later the cells were either exposed to 40 mJ/cm² UVB, 10 nM FICZ (6-formylindolo[3,2-b]carbazole), 10 µM OT (octinoxate), or combinations of those for 6 h. For cells that were treated with both UVB and OT, OT treatment occurred immediately after UVB exposure. qPCR was performed to analyze (A) AhR, (B) CYP1A1, and (C) CYP1B1 mRNA levels. * indicates significant when compared with groups treated with negative siRNA + 40 mJ/cm² UVB using a 1-way ANOVA with a Bonferroni posttest (p < .0001), n = 4 samples per group.

OT and CYP1A1/1B1 Inhibition

After observing the potentiation effect that OT exerts on FICZ activation of the AhR, we examined the mechanism of potentiation. We initially conducted preliminary UV-VIS and fluorimetry experiments that strongly suggested that FICZ and OT do not physically interact (data not shown). We hypothesized that OT may be having an inhibitory effect on the CYP1A1 and 1B1 enzymes. With regards to the canonical signaling pathway, AhR activation leads to gene transcription and translation of these enzymes, which serve to break down the ligand that activated the AhR and negatively regulate signaling. If OT inhibits these enzymes, the half-life of FICZ could be increased, potentially leading to sustained AhR activation. It has been shown previously that various compounds with weak-binding affinity for the AhR are able to indirectly activate the AhR via inhibition of CYP enzymes (Wincent et al., 2012).

To test for CYP inhibition by OT we first performed the cell-based EROD assay. Some cytochrome P450 enzymes can catalyze the de-ethylation of 7-ER to the fluorescent compound resorufin, and alterations in this may indicate enzyme inhibition (Guengerich, 1994; Shimada et al., 1998). We treated the H1L1.1c2 hepatoma cells with either FICZ, OT, or the FICZ+OT together and performed the EROD assay after a 24-h treatment. The relative amount of fluorescence measured after treatment corresponds to the catalytic activity of CYP450 enzymes present, and alterations in fluorescence values below positive control levels could suggest CYP450 enzyme inhibition. Our results (Figure 4A) showed that cells cotreated with FICZ and OT exhibited significantly reduced fluorescence when compared with cells treated with FICZ alone, suggesting that OT may function as a CYP inhibitor.

Figure 4.

EROD (7-ethoxyresorufin-O-deethylase) assay, CYP1A1 and CYP1B1 inhibition studies. A, The cell-based EROD assay was performed using H1L1.1c2 cells. The conversion of 7-ethoxyresorufin to the fluorescent compound resorufin was measured using UV-VIS spectrometry in cells treated with either 6-formylindolo[3,2-b]carbazole (FICZ), octinoxate (OT), or both. * indicates significant when compared with all other treatment groups using a 1-way ANOVA with a Bonferroni posttest (p = .0063), n = 3. B–E, The CYP450 Glo Assays were used to determine if OT is a CYP1A1 and/or CYP1B1 inhibitor. Estimated IC50 values were calculated for (B) CYP1A1 and (C) CYP1B1 by log transforming the concentration of OT and performing nonlinear regression using the log(inhibitor) protocol in GraphPad Prism 5. The process was repeated using α-naphthoflavone (a-NF) instead of OT to compare the relative IC50 values for (D) CYP1A1 and (E) CYP1B1.

Although the EROD assay is a useful tool to examine potential effects of our treatment on CYP450 enzymes, it is unable to discriminate between changes in the amount of enzyme present and the functional capacity of the enzymes. To confirm that the potentiation effect is due to CYP1A1 and/or CYP1B1 inhibition, we next performed a cell-free inhibition assay using the P450-Glo Assay 1A1 and 1B1 Systems (Promega). This assay utilizes enzyme-specific substrates that are linked with a luminescent compound. When the enzyme acts on the substrate, the compound is cleaved and the resulting luminescence correlates with enzyme activity. Compounds that are suspected inhibitors are added to a fixed mixture of enzyme plus substrate and the relative luminescent signals are used to assess inhibition activity as well as to calculate estimated IC50 values. Our results indicated a dose-dependent decrease in luminescence with increasing OT concentration, which indicates a decrease in catalytic function of both CYP1A1 and CYP1B1 (Supplementary Figs. 5A and 5B). We log transformed the concentrations of OT and performed nonlinear regression to calculate IC50 values for both enzymes. Results indicated that the IC50 of OT is approximately 1.0 µM and 586 nM for CYP1A1 and CYP1B1, respectively (Figs. 4B and 4C). OT’s function as a CYP1A1 and CYP1B1 inhibitor was compared, using the P450-Glo Assay system, to the IC50 values measured for α-naphthoflavone (α-NF), a known CYP1A1 and CYP1B1 inhibitor. ɑ-NF has literature reported IC50 values of 60 nM and 5 nM for CYP1A1 and CYP1B1, respectively (Shimada et al., 1998). Our studies measured ɑ-NF IC50 values of 33.4 nM for CYP1A1 and 2.35 nM for CYP1B1 (Figs. 4D and 4E). These values are close to the literature reported values with similar magnitude differences between CYP1A1 and CYP1B1. These results confirm that OT is a CYP1A1 and CYP1B1 inhibitor but is estimated to be x10–100 weaker than α-NF. Nonetheless, the data suggest the potential for OT skin exposure to induce off-target effects by modulating AhR in skin as discussed below.

OT’s Effect on CYP1A1 and CYP1B1 in Mouse Skin

After we observed the potentiation effect in vitro and identified the mechanism of action, we examined whether or not this effect was also observable in vivo. Literature has previously shown that topical application of FICZ on mouse ears elevates CYP1A1 mRNA expression between 3- and 12-h postapplication (Wincent et al., 2012). Mice approximately 4 months old were treated with 20 µl of either vehicle (1:1 DMSO: Acetone), 10 µM FICZ, 7.5% OT, or FICZ and OT together their ears. After 9 h, the mice were euthanized and the ears were taken and processed for qPCR. Our results showed that FICZ alone was able to slightly elevate the levels of both transcripts above vehicle control, and that the cotreatments of FICZ and OT significantly elevated them when compared with FICZ alone (Figs. 5A and 5B). Our results confirm that OT is also able to potentiate the effects of FICZ in an in vivo model.

Figure 5.

Octinoxate (OT) potentiates FICZ (6-formylindolo[3,2-b]carbazole ) elevation of CYP1A1 and CYP1B1 mRNA levels in mouse skin. The ears of mice between 4 and 5 months of age were treated with either vehicle, 10 µM FICZ, 7.5% OT, or both FICZ + OT together. Nine- hour posttreatment the mice were euthanized and the ears were taken for RNA extraction and qPCR analysis of (A) CYP1A1 and (B) CYP1B1 mRNA levels relative to the housekeeping gene GAPDH. * indicates significant when compared with vehicle, ^ indicates significant when compared with FICZ alone using a 1-way ANOVA with a Bonferroni posttest (p < .0001), n = 5 samples per group.

It has also been reported that UVR exposure can activate the AhR in vivo, as evidenced by increased CYP1A1 and CYP1B1 mRNA levels. We wanted to determine if cotreatments with UVR and OT would produce observable potentiation effects in vivo. We exposed mice to 1 dose of 270 mJ/cm² of UVB and applied a 7.5% OT solution to their ear immediately after exposure. To our surprise, qPCR analysis of CYP1A1 and CYP1B1 mRNA levels showed no change due to UVR exposure or UVR+OT exposure (Supplementary Figs. 6A and 6B). We performed a dose response to determine if this was due to too low UVR exposure, yet increasing the dose had no significant effect on either transcript (Supplementary Figs. 6C and 6D). It is plausible that this particular strain of mice does not exhibit the expected AhR response to UVR at 9-h post-UVR. Other possibilities include insufficient dietary level of tryptophan or allelic variations in the AhR gene in this mouse model as discussed below.

DISCUSSION

Our study has discovered a previously unappreciated off-target effect of OT on AhR signaling. OT is a common UVR filter formulated in consumer products at a concentration up to 7.5 wt% and is used either as the sole active ingredient or in mixtures to extend the range of UVR protection. It is also often formulated in products targeting infants and children. We have shown that cotreating a hepatoma reporter cell line, a keratinocyte cell line, and mouse skin with OT and the AhR ligand FICZ leads to an increase in CYP1A1 and CYP1B1 mRNA levels that is greater than what FICZ alone elicits. Our mechanistic studies indicate that OT is a CYP1A1 and CYP1B1 inhibitor. These results suggest that OT’s inhibition of these enzymes may lead to an accumulation of AhR ligands in the skin, potentially resulting in sustained AhR activation.

Although this is the first report of a UVR filter modulating AhR activity via enzyme inhibition, there is literature precedent for AhR activation from CYP450 inhibition. Several AhR activating compounds, particularly those that exhibit low AhR-binding affinity, have been shown to be CYP enzyme inhibitors rather than direct AhR activating ligands (Wincent et al., 2012). Hence, in the presence of inhibitors, AhR activation can occur even under conditions of low ligand concentration. Our assays showed that OT alone had minimal effects on both the AhR activation of the H1L1.1c2 cell line and CYP1A1 and CYP1B1 mRNA levels in HaCaTs and mouse skin. However, when OT was cotreated with FICZ, we saw an increase in AhR activation that was greater than what FICZ alone was able to elicit. Our results suggest that, in the presence of OT, FICZ metabolism is reduced, thereby increasing AhR activation when compared with FICZ only treatments (Figs. 6A and 6B).

Figure 6.

Proposed mechanism for octinoxate (OT) observed potentiation. A, Canonical aryl hydrocarbon receptor (AhR) signaling involves a negative feedback loop where ligand-activated AhR induces AhR-ligand complex translocation to the nucleus and interaction with xenobiotic response elements on AhR target genes. This promotes gene transcription of CYP450 enzymes that metabolize the ligand, eventually reducing AhR signaling. B, In the presence of OT, we propose that, after ligand-mediated AhR activation, OT will reduce ligand metabolism via CYP1A1 and CYP1B1 inhibition, subsequently increasing the duration of AhR activation. Abbreviations: AhRR, aryl hydrocarbon receptor repressor; ARNT, aryl hydrocarbon receptor nuclear translocator.

Although it has repeatedly been shown that cutaneous UV exposure in rodents, fish, and humans induces AhR responses locally in skin and internally (Fritsche et al., 2007; Ma, 2011; Memari et al., 2019), we did not observe that UVR exposure increased CYP1A1 and CYP1B1 mRNA levels in the skin of our mouse colony 9-h post-UVR exposure. It is possible that the lack of an AhR response from UVR could be due to allelic variations in the mouse AhR gene. It has been reported that there exists a low affinity allelic variation of the AhR gene (Smith et al., 2018). If our mice predominantly have the low affinity variant, it could explain why we did not observe changes in CYP1A1 and CYP1B1 mRNA levels in response to UVR exposure. These in vivo experiments are important to the translational relevance of our findings and could provide useful information with regards to systemic effects, and so we are currently examining other experimental approaches to study the potentiation of UVR by OT in mouse skin. These include more in-depth dose response and time course experiments to determine the optimal dose and timepoint for our particular mouse strain. Studying additional tissue types, such as liver, could also provide useful information and help us ascertain whether or not this is a skin-specific issue. In addition to studying these tissues after in vivo exposure, we can also isolate cells from different tissue types and expose them in vitro. Different strains of mice may require different UVR doses to elicit effects on the AhR, especially if their skin pigmentation differs greatly. By isolating cells from different tissue types and exposing them to UVR in vitro, we would be able to more accurately compare the responsiveness of our mouse colony to other strains of mice with less concern over different dose requirements.

It is important to recognize that the experiments performed in this study primarily tested a single dose of the UVR filters. Future studies should include more thorough dose response studies on these UVR filters, as it is possible that, at other concentrations, other UVR filters may modulate AhR signaling. For example, studies on benzophenone-3 (oxybenzone) showed that it is metabolized by various CYP450 enzymes, particularly CYP1A1 and CYP1B1, in both human and rat liver microsomes (Watanabe et al., 2015). It is possible that, at higher concentrations, some of these other UVR filters may impact the functionality of those enzymes, and subsequently influence AhR signaling. This need for thorough dose response studies also carries over to our studies on mixtures of UVR filters. It is possible that other concentrations of UVR filters may reduce the potentiation effect or possibly act synergistically with OT to produce an even greater potentiation. In addition, time course experiments would also provide additional important information, particularly in regards to our qPCR experiments. The majority of our qPCR experiments used a 6-h timepoint. Although we were able to see significant differences at this time, it would be useful to test additional timepoints so we can examine the duration of the effect. It would also be useful with regards to the UVR data, as we observed the potentiation effect of UVR+OT for CYP1B1 and not CYP1A1. It is entirely possible that, if the duration of treatment was extended, that we would see an effect.

It has been shown that UVR filters reduce the overall exposure of keratinocytes to UVR (Duale et al., 2010; Gordon et al., 2009). In light of this, for our cotreatment studies we first exposed the cells to UVR and then immediately applied OT after exposure. In the future, it would be interesting to repeat the experiments but instead pretreat cells with OT and then expose them to UVR. We would hypothesize that this rearrangement of treatment order would reduce or eliminate the potentiation effect, as there would be little or no opportunity for FICZ to form. Such studies would convey important translational data because sunscreen consumers are encouraged to apply the product prior to UVR exposure. However, consumers do not always follow these guidelines, which may allow for the activation of the AhR via UVR exposure and potential OT inhibition of CYP450 enzymes.

This variability in consumer application of sunscreen makes it difficult to estimate human exposure to OT and to determine if the concentration is high enough for concern over CYP1A1/1B1 inhibition. Our cellular studies with OT were performed at 10 µM, which is approximately 100-fold lower than the anticipated human exposure after 1 dose of OT. This suggests that humans apply levels of OT greater than the IC50 for both CYP1A1 and CYP1B1 in vitro. In addition, a single application of OT on mouse skin was sufficient to upregulate CYP1A1 and CYP1B1 mRNA levels with simultaneous FICZ application. Literature suggests that UVR filters, including OT, readily penetrate human skin and reach systemic circulation (Janjua et al., 2004; Matta et al., 2019). Under maximal use conditions circulating OT levels in human plasma can reach approximately 10 ng/ml, with basal levels increasing with daily application (Wang and Ganley, 2019). It is reasonable to expect that OT levels in the skin would be much higher. Such accumulation may have impacts on systemic organs as well as the skin. Topical application of OT on Wistar rats resulted in accumulation in the liver (Mota Ade et al., 2013). The liver highly expresses the AhR and CYP450 enzymes and plays an essential role in metabolism. Although more research is required, the liver could be significantly impacted by OT’s inhibition of CYP1A1 and CYP1B1 enzymes. The liver also expresses other CYP450 enzymes, and the effects of OT on those enzymes have not yet been studied.

Dysregulation of CYP450 enzyme activity may have far-reaching implications. The AhR plays an important, although not fully understood, role in skin integrity. In 2005, Tauchi et al. (2005), published work strongly suggesting that constitutive activation of the AhR may have significant negative consequences for skin barrier function. They generated a transgenic mouse model that expressed a constitutively active form of the AhR in skin. Starting between 3 and 5 weeks of age, the mice developed skin lesions and hair loss. This was accompanied by lymphocyte infiltration and a loss of subcutaneous fat. Activation of the AhR by the environmental toxicant 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) has been shown to increase dermal thickening and mast cell infiltration in a mouse model of atopic dermatitis (Ito et al., 2008). However, increased AhR activation may have beneficial outcomes in the skin. Coal tar, that contains AhR agonists, is a therapy for atopic dermatitis and has been shown to increase essential skin barrier proteins in an AhR-dependent manner (van den Bogaard et al., 2013). AhR activation via FICZ has been shown to reduce psoriasis symptoms and promote filaggrin expression (Di Meglio et al., 2014; Kiyomatsu-Oda et al., 2018).

These seemingly conflicting consequences of AhR activation in the skin could be explained by duration of AhR activation. FICZ is rapidly metabolized by CYP450 enzymes, whereas TCDD is relatively resistant to metabolism and has a long half-life. This disparity of half-life could result in TCDD and other slowly metabolized AhR ligands having prolonged effects when compared with rapidly metabolized ligands. For example, it has been previously suggested that treatment with TCDD and FICZ elicit different effects on CD4+ T cells; however, when duration of activation was controlled for these differences were nonexistent (Ehrlich et al., 2018). Other studies have been performed CD8+ T cells and their response to influenza; when the exposure dose of FICZ was increased and the metabolism of FICZ was decreased, the effects of FICZ on CD8+ cells resembled the effects of TCDD (Wheeler et al., 2014). While experiments studying such effects have yet to be done in a skin-specific context, it is reasonable to hypothesize that the different effects AhR activation appears to have in the skin may be due to duration of activation. It is also important to keep in mind that studies have shown that AhR modulating compounds appear to have different effects depending on the inflammatory condition of the skin (Haarmann-Stemmann et al., 2015). Although some ligands, such as FICZ, are metabolized quickly, the presence of a CYP1A1/1B1 inhibitor, such as OT, may increase the half-life of the ligand, leading to sustained AhR activation. In the skin specifically, this altered metabolism may promote various kinds of effects, which would likely at least partially be influenced by the inflammatory state of the skin.

The AhR, whereas often associated with the metabolism of compounds, also plays roles in various other biological processes, such as cell cycle progression. In vitro treatment with an AhR ligand has been shown to alter the cell cycle at varying points, but it also appears that the AhR exerts effects even in the absence of exogenous ligands; the authors of such studies have interpreted these effects as caused by interactions with the metabolic breakdown of endogenous ligands (Levine-Fridman et al., 2004; Ma and Whitlock, 1996; Puga et al., 2000; Weiss et al., 1996). siRNA knockdown of the AhR or CYP1A1 in HaCaTs leads to reduced DNA synthesis and an increased percentage of cells in the G₀/G₁ phase (Kalmes et al., 2011). It is reasonable to postulate that topical application of a CYP1A1 and 1B1 inhibitor, like OT, may alter keratinocyte’s normal progression through the cell cycle by causing sustained levels of endogenous AhR ligands. The AhR has also been shown to negatively regulate nucleotide excision repair via modulation of p27, and AhR knockout mice are more resistant to UVB-induced tumors than AhR competent mice (Pollet et al., 2018). In addition, the AhR has also a direct antiapoptotic function by directly inhibiting p27, leading to alterations in the activation and phosphorylation of retinoblastoma protein (Frauenstein et al., 2013). DNA damage repair and apoptosis are important defense mechanisms against the propagation of harmful mutations that may occur after UVR exposure, and changes in either of them may have drastic consequences to an organism as a whole. There is a clear, although not yet fully understood, role of the AhR in these mechanisms, and the presence of an inhibitor that can elevate AhR signaling by reducing the metabolism of endogenous ligands may disrupt the balance of these processes and alter skin homeostasis.

In conclusion, we have shown that the UVR filter OT is a CYP1A1 and CYP1B1 inhibitor, and that this inhibition results in a measurable elevation of AhR activation, which may have implications for skin homeostasis. Further research is necessary to begin to fully understand such implications. We are currently investigating whether or not this potentiation also occurs in vivo with cotreatments of UVR and OT, as well as downstream consequences this potentiation may have on the skin. In addition, it is important to consider if OT also influences other regulators of the AhR, particularly AhR proteolysis or negative regulation by the protein aryl hydrocarbon receptor repressor. Commercially available sunscreens are a necessary part of life for many individuals, and it is essential that we know as much as possible about the UVR filters in them to ensure that people are applying safest and targeted protection.

SUPPLEMENTARY DATA

Supplementary data are available at Toxicological Sciences online.