The skin forms a barrier between the body and its external environment in order to prevent the intrusion of pathogens and the uncontrolled loss of internal water and solutes. The epidermis is the outermost layer of skin and provides the first line of defense against ultraviolet (UV) radiation and other environmental factors. UV is divided into three ranges based on wavelength: UVA (320–400 nm), UVB (280–320 nm), and UVC (100–290 nm). Among these, UVC is unlikely to be present in terrestrial sunlight because it is blocked by the ozone layer, whereas UVA and UVB can come into contact with the skin. The skin can be separated by the basement membrane into two layers, the dermis and epidermis. The mammalian epidermis is composed of four layers, including the basal, spinous, granular, and cornified layers [1
]. UVA can penetrate the dermal layer of the skin and reach the capillaries, whereas UVB is blocked in the upper dermis. According to wavelength-dependent studies, UVB radiation has more cytotoxic and mutagenic effects than UVA does [2
]. Long-term exposure to UVB induces damage to both the dermal and epidermal skin [4
]. UVB-induced cell death is caused through directly induced DNA damage and indirect action mediated by the generation of reactive oxygen species (ROS) and nitric oxide [5
]. The main causes of UV-induced DNA damage are 2,3-cyclobutane pyrimidine dimer and a pyrimidine (6-4) pyrimidone photoproduct [6
]. In contrast, it is not fully understood what mechanisms are involved in damage to the skin barrier under low toxic conditions. An investigation of the effect of each toxic factor on the tight junction (TJ) barrier may be needed in order to clarify the molecular mechanism.
Human skin keratinocytes form a TJ at the most apical pole of the lateral membrane between neighboring cells. The TJ prevents the abnormal paracellular movement of water, solutes, and pathogens. Claudins (CLDNs) and occludin are integral membrane proteins in TJ, and they comprise a large family of 27 subtypes in mammals [7
]. Among them, CLDN1 and CLDN4 are highly expressed in human keratinocytes [9
]. Most CLDNs contain carboxyl-terminal PSD95/Dlg/ZO-1 (PDZ)-binding motifs that mediate interactions between CLDNs and scaffolding proteins such as ZO-1 and ZO-2. These scaffolding proteins indirectly link CLDNs to actin filaments. CLDN1-deficient mice develop aberrant stratum corneum barrier functions in the skin [10
]. A premature stop codon of CLDN1, resulting in a lack of CLDN1 protein, has been identified in neonatal ichthyosis and sclerosing cholangitis syndrome, a disorder characterized by scalp hypotrichosis, ichthyosis, scarring alopecia, and sclerosing cholangitis [11
]. Therefore, CLDN1 may have an important role in maintaining the barrier function in the skin.
An antioxidant effect is one of the key factors in reducing skin injury, aging, and cancer risk, and antioxidant activities have been reported in components extracted from plants, fruits, vegetables, and bee propolis [12
]. Propolis contains many classes of compounds, including flavonoids, phenolic acids, and others. The components of propolis are different in each area [14
]. The ethanol extract of Brazilian green propolis (EBGP) exerts strong antioxidant activity in mouse skin [15
] and protects human keratinocytes against UV-induced apoptosis [16
]. In addition, hydroalcoholic extracts of Brazilian propolis improve dermal burn healing [17
]. EBGP may be useful in protecting skin damage from various external stimuli, but the effect of EBGP on the TJ barrier has not been examined.
Human keratinocyte-derived HaCaT cells can easily be plated as a monolayer and form the TJ [18
]. Therefore, they may be useful in examining the function of TJ in the skin. In the present study, we investigated the effects of ROS and EBGP on the cellular localization and function of CLDN1 in HaCaT cells. In addition, the molecular mechanism of tight junctional localization of CLDN1 was assessed by immunoprecipitation, immunoblotting, and immunofluorescence measurements.
Disruption of the TJ barrier in the skin is caused by UV exposure and oxidative stress. In previous research, TJ integrity has been examined using high doses of UV, which can induce noticeable cell damage [24
]. The production of ROS and H2
was increased by UVB in a dose-dependent manner (Figure 1
A,B). ROS production has been reported to be increased by UVB irradiation in a dose- and time-dependent manner [25
]. Park et al. [26
] reported that ROS production was increased 1.3-fold by UVB irradiation (10 mJ/cm2
) and incubation for 8 h in HaCaT cells, which is similar to our data. Transient H2
treatment increased the production of ROS and decreased cell viability in a dose-dependent manner (Figure 1
D,E). The percentage of damaged cells was about 10%–20% in the present experimental conditions of transient H2
treatment. Transient H2
(200 μM) treatment induced the mislocalization of CLDN1 without affecting the amount of CLDN1 protein (Figure 2
). There was no apparent change in CLDN4 localization, but the amount of CLDN4 protein increased transiently after 6 h of H2
treatment. Recently, El-Chami et al. [27
] reported that H2
(1 mM, a higher concentration than in our study) induced the mislocalization of CLDN1, CLDN4, and occludin in a continuous cell line of rat epidermal keratinocytes. We suggest that oxidative stress selectively induces the mislocalization of CLDN1 in low-level cell toxic conditions.
The protein level of CLDN1 is decreased in atopic dermatitis, whereas that of CLDN4 is increased [28
]. The loss of CLDN1 localization in the TJ may induce a compensatory elevation of CLDN4 expression. Nevertheless, the barrier function of TJ was decreased after 6 h of transient H2
treatment (Figure 3
A). The TJ barrier function was reduced by the knockdown of CLDN1 by siRNA (Figure 4
A,B). Furthermore, the rescue effects of EBGP on the decrease in TER and increase in LY flux by H2
were blocked by the knockdown of CLDN1 (Figure 4
C). There results indicate that CLDN1 has an important role in the maintenance of the TJ barrier. Previously, Furuse et al. [29
] have reported that a continuous TJ was observed in the stratum granulosum of the epidermis in CLDN1-deficient mice and in WT mice using ultrathin section electron microscopy. However, a small molecule tracer (~600 D) passes through the TJ of the epidermis in CLDN1-deficient mice, whereas the diffusion is prevented at the TJ in WT mice. We suggest that CLDN1 cannot be replaced by CLDN4 in the skin.
-induced mislocalization of CLDN1 was inhibited by a clathrin-dependent endocytosis inhibitor (Figure 5
A), suggesting that H2
enhances the endocytosis of CLDN1 from the TJ to intracellular compartments. Although the protein levels of CLDN1 did not change, we did not detect the subcellular localization of CLDN1 in the organelle (Figure 2
). We have to clarify the subcellular localization of CLDN1 using organelle markers in further studies. The p-Thr level of CLDN1 was decreased by H2
A). The phosphorylation status of proteins is regulated by various protein kinases and PPs. Go6976 induced the dephosphorylation and mislocalization of CLDN1 (Figure 7
). On the contrary, the H2
-induced mislocalization of CLDN1 was rescued by the PKC activator PMA. These results suggest that the TJ localization of CLDN1 is upregulated by PKC-dependent phosphorylation. Dephosphorylation of CLDN1 is caused by the activation of PP2A [23
increased PP activities, and the mislocalization of CLDN1 was rescued by a PP inhibitor, cantharidin, indicating that PPs may also be involved in the dephosphorylation of CLDN1 by H2
. CLDN1 is phosphorylated in both renal tubular MDCK I and colonic HT29 cells, but PKC-induced changes in the phosphorylation state were detected only in MDCK I cells [21
]. The regulatory mechanism for the phosphorylation of CLDN1 may be different in each tissue.
Threonine phosphorylation sites of CLDN1 by PKC were predicted at T191 and T195 using the NetPhos 2.0 and Disphos 1.3 servers [22
]. A phosphorylation mimic T191E mutant was localized to the TJ and maintained the TJ barrier in cells treated with H2
, whereas WT and a T195E mutant lost their localization and barrier function in the TJ (Figure 6
C,D). These results indicate that phosphorylation at T191 may be necessary for CLDN1 to localize to the TJ in HaCaT cells. The necessity of phosphorylation at T191 has been reported using human embryonic kidney 293 cells [30
] and MDCK cells [31
caused a decrease in CLDN1 localization in TJ and an increase in paracellular permeability, which were rescued by EBGP (Figure 3
C,D). Similar effects were observed in the UVB-treated cells. Although the effect of EBGP may be due to antioxidant activity, some cinnamic acid derivatives and flavonoids, which are contained in propolis, have beneficial effects on PKC. Artepillin C enhances adipocyte differentiation and glucose uptake mediated by the activation of PKC [32
]. Chlorogenic acid prevents α-amino-hydroxy-5-methyl-isoxazole-4-propionate-mediated excitotoxicity in optic nerve oligodendrocytes through a PKC-dependent pathway [33
]. Caffeic acid phenethyl ester inhibits the expression and activity of PP2A [34
]. Our preliminary data indicate that kaempferide, which is abundant in EBGP [35
], had lower antioxidant activity compared to EBGP, but it rescued the H2
-induced mislocalization of CLDN1 and the reduction in the TJ barrier function. The components of propolis, including kaempferide, vary depending on the area [14
]. A comparison of the effects of propolis produced in various places may be good for the identification of active components. Further studies are needed on which components of EBGP can rescue the H2
-induced mislocalization of CLDN1 using human keratinocytes and what doses are effective. The identification of active components could lead to expanding the range of raw substances beyond green propolis, which could prove useful for the same functions.
In conclusion, we found that UVB irradiation increased ROS production, including H2O2, and both UVB and H2O2 caused the mislocalization of CLDN1 in HaCaT cells. The H2O2-induced mislocalization of CLDN1 was rescued by EBGP, PMA, and cantharidin. H2O2 decreased PKC activity and increased PP activities, which were inhibited by EBGP. Go6976 decreased p-Thr levels and the TJ localization of CLDN1. These results suggest that the TJ localization of CLDN1 is controlled by PKC. H2O2 decreased the p-Thr level of CLDN1, but this effect was blocked by EBGP. The T191E CLDN1 mutant was localized to the TJ after treatment with H2O2. We suggest that the phosphorylation of CLDN1 at T191 is necessary for its localization in the TJ. EBGP and its components may be useful in preventing the destruction of the TJ barrier by UVB and oxidative stress.
4. Materials and Methods
Rabbit anti-CLDN1, mouse anti-CLDN4, and rabbit anti-ZO-1 antibodies were obtained from Thermo Fisher Scientific (San Diego, CA, USA). Goat anti-β-actin and mouse anti-FLAG antibodies were from Santa Cruz Biotechnology (Santa Cruz, CA, USA) and Wako Pure Chemical (Osaka, Japan), respectively. Mouse anti-p-Ser and anti-p-Thr antibodies were from Sigma-Aldrich (Saint Louis, MO, USA). EBGP, LY, H2DCFDA, and PMA were from Yamada Bee Company, Inc. (Lot: LY-009, Okayama, Japan), Biotium (Fremont, CA, USA), Thermo Fisher Scientific, and LC Laboratories (Woburn, MA, USA), respectively. OxiOrange and Hydrop were from Goryo Chemical (Hokkaido, Japan). All other reagents were of the highest grade of purity available.
4.2. Cell Culture
HaCaT cells, an immortalized nontumorigenic human keratinocyte-derived cell line [36
], were grown in Dulbecco’s Modified Eagle’s Medium (Sigma-Aldrich) supplemented with 5% fetal bovine serum (FBS, Sigma-Aldrich), 0.07 mg/mL penicillin-G potassium, and 0.14 mg/mL streptomycin sulfate in a 5% CO2
atmosphere at 37 °C. One day before experiments, cells were transferred to FBS-free medium. Cell viability was examined using a WST-1 assay.
4.3. UVB Irradiation
UVB irradiation was carried out using a UV Crosslinker CL-1000M (Analytik Jena, Upland, CA, USA), which emits most of its energy within the UVB range (peaking at 302 nm). HaCaT cells were irradiated at a dose of 5–50 mJ/cm2 in Hank’s balanced salt solution. After UVB radiation, cells were incubated in fresh medium until analysis.
4.4. Production of Reactive Oxygen Species
H2DCFDA can detect several ROS, including H2O2, ∙OH, and peroxy radical, whereas OxiOrange and Hydrop selectively detect ∙OH and H2O2, respectively. HaCaT cells were incubated with these ROS-sensitive fluorescent probes for 30 min. The fluorescence intensity of each probe was detected using an Infinite F200 Pro microplate reader (Tecan, Mannedorf, Switzerland).
4.5. Confocal Microscopy
Cells were plated on cover glasses. After forming a confluent monolayer, the cells were incubated with cold methanol for 10 min at −30 °C and then permeabilized with 0.2% Triton X-100 for 10 min. Following permeabilization, the cells were blocked with 4% Block Ace (Dainippon Sumitomo Pharma, Osaka, Japan) for 30 min and incubated with anti-CLDN1, anti-CLDN4, anti-FLAG, or anti-ZO-1 antibodies (1:100 dilution) for 16 h at 4 °C, followed by incubation with Alexa Fluor 488- or 555-conjugated secondary antibodies (1:100 dilution). The fluorescence images were observed using an LSM 700 confocal microscope (Carl Zeiss, Jena, Germany).
4.6. SDS-Polyacrylamide Gel Electrophoresis and Immunoblotting
Cells were scraped into cold phosphate-buffered saline and precipitated by centrifugation. They were lysed in a RIPA buffer containing 150 mM NaCl, 0.5 mM EDTA, 1% Triton X-100, 0.1% SDS, 50 mM Tris-HCl (pH 8.0), and a protease inhibitor cocktail (Sigma-Aldrich) and were sonicated for 20 s. After centrifugation at 6000× g for 5 min, the supernatants were collected and used as cell lysates, which included membrane and cytoplasmic proteins. Samples were applied to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and blotted onto a polyvinylidene fluoride membrane. The membrane was then incubated with the respective primary antibody (1:1000 dilution) at 4 °C for 16 h, followed by a peroxidase-conjugated secondary antibody (1:3000 dilution) at room temperature for 1.5 h. Finally, the blots were incubated in EzWestLumi Plus (Atto, Tokyo, Japan) or ImmunoStar Basic (Wako Pure Chemical) and scanned with a C-DiGit Blot Scanner (LI-COR Biotechnology, Lincoln, NE, USA). The blots were stripped and reprobed with an anti-β-actin antibody. Band density was quantified using ImageJ software (National Institute of Health software). The signals were normalized using a β-actin loading control.
4.7. Measurement of Paracellular Permeability
Cells were plated on Transwell plates (0.4 μm pore size, Corning Inc., Corning, NY, USA). After forming a confluent monolayer, TER was measured using a Millicell-ERS epithelial volt-ohmmeter (Millipore, Billerica, MA, USA). TER values (ohms × cm2) were normalized by the area of the monolayer and were calculated by subtracting the blank values from the filter and the bathing medium. The paracellular permeability to LY, a fluorescent paracellular flux marker, for 1 h from upper to lower compartments was measured with an Infinite F200 Pro microplate reader.
4.8. Isolation of Total RNA and Quantitative Real-Time Polymerase Chain Reaction
Total RNA was extracted using a TRI reagent (Sigma-Aldrich). Reverse transcription and quantitative real-time polymerase chain reaction (PCR) was performed as described previously [37
]. The primer pairs used for PCR were human CLDN1 (sense: 5′-ATGAGGATGGCTGTCATTGG-3′; antisense: 5′-ATTGACTGGGGTCATAGGGT-3′) and human β-actin (sense: 5′-CCTGAGGCACTCTTCCAGCCTT-3′; antisense: 5′-TGCGGATGTCCACGTCACACTTC-3′).
4.9. PKC and Serine/Threonine Protein Phosphatase Activity Assays
Cells were harvested in 1× Passive Buffer (Promega, Madison, WI, USA). PKC activity was measured using a CycLex PKC Super Family Kinase Assay Kit (Medical & Biological Laboratories, Nagoya, Japan) in accordance with the manufacturer’s protocol. Serine/threonine PP activities were investigated using paranitrophenylphosphate (pNPP) as a substrate at pH 7.5 (neutral condition) and pH 8.4 (alkaline condition). The hydrolysis of pNPP was measured by monitoring the absorbance at 405 nm.
4.10. Vector Construction and Transfection
A vector containing human CLDN1 cDNA was prepared as described previously [31
] and was called CLDN1/pCMV, which encoded a FLAG tag at the amino terminus. Mutants of T191E and T195E were generated as described previously [31
]: siRNAs against CLDN1 (SASI_Hs01_00211216) and a negative control (SIC-001) were purchased from Sigma-Aldrich. Plasmid vector and siRNAs were transfected into cells using Lipofectamine 2000, as recommended by the manufacturer.
Results are presented as means ± S.E.M. Differences between groups were analyzed using one-way analysis of variance, and corrections for multiple comparison were made using Tukey’s multiple comparison test. Comparisons between two groups were made using Student’s t-test. Statistical analyses were performed using KaleidaGraph version 4.5.1 software (Synergy Software, Reading, PA, USA). Significant differences were assumed at p < 0.05.