The Effects of Mucopolysaccharide Polysulfate on Steroid-Induced Tight Junction Barrier Dysfunction in Human Epidermal Keratinocytes and a 3D Skin Model

Atopic dermatitis (AD) is a chronic skin disease characterized by skin barrier dysfunction, immune abnormalities, and pruritus [1]. The skin barrier dysfunction in AD results from the downregulation of epidermal barrier components, such as filaggrin, acylceramides, cornified envelope precursors, and claudin-1 [2, 3]. Claudin-1, a major component of tight junctions (TJs), is downregulated in lesional and nonlesional skin of patients with AD [4‒7], leading to an increase in permeability to environmental antigens, allergens, irritants, or pollutants [4, 8].

Topical corticosteroids (TCS), which represent the current standard therapy for AD, have been used to reduce inflammation in lesional skin [9‒11]. However, the long-term use of TCS is associated with several cutaneous side effects, such as skin atrophy and barrier impairment in both humans and animals [12‒16]. Mechanisms of corticosteroid-induced barrier impairment are associated with decreased epidermal lipid synthesis [12], decreased expression of structural proteins associated with epidermal differentiation like involucrin, filaggrin, and loricrin [17], and inhibition of epidermal TJ-related factors like claudin-1 [18, 19]. Claudin-1 knockout mice were found to die, due to water loss in the epidermis [20], and also showed the abnormal formation of the stratum corneum and loss of the stratum corneum barrier [21]. It has also been reported that the knockdown of claudin-1, claudin-4, occludin, and zonula occludens-1 (ZO-1) increased paracellular permeabilities for ions and larger molecules, demonstrating that all of these TJ proteins contribute to barrier formation [22].

When skin inflammation of patients with AD is in remission, a TCS or tacrolimus ointment is applied intermittently. Additionally, skin care products, such as moisturizers, are used to maintain the remission status [23, 24]. Although moisturizers have been used to reduce the frequency of AD relapses, the mechanisms of action of these products are poorly understood. Mucopolysaccharide polysulfate (MPS) is the active pharmaceutical ingredient of Hirudoid^®, which is used as a moisturizer for patients with xerosis in Japan. Although MPS may improve epidermal barrier function, their effects on the TJ barrier, which is important for epidermal barrier function, are not completely understood.

In the present study, we evaluated the effects of clobetasol 17-propionate (CP), one of the most potent TCS currently available, on the TJ barrier in human epidermal keratinocytes (HEKa); then, we evaluated the effects of moisturizing ingredient (MPS and urea) on the TJ barrier in combination with CP. Furthermore, we examined the effects of MPS in combination with CP on TJ barrier function in a 3D cultured skin model.

MPS was obtained from Maruho Co., Ltd. (Osaka, Japan). Commercially available CP (Tokyo Chemical Industry Co., Ltd., Tokyo, Japan) and urea (FUJIFILM Wako Pure Chemical Corporation, Osaka, Japan) were used.

HEKa (Thermo Fisher Scientific, Waltham, MA, USA) were cultured in HuMedia-KG2 (Kurabo Industries Ltd., Osaka, Japan) and maintained at 37°C in a humidified CO₂ incubator set to 5% CO₂. Next, the HEKa were plated at a density of 1.25 × 10⁵ cells/cm² and incubated for attachment for 6 h. Thereafter, the cells were incubated with CP (1, 3, or 10 μg/mL), or a combination of CP (3 μg/mL) and MPS (1, 10, or 100 μg/mL), or a combination of CP (3 μg/mL) and urea (0.06, 0.6, or 6 mg/mL) or urea (0.06, 0.6, or 6 mg/mL) in a medium containing 1% dimethyl sulfoxide (DMSO; FUJIFILM Wako Pure Chemical Corporation) for 72 h. Vehicle control was treated with a medium containing 1% DMSO only. The cytotoxicity was determined using a lactate dehydrogenase cytotoxicity detection kit (Takara Bio, Inc., Shiga, Japan). The measurement proceeded according to the manufacturer’s instructions.

HEKa were lysed in a RIPA buffer containing protease inhibitors (Nacalai Tesque, Inc., Kyoto, Japan). Equal amounts of proteins were loaded on SDS-PAGE gels (XV PANTERA MP GEL, D. R. C. Co., Ltd., Tokyo, Japan) and transferred to PVDF membranes (Bio-Rad Laboratories, Inc., California, USA). After blocking with 5% skimmed milk in Tris-buffered saline with 0.1% Tween^® 20 detergent (TBS-T; Nacalai Tesque) at room temperature for 1 h, the membranes were washed with TBS-T. The membranes were then incubated with primary antibodies against claudin-1 (1:5,000 dilution, Invitrogen, Carlsbad, CA, USA) or β-actin (1:5,000 dilution, Cell Signaling Technology, MA, USA) at 4°C overnight. After washing TBS-T, the membranes were incubated with HRP-linked anti-rabbit IgG or HRP-linked anti-mouse IgG secondary antibodies (1:10,000 dilution, Cell Signaling Technology) at room temperature for 1 h. After washing TBS-T, the SuperSignal West Dura Extended Duration Substrate (Thermo Fisher Scientific) was used to visualize the proteins of interest in LAS-4000EPUVmini (Fujifilm Co., Ltd., Tokyo, Japan). Quantitative analyses were performed with Multi Gauge (Version 3.2, Fujifilm), and values were normalized against β-actin expression.

Cells were fixed in a 4% paraformaldehyde phosphate buffer solution (FUJIFILM Wako Pure Chemical Corporation) for 15 min at room temperature. After washing with PBS (Sigma-Aldrich, Steinheim, Germany), cells were incubated with the Blocking One (Nacalai Tesque) at room temperature for 1 h. Subsequently, the cells were incubated with primary antibodies against claudin-1 (1:1,000 dilution, Abcam, Cambridge, MA, USA) at 4°C overnight. After washing with PBS, the cells were incubated with a fluorophore-conjugated secondary antibody (1:400 dilution, Thermo Fisher Scientific) at room temperature for 1 h. Cells were washed with PBS and then mounted using ProLong Diamond Antifade Mountant with 4′,6-diamidino-2-phenylindole (Invitrogen) on coverslips. Photomicrographs were obtained using a confocal laser scanning microscopy (FLUOVIEW FV3000 instrument, Olympus, Tokyo, Japan).

The barrier function of HEKa was assayed by measuring the resistance of a cell-covered electrode using an ECIS Zθ instrument (Applied Biophysics, NY, USA), as previously described [25, 26]. Briefly, an ECIS 96W20idf plate was treated with a 10 mml-cysteine solution (room temperature, 15 min) and then washed with distilled water. Thereafter, CP (1, 3, or 10 μg/mL) or a combination of CP (3 μg/mL) and MPS (1, 10, or 100 μg/mL) was added 6 h after cell seeding. The change in TEER values, following the addition of test samples, was measured at a low frequency (4 kHz) for 72 h.

EpiDerm™ Skin Model EPI-200 (Kurabo) cultures were maintained in an EPI-100 assay medium (Kurabo) at 37°C in a humidified CO₂ incubator set to 5% CO₂. EpiDerm™ Skin Model EPI-200 cultures were incubated with CP (10, 20, or 30 μg/mL), and a combination of CP (30 μg/mL) and MPS (1, 10, or 100 μg/mL) in an EPI-100 assay medium containing 1% DMSO for 48 h. The EZ-Link™ Sulfo-NHS-LC-Biotin (Thermo Fisher Scientific) in an EPI-100 assay medium was added to the basolateral side of the EPI-200 at a final concentration of 1 mg/mL for 2 h before skin tissue sampling. The skin sample was divided into two parts, one for western blot analysis and the other for the TJ permeability assay. For western blot analysis, the skin samples were lysed in a RIPA buffer containing protease inhibitors (Nacalai Tesque). SDS-PAGE, antibody staining, and expression analysis were performed as described above (refer to Western Blot Analysis).

The permeability assay was performed following the protocol described by Yuki et al. [27] with slight modifications. The frozen fragments with a thickness of 5 μm were fixed with ice-cold 95% ethanol for 30 min. After washing with PBS (Nacalai Tesque), the sections were blocked with 1% BSA/PBS at room temperature for 1 h. The sections were incubated with the primary antibodies: ZO-1 Polyclonal Antibody (1:500 dilution, Invitrogen) diluted in 1% BSA/PBS at 4°C overnight. After washing with PBS, the sections were incubated with Alexa Fluor 488-conjugated goat anti-rabbit IgG (H + L) Cross-Adsorbed Secondary Antibody (1:1,000 dilution, Invitrogen), and Alexa Fluor 568-conjugate Streptavidin (1:400 dilution, Invitrogen), diluted in 1% BSA/PBS at room temperature for 1 h. After washing with PBS, the sections were sealed with Mowiol 4-88 (Sigma-Aldrich) and observed under a confocal laser scanning microscope (FLUOVIEW FV3000 instrument, Olympus). ZO-1-positive sites with or without biotin stop were quantified for each sample. The ratio of biotin leakage from the basolateral side to the apical side at the localization of ZO-1 was calculated under blinded condition of sample number, referring to the evaluation method described in a previous report [28].

Data were expressed as mean ± standard error of the mean. A significant difference between the two groups was evaluated using an F-test followed by Student’s t test or Aspin-Welch’s t test. Significant differences between three or more groups were evaluated using Dunnett’s multiple-comparison test. Statistical analyses were performed using EXSUS software (EPS Corporation, Tokyo, Japan). Statistical significance was set at p < 0.05.

To evaluate whether TJ barrier function was disrupted by CP, we analyzed the TEER, as an index of TJ barrier function, and expression levels of claudin-1 in CP-treated HEKa. CP at 1–10 μg/mL significantly reduced claudin-1 expression levels in HEKa in a concentration-dependent manner (Fig. 1a and online suppl. Fig. 1; for all online suppl. material, see https://doi.org/10.1159/000529962). Fluorescence immunostaining showed that claudin-1 was localized on the cell membrane in the vehicle control group; the fluorescence intensity of claudin-1 decreased following CP treatment (online suppl. Fig. 2). Further, CP treatment resulted in a significant decrease in TEER as compared to the vehicle control (Fig. 1b). CP-treated cells showed no cytotoxicity after 72 h; however, the cells were morphologically different from those in the vehicle control group (online suppl. Fig. 3).

Since CP reduced claudin-1 levels and TEER in HEKa, the effects of MPS and urea on the CP-induced reduction of TJ barrier function were investigated. As compared to the vehicle control, MPS inhibited the decrease in claudin-1 levels induced by 3 μg/mL CP treatment and reached a recovery effect of about 80% at a concentration of 10 μg/mL. In contrast, urea failed to show the same recovery effect as those produced by MPS (Fig. 2a and online suppl. Fig. 4). Treatment with 3 μg/mL CP resulted in a significant decrease in TEER, while treatment with 1–100 μg/mL MPS and 3 μg/mL CP significantly increased TEER compared to 3 μg/mL CP (Fig. 2b). Moreover, the morphology of cells concomitantly treated with MPS and 3 μg/mL CP was similar to that of cells in the vehicle control group (online suppl. Fig. 3). Fluorescence immunostaining also showed that claudin-1 expression levels were lower in CP-treated cells than in vehicle control cells, but cells treated with 100 μg/mL MPS and 3 μg/mL CP had the same expression levels as those in the vehicle control group. On the other hand, the fluorescence intensity after 6 mg/mL urea and 3 μg/mL CP treatment was similar to that observed after 3 μg/mL CP treatment (Fig. 3).

MPS improved the TJ barrier impairment induced by CP in HEKa. Therefore, we investigated the effect of CP treatment and, subsequently, MPS in the presence of CP on claudin-1 expression in a 3D cultured skin model. The results showed that CP treatment alone induced a concentration-dependent decrease in claudin-1 expression levels (Fig. 4a) and that MPS, in the presence of CP, improved claudin-1 expression (Fig. 4b).

As MPS improved TJ barrier dysfunction in the presence of CP in 3D cultured skin, we investigated TJ permeability to biotin to assess the effect of MPS on the inside-out barrier function in CP-treated 3D cultured skin. In the vehicle control, the tracer biotin (red fluorescence) was not detected above the localization of ZO-1, a TJ barrier marker; however, with 30 μg/mL CP treatment, the fluorescence of biotin was observed above ZO-1. After co-treatment with 30 μg/mL CP and 100 μg/mL MPS, the fluorescence of biotin above ZO-1 was not observed (Fig. 5a). The quantitative degree of leaked biotin fluorescence above ZO-1 after CP treatment in the presence or absence of MPS was examined. In the presence of CP, a dose-dependent leakage of biotin was observed (Fig. 5b). In contrast, MPS treatment improved the CP-induced biotin leakage in a dose-dependent manner (Fig. 5c).

TCS has been used to reduce inflammation in the lesional skin of patients with AD [9‒11]. However, while the long-term effects of these drugs on nonlesional skin, such as skin atrophy and barrier permeability dysfunction, have been reported, their effects on the TJ barrier are not well understood [12‒15]. Although it has been shown that the topical application of CP-containing products inhibits TJ-related proteins in vivo [18, 19], such effects have rarely been reported in in vitro studies using keratinocytes. In this study, we demonstrated that CP treatment was associated with a decreased protein expression of claudin-1 in human keratinocytes and 3D cultured skin. CP treatment resulted in decreased TEER in HEKa and increased leakage of tracer from the inside to the outside in 3D cultured skin. Bergmann et al. [8] reported that TEER and the degree of biotin leakage in 3D cultured skin were highly correlated with claudin-1 expression levels. Thus, the CP-induced TJ barrier dysfunction may be partially responsible for the inhibition of claudin-1. CP is known to inhibit not only the TJ-related protein claudin-1 but also a variety of epidermal differentiation-linked structural proteins [17, 29]. Indeed, the morphology of keratinocytes was also found to be altered by CP treatment; however, there was no CP-induced cytotoxicity (online suppl. Fig. 3). These changes in keratinocyte morphology may be involved in the atrophic activity of CP in keratinocytes. Barners et al. [29] reported that depolymerization of the epidermal F-actin cytoskeleton in keratinocytes was one of the mechanisms underlying the skin atrophy induced by CP. Moreover, F-actin depolymerization leads to functional and morphological TJ disruption [30]. Thus, CP may impair TJ barrier function by acting not only on TJ-related proteins but also on various factors related to the epidermal differentiation and cytoskeleton in keratinocytes.

Next, we examined the effects of MPS on CP-induced TJ barrier dysfunction and found that it inhibited the decrease in claudin-1 expression and TEER and improved biotin leakage. In monolayer keratinocytes, the increasing effect of MPS on TEER was maximal at 1 μg/mL, whereas the expression of claudin-1 showed a slight increase at 1 μg/mL. This discrepancy may be because MPS also affects barrier function-related factors, like TJ-related proteins except for claudin-1 and differentiation markers, in keratinocytes. In our previous report [31], MPS increased the protein levels of claudin-1 and ZO-1, a TJ constituent, in normal HEKa. Wen et al. [32] reported that topical MPS-containing products increased mRNA expressions of epidermal differentiation markers and lipid synthetic enzymes in clobetasol propionate-containing product-induced skin barrier impairment in mice. Indeed, the abnormal cell morphology of keratinocytes induced by CP also tended to normalize after 72 h of coadministration of 1 μg/mL MPS (data not shown). This suggested that MPS may sufficiently recover the barrier function impaired by CP even with treatment at 1 μg/mL by enhancing the expression of TJ-related proteins and other barrier-related factors. Therefore, increased expression of claudin-1 may be partially responsible for inhibiting the effect of MPS on CP-induced TJ barrier dysfunction.

Urea did not affect the expression of claudin-1 decreased by CP but rather tended to decrease further. Urea, an endogenous metabolite, is a natural moisturizing factor that enhances the hydration of the stratum corneum. Topical urea improves the barrier function in human volunteers and a murine model of AD [33]. However, this report showed that the treatment of keratinocytes with urea did not alter claudin-1 mRNA levels. Moreover, urea caused a dose-dependent decrease in TEER and TJ protein levels in human enterocytes (T84 cells) [34]. In our preliminary study, we also observed a decrease in claudin-1 expression after treatment of keratinocytes with urea (online suppl. Fig. 5). Therefore, despite its moisturizing effects, urea may have no improving effects on the TJ barrier dysfunction caused by CP treatment in keratinocytes.

MPS, a polysulfated chondroitin sulfate, is classified as a sulfated glycosaminoglycan [35]. Hyaluronic acid (50 kDa), a nonsulfated glycosaminoglycan, is known to increase the expression levels of TJ-related proteins, such as claudin-1, occludin, and ZO-1 in the reconstituted epidermis [36]. The interaction of hyaluronic acid with its receptor CD44 has also been reported to inhibit CP-induced skin atrophy [29], and in the CD44 knockout keratinocytes, claudin-1 expression is reduced and TJ barrier formation is delayed [37]. Recently, it has been reported that MPS-containing creams suppress CP-induced skin atrophy in mice and show efficacy against CP-induced impairment of the epidermal permeability barrier [32] and that treatment of keratinocytes with MPS increases TJ-related proteins such as claudin-1 and ZO-1 [31]. In consideration of these reports, it is suggested that CD44 activation may be involved in the increased expression of TJ-related proteins in MPS, but the relationship with CD44 signaling needs to be examined in future studies.

In a clinical study, treatment with moisturizers such as formulations containing urea or MPS was reportedly useful for maintaining remission in patients with AD after TCS treatment [38, 39]. Although the detailed mechanism by which MPS prolongs the duration of remission in patients with AD is unclear, MPS may be partially responsible for the improvement of TJ barrier impairment caused by TCS and inflammatory cytokines such as IL-4, IL-13, and IL-33. Recently, TJ barriers have been shown to regulate itch by pruning nerve terminals in the epidermis; however, this function was lost in the skin of patients with AD and a mouse model of chronic dry skin itch [40]. In AD and aging skin with decreased expression of claudin-1 in the epidermis [5, 41], the chronic itch may be induced by TJ barrier dysfunction. Taken together, the improvement of TJ barrier function by MPS may partially be responsible for prolonged remission periods in AD patients, which might reduce chronic pruritus in AD and senile xerosis and prevent the worsening of symptoms.

A limitation of this study is that the effect of MPS on CP-induced TJ barrier dysfunction was only evaluated in vitro. In addition, the detailed mechanism of action of MPS, such as the interaction of CD44 with the TJ barrier function, needs to be examined. In the future, clinical studies should be conducted to investigate the effects of MPS combined with TCS therapy on the TJ barrier function.

In conclusion, our results indicated that CP treatment decreased the expression of claudin-1 and epidermal barrier function. We also found that MPS improved CP-induced TJ barrier impairment, at least partially, by enhancing the expression of claudin-1. Taken together, our findings suggest that the combination of TCS and MPS may maintain a normal TJ barrier function and prolong remission in patients with AD.

The authors are grateful to Prof. Akiharu Kubo for his support and experimental suggestions.

The primary cell and 3D skin model used in this study were purchased from Thermo Fisher Scientific and Kurabo, respectively. Ethical approval and consent for the use of these cells are not required in accordance with local/national guidelines.

The authors have no conflict of interest to declare and are employees of Maruho Co., Ltd.

Akira Koda designed and planned the experiments, analyzed the data, and drafted the manuscript. Yuko Ishii, Mika Fujikawa, and Keisuke Kikuchi performed the research, designed and planned the experiments, analyzed the data, and provided comments on the manuscript. Ryota Hashimoto and Ayu Kashiwagi performed the research and analyzed the data. Yuhki Ueda and Takaaki Doi reviewed the manuscript, provided comments, and revised and approved the manuscript.

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

All data generated or analyzed during this study are included in this article and its online supplementary material. Further inquiries can be directed to the corresponding author.