RASSF1A regulates the abnormal cell proliferation in psoriasis via inhibition of YAP

Background: Psoriasis is a chronic, inammatory skin disease with high incidence, treatment resistance, and high recurrence. Currently, the exact etiology and pathogenesis of psoriasis are unclear. The goal of this study was to characterize the effect of the upstream negative regulator RAS-association domain family 1A (RASSF1A) on Yes-associated protein (YAP) in psoriasis. Methods: Skin lesions of 22 patients with psoriasis and 19 controls with normal skin tissue were used. Human epidermal keratinocytes were stimulated with M5 (IL-1 α, IL-17, IL-22, TNF-α, oncostatin M) to establish the psoriatic cell model. Methylation inhibitor 5-Aza-CdR was prepared at different concentrations (5, 10, 20 μmol/L). Cells were infected with lentivirus vector overexpressing RASSF1A. Twenty-ve 6–8-week-old female BALB/c mice were used to establish the psoriatic mouse model. Mice were randomly divided into ve groups: control (Vaseline applied daily), psoriasis (imiquimod applied daily), and the three different 5-Aza-CdR concentrations (applied daily with imiquimod). Methylation-specic PCR (MSP) was used to detect RASSF1A methylation and immunohistochemistry was used to detect RASSF1A expression in skin lesions. After adding 5-Aza-CdR or lentivirus vector overexpressing RASSF1A, YAP expression, cell proliferation, cell cycle, apoptosis, inammatory cytokines, and related signal pathway activity were investigated. Results: methylation level increased, its expression in patients with psoriasis and mice with skin lesions of 5-Aza-CdR or lentivirus vector overexpressing RASSF1A increased the expression of reduced the expression of YAP and inammatory cytokines, well as AKT, ERK, STAT3, and NF-κB signaling pathway activities, induced cell cycle arrest in G0/G1 phase, increased apoptosis, and improved skin lesions. Conclusions: RASSF1A inhibited the proliferation of psoriatic cells, induced apoptosis, and reduced the expression of inammatory factors by inhibiting YAP expression. Based on our ndings, targeted drugs that can inhibit RASSF1A methylation and increase its expression may be useful in the treatment of psoriasis.


Introduction
Psoriasis is a chronic, in ammatory, and recurrent skin disease with a population incidence of 0.1-4%.
Psoriasis affects all ages and genders equally 1,2 . Due to the high incidence, treatment resistance, recurrence, and often visible skin lesions, psoriasis is considered a persistent skin disease that affects the physical and mental health of the patients 3,4 . Currently, the exact etiology and pathogenesis of psoriasis are unclear. The over-proliferation of keratinocytes (KC) and shortening of cell cycle are important pathophysiological characteristics of psoriasis. Research shows that KC in psoriasis are apoptosisresistant; it is, therefore, important to explore the causes of KC apoptosis imbalance and cell cycle acceleration while studying the pathogenesis of psoriasis 5 . In our previous study, we found that Yesassociated protein (YAP) was highly expressed in psoriatic lesions and was involved in its pathogenesis through regulation of KC proliferation and apoptosis 6 . RAS-association domain family 1A (RASSF1A) is the most common negative regulator of RAS. It is expressed in almost all normal tissues, but is downregulated or absent in some tumor tissues and cell lines 7 .
The core of the kinase cascade of the Hippo signaling pathway includes Mst1/2, Sav1, Lats1/2, Mob1, YAP, and transcribed activator with PDZ binding motif, and the signaling transduction process is composed of a series of kinase cascade phosphorylation reactions. YAP is the downstream effector of this signaling pathway 14 , which regulates cell proliferation and apoptosis, and regulates the growth, development, and size of tissues and organs. When phosphorylated by upstream kinases at serine 127 site, pS127-YAP is inactivated and retained in the cytoplasm, so that cannot play its role as a transcription factor 14,15 . However, the relationship between RASSF1A and the Hippo-YAP signaling pathway is still unclear. Studies show that RASSF1A rarely mutates, and its low or loss of expression is mainly due to hypermethylation of the CpG island in the promoter region 8 . RASSF1A can be used as a negative upstream regulator of YAP [9][10][11] , which may play an important role in psoriasis pathogenesis. It is suggested that RASSF1A can be used as a promoter of the upstream kinase cascade to regulate the Hippo-YAP signaling pathway. Several research groups have found that in mammalian cells, RASSF1A can interact with Mst, the core of the kinase cascade in the Hippo-YAP signaling pathway. Furthermore, RASSF1A can interact with Mst1/2 through the SARAH domain 9,10 . However, before 2007, the role of RASSF1A in combination with Mst1/2 was unclear. Matallanas et al. 11 found that the oncogene Raf can bind and inhibit the activity of Mst2, thereby inhibiting the apoptotic signaling pathway. RASSF1A inhibits the interaction between Raf and Mst2, leading to phosphorylation of Mst2, and further to phosphorylation of Lats1/2 and downstream YAP. Phosphorylated YAP binds to p73 in the nucleus and promotes the transcription of PUMA, which promotes apoptosis. In addition, RASSF1A also activates the Hippo signaling pathway to induce Fas, TNF-α, or BAX-mediated apoptosis. Knockdown of RASSF1A leads to a decrease in Mst1/2 activation and to blockage of apoptosis induced by Fas or TNF-α, indicating that RASSF1A and Mst1/2 belong to the same signaling pathway. This study examined RASSF1A expression and methylation in psoriasis, analyzed its correlation with YAP expression, and veri ed its effect on YAP expression, KC proliferation, cell cycle, and apoptosis by modifying RASSF1A methylation level in vitro and in vivo.

Patient samples
Skin lesions of 22 psoriasis patients (12 males and 10 females) were collected from the Department of Dermatology tissue bank at the Second A liated Hospital of Xi'an Jiaotong University from September 2016 to August 2019, and 19 normal skin tissues (10 males and 9 females) from cosmetic surgeries were selected as controls. The study was approved by the ethics committee of the Second A liated Hospital of Xi'an Jiaotong University. Specimens were obtained with informed consent, and were con rmed by pathomorphological examination. Tissue samples were para n-embedded and sectioned for immunohistochemical staining. The results were classi ed as negative (-), weakly positive (+), moderately positive (++), and strongly positive (+++) after double blind examination by 2 pathologists. Additionally, 10 psoriasis samples and 10 healthy controls were frozen in liquid nitrogen for subsequent protein and methylation level detection.

Establishing the psoriatic cell model
Human epidermal KC (ScienCell) were routinely cultured in KC medium (ScienCell). We used 10 ng/mL M5 (mixture of IL-1 α, IL-17, IL-22, TNF-α and oncostatin M, PeproTech) to stimulate the cells for 48 h using the scheme for establishing a psoriatic cell model 12 . Three cell groups, each treated with a different concentration of 5-Aza-CdR (5, 10, 20 μmol/L; MedChemExpress), a widely used methylation inhibitor, were set-up after the cells were treated with M5.

Lentivirus infection
RASSF1A overexpression vector was constructed by Shenyang Wanleibio Technology. Cells were adjusted in the logarithmic growth period and infected with lentivirus according to the manufacturer's instructions. RASSF1A and YAP expression was detected 48 h after infection.

Establishing the psoriatic mouse model
Twenty-ve 6-8-week-old female BALB/c mice (18-20 g) were selected and housed under speci c pathogen-free conditions. Experiments were approved by the ethics committee of the Second A liated Hospital of Xi'an Jiaotong University. After 1 week of adaptive feeding, the mice were randomly divided into 5 groups: control, psoriasis, and three 5-Aza-CdR groups, each treated with a different concentration of 5-Aza-CdR. In the psoriasis group, 62.5 mg of imiquimod cream (Hubei Keyi Pharmaceutical Co., Ltd.) was applied daily to the back 13 . In the control group, equal amount of vaseline (Nanchang Baiyun Pharmaceutical Co., Ltd.) was applied daily. In the 5-Aza-CdR group, different concentrations (5, 10, 20 μmol/L) of 5-Aza-CdR were prepared with dimethyl sulfoxide (DMSO) and applied with imiquimod. Seven days later, the mice were anesthetized with phenobarbital (Shenyang Wanleibio Co., Ltd) euthanized, and then skin lesions were taken.

Quantitative real-time PCR (qRT-PCR)
TRIzol (Invitrogen) was used to extract the total RNA from cells in each experimental group, and cDNA was obtained by reverse transcription. PCR primers were synthesized by Shanghai Biotechnology Co., Ltd.
The primer sequence is shown in Supplemental Table S1. The standard curve for each group was obtained by qRT-PCR, and the CT value was computer-analyzed. After standardization with glyceraldehyde-3-phosphate dehydrogenase (GAPDH), relative expression was calculated with the 2-ΔΔCT method.

Cell cycle
Cells were digested according to the cell cycle kit manufacturer's instructions (Jiangsu Kaiji Biotechnology Co., Ltd.), washed twice with precooled phosphate buffered saline (PBS), and xed overnight with 75% ethanol at -20 °C. Then, 100 μg/mL RNaseA and 50 μg/mL propidium iodide (PI) staining solution were added, and cells were incubated at 37 °C in the dark for 30 min. Cell cycle was measured through ow cytometry (FACSCalibur, BD Biosciences).

Cell apoptosis
Cells were digested according using a cell apoptosis kit following the manufacturer's instructions (Jiangsu Kaiji Biotechnology Co., Ltd.) and were washed twice with precooled PBS. Cells (1-5 × 10 5 ) suspended in 500 μL binding buffer were collected, stained with 5 μL Annexin V-FITC and 5 μL PI staining solution, and then incubated at room temperature in the dark for 5-15 min. Cell apoptosis was measured through ow cytometry.

ELISA
Using the ELISA Kit's instructions (Lianke Biotech), supernatant or mouse serum samples were detected by standard ELISA. Optical density (OD) 450 nm values were then obtained using a spectrophotometer to draw the standard curve. The concentrations of IL-1 β, IL-6, IL-17, and IL-23 were then measured.

Statistical analysis
All data are presented as mean ± standard error of the mean (SEM). SPSS 19.0 software was used for statistical analysis. Pearson chi-square and Mann-Whitney U tests were used to analyze the immunohistochemistry results. Student t-test was used for comparisons between 2 groups. One-way analysis of variance (ANOVA) was used for experiments with more than 2 groups and least signi cant difference (LSD) test was used as a post-hoc for multiple comparisons. P<0.05 was considered statistically signi cant.

Expression and methylation of RASSF1A in psoriasis
Immunohistochemistry results showed that the rate of RASSF1A protein expression was 78.95% (15/19) in normal skin tissues and 50.00% (11/22) in psoriatic lesions. The difference was statistically signi cant (Table 1, Fig. 1a). The qRT-PCR and western blot results showed that RASSF1A mRNA and protein expression in normal skin was higher than that in psoriatic lesions (Fig. 1b, c). Results of MSP assays performed for methylation detection showed that the methylation level of RASSF1A increased gradually in normal skin tissue, normal tissue beside psoriatic lesions, and psoriatic tissues, indicating that low expression may be due to promoter methylation (Fig. 1d).

Validation of RASSF1A overexpression e ciency and effect on YAP expression
To con rm the effect of RASSF1A overexpression on YAP expression, we infected cells with a RASSF1A overexpressing lentivirus. After 48 h of lentivirus infection, RNA and protein were extracted to verify the RASSF1A overexpression e ciency. Compared with those in the control group (Ctrl) and the group infected with empty virus vector (BLK), RASSF1A mRNA and protein were overexpressed in the group infected with RASSF1A overexpression lentivirus. RASSF1A overexpression reduced YAP expression and induced its phosphorylated protein pS127-YAP expression ( Fig. 1e and Supplemental Fig. S1a, b).

Effect of methylation inhibitor 5-Aza-CdR on RASSF1A and YAP expression in psoriatic cells
5-Aza-CdR is a methylation inhibitor that can inhibit the methylation of RASSF1A gene and induce its expression. MSP results showed that different concentrations of 5-Aza-CdR reduced RASSF1A methylation level in psoriatic cells induced by M5 dose-dependently (Fig. 1f). Western blot showed that 5-Aza-CdR dose-dependently increased RASSF1A expression in psoriatic cells, with the level not returning to pre-M5 levels. Simultaneously, 5-Aza-CdR gradually reduced YAP expression and induced pS127-YAP expression in M5-induced psoriatic cells. RASSF1A overexpression in psoriatic cells had the similar effect with 5-Aza-CdR (Fig. 1g).

Effect of RASSF1A overexpression and 5-Aza-CdR on cell proliferation
To verify RASSF1A overexpression upon lentivirus infection or upon treatment with the methylation inhibitor 5-Aza-CdR, MTT assay was performed. Results showed that cell proliferation increased at all time points after adding M5, consistent with the observation in the abnormal psoriasis cell proliferation model. After adding 5-Aza-CdR, cell proliferation was inhibited, and was more obvious with increasing 5-Aza-CdR concentration, but did not return to pre-M5 level. The proliferation ability of cells also increased following BLK infection and incubation with M5, but there was no signi cant difference compared with the uninfected group. However, after RASSF1A overexpression, the proliferation ability of cells decreased signi cantly but did not return to pre-M5 level (Fig. 2a). 3.5 Effect of RASSF1A overexpression and 5-Aza-CdR on cell cycle PI staining and ow cytometry were used to detect cell cycle after RASSF1A was overexpressed through lentivirus infection or through treatment with the methylation inhibitor 5-Aza-CdR. Results showed that the cell proportion decreased in the G0/G1 phase and increased in the S phase after adding M5, indicating that the proportion of cells in active replication increased. This was consistent with the abnormal psoriasis cell proliferation model (Fig. 2a). Adding 5-Aza-CdR increased the G0/G1 phase ratio and decreased the S phase ratio, and this was more obvious with increasing 5-Aza-CdR concentration, but did not return to pre-M5 level. There was no signi cant difference in cell cycle distribution between the BLK transfected group and the non-transfected group. However, after RASSF1A overexpression, the G0/G1 phase proportion increased and S phase proportion decreased, but did not return to pre-M5 level. Although G2/M ratio also changed, there was no statistical difference ( Fig. 2b and Supplemental Fig.   S1c). Western blot showed that the expression of cyclin A, cyclin B1, cyclin D1, cyclin E, and CDK1 increased at different levels when M5 was added. Their expression decreased in a dose-dependent manner when 5-Aza-CdR was added, and RASSF1A overexpression had the same effect (Fig. 2c). Cell cycle regulation is complicated because of different regulating factors that are involved in each cell cycle. Our results showed that altering RASSF1A expression affected the cell cycle process of psoriatic cells by affecting the expression of each cell cycle regulation factor.

Effect of RASSF1A overexpression and 5-Aza-CdR on cell apoptosis
V-FITC/PI double staining and ow cytometry were used to detect cell apoptosis after RASSF1A was overexpressed through lentivirus infection or through treatment with the methylation inhibitor 5-Aza-CdR. Results showed that KC apoptosis rate decreased when M5 was added, and increased when 5-Aza-CdR or RASSF1A was added, but did not return to pre-M5 level ( Fig. 2d and Supplemental Fig. S1d). Further, western blot results showed that the expression of apoptosis-related proteins, cleaved-caspase-3 (ccaspase-3), BAX, p53, and p21 decreased, while that of Bcl-2 and p73 increased. RASSF1A overexpression or 5-Aza-CdR methylation inhibitor partially restored their expression at different levels (Fig. 2e).

Effect of methylation inhibitor 5-Aza-CdR on psoriatic mouse model
Our results in vitro con rmed that RASSF1A overexpression through lentivirus infection or through treatment with the methylation inhibitor 5-Aza-CdR could inhibit cell proliferation and induce G0/G1 cell cycle arrest and apoptosis of M5-induced psoriatic cell model. To verify the results in vivo, we established imiquimod-induced psoriatic mouse model. The skin of psoriatic model group (Pso) mice was thickened with erythema and scales, while the skin of Ctrl mice was normal, indicating that the model induction was successful. The erythema and scales on the mouse's back decreased, and the skin lesions became thinner with increasing 5-Aza-CdR concentration ( Table 2, Fig. 3a, b).

Effect of methylation inhibitor 5-Aza-CdR on YAP and RASSF1A expression in psoriatic mouse model
Immunohistochemistry con rmed that RASSF1A expression in the Pso was lower than that in the Ctrl in mice similar to that in the human samples (Fig. 1a). The addition of 5-Aza-CdR increased the positive rate of RASSF1A expression, and the trend was more obvious with increasing 5-Aza-CdR concentration. The semiquantitative analysis results showed that the intensity of RASSF1A expression in the epidermis of the Ctrl group was signi cantly different from that in the Pso group (P=0.032). Moreover, the expression of YAP in the Pso group was higher than that in the Ctrl group. The positive rate of YAP expression decreased after adding 5-Aza-CdR, and the trend was more obvious with increasing 5-Aza-CdR concentration. The semiquantitative analysis of expression intensity showed that RASSF1A expression in the Ctrl group was signi cantly higher than that in the Pso group (P = 0.009), and there was a signi cant difference between the Pso group and 10 μM 5-Aza-CdR group (P = 0.031) or 20 μM 5-Aza-CdR group (P = 0.013) ( Table 3, Fig. 4a, b). MSP con rmed 5-Aza-CdR could reduce RASSF1A methylation (Fig. 4c), and western blot con rmed 5-Aza-CdR could induce RASSF1A expression, while reducing YAP expression and driving its phosphorylated protein pS127-YAP in a dose-depend manner (Fig. 4d).

Effects of RASSF1A overexpression and 5-Aza-CdR on the expression of cytokines
To detect changes in different cytokines after RASSF1A overexpression through lentivirus infection or through treatment with the methylation inhibitor 5-Aza-CdR in vitro and in vivo, ELISA was performed. Results showed that the concentrations of IL-1β, IL-6, IL-17, and IL-23 in the supernatant of the psoriatic model of cells or mouse serum increased. RASSF1A overexpression or addition of 5-Aza-CdR in cells and topical 5-Aza-CdR administration on the back of mice reduced the expression of the cytokines to different levels, but the levels did not return to the Ctrl level (Fig. 5a, b). The results suggested that RASSF1A may reduce the expression of these cytokines to a certain extent to reduce the in ammatory response in psoriasis.

The effect of RASSF1A overexpression and 5-Aza-CdR on the expression of various signaling pathways
To explore which signaling pathways are involved in the effect of RASSF1A on psoriasis, the expression of the key members of AKT, ERK, STAT3, and NF-κB pathways was detected through western blotting. Results showed that the expression of p-ERK, p-STAT3, NF-κB P65 in the nucleus (NF-κB P65[n]) increased and that of NF-κB P65 in the cytoplasm (NF-κB P65[p]) decreased with M5 addition to KC cells under conditions where the total AKT, ERK, and STAT3 were basically unchanged, indicating increased activities of these signaling pathways. However, RASSF1A overexpression or addition of methylation inhibitor 5-Aza-CdR reduced the expression of p-ERK, p-STAT3, NF-κB P65 (n) and increased that of NF-κB P65 (p), with not so obvious changes in the expression of p-AKT (Fig. 5c).

Discussion
Our previous study showed that YAP expression increased in psoriasis patients and mouse model of psoriasis skin lesions. In vitro, after YAP knockdown, the proliferation of the KC cell line HaCaT slowed down, the cell cycle was blocked in the G0/G1 phase and cell apoptosis increased 6 . However, it is still unknown that which upstream regulator regulate its expression.
In this study, we found that the expression of RASSF1A decreased whereas its methylation increased in psoriasis lesions compared to those in normal skin tissues. So we speculated that there might be a relationship between the expression of RASSF1A and YAP. As it was assumed that low or loss of RASSF1A expression is mainly caused by hypermethylation of the CpG island in the promoter region 8 , we overexpressed RASSF1A by inhibiting its methylation or by lentivirus infection. After RASSF1A was overexpressed through lentivirus infection or through treatment with the methylation inhibitor 5-Aza-CdR, the expression of YAP was found to be downregulated. Simultaneously, the proliferation rate of the M5induced psoriatic cells decreased; the cell cycle was blocked in the G0/G1 phase and the mRNA expression of the cell cycle regulators cyclin A, cyclin B1, cyclin D1, cyclin E, and CDK1 was decreased. Further, under this same condition, apoptosis increased and the expression of the apoptosis-promoting proteins c-caspase-3, BAX, p53, and p21 increased, while that of the apoptosis-inhibiting protein Bcl-2 decreased. The study of related signaling pathways showed that the expression of key molecules in the ERK, STAT3, and NF-κB signaling pathways was downregulated to varying degrees, with not so obvious changes in the AKT signaling activity, after RASSF1A overexpression. In vivo study of the imiquimodinduced psoriatic mice con rmed that the methylation inhibitor 5-Aza-CdR could improve skin lesions, increase RASSF1A expression, and reduce YAP expression at the same time. Meanwhile, levels of IL-1β, IL-6, IL-17, and IL-23 in the supernatant of the psoriatic model of cells or mouse serum decreased, indicating the suppression of in ammatory response. The changes in the psoriatic cell model after RASSF1A overexpression were consistent with those observed after YAP knockdown 6 . As overexpression of RASSF1A inhibited YAP expression in both cell and animal experiments, we believe that RASSF1A can inhibit psoriatic cell proliferation, induce apoptosis, and improve the expression of in ammatory factors by inhibiting YAP expression.

Conclusions
Our study found that as RASSF1A methylation level increased, its expression in psoriasis patients and mice with skin lesions decreased. Adding methylation inhibitor 5-Aza-CdR or RASSF1A overexpression lentivirus inhibited cell proliferation, induced cell cycle arrest at G0/G1 phase, increased apoptosis, reduced the expression of in ammatory cytokines and activity of AKT, ERK, STAT3 and NF-κB signaling pathways, and inhibited the expression of YAP. Increasing RASSF1A expression through methylation inhibitors may provide new directions, ideas, and strategies for developing new drug therapies for treating psoriasis. However, the speci c regulation process of the Hippo-YAP signaling pathway, particularly the in uence of upstream factors such as Mst1/2, Sav1, Lats1/2 and Mob1 in the core of the kinase cascade, needs further study.

Consent for publication
Written informed consent for tissue collection and use was obtained from all patients.

Availability of data and materials
All data generated or analyzed during this study are included in this published article and its supplementary information les.