Analysis of changes in microRNA expression profiles in response to the troxerutin-mediated antioxidant effect in human dermal papilla cells

Dermal papilla (DP) cells function as important regulators of the hair growth cycle. The loss of these cells is a primary cause of diseases characterized by hair loss, including alopecia, and evidence has revealed significantly increased levels of reactive oxygen species (ROS) in hair tissue and DP cells in the balding population. In the present study, troxerutin, a flavonoid derivative of rutin, was demonstrated to have a protective effect against H2O2-mediated cellular damage in human DP (HDP) cells. Biochemical assays revealed that pretreatment with troxerutin exerted a protective effect against H2O2-induced loss of cell viability and H2O2 induced cell death. Further experiments confirmed that troxerutin inhibited the H2O2-induced production of ROS and upregulation of senescence-associated β-galactosidase activity. Using microRNA (miRNA) microarrays, the present study identified 24 miRNAs, which were differentially expressed in the troxerutin pretreated, H2O2-treated HDP cells. Subsequent prediction using bioinformatics analysis revealed that the altered miRNAs were functionally involved in several cell signaling pathways, including the mitogen-activated protein kinase and WNT pathways. Overall, these results indicated that ROS-mediated cellular damage was inhibited by troxerutin and suggested that the use of troxerutin may be an effective approach in the treatment of alopecia.


Introduction
Trihydroxyethylrutoside (troxerutin) is one of the flavonoid rutoside derivatives. It exhibits non-mutagenic properties and has a functional role in the treatment of chronic venous insufficiency (CVI) (1,2). A number of studies have demonstrated other beneficial effects of troxerutin, in vitro and in vivo, and it may be effective in reducing different cytotoxicities. In particular, troxerutin has been observed to exhibit an inhibitory effect on the neurotoxicity induced by high cholesterol-mediated cognitive deficits, kainic acid-triggered excitotoxic damage and β-amyloid oligomerization (3)(4)(5). In addition, troxerutin has a photoprotective effect against ultraviolet B (UVB) radiation in human skin cells, including dermal fibroblasts and keratinocytes (6,7). Troxerutin also exerts a protective effect against γ-radiation in mice (8,9). Although the precise cellular mechanisms underlying the effects of troxerutin remain to be fully elucidated, these reports can be summarized as a single meaningful finding, that troxerutin inhibits the production of reactive oxygen species (ROS). In vivo investigations have demonstrated that CVI-bearing patients have increased levels of ROS, and troxerutin has a protective effect against oxygen-derived free radical scavengers on the endothelium in these patients (10,11). In addition, the aforementioned neurotoxicities are inhibited following troxerutin application by reducing the production of ROS (3,4,12). UVB and γ-radiation are known ROS stimulators (13,14), and a previous study demonstrated that troxerutin protects against radiation-induced lipid peroxidation (9). These studies suggest that this toxerutin may offer a novel therapeutic strategy for ROS-induced diseases.
Dermal papilla (DP) cells are located at the base of hair follicles and are important in the induction of growth and maintenance of epithelial cells, which are the predominant components of hair follicles (15). In response to hormonal changes, DP cells direct the follicular epithelial cells to enter the hair growth cycle, which involves anagen, an active growing phase; catagen, a short transitionary regressive phase; and telogen, a dormant resting phase (15). An increasing body of evidence has demonstrated excessive loss of viability and death of DP cells in balding regions of the scalp,

Analysis of changes in microRNA expression profiles in response to the troxerutin-mediated antioxidant effect in human dermal papilla cells
compared with non-balding regions, due to increased levels of 5α-reductase (16), a converting enzyme for androgenic hormones and intracellular ROS (17). In addition, previous reports have indicated that oxidative stress is generated by the exposure of androgen-sensitive prostate cancer cells to high levels of androgens (18), and that lipid peroxides increase the levels of ROS and apoptosis of the hair follicle cells (19). Furthermore, DP cells in the balding scalp grow more slowly in vitro, compared with cells from the non-balding scalp. The reduced proliferative activity of balding DP cells is associated with changes in the expression levels of senescence-associated (SA) β-galactosidase, oxidative stress markers, superoxide dismutase and catalase (20). These findings indicate that oxidative stress is important in the loss of DP cells and in hair production.
In the present study, the hypothesis that troxerutin inhibits ROS-mediated cellular dysfunction in human DP (HDP) cells was investigated. In addition, using micro (mi)RNA microarrays and bioinformatics analysis, the role of troxerutin in the regulation of the expression and mechanisms of specific miRNAs was evaluated. The present study aimed to examine troxerutin as a potential novel chemical agent for the prevention and/or treatment of alopecia.

Materials and methods
Cell culture and viability. The HDP cells were purchased from Innoprot (Biscay, Spain) and cultured in Dulbecco's modified Eagle's medium, containing 10% fetal bovine serum (FBS; Thermo Fisher Scientific, Waltham, MA, USA) and 1% penicillin-streptomycin (Gibco Life Technologies, Grand Island, NY, USA) at 37˚C and 5% CO 2 . The cells were plated at a density of 4x10 3 /well in a 96-well plate. At 70-80% confluence, the cells were treated with troxerutin (Sigma-Aldrich, St. Louis, MO, USA) at concentrations ranging between 0 and 60 µM for 24 h at 37˚C. Subsequently, 10 µl water-soluble tetrazolium salt assay solution (EZ-Cytox Cell Viability Assay kit; Itsbio, Seoul, Korea) was added to each well and, following incubation for 30 min at 37˚C, the optical density was measured at 490 nm using an iMark microplate reader (Bio-Rad Laboratories, Inc., Hercules, CA, USA). To examine troxerutin-mediated ROS protection, the cells were pretreated with troxerutin at the following concentrations: 0, 5, 10 and 15 µM for 8 h. Subsequently, 750 µM H 2 O 2 was added to each well. Following incubation for 24 h at 37˚C, cell viability was evaluated using an EZ-Cytox Cell Viability Assay kit. The level of cell viability (%) was normalized to that of 0.1% dimethyl-sulfoxide (DMSO; Sigma-Aldrich)-treated cells. Each experiment was repeated at least three times. The P-value was determined using Student's t-test and P<0.05 was considered to indicate a statistically significant difference.
Analysis of cell cycle. The HDP cells (2x10 6 ), which had been treated with troxerutin and/or H 2 O 2 were trypsinized using 0.25% trypsin-EDTA (Gibco Life Technologies), washed once with phosphate-buffered saline (PBS), and used for analysis. The cell cycle distribution was measured using propidium iodide (PI; Sigma-Aldrich) staining solution, containing 50 µg/ml PI, 0.5% Triton X-100 (Sigma-Aldrich) and 100 µg/ml RNase (Qiagen, Hilden, Germany). The cells were fixed in 70% cold ethanol (Merck Millipore, Darmstadt, Germany) and incubated for 1 h at -20˚C. Subsequently, PI staining solution was added to the fixed cells, followed by incubation for 1 h in the dark at 37˚C. The PI fluorescence intensity was detected using a BD FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA). The mean PI fluorescence intensity was calculated based on the measurements of 10,000 cells using the FL2-H channel.
Analysis of intracellular levels of ROS. The HDP cells (2x10 6 ), which had been treated with troxerutin and/or H 2 O 2 were washed with PBS and trypsinized. Intracellular ROS levels were measured using 2'7'-dichlorofluorescein diacetate fluorescent dye (DCF-DA; Sigma-Aldrich), as described previously (21). The cells were resuspended in 10 µM DCF-DA and further incubated at room temperature for 1 h in the dark. The intensity of the resulting fluorescence was measured using a BD FACSCalibur flow cytometer (BD Biosciences). The mean DCF fluorescence intensity was calculated based on measurements of 10,000 cells using the FL1-H channel. The M1 range was calculated as the percentage of each subpopulation of cells exhibiting increased DCF-DA fluorescence.
Analysis of cellular senescence. The HDP cells (2x10 6 ), which had been treated with troxerutin and/or H 2 O 2 were washed in PBS and fixed for 5 min at room temperature in 2% formaldehyde/0.2% glutaraldehyde (Sigma-Aldrich). The cells were washed with PBS and incubated for 15 h at 37˚C with SA β-galactosidase staining solution (BioVision, Milpitas, CA, USA). The stained cells (blue) were observed using a bright-field microscope (CKX41; Olympus Corporation, Tokyo, Japan; magnification, x200), counted in three different fields and the percentage of stained cells was determined.
Analysis of miRNA expression profiles. The HDP, which had been cells treated with troxerutin and/or H 2 O 2 were washed in cold PBS and trypsinized for RNA purification. The total RNA was extracted and purified from the cells using TRIzol ® reagent (Invitrogen Life Technologies, Carlsbad, CA, USA), according to the manufacturer's instructions. The integrity (RNA integrity number >8.0) and purity (A260/280 and A260/230 values >1.8) were confirmed using an Agilent 2100 Bioanalyzer ® (Agilent Technologies, Inc., Santa Clara, CA, USA) and a MaestroNano ® microvolume spectrophotometer (Maestrogen, Las Vegas, NV, USA), respectively. Samples (100 ng) of RNA meeting these criteria were first dephosphorylated by incubation with calf intestinal alkaline phosphatase (Agilent Technologies, Inc.) at 37˚C for 30 min. Subsequently, cyanine 3-pCp labeling solution (Agilent Technologies, Inc.) and T4 RNA ligase (Agilent Technologies, Inc.) were added to the dephosphorylated RNA samples and incubated at 16˚C for 2 h. Following the labeling reaction, the samples were dried and treated with GE Blocking Agent (Agilent Technologies, Inc.). The samples were hybridized to the SurePrint G3 Human v16 miRNA 8x60 K (Agilent Technologies, Inc.) microarray at 55˚C, with constant rotation at 20 rpm in an Agilent Microarray Hybridization Chamber(Agilent Technologies, Inc.) for 20 h. The array was then washed and scanned using an Agilent SureScan Microarray scanner and the images captured were quantified using Agilent Feature Extraction software (version 10.7; Agilent Technologies, Inc.). The data were analyzed with the assistance of GeneSpring GX software version 7.3 (Agilent Technologies). In addition, fold-change analysis was performed to select those with ≥2.0-fold between the H 2 O 2 -treated control cells and those treated with troxerutin and H 2 O 2 .
Bioinformatic analysis of altered miRNAs. Analysis of the biological significance of the altered miRNAs in the present study was performed, as previously described (21). First, the putative target genes of the altered miRNAs were predicted using Micro Cosm Targets Version 5 (http://www.ebi.ac.uk/enright-srv/microcosm/htdocs/targets/v5/). Subsequently, the target genes were grouped into four categories: Aging, skin development, apoptosis and cell proliferation, based on the AmiGo 2 Gene Ontology (GO) analysis tool (amigo.geneontology. org/cgi-bin/amigo/browse.cgi). The putative target genes of each miRNA were further analyzed for biologic function using the Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway within the Database for Annotation, Visualization and Integrated Discovery, (DAVID; http://david.abcc.ncifcrf. gov/home.jsp) bioinformatics resources (version 6.7), according to the standard procedures (22). The 'KEGG_pathway' category was processed by setting the threshold of the EASE score, a modified Fisher's exact P-value, to 0.1. The KEGG pathways, identified as having a percentage of involved target genes / total target genes in each pathway >1% were selected.

H 2 O 2 -induced cell damage is inhibited by troxerutin in HDP cells. HDP cells are a major component of skin and direct
hair growth, and loss or senescence of these cells is a key cause of hair loss (15). Previous reports have demonstrated that cisplatin-and androgen overexposure-mediated cellular dysfunction, including apoptosis, can induce alopecia and occur predominantly by stimulating the production ROS in HDP cells (16,23,24). In our previous study, pretreatment of human dermal fibroblasts and HaCaT keratinocytes with troxerutin was observed to protect against UVB-mediated cell death (6,7). UVB light is an important inducer of ROS production in several types of cell (13); therefore, the present study examined the possible role of troxerutin in protecting against ROS-induced cell stress and damage in HDP cells. Initially, troxerutin-mediated cytotoxicity in HDP cells was screened for. No significant changes in cell viability were detected following treatment with between 0 and 60 µM troxerutin for 24 h (Fig. 1A), indicating that troxerutin was non-toxic in the HDP cells. Subsequently, the protective role of troxerutin against ROS-induced cell damage was determined using H 2 O 2 , a ROS inducer. The HDP cells were pretreated with several concentrations (0-60 µM) of troxerutin for

H 2 O 2 -induced cell death is inhibited by troxerutin in HDP cells.
Our previous study demonstrated that troxerutin effects the level of cell death induced by UVB irradiation (6,7) and Fig. 1 shows the protective role of troxerutin against ROS-induced cell damage, observed in the present study. Therefore, to examine whether troxerutin is involved in the response to ROS stress, which is known to inhibit cell-cycle progression and induce cell death, the present study investigated changes in the cell cycle and in cell death in the HDP cells pretreated with different concentrations of troxerutin prior to H 2 O 2 exposure. The HDP cells were pretreated with 0, 5 or 10 µM troxerutin for 6 h, followed by treatment of the cells with 750 µM H 2 O 2 for an additional 24 h. Subsequently, the cells were analyzed by flow cytometry. At concentrations of 5 and 10 µM, pretreatment with troxerutin caused a decrease in the number of cells in the sub-G1 phase, indicative of cell death ( Fig. 2A). H 2 O 2 increased the

A B
percentage of the non-pretreated cells in the sub-G1 phase to 10.63±0.43%; however, this value increased to only 9.38±0.11 and 4.53±0.53% in the cells pretreated with 5 and 10 µM troxerutin, respectively (P<0.05; Fig. 2A and B). Therefore, these results suggested that troxerutin overcame the effect of H 2 O 2 -mediated cell death, resulting in a diminution of cells in the sub-G1 phase.

A B A B
The cells pretreated with troxerutin were 19.92±1.95% DCF-positive following H 2 O 2 treatment, indicating that troxerutin reduced the H 2 O 2 -induced production of ROS in the HDP cells.
Hydrogen peroxide-induced senescence is inhibited by troxerutin in HDP cells. Increased ROS are one of the key mediators of cellular senescence (25). A previous report demonstrated that the premature senescence of balding DP cells is associated with changes in the expression of SA β-galactosidase and also suggested that oxidative stress may be involved in the premature senescence of these cells (20 (Fig. 4). These data indicated that troxerutin has the potential to inhibit cellular senescence in HDP cells.
Troxerutin-mediated protective effects against ROS are involved in changes in miRNA expression in HDP cells. The present study also analyzed the miRNA expression profiles of non-pretreated and troxerutin-pretreated HDP cells treated with H 2 O 2 , as miRNAs can be involved in cell death, ROS scavenging and senescence (21,(26)(27)(28). In the microarray, a total of 24 miRNAs were detected with a ≥2.0-fold change in expression levels between the two groups. Among these, 10 miRNAs were upregulated and 14 were downregulated in the H 2 O 2 -treated cell, which had been pretreated with troxerutin (Table I). To investigate the biological value of the microarray data, several bioinformatic analyses were performed to predict the putative target genes of the altered miRNAs, and the GO and signaling pathways of the target genes. The putative target genes of each miRNA were determined using MicroCosm Targets Version 5, following which GO analysis was performed for the target genes. Subsequently, the target genes of each miRNA were categorized into four biological functions: Aging, skin development, apoptosis and cell proliferation. Several target genes of each miRNA were found to be involved in these four biological functions at different levels (Tables II and III). For example, hsa-miR-602, which was the most highly upregulated miRNA (6.91-fold) based on the microarray data, potentially targets 34 genes, six of which were involved in aging, and one of the remaining 28 target genes was involved in skin development (Table II). Similarly, the target genes of the downregulated miRNAs were also differentially involved in the four functions (Table III), indicating that the altered miRNAs identified by the microarray analysis had distinct biological roles associated with the protective effect of troxerutin in H 2 O 2 -treated HDP cells. Therefore, the present study further analyzed the signaling pathways associated with the upregulated and downregulated miRNAs using KEGG pathway analysis and the DAVID bioinformatics tool (22), the results of which are presented in Tables IV and V, respectively. The results demonstrated that the altered miRNAs are functionally involved in shared and unique pathways among the miRNAs. For example, hsa-miR-602 was identified to be functionally involved in MAPK, insulin, and calcium signaling pathways, whereas hsa-miR-205-3p was found to be involved in cancer, MAPK, Wnt and cell adhesion signaling pathways. Overall, these results indicated that the miRNA expression patterns of non-pretreated and troxerutin-pretreated H 2 O 2 -treated HDP cells can be distinguished, and those which are significant changed may be involved in troxerutin-mediated protection against H 2 O 2 -induced cellular stress through the regulation of multiple signaling pathways.

Discussion
In the present study, the protective effect of troxerutin against H 2 O 2 -induced oxidative stress in HDP cells was confirmed using biochemical assays. Notably, pretreatment with troxerutin decreased the cell death, ROS production and cellular senescence, which was mediated by exposure to H 2 O 2 .
Although the specific signaling pathways involved in the protective effect were not demonstrated, the findings of the present study are important in that they identify troxerutin as a candidate agent for use in the prevention and/or treatment of    alopecia. A growing body of evidence suggests the role of oxidative stress in alopecia, and that the prevention of oxidative stress may offer novel strategies for the intervention and reversal of alopecia and even graying of hair (17). A previous case study confirmed increased oxidative stress in alopecia areata patients compared with healthy individuals (29). In addition, a study using a mouse model demonstrated that hair dye-induced hair loss is caused predominantly by H 2 O 2 -induced oxidative stress (30). Oxidative stress stimulates the production of a known inhibitor of hair follicles, tumor growth factor-β, in DPC cells, which induces the onset of androgenic alopecia (24). Our previous studies demonstrated that troxerutin has a photoprotective effect against UV radiation on dermal fibroblasts and keratinocytes (6,7), and several clinical and theoretical reports have revealed that UV radiation has negative effects on hair growth through the induction of oxidative stress, acute telogen effluvium and follicular micro-inflammation in follicular stem cells (31)(32)(33). Therefore, countering oxidative stress can be considered an important strategy to overcome stress-or androgen-dependent alopecia, and the results of the present study confirmed that troxerutin inhibited oxidative stress-induced cellular damage in the DPC cells. In addition, our previous studies and the present studies demonstrated low levels of cytotoxicity of troxerutin on dermal fibroblasts, keratinocytes and DP cells (6,7). Therefore, further investigation of the clinical effect of topical application of troxerutin to the scalp is required. Using miRNA microarray analysis, the present study identified 24 miRNAs in the HDP cells treated with troxerutin and H 2 O 2 , which were differentially expressed compared with the cells treated with H 2 O 2 only. Of these, hsa-miR-602 was the most markedly upregulated by troxerutin in the H 2 O 2 -treated HDP cells (6.91-fold), and has been reported to downregulate the expression of the RASSF1A and TP73 tumor suppressor genes (34). Several reports have revealed that H 2 O 2 induces the expression of TP73 (35) and that the anticancer drug cisplatin, which has been reported to induce alopecia in patients, stimulates ROS-induced apoptosis and functionally upregulates the expression of p73 (36)(37)(38). Although the cellular functions of miR-602, RASSF1A and TP73 have not been investigated in DP cells, the data of the present study suggested that the interaction between miR-602 and the two genes may be functionally involved in ROS-induced cellular stress and even alopecia-associated mechanisms. The biological functions of hsa-miR-575, which was the most downregulated miRNA in the results of the present study, have not been reported previously; however, it may regulate H 2 O 2 -mediated cellular stress. PDPK1, also termed PDK1, is a putative target of miR-575 (Table III) and a well-known kinase, which phosphorylates and activates Akt1 kinase, induces cell proliferation and protects against H 2 O 2 -induced apoptosis (39,40). The present study also classified the biological functions of differentially expressed miRNAs by troxerutin in the H 2 O 2 -treated HDP cells. KEGG pathway analysis of the target genes of the upregulated and downregulated miRNAs revealed 18 and 23 pathways, respectively, were statistically enriched. Among these, the WNT and MAPK signaling pathways, which were the most markedly enriched pathways associated with the target genes of the upregulated and downregulated miRNAs, are involved in the regulation of H 2 O 2 -mediated cellular stress, including apoptosis and antioxidative mechanisms (41)(42)(43)(44)(45), suggesting that the miRNAs altered by troxerutin may be involved in protective mechanisms against H 2 O 2 -induced damage through the regulation of these pathways.
In conclusion, the present study revealed a novel role of troxerutin as a putative antioxidant agent in HDP cells. In addition, the results revealed 24 differentially expressed miRNAs and determined the putative involvement of 18 signaling pathways associated with upregulated miRNAs and 23 signaling pathways associated with downregulated miRNAs in the troxerutin-mediated protective effect against H 2 O 2 -induced cell damage. Although further experiments are required to confirm the differentially expressed miRNAs and their target genes, the results of the present study may assist in elucidating the mechanism underlying the troxerutin-mediated protection and miRNA-associated signaling pathways in HDP cells.