Normal human neonatal skin fibroblasts (NB1RGB cells) were purchased from the RIKEN Cell Bank (Ibaraki, Japan). The cells were cultured in alpha-Minimum Essential Medium (MEM) medium (Thermo Fisher Scientific, Waltham, MA, United States) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific), penicillin 100 U/mL, and streptomycin 2.5 μL/mL. Cells were plated at a concentration of 1 × 10⁵ cells/dish (60 mm), and the plates were supplemented with a 1% alpha-arbutin (GLICO Co. Ltd.) solution or distilled water (Gibco) as a negative control. This dose was selected because 1% solution is the maximum permitted dose in over-the-counter cosmetic products. After incubating for 48 h at 37°C in humidified air with 5% CO₂, the total RNA was isolated from the fibroblasts using the RNeasy kit (Qiagen) according to the manufacturer’s protocol. The extracted RNA was stored at −80°C.

A reverse transcription of the total RNA was carried out using the SuperScript II Reverse Transcriptase (Thermo Fisher Scientific), which was stored at 4°C. A quantitative real-time polymerase chain reaction (PCR) of the MMP3, EGFR, FOXO1, SIRT1, COL1A1, IGF1R1, ELOVL3, and NFE2L2 (Nrf2) genes was performed using the StepOnePlus Real-Time PCR System (Thermo Fisher Scientific) with TaqMan gene expression assays. A relative quantification of gene expression was performed using the standard curve method, with glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene as an endogenous control.

An unpaired statistical t test was performed to verify the significance between expression in control and arbutin supplementation condition. Genes were normalized to GAPDH housekeeping genes before performing the t test.

After using RT-PCR to verify changes in gene expression, we assessed the total antioxidant profile of the cells incubated for 48 h using the Human Oxidative Stress RT2 Profiler PCR Array (Qiagen NV), which contains 84 key genes related to the oxidative stress response. All reactions were carried out using the StepOnePlus Real-Time PCR System according to the manufacturer’s protocol and statistical analysis. The relative expression values were calculated using the 2^–ΔΔCt method in which the cutoff was set to 35 Ct. Genes with low absolute expression levels were excluded from further analyses. All values from both the TaqMan assay and PCR array experiments are presented as the arithmetic means of two biological replicates.

Phylogenetic footprinting using the SHOE software (Polouliakh et al., 2018) previously developed by the authors was performed on genes overexpressed in the PCR oxidative stress array and gene expression assay. An analysis workflow was performed on the results of SHOE using the REACTOME database (Caley et al., 2015) and CellDesigner pathway editor (Funahashi et al., 2007).

This study represents an initial step toward investigating the beneficial effects of alpha-arbutin on human dermal fibroblast cell cultures. A gene expression assay revealed that treatment with alpha-arbutin enhanced the mRNA expression of type I procollagen (COL1A1), matrix metallopeptidase 3 (MMP3), ELOVL fatty acid elongase 3 (ELOVL3), insulin-like growth factor 1 receptor (IGF1R), and epidermal growth factor receptor (EGFR). On the other hand, forkhead box protein O1 (FOXO1) and sirtuin 1 (SIRT1) were downregulated by the supplementation of alpha-arbutin (Figures 1A,B). The RT2 Profiler PCR Array analysis of the oxidative stress pathway revealed 34 genes that doubled or halved their expression, and among them, the following 11 genes are involved in the downstream regulation of NFE2L2 (Nrf2) gene (Table 1) upon exposure to hydroquinone. They are TXN, GSTP1, TXNRD1, SQSTM1, PRDX4, GPX3, GSTZ1, NQO1, HMOX1, GCLM, and FTH1 genes. NFE2L2 has also been reported to be activated by hydroquinone and tert-butylhydroquinone (Papaconstantinou, 2009). Arbutin may release hydroquinone after its glycosidic bonds are cleaved or through hydrolysis by human skin bacteria (Tada et al., 2014). The arbutin and hydroquinone structures are shown in Supplementary Figure 1. The only difference between the alpha-arbutin and hydroquinone structures is the glycosidic bonds (sugar chain). In our study, NFE2L2 was upregulated at 48 h (Figure 2). From this, we concluded that alpha-arbutin may activate NFE2L2, which consequently activates target genes that reduce ROS.

The upregulated genes in the oxidative stress pathway include 24-dehydrocholesterol reductase (DHCR24), sirtuin 2 (SIRT2), thioredoxin reductase 1 (TXNRD1), and heme oxygenase 1 (HMOX1). Members of the peroxiredoxin system (PRDX4, PRDX5) and a member of the glutathione peroxidase system (GPX3) were also upregulated, and they each represent two different systems in the oxidative stress response. The glutathione metabolic pathway was prominently upregulated (GPX3, GSS, GSTP1, GCLM, GSTZ1 genes), and its fold changes ranged from 2.37 to 2.39.

Other noticeably upregulated genes directly or indirectly involved in the antioxidative status of the cell include sulfiredoxin 1 (SRXN1), which works with PRDX4; HSPA1A, which stabilizes proteins against aggregation; NQO1, which reduces antioxidant molecules; and FTH1, which is involved in iron storage. The prominently downregulated genes contained selenoprotein P (SEPP1), dual oxidase 2 (DUOX2), apolipoprotein E (APOE), eosinophil peroxidase (EPX), the NADPH oxidase 5 (NOX5), lactoperoxidase (LPO), and nitric oxide synthase 2 (NOS2).

First, we assumed that alpha-arbutin is involved in wound healing and the oxidative stress response. This is because the MMP3 gene regulates wound healing rate through its role in wound contraction, as shown in a previous study (Caley et al., 2015). The upregulation of MMP3 and EGFR was previously identified (Kusukawa et al., 1996) (Figure 3, reaction 1) as well. In our study, we also observed upregulation for EGFR (Figure 1A). The downregulation of FOXO1 at 48 h (Figure 1B) provided further evidence of arbutin’s involvement in wound healing. Enhanced wound repair and reduced scarring were observed in a previous study on FOXO1± mice (Mori et al., 2014), which supports our findings.

The gene expression of type I collagen α1 (COL1A1) was found to significantly decrease at the wound sites of FOXO1± mice 7 days after injury, which is believed to be the reason for less scarring (Caley et al., 2015). COL1A1 was downregulated after 24 h (data not shown) and upregulated after 48 h in our study (Figure 1A).

FOXO1 has been reported to target genes coding for intracellular and extracellular antioxidant proteins and is deacetylated by SIRT1 (Figure 3, reaction 2), which results in its activation in most cases (Kusukawa et al., 1996; Caley et al., 2015). In our findings, SIRT1 was also downregulated, supporting the notion that FOXO1 is inactive. SIRT1 downregulation phosphorylates FOXO1, which is consequently inactivated by the serine/threonine-specific protein kinase B (Akt), as shown in Figure 3, reaction 3. Kinase Akt may be inactivated by stimulating cells with insulin (Caley et al., 2015). Akt is downstream of insulin/IGF-1 signaling and is activated by PI3K through the phosphorylation of insulin receptor substrates-1 and -2 (Kusukawa et al., 1996).

IGF1R was upregulated in our findings; this further suggests that the insulin pathway is active, meaning that Akt phosphorylates and inactivates FOXO1. EGFR and IGF1R were found to interact on multiple levels in the early stages and activated each other (Figure 3, reaction 2) directly or indirectly (Zellmer et al., 2010; Mori et al., 2014).

The EGFR-mediated repression of GATA3 activated the transcription of ELOVL3 (Sugimoto et al., 2004) (Figure 3, reaction 5), supporting our finding of increased ELOVL3 expression. ELOVL3 is a fatty acid chain elongation enzyme necessary for skin functions. IGF1R was found to regulate LARP6 expression in an Akt signaling–dependent manner. The activation of the pathway led to an upregulation of LARP6, causing an increase in COL1A1 expression (Hughes et al., 2011) (Figure 3, reaction 2). The Ins/IGF-1 signaling pathway itself is of particular interest because of the changes in its responsiveness to the ROS environment; lower levels of ROS activate the pathway, whereas higher ROS levels inhibit its signaling processes (Gross et al., 2009). Alpha-arbutin has been found to inhibit the formation of hydroxyl radicals via L-tyrosine-tyrosinase (Figure 3, reaction 7), and alpha-arbutin is expected to alleviate oxidative stress derived from the melanogenic pathway in the skin (Kousteni, 2011). This supports the notion that the lower ROS levels were a result of the alpha-arbutin treatment and its inhibitory effect on tyrosinase.

PTGS2 was found to be induced by Akt and hypothesized to be active throughout the NF-κB pathway in mutated PTEN endometrial cancer cells (van der Veeken et al., 2009) (Figure 3, reaction 4). It is unclear why NOS and SEPP1, both expressed predominantly in the liver, are downregulated. It is also unclear why NOX5 predominantly expressed in the testis and lymphocyte-rich areas of the spleen and lymph nodes is downregulated in alpha-arbutin–treated fibroblasts. Among the downregulated genes, apolipoprotein 4 (APOE4) is of particular interest (Table 1). It was previously found that the downregulation of APOE4 at a mature age attenuates Parkinson disease (Jin et al., 2016; Sugita et al., 2019); however, its downregulation by alpha-arbutin was unknown.

Twenty-eight out of the 42 genes (34-gene PCR array + 8-gene gene assay) analyzed with SHOE had orthologs in mice and rats. Fifteen genes were determined by SHOE as those having cross-species conserved transcription factors binding motifs (Supplementary Table 1). The motifs that matched the consensus sequence to ≥70% were selected for Supplementary Table 1 and sorted accordingly. Examples of the discovered motifs can be seen in Figure 4. Supplementary Figure 2 shows the further analysis workflow after identifying the transcription factor motifs, and Supplementary Figure 3 shows genes visualized by CellDesigner having cross-species conserved transcription factors binding motifs (Kusukawa et al., 1996).

Here, select genes and transcription factor motifs are discussed; the rest can be viewed and downloaded on the SHOE site¹.

The IGFR1 gene had the most cross-species conserved promoters between human–mouse–rat (85 sites), with top candidate transcription factors such as E2F (2 sites), TFII-I (2 sites), SP1 (6 sites), ETF (10 sites), NF-κB (1 site), and c-Rel (1 site). (Hereafter, the numbers of copies of the site per motif name are indicated by numbers only.)

The second most substantial conserved region was located in the APOE gene (27 sites), with the top motifs MAZ (4), E2F (2), ETF (2), SP1 (4), CAC-binding (2), and ZF5 (2). The third was the GPX1 gene (22 sites), with top-scoring AML1 (2), Ets (3), E2F, AP-1, MAZ, NF-E2, and USF.

The GSS gene (17 sites) had E2F (2), SP1 (3), MAZ, NRF1(2), LF-A1, CAC binding, KROX (4) motifs, and the GCLM gene (15 sites) appeared to have Lyf-1, STATx, SP1, E2F (2), and NFE2L2 motifs in the promoter, which emphasizes our hypothesis of regulation overexpressed genes via NFE2L2.

E2F, ETF, and SP-1 sites were present in nearly all genes. E2F transcription factors are known to regulate many genes involved in DNA synthesis and cell cycle entry (Mori et al., 2014). Together with ETF and SP-1, E2F1 was found to have an essential role in murine hepatocytes in the process of proliferation-dependent differential gene expression (Zellmer et al., 2010). E2F1 and SP1 were previously demonstrated to functionally and physically interact with each other and may regulate the transcription of genes that contain one or both binding sites for the respective transcription factors (Hughes et al., 2011).