Excessive All-Trans Retinoic Acid Inhibits Cell Proliferation Through Upregulated MicroRNA-4680-3p in Cultured Human Palate Cells

Cleft palate is the second most common congenital birth defect, and both environmental and genetic factors are involved in the etiology of the disease. However, it remains largely unknown how environmental factors affect palate development. Our previous studies show that several microRNAs (miRs) suppress the expression of genes involved in cleft palate. Here we show that miR-4680-3p plays a crucial role in cleft palate pathogenesis. We found that all-trans retinoic acid (atRA) specifically induces miR-4680-3p in cultured human embryonic palatal mesenchymal (HEPM) cells. Overexpression of miR-4680-3p inhibited cell proliferation in a dose-dependent manner through the suppression of expression of ERBB2 and JADE1, which are known cleft palate-related genes. Importantly, a miR-4680-3p-specific inhibitor normalized cell proliferation and altered expression of ERBB2 and JADE1 in cells treated with atRA. Taken together, our results suggest that upregulation of miR-4680-3p induced by atRA may cause cleft palate through suppression of ERBB2 and JADE1. Thus, miRs may be potential targets for the prevention and diagnosis of cleft palate.


INTRODUCTION
The prevalence of cleft lip with or without cleft palate (CL/P) is ∼1 in 700 live births worldwide (Ferguson, 1988), and affected individuals require surgical repairs, speech therapy, and dental treatments (Ferguson, 1988). Palate development (which consists of the growth, elevation, and fusion of the palatal shelves) starts at 6-8 weeks of gestation in humans; any failure in these process results in cleft palate (Habib, 1978;Mossey et al., 2009). The etiology of cleft palate is complex and involves environmental and genetic factors and their interactions (Dhulipala et al., 2006;Havasi et al., 2013;Buser and Pohl, 2015;Liu et al., 2018). Maternal use and exposure to tobacco, alcohol, drugs (e.g., retinoic acid and dexamethasone), and chemicals (e.g., dioxin and heavy metals), as well as mutations in genes related to the degradation/metabolism/release of these teratogens, are considered to be risk factors (Prescott et al., 2002;Chevrier et al., 2005;Ramirez et al., 2007). An increasing number of studies suggest that microRNAs (miRs), which are endogenous small non-coding RNAs (∼22 nucleotides long) that negatively regulate the expression of their target genes (Bartel, 2004;Obernosterer et al., 2006), play important roles in normal palate development and CL/P in humans and mice (Karsy et al., 2010;Shin et al., 2012;Seelan et al., 2014;Chung et al., 2016;Schoen et al., 2017Schoen et al., , 2018Wang et al., 2017;Mukhopadhyay et al., 2019); however, it remains elusive how and which miRs are crucial roles in CL/P. Our recent studies show that overexpression of either miR-140-5p, miR-133b, miR-374a-5p, miR-381-3p, or miR-4680-3p suppresses cell proliferation in cultured human embryonic palatal mesenchymal (HEPM) cells Suzuki et al., 2019), suggesting that these miRs may be involved in the pathogenesis of cleft palate.
All-trans retinoic acid (atRA), a derivative of vitamin A, is the most active retinoid that plays important roles in a variety of biological processes, including cell proliferation, differentiation, and extracellular matrix production (Wang and Kirsch, 2002;Lai et al., 2003;Rhinn and Dolle, 2012;Cunningham and Duester, 2015). While atRA is widely used in the treatments of skin diseases and cancers (Karsy et al., 2010;Siddikuzzaman and Berlin Grace, 2011;Mihaly et al., 2012), excessive intake of atRA, a known teratogen, disrupts embryogenesis, causing birth defects (Abbott et al., 1989;Ross et al., 2000;Roberts, 2020) including cleft palate (Abbott et al., 1989;Ross et al., 2000;Wang and Kirsch, 2002;Lai et al., 2003;Yao et al., 2011;Havasi et al., 2013;Hu et al., 2013;Hou et al., 2019;Roberts, 2020). Recent studies show that atRA alters the expression of miRs in human cancer cell lines (Liu et al., 2018(Liu et al., , 2019. In this study, we will determine whether and how atRA can alter miR expression, which suppresses the expression of genes related to CL/P.

Cell Culture
HEPM cells were obtained from American Type Culture Collection (CRL-1486; ATCC) and maintained under Minimum Essential Medium Eagle-alpha modification (αMEM), supplemented with 10% fetal bovine serum (FBS), penicillin/streptomycin, and L-glutamine, at 37 • C in a humidified atmosphere with 5% CO 2 .

Bromodeoxyuridine (BrdU) Incorporation Assay
The cells were plated onto 35-mm dishes at a density of 25,000/dish and treated with 30 µM atRA, or vehicle (dimethyl sulfoxide). After 72 h, the cells were incubated with BrdU for 1 h. Incorporated BrdU was stained with a rat monoclonal antibody against BrdU (ab6326; Abcam, 1:1,000), as previously described (Suzuki et al., 2018a). A total of nine fields, which were randomly selected from three independent experiments, was used for the quantification of BrdU-positive cells.
Terminal 2 ′ -Deoxyuridine, 5 ′ -Triphosphate (dUTP) Nick-End Labeling (TUNEL) Staining The cells were plated onto 35-mm dishes at a density of 25,000/dish and treated with 30 µM atRA or vehicle for 72 h. The Click-iT Plus TUNEL Assay with Alexa 594 (C10618, Molecular Probes) was used to detect apoptotic cells, as previously described (Suzuki et al., 2018b). A total of four fields, which were randomly selected from two independent experiments, was used for the quantification of TUNEL-positive cells.

Knockdown and Overexpression of ERBB2 and JADE1
For the siRNA experiments, the cells were plated onto 35-mm dishes at a density of 20,000/dish. When the cells reached 70% confluency, they were treated with siRNA for either ERBB2, JADE1, or negative control, at 3 pmol in 6 µL of transfection reagent (TransIT-X2 system) in 2 mL of αMEM per dish.

Rescue Experiments
The cells were plated onto 60-mm dishes at a density of 250,000/well and treated with 30 µM atRA or vehicle control.
After 24 h, the cells were transfected with either miR-4680-3p inhibitor (3 pmol) or control miR inhibitor (3 pmol; mirVana, Thermo Fisher Scientific), and either ERBB2, JADE1, or control overexpression vector, using the TransIT-X2 system (Mirus Bio LLC), according to the manufacturer's protocol (at 1 µg in 12 µL of transfection reagent in 4 mL of αMEM per dish). The cells were harvested 24 h after transfection and used for further experiments.

Statistical Analysis
All experiments were performed independently at least two times. The statistical significance of the differences between two groups was evaluated using a two-tailed Student t-test. Multiple comparisons were made by one-way analysis of variance with the post-hoc Tukey-Kramer's test. A p < 0.05 was considered to be statistically significant. Data are represented as mean ± standard deviation in the graphs.

atRA Inhibits Cell Proliferation in a Dose-Dependent Manner in HEPM Cells
To determine the dose-dependent effects of atRA on cell proliferation in HEPM cells, we performed cell proliferation assays using HEPM cells treated with atRA at various concentrations (0, 1, 3, 10, and 30 µM). We found that cell proliferation activity was decreased by atRA treatment in a dose-dependent manner ( Figure 1A). BrdU incorporation assays confirmed that cell proliferation was significantly decreased in cells treated with 30 µM atRA (Figures 1B,C). While excessive atRA is known to induce apoptosis in several tissues and cells (Okano et al., 2007;Mercader et al., 2008;Nelson et al., 2019;Quan et al., 2019), atRA failed to induce apoptosis in HEPM cells at 30 µM (Supplementary Figure 1). Previous studies show that atRA inhibits cell proliferation through downregulation of cyclin D1 (CCND1) and upregulation of cyclin-dependent kinase inhibitor 1B (CDKN1B; a.k.a. p27, KIP1) in HEPM cells (Dong et al., 2017) as well as in MCF-7 cells, a human breast cancer cell line (Teixeira and Pratt, 1997). We therefore evaluated the expression of CCND1 and CDKN1B by immunoblotting and confirmed that CCND1 expression was downregulated, and CDKN1B expression was upregulated, with atRA treatment (Figure 1D).
atRA Suppresses ERBB2 and JADE1 Expression Through Upregulation of miR-4680-3p in HEPM Cells Our previous studies showed that overexpression of either miR-140-5p, miR-133b, miR-374a-5p, miR-381-3p, or miR-4680-3p inhibits proliferation of HEPM cells through the suppression of genes that are crucial for palate development Suzuki et al., 2019). We therefore hypothesized that atRA induces the expression of these miRs. To test that hypothesis we analyzed the expression of these miRs after treatment of HEPM cells with atRA for 24 and 72 h, and found that expression of miR-4680-3p, but not miR-140-5p, miR-133b, miR-374a-5p, or miR-381-3p, was specifically and significantly upregulated with atRA treatment (Figure 2A). Next, to identify target genes suppressed by a miR-4680-3p mimic, we performed quantitative RT-PCR (qRT-PCR) analyses for the target genes (ERBB2, JADE1, MTHFD1, and WNT5A), which were predicted through bioinformatic analysis . We found that expression of ERBB2 [a.k.a. HER2, a member of the epidermal growth factor receptor (EGFR) family of transmembrane tyrosine kinase-type receptors (Schechter et al., 1984)] and JADE1 [a.k.a. PHF17, a member of the extended plant homeodomain (PHD) finger protein subfamily (Tzouanacou et al., 2003)] was significantly downregulated in HEPM cells treated with atRA for 24 and 72 h ( Figure 2B). This suppression of ERBB2 and JADE1 in atRAtreated cells was confirmed with immunoblotting after 72 h of atRA treatment (Figure 2C). Previous studies showed that ERBB2 stimulates several intracellular pathways such as MAPK/ERK and PI3K/AKT/mTOR (Yarden and Pines, 2012;Croessmann et al., 2019). For this reason, we analyzed the ERK1/2 and mTOR pathways in cells treated with atRA and found that ERK1/2 phosphorylation was downregulated, while mTOR phosphorylation was not altered, with atRA treatment ( Figure 2C). ERBB2 was detected in the plasma membrane and cytosol in controls, as previously reported (Chung et al., 2016); by contrast, its expression was significantly decreased in atRA-treated cells (Figure 2D). JADE1 was detected in the nucleus and cytosol in controls, as previously reported (Panchenko et al., 2004;Havasi et al., 2013;Siriwardana et al., 2015); however, JADE1 expression was significantly decreased in atRA-treated HEPM cells ( Figure 2D). Thus, our results indicate that atRA induces miR-4680-3p expression, leading to the suppression of ERBB2 and JADE1 via ERK1/2 signaling in HEPM cells.
Next, to evaluate the effect of expression of ERBB2 and JADE1 on cell proliferation, we treated HEPM cells with siRNAs for ERBB2 and JADE1. We confirmed that siRNA knockdown of either ERBB2 or JADE1 suppressed their expression at the mRNA and protein levels (Figures 3A-D). Under these conditions, cell proliferation was significantly suppressed by either ERBB2 or JADE1 siRNA knockdown. In addition, additional suppression was observed with a combination of ERBB2 and JADE1 siRNAs ( Figure 3E). Furthermore, we confirmed that knockdown of ERBB2 and JADE1 in HEPM cells resulted in downregulated CCND1 and upregulated CDKN1B (Figure 3F). To evaluate the functional significance of ERBB2 and JADE1, we conducted rescue experiments by overexpressing ERBB2 and JADE1 in cells treated with atRA. We first confirmed that expression of ERBB2 and JADE1 was significantly upregulated following overexpression of these genes (Figures 3G,H). Under these conditions, we found that overexpression of ERBB2 and JADE1 partially rescued the cell proliferation inhibited by atRA (Figure 3I). Taken together, our results indicate that atRA inhibits cell proliferation through dysregulation of the ERBB2/JADE1-mediated CCND1/CDKN1B pathway in HEPM cells.
Inhibition of miR-4680-3p Can Partially Restore the Decreased Cell Proliferation Induced by atRA To examine whether normalization of upregulated miR-4680-3p can restore decreased cell proliferation under atRA treatment conditions, we treated HEPM cells with a miR-4680-3p inhibitor, with or without atRA treatment, and found that a miR-4680-3p inhibitor could partially normalize the reduced cell proliferation (Figures 4A-C). As expected, suppression of ERBB2 and JADE1 was normalized by the miR-4680-3p inhibitor under atRA treatment conditions (Figures 4D,E). In addition, we confirmed that phosphorylation of ERK1/2 and expression of CCND1 and CDKN1B were normalized with miR-4680-3p inhibitor ( Figure 4E). Taken together, our results indicate that atRA inhibits cell proliferation through miR-4680-3p expression in HEPM cells.

DISCUSSION
Excessive intake of atRA, a teratogenic reagent, induces cleft palate in humans and mice. Previous studies indicate that atRA inhibits cell proliferation and induces apoptosis in HEPM cells and mouse embryonic palatal mesenchymal (MEPM) cells (Yu et al., 2005;Dong et al., 2017). Since miRs are postulated to be essential in various biological processes, such as cell proliferation and apoptosis (Chen et al., 2004;Buser and Pohl, 2015;Shirjang et al., 2019;Akkoc and Gozuacik, 2020), we hypothesized that excessive atRA induces expression of miRs, which suppress genes crucial for palate development, leading to decreased cell proliferation. Our previous studies show that overexpression of miR-140-5p, miR-133b, miR-374a-5p, miR-381-3p, and miR-4680-3p suppresses cell proliferation through downregulation of genes related to cleft palate in HEPM cells Suzuki et al., 2019). Among them, we found that atRA specifically induced miR-4680-3p expression, which in turn suppressed expression of ERBB2 and JADE1. While expression of some predicted genes targeted by miR-4680-3p was not changed by miR-4680-3p overexpression in HEPM cells, these genes may be regulated by a combination of other miRs or through feedback loops reported in several cancer cell lines and in C2C12 cells, an immortalized mouse myoblast cell line (Hou et al., 2019;Liu et al., 2019;Quan et al., 2019).
ERBB2 is a member of the ERBB receptor tyrosine kinase family including epidermal growth factor receptor (EGFR) (Yarden and Shilo, 2007). The binding of ligands to receptors induces the homo-or hetero-dimerization of receptors and activates the kinase domain that induces downstream signaling cascades such as MAPK/ERK and PI3K/AKT/mTOR pathways, known to be crucial for cell proliferation, migration, and differentiation (Avraham and Yarden, 2011;Arteaga and Engelman, 2014). Our previous bioinformatic study suggests that ERBB signaling pathway may play a substantial role in palate formation (Yan et al., 2020). In fact, the activation of the ERBB2 pathway upregulates CCND1 expression, decreases CDKN1B stability Yang et al., 2000;Lenferink et al., 2001), and promotes cell proliferation and angiogenesis in tumors (Le et al., 2005). In addition, inhibition of the ERBB2 pathway with a neutralizing antibody or small-molecule inhibitor for ERBB2 normalizes CDKN1B expression, leading to cell cycle arrest in human breast cancer cells (Le et al., 2003). atRA inhibits cell proliferation through downregulation of CCND1 and upregulation of CDKN1B in HEPM cells (Yu et al., 2005;Dong et al., 2017). Our results show that atRA inhibits cell proliferation through upregulated miR-4680-3p expression, which suppresses ERBB2 expression and its downstream ERK1/2 signaling pathway in HEPM cells. Taken together, our findings for the regulation of miRs by atRA shed light on the link between environmental and genetic factors, and explain how cell proliferation is inhibited by atRA.
JADE1 (a.k.a. PHF17), a transcription factor, contains two variants: JADE1-L (a long form with 842 amino acids) and JADE1-S (a short form without a C-terminal fragment of 333 amino acids) (Borgal et al., 2014;Siriwardana et al., 2015). Although the role of JADE1 remains elusive, the protein exhibits histone acetyltransferase (HAT) activity and acts as a co-factor of the HBO1 complex in histone H4 acetylation during gene regulation, which plays crucial roles in cell cycle regulation (Panchenko et al., 2004;Foy et al., 2008;Panchenko, 2016;Han et al., 2018). The suppression of JADE1 (both JADE1-L and JADE1-S) by siRNA knockdown results in suppression of DNA synthesis in cultured epithelial cell lines and primary fibroblasts (Havasi et al., 2013). Our results indicate that atRA treatment downregulates JADE1 expression, leading to decreased cell proliferation in HEPM cells through downregulated CDKN1 expression and upregulated CDKN1B expression. In future studies, we will characterize JADE1, namely whether and how JADE1 expression is regulated through the ERK1/2 pathway, and whether and how JADE1 regulates expression of CCND1 and CDKN1B. With those data in hand, we will then be able to conclude that JADE1 expression is directly regulated by miR-4680-3p, or regulated through ERBB2-mediated ERK1/2 signaling. In case treatment with both miR-4680-3p mimic and ERK1/2 inhibitor induces additional inhibition of cell proliferation, as well as additional suppression of CCND1 expression, JADE1 expression may be regulated by miR-4680-3p through the ERK1/2 pathway. (E) Cell proliferation assays in HEPM cells treated with an ERBB2 or JADE1 siRNA for 24, 48, or 72 h. **p < 0.01, ***p < 0.001. Each treatment group was compared with a control siRNA group at each indicated day. (F) Immunoblotting of CCND1, CDKN1B, and GAPDH in HEPM cells treated with siRNA for ERBB2 or JADE1 for 48 h. Representative images from two independent experiments are shown. (G) ERBB2 expression following overexpression in HEPM cells. **p < 0.01. (H) JADE1 expression following overexpression in HEPM cells. ***p < 0.001. (I) Cell proliferation assays in HEPM cells overexpressing ERBB2 and/or JADE1 for 24, 48, or 72 h. **p < 0.01, ***p < 0.001. Each treatment group was compared with control vector (mock) group at each indicated day. ## p < 0.01 vs. mock + atRA at day 3. FIGURE 4 | were counterstained with DAPI (blue). (C) Graph shows the quantification of BrdU-positive cells. **p < 0.01 vs. control inhibitor + vehicle. ## p < 0.01 vs. control inhibitor + atRA. (D) Quantitative RT-PCR for the indicated genes after treatment with atRA, with/without miR-4680-3p inhibitor, for 24 h in HEPM cells. **p < 0.01 vs. control inhibitor + vehicle. ## p < 0.01 vs. control inhibitor + atRA. (E) Immunoblotting of ERBB2, JADE1, phosphorylated ERK1/2 (P-ERK1/2), ERK1/2, CCND1, CDKN1B, and GAPDH in HEPM cells treated with 30 µM atRA, with/without miR-4680-3p inhibitor, for 72 h. Representative images from two independent experiments are shown.
In summary, our findings and those of others suggest that excessive intake of retinoic acid during pregnancy leads to reduced cell proliferation through dysregulation of CCND1 and CDKN1B mediated by the miR-4680-3p-ERBB2-ERK1/2-CCND1/CDKN1B cascade in palatal mesenchymal cells, ultimately leading to cleft palate.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding author/s.