MicroRNA-124-3p Plays a Crucial Role in Cleft Palate Induced by Retinoic Acid

Cleft lip with/without cleft palate (CL/P) is one of the most common congenital birth defects, showing the complexity of both genetic and environmental contributions [e.g., maternal exposure to alcohol, cigarette, and retinoic acid (RA)] in humans. Recent studies suggest that epigenetic factors, including microRNAs (miRs), are altered by various environmental factors. In this study, to investigate whether and how miRs are involved in cleft palate (CP) induced by excessive intake of all-trans RA (atRA), we evaluated top 10 candidate miRs, which were selected through our bioinformatic analyses, in mouse embryonic palatal mesenchymal (MEPM) cells as well as in mouse embryos treated with atRA. Among them, overexpression of miR-27a-3p, miR-27b-3p, and miR-124-3p resulted in the significant reduction of cell proliferation in MEPM cells through the downregulation of CP-associated genes. Notably, we found that excessive atRA upregulated the expression of miR-124-3p, but not of miR-27a-3p and miR-27b-3p, in both in vivo and in vitro. Importantly, treatment with a specific inhibitor for miR-124-3p restored decreased cell proliferation through the normalization of target gene expression in atRA-treated MEPM cells and atRA-exposed mouse embryos, resulting in the rescue of CP in mice. Taken together, our results indicate that atRA causes CP through the induction of miR-124-3p in mice.


INTRODUCTION
Cleft lip with/without cleft palate (CL/P) is the second most common congenital birth defect worldwide, with a prevalence of as high as 1 in 700 live births (IPDTOC Working Group, 2011). CL/P impacts on various physiological functions such as swallowing, feeding, speech, and hearing, even after multiple surgical corrections, orthodontic treatment, and speech therapy (Ferguson, 1988); therefore, the quality of life of both patients and their families is strongly diminished (Sischo et al., 2017;Kummer, 2018;De Cuyper et al., 2019). Palate development is regulated through fine-tuned spatiotemporal gene regulatory networks that control the growth, elevation, and fusion of the palatal shelves through cell migration, proliferation, apoptosis, differentiation, and extracellular matrix secretion and arrangement. Dysregulation of each step of palatogenesis results in a failure in normal palate development and causes cleft palate (CP). Our previous studies showed that at least 223 genes [a total of 198 genes related to cleft palate only (CPO), including cleft in the secondary palate, primary palate, soft palate, and submucous CP, and a total of 45 genes related to cleft lip with cleft palate (CLP), with 20 genes duplicated in CPO and CLP adjusted] in mice and 185 genes (a total of 27 genes related to CPO and a total of 177 genes related to CL/P) in humans are associated with CP (Suzuki et al., 2018a(Suzuki et al., ,b, 2019a. Thus, a large number of genes play crucial roles in palate development. The etiology of CL/P is further complicated with interactions between genetic and environmental factors (Murray, 2002;Beaty et al., 2016;Gonseth et al., 2019). As for environmental factors, maternal exposures to smoking and alcohol consumption are known to be a risk for CL/P (Ericson et al., 1979;Munger et al., 1996;Romitti et al., 1999;Garland et al., 2020). In addition, several chemicals are known to be teratogens that cause CL/P [e.g., dexamethasone, dioxins, and heavy metals (Bove et al., 1995;Buser and Pohl, 2015;Suhl et al., 2018;Pi et al., 2019)]. Malnutrition and mutations in genes related to nutritional metabolic/signaling pathways are also associated with CL/P in humans and animal models. For example, retinoic acid (RA), a metabolite of vitamin A, plays important roles in cell fate determination, cellular patterning, and cell differentiation in development (Rhinn and Dolle, 2012;Roberts, 2020). Excessive RA intake causes CPO in mice (Zhang et al., 2003;Kuriyama et al., 2008;Wang et al., 2019). Mice with deletion in Cyp26b1, a key enzyme for RA degradation, cause CPO, micrognathia, truncation of the fore/hind limbs, and ossification defects in calvaria bones (Maclean et al., 2009). In addition, mice with a mutation in retinol dehydrogenase 10, a key enzyme for RA synthesis (Rdh10 m366Asp mice), show reduced RA levels and exhibit midline cleft, syndactyly, and a malformed forebrain (Ashique et al., 2012). Mice with dominant negative mutations in retinoic acid receptor alpha (Rara 403 * ) and mice with deficiency of both RA receptors alpha and gamma (Rara −/− ;Rarg −/− ) exhibit midline cleft (Damm et al., 1993;Lohnes et al., 1994;Mark et al., 1995). Thus, an appropriate amount of RA is crucial for normal embryonic development, with either too much or too low RA levels causing CP in mice. In humans, decreased serum levels of vitamin A and RA binding protein 4 (RBP4), a RA translocator, have been reported in non-syndromic CL/P patients . In addition, excessive vitamin A intake is known to be associated with multiple birth defects (Lammer et al., 1985;Martinez-Frias and Salvador, 1990;Werler et al., 1990); however, the minimum teratogenic dose appears to be well above the level consumed by most women through multivitamin and vitamin A supplements during pregnancy (Mills et al., 1997;Skare et al., 2012).
A number of microRNAs (miRs), which are small non-coding RNAs (21-25 nucleotides) that regulate the expression of target genes at the post-transcriptional level (Hudder and Novak, 2008;Hou et al., 2011), play important roles in a wide array of cellular functions during the development of various tissues, including the upper lip and the palate (Shin et al., 2012;Seelan et al., 2014;Warner et al., 2014;Mukhopadhyay et al., 2019). For instance, loss of a miR-processing enzyme, such as DROSHA and DICER, results in craniofacial developmental defects in mice (Zehir et al., 2010;Nie et al., 2011;Schoen et al., 2017), and polymorphisms in DROSHA are associated with risk of CL/P in humans (Xu et al., 2018). In addition, mice with a deletion of miR-17-92 cluster, which is located on chromosome 14 in mice and chromosome 13 in humans, exhibit either bilateral or unilateral CLP and delayed endochondral ossification, hypoplastic lung, and cardiac ventricular septal defect (Ventura et al., 2008;de Pontual et al., 2011;Wang et al., 2013). By contrast, overexpression of miR-17-92 in mouse palatal mesenchymal cells results in increased cell proliferation . Thus, miRs can regulate cellular functions through the regulation of their target's gene expression. Currently, while the importance of miRs in development is well recognized through loss-of-function studies for miRs, it remains largely unknown which miRs are elevated by environmental factors to suppress genes that are crucial for palate development.
Our previous bioinformatic studies showed that 18 miRs are possibly involved in the regulation of CP-associated genes in mice (Suzuki et al., 2018a). Recent studies show that all-trans RA (atRA) modulates miR expressions in human cancer cell lines . However, it is still unclear which miRs are functionally relevant in palate development and whether expression of these miRs is altered by atRA exposure. In this study, we investigated the mechanism of how miRs contribute to pathogenesis of CPO induced by excessive atRA intake.

Cell Culture
MEPM cells were isolated from the palatal shelves of E13.5 C57BL/6J mice, as previously described (Iwata et al., 2014). The palatal shelves from one embryo were used for each cell culture, and three independent experiments were conducted using cells from different litters. MEPM cells were maintained under Dulbecco's modified Eagle's medium (DMEM; Sigma Aldrich, St. Louis, MO, United States) supplemented with 10% fetal bovine serum (FBS), penicillin/streptomycin (Sigma Aldrich), 2-mercaptoethanol (Gibco, Waltham, MA, United States), and non-essential amino acids (Sigma Aldrich) at 37 • C in a humidified atmosphere with 5% CO 2 . Animals C57BL/6J mice were obtained from The Jackson Laboratory, Bar Harbor, ME, United States. Pregnant female mice were orally administered 40-70 mg/kg atRA (R2625, Sigma-Aldrich) suspended in 10% ethanol and 90% corn oil emulsion at E11.5. Control mice received an equivalent amount of emulsion without atRA (0.1 ml/10 g body weight). For the rescue experiments, 50 mg/kg atRA was orally administered at E11.5, and then the miR-124-3p inhibitor (Integrated DNA Technologies, Coralville, IA, United States) was intraperitoneally injected at 5 mg/kg at E12.5 and E13.5. The protocol was approved by the Animal Welfare Committee (AWC) and the Institutional Animal Care and Use Committee (IACUC) of UTHealth (AWC 19-0079). All mice were maintained at the animal facility of UTHealth.

Cell Proliferation Assay
MEPM cells were plated onto 96-well cell culture plates at a density of 5,000 cells per well and treated with a mimic for negative control (4464061) [mirVana miRNA inhibitor (chemically modified, singlestranded oligonucleotides with patented secondary structure), Thermo Fisher Scientific], using the Lipofectamine RNAiMAX transfection reagent (Thermo Fisher Scientific), according to the manufacturer's protocol (3 pmol of the mimic or the inhibitor and 0.3 µl of the transfection reagent in 100 µl DMEM per well). Cell proliferation was measured using the Cell Counting Kit 8 (Dojindo Molecular Technologies Inc., Gaithersburg, MD, United States) 24, 48, or 72 h after treatments (n = 6 per group). For the atRA exposure experiments, MEPM cells were plated onto 96-well cell culture plates at a density of 5,000 cells per well and treated with 30 µM atRA. After 24, 48, or 72 h of treatment, cell numbers were determined as described above.

Bromodeoxyuridine (BrdU) Incorporation Assay
MEPM cells were plated onto 35-mm dishes at a density of 25,000/dish and treated with 30 µM atRA or control vehicle (dimethyl sulfoxide). After 24 h, the cells were incubated with 100 µg/ml BrdU (Sigma Aldrich) for 1 h. Incorporated BrdU was stained with a rat monoclonal antibody against BrdU (ab6326; Abcam, Cambridge, MA, United States 1:1,000), as previously described (Iwata et al., 2014). A total of six fields, which were randomly selected from three independent experiments, were used for the quantification of BrdU-positive cells.
Immunoblotting MEPM cells were plated onto a 60-mm dish at a density of 50,000 cells per dish and treated with either 30 µM atRA or vehicle for 72 h, or with each miR mimic or control miR for 48 h. The treated cells were lysed with RIPA buffer (Cell Signaling Technology, Danvers, MA, United States) with a protease inhibitor cocktail (Roche, Indianapolis, IN, United States). The cells were harvested and centrifuged at 21,130 × g for 10 min at 4 • C. The supernatant of each sample was collected, and protein concentration was determined using the BCA protein kit (Pierce). Protein samples were applied to Mini-PROTEAN TGX Gels (Bio-Rad, Hercules, CA, United States) and transferred to a polyvinylidene difluoride (PVDF) membrane. A mouse monoclonal antibody against GAPDH (MAB374, Millipore, Burlington, MA, United States, 1:6,000), a rabbit monoclonal antibody against CCND1 (2978, Cell Signaling Technology, 1:1,000), and a rabbit polyclonal antibody against cleaved caspase 3 (9661, Cell Signaling Technology, 1:1,000) were used. Peroxidase-conjugated anti-mouse IgG (7076, Cell Signaling Technology, 1:100,000) and anti-rabbit IgG (7074, Cell Signaling Technology, 1:100,000) were used as secondary antibodies.
Quantitative RT-PCR MEPM cells were plated onto a 60-mm dish at a density of 40,000 cells per dish. When the cells reached 80% confluence, they were treated with a mimic or an inhibitor for miR-27a-3p, miR-27b-3p, miR-124-3p, or a negative control, at 3 pmol in 6 µl of transfection reagent (Lipofectamine RNAiMAX transfection reagent in 4 ml DMEM per dish). After 24 h of treatment, total RNA was extracted with the QIAshredder and miRNeasy Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer's instructions. For the atRA experiments, the cells were plated onto a 60-mm dish at a density of 50,000 cells per dish and treated with 30 µM atRA for 24 h, and total RNA from MEPM cells (n = 6 per group) was isolated as described above. For the animal experiments, palatal shelves were microdissected at E13.5 and E14.5. Total RNA (1 µg) from each sample was reverse-transcribed using iScript Reverse Transcription Supermix for qRT-PCR (Bio-Rad), and the cDNA was amplified with iTaq Universal SYBR Green Supermix (Bio-Rad) using the CFX96 Touch Real-Time PCR Detection system (Bio-Rad). The PCR primers used in this study are listed in Supplementary Table 1. The amount of each mRNA was normalized by Gapdh. miR expression was measured with Taqman Fast Advanced Master Mix and Taqman Advanced miR cDNA Synthesis Kit (Thermo Fisher Scientific), according to the manufacturer's instructions. Probes for miR-27a-3p (mmu478384_mir), miR-27b-3p (mmu478270_mir), miR-124-3p (mmu480901_mir), and miR-26a-5p (477995_mir) were purchased from Thermo Fisher Scientific.

Craniofacial Tissue Explant Culture
Timed-pregnant mice were euthanized at E13.5 and decapitated in PBS. The mandible and tongue were removed from the embryos, and each explant, including the upper half of the head, was placed in a glass tube containing BGJb medium (Gibco, 12591) supplemented with 50% fetal bovine serum, 0.1% ascorbic acid, and antibiotics. The tubes were placed in a rotary apparatus rotating at 50 rpm in an incubator at 37 • C and 5% CO 2 . After 3 days in culture with/without 30 µM atRA, the explants were fixed in 4% PFA and processed.

Flow Cytometry
MEPM cells were plated onto 60-mm dishes at a density of 50,000/dish and treated with 30 µM atRA or control vehicle (dimethyl sulfoxide). After 24 h, the cells were harvested by trypsin and washed twice with cold BioLegend's Cell Staining Buffer (BioLegend, San Diego, CA, United States; 420201). Cells were centrifuged at 500 g for 5 min at 4 • C. The cell pellets were resuspended with Annexin V Binding Buffer (BioLegend; 422201) at a concentration of 1.5 × 10 6 cells/ml. Resuspended cells (100 µl) were transferred into a Falcon tube and incubated with 5 µl of Annexin V (BioLegend: 421301) and 10 µl of propidium iodide (BioLegend: 421301) for 15 min at room temperature in the dark. A volume of 400 µl of Annexin V Binding Buffer was added to the Falcon tube, and the samples were analyzed with FACS Aria II (BD Biosciences, San Jose, CA, United States).

Statistical Analysis
All experiments were performed independently three times. All statistical analyses were performed using the SPSS software (version 26.0, IBM, Armonk, NY, United States). The statistical significance of the differences between two groups (control and treated groups) was evaluated using independent t-tests. The statistical significance for multiple two groups was evaluated using multiple t-tests after Bonferroni correction. An adjusted p-value after Bonferroni correction (equivalent to non-adjusted p < 0.05) was considered to be statistically significant. For a comparison among multiple groups (e.g., control, treated, and rescued groups) with one factor such as gene or positive cell, a one-way analysis of variance (ANOVA) with Tukey's honest significant difference test was used for assessment. For a comparison among multiple groups (e.g., control, treated, and rescued groups) with multiple factors, a two-way ANOVA with Tukey's honest significant difference test was used for assessment.
Cell proliferation assays were analyzed using a two-way ANOVA with Dunnett's (vs. control) or Tukey's (between all groups) honest significant difference test. A p < 0.05 was considered to be statistically significant. Data are represented as mean ± standard deviation in the graphs.
Frontiers in Cell and Developmental Biology | www.frontiersin.org atRA Induces miR-124-3p Expression in MEPM Cells We conducted cell proliferation assays using MEPM cells treated with atRA. To determine the atRA concentration that affects cell proliferation, the cells were treated with 10 and 30 µM atRA. We found that 30 µM atRA significantly inhibited cell proliferation in MEPM cells (Figure 3A and Supplementary Figure 2A). The reduced cell proliferation by atRA was confirmed with BrdU incorporation assays ( Figure 3B) and expression of Cyclin D1 (CCND1), a cell cycle accelerator, in MEPM cells treated with atRA ( Figure 3C). In addition, cleaved caspase 3, an indicator of apoptosis, was undetectable in cells treated and untreated with atRA (Supplementary Figure 2B). Moreover, flow cytometry analysis showed no significant change in the profile (cell death vs. healthy cells) of cells treated with atRA (Supplementary Figure 2C). Taken together, these data indicate that atRA inhibits the proliferation of MEPM cells. Interestingly, the expression of miR-124-3p was specifically induced by atRA treatment ( Figure 3D); by contrast, expression of miR-27a-3p and miR-27b-3p was not altered by atRA treatment. As expected, the miR-124-3p target genes (Axin1, Fst, Vcan, and Zeb1) were significantly downregulated with atRA treatment (Figure 3E). Taken together, these observations suggest that the expression of these genes was downregulated by atRA through miR-124-3p. We confirmed that the expression of all the other predicted genes regulated by miR-124-3p (additional 39 genes) was not correlated with the atRA condition (Supplementary Figure 3). For instance, although three genes (Bmpr1a, Gli3, and Snal2) were significantly downregulated and Six4 was significantly upregulated under atRA treatment, the expression of these genes was not altered with treatment with the miR-124-3p mimic. Therefore, the use of both a mimic and an inhibitor for the identification of genes regulated by miR-124-3p was helpful to identify genes directly regulated by miR-124-3p.
indicate that atRA inhibits cell proliferation through miR-124-3p expression in MEPM cells.
atRA Induces miR-124-3p Expression in the Developing Palate in Mice The oral administration of atRA to pregnant mice is known to induce CP (Abbott et al., 1989;Ross et al., 2000;Wang et al., 2002;Lai et al., 2003;Yao et al., 2011;Havasi et al., 2013;Hu et al., 2013;Hou et al., 2019;Roberts, 2020). However, the amount and timing of atRA administration that induces CP may differ in each mouse strain and protocol. Therefore, we tested various doses (40, 50, 60, and 70 mg/kg) of oral atRA in C57BL/6J mice. We found that atRA administration at 40 mg/kg induced CP with 57% penetrance (12/21 embryos), while more than 50 mg/kg atRA induced CP with 100% penetrance (Table 1). Therefore, we administered a single dose of 50 mg/kg atRA at E11.5 in this study ( Figure 5A). The palatal shelves of embryos treated with vehicle were normally elevated and fused at E14.5, while the palatal shelves of embryos treated with atRA failed to elevate at E14.5, causing CP at E18.5 ( Figure 5B). To confirm the reduction of cell proliferation in the developing palates of mice treated with atRA, we evaluated cell proliferation by Ki-67 immunostaining. As expected, the number of Ki-67-positive cells (i.e., proliferating cells) was significantly reduced in the palatal shelves of atRA-treated embryos ( Figure 5C). Next, we microdissected the palatal shelves at E13.5 and E14.5 from mice treated with either vehicle or atRA and measured the miR expression. In the vehicle control group, the expression of miR-27a-3p and miR-27b-3p was not changed between E13.5 and E14.5, while the expression of miR-124-3p at E14.5 was upregulated compared to that of E13.5 ( Figure 5D). The expression of miR-124-3p was significantly upregulated with atRA administration at both E13.5 and E14.5 compared to controls, while the expression of miR-27a-3p and miR-27b-3p was comparable to the controls ( Figure 5D). Furthermore, quantitative RT-PCR analysis for genes targeted by miR-124-3p confirmed that a total of three genes (Fst, Vcan, and Zeb1) were significantly downregulated in the palatal shelves of mice given atRA compared to controls (Figure 5E). To confirm that the hypoplastic mandible secondarily caused CP in mice treated with atRA, we cultured E13.5 craniofacial explants, which were extracted from the mandible and tongue, with/without atRA for 3 days (Figure 5F). In controls, the palatal shelves were elevated and almost completely fused. By contrast, the explants treated with atRA showed a widely opened palate. These results indicate that atRA primarily causes CP in mice treated with atRA.

Normalization of miR-124-3p Expression Can Rescue CP Induced by atRA in Mice
Finally, we attempted to rescue CP induced by atRA in mice through treatment with an inhibitor specific for miR-124-3p ( Figure 6A). The administration of a miR-124-3p inhibitor reduced CP penetrance from 100 to 35% in the atRA-induced CPO mouse model (Figures 6B,C). The reduced cell proliferation was partially normalized in mice injected with a miR-124-3p inhibitor under atRA-treated conditions ( Figure 6D). Moreover, the expression of Vcan and Zeb1 was partially normalized with a miR-124-3p inhibitor upon atRA exposure ( Figure 6E). To analyze the expression pattern of extracellular matrix (ECM) formation in mice treated with atRA, we performed Azan staining for the visualization of overall ECM production and pattern. We did not detect any alteration in ECM patterning. Furthermore, to confirm that there was alteration in the patterning, but not in the level, of expression of the molecules, we conducted immunohistochemical analyses for VCAN and CDC42 in mice treated with atRA and confirmed that there was no change in the expression pattern ( Figure 6F). Taken together, our results indicate that atRA induces CPO through the upregulation of miR-124-3p expression in mice.
In this study, we found that the overexpression of miR-27a-3p, miR-27b-3p, and miR-124-3p could inhibit cell proliferation through the suppression of genes associated with mouse CP. miR-27 is a small somatic-enriched miR family comprising two paralogous members, miR-27a and miR-27b, in vertebrates (Kozomara and Griffiths-Jones, 2014). In zebrafish, miR-27 is highly expressed in the pharyngeal arches, and knockdown of miR-27 results in a complete loss of pharyngeal cartilage due to downregulated cell proliferation (Kara et al., 2017). miR-27a-3p, a member of the miR-23a/27a/24-2 cluster, is significantly upregulated in laryngeal carcinoma (Tian et al., 2014) and plays important roles in cell proliferation and migration in human gastric and breast cancers (Mertens-Talcott et al., 2007;Zhou et al., 2016). miR-27b-3p, a member of the miR-23b/27b/24-1 cluster, promotes cell proliferation and invasion in breast cancers (Jin et al., 2013), while miR-27b-3p inhibits cell growth and tumor progression in neuroblastoma , suggesting that miR-27b-3p functions are dependent on cancer types. In addition, a recent study showed that circulating miR-27b might be a potential biomarker for non-alcoholic fatty liver disease (Tan et al., 2014), and another study showed that miR-27b-3p is associated with CP in non-syndromic CL/P (Wang et al., 2017). miR-124-3p is abundant in the human and mouse brain (Lagos-Quintana et al., 2002;Landgraf et al., 2007). In addition, miR-124-3p acts as a tumor suppressor against breast cancer and hepatocellular carcinoma (Liang et al., 2013;Long et al., 2018). Our previous study showed that the overexpression of miR-124-3p suppresses cell proliferation in primary mouse lip mesenchymal cells (Suzuki et al., 2019b). miR-124-3p regulates cell cycle by targeting integrin subunit beta 1 (ITGB1) in oral squamous cell carcinoma cells (Hunt et al., 2011), suggesting that miR-124-3p expression might remain at lower levels during normal palate development. Therefore, miR-124-3p may be a potential biomarker and a therapeutic target to prevent and diagnose CP.
In this study, we found that Bmi1 and Tgfbr3 in miR-27a-3p, Spry2 in miR-27b-3p, and Axin1, Fst, Vcan, and Zeb1 in miR-124-3p were regulated in MEPM cells through both gain-of-function and loss-of-function assays for miR in gene regulation. Because these genes are involved in various signaling pathways, including Wnt/β-catenin, BMP, TGF-β, and epidermal growth factor (EGF) signaling (Iwata et al., 2011;Reynolds et al., 2020), the overexpression of these miRs might lead to the suppression of multiple CP-associated genes through the dysregulation of these signaling pathways. In addition, we found that atRA specifically induced miR-124-3p, which suppressed the expression of genes associated with CP. While the expression of some predicted genes (Dlx2, Fign, Ski, Spry1, Tgfbr1, and Vcan) targeted by miR-124-3p was not changed by atRA in MEPM cells and mouse palatal shelves, these genes may be regulated by a combination of other miRs or through feedback loops. In addition, since four genes (Axin1, Fst, Vcan, and Zeb1) were downregulated in both miR-124-3p overexpression and atRA treatment conditions, these genes may be closely associated with atRA-induced CP in mice. AXIN1 (a WNT signaling negative regulator), FST (a BMP signaling inhibitor), and VCAN (an extracellular matrix) regulate growth factor signaling, which plays crucial roles in embryonic development; therefore, the fine-tuning regulation of the signaling pathways is important for proper embryogenesis. Homeobox gene ZEB1 regulates the epithelial-mesenchymal transition (EMT) in cancer and embryonic development (Schmalhofer et al., 2009;Gheldof et al., 2012); miR-200 family plays a role in EMT by regulating the ZEB1 expression in the developing palate in mice (Carpinelli et al., 2020).
Our rescue experiments show that a miR-124-3p inhibitor can reduce CP penetrance in the atRA-induced CP mouse model. Thus, our study is useful to identify the causative molecular mechanism of CP. However, there are some limitations in this study. First, it is unclear whether any combination of miR inhibitors can achieve better outcomes. In a future study, we will investigate which miRs are the most significantly associated with atRA-induced CP. Second, the expression of most genes was tested only by quantitative RT-PCR analysis. A future study may include not only the level but also the patterning of gene expression associated with atRA-induced CP. Finally, atRA is known to induce tongue abnormalities (Cong et al., 2014), and this study did not exclude the possibility that atRAinduced CP was secondarily caused by tongue abnormalities. A future study will test whether or not tongue anomalies due to atRA administration cause CP using a tissue explant culture system. Despite these limitations, the understanding of miR dysregulation by atRA will shed light on the link between environmental and genetic factors in CP.

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.

ETHICS STATEMENT
The animal study was reviewed and approved by the Animal Welfare Committee (AWC) and the Institutional Animal Care and Use Committee (IACUC) of UTHealth (AWC19-0079).