Skip to main content
NCBI home page
Journal of Dental Research logoLink to Journal of Dental Research
. 2022 Jun 30;101(11):1398–1407. doi: 10.1177/00220345221105816

Spatiotemporal MicroRNA-Gene Expression Network Related to Orofacial Clefts

F Yan 1, LM Simon 2, A Suzuki 3,4, C Iwaya 3,4, P Jia 1, J Iwata 3,4,5, Z Zhao 1,5,6
PMCID: PMC9516630  PMID: 35774010

Abstract

Craniofacial structures change dynamically in morphology during development through the coordinated regulation of various cellular molecules. However, it remains unclear how these complex mechanisms are regulated in a spatiotemporal manner. Here we applied natural cubic splines to model gene and microRNA (miRNA) expression from embryonic day (E) 10.5 to E14.5 in the proximal and distal regions of the maxillary processes to identify spatiotemporal patterns of gene and miRNA expression, followed by constructing corresponding regulatory networks. Three major groups of differentially expressed genes (DEGs) were identified, including 3,927 temporal, 314 spatial, and 494 spatiotemporal DEGs. Unsupervised clustering further resolved these spatiotemporal DEGs into 8 clusters with distinct expression patterns. Interestingly, we found 2 clusters of differentially expressed miRNAs: 1 had 80 miRNAs monotonically decreasing and the other had 97 increasing across developmental stages. To evaluate the phenotypic relevance of these DEGs during craniofacial development, we integrated data from the CleftGeneDB database and constructed the regulatory networks of genes related to orofacial clefts. Our analysis revealed 2 hub miRNAs, mmu-miR-325-3p and mmu-miR-384-5p, that repressed cleft-related genes Adamts3, Runx2, Fgfr2, Acvr1, and Edn2, while their expression increased over time. On the contrary, 2 hub miRNAs, mmu-miR-218-5p and mmu-miR-338-5p, repressed cleft-related genes Pbx2, Ermp1, Snai1, Tbx2, and Bmi1, while their expression decreased over time. Our experiments indicated that these miRNA mimics significantly inhibited cell proliferation in mouse embryonic palatal mesenchymal (MEPM) cells and O9-1 cells through the regulation of genes associated with cleft palate and validated the role of our regulatory networks in orofacial clefts. To facilitate interactive exploration of these data, we developed a user-friendly web tool to visualize the gene and miRNA expression patterns across developmental stages, as well as the regulatory networks (https://fyan.shinyapps.io/facebase_shiny/). Taken together, our results provide a valuable resource that serves as a reference map for future research in craniofacial development.

Keywords: developmental biology, cleft lip, spatiotemporal analysis, gene expression profiling, gene regulatory networks, nonlinear dynamics

Introduction

Orofacial clefts, including cleft lip with/without cleft palate, midline cleft, oblique facial cleft, and transverse facial cleft, severely influence patients’ quality of life due to a list of complications (Suzuki et al. 2016; Suzuki et al. 2018). Both genetic and environmental factors contribute to the complex etiology of orofacial clefts and dysmorphism (Grubb et al. 2021). Despite advances in human genome research and sequencing technology, the causes of nearly 70% of all birth defects remain unknown (Warner et al. 2014).

The early development of the orofacial region in humans and mice depends on the growth and fusion of several prominences, including the frontonasal process, the lateral and medial nasal processes, and the maxillary process (Bush and Jiang 2012). The maxillary process developed from the first branchial arch grows medially and is attached with the lateral and medial nasal processes, together forming the upper lip with the medial nasal process and pushing the medial nasal process toward the midline. The horizontal merging of the 2 medial nasal processes forms the primary palate, which carries the incisors and canines. Secondary palate development is initiated with vertical elongation of the lateral maxillary processes at embryonic day (E) 11.5 in mice and the sixth week of gestation in humans. The palatal shelves grow vertically along with the side of the tongue by E13.5 in mice. Following downward growth of the tongue and jaw, the palatal shelves elevate horizontally above the dorsal surface of the tongue and attach at the midline of the oral cavity by E14.5 in mice and seventh to ninth week of gestation in humans.

The entire morphogenesis process relies on tightly coordinated regulation of gene expression. Failure in any processes of midfacial morphogenesis results in developmental defects (Suzuki et al. 2016). A major epigenetic regulator of gene expression is microRNA (miRNA), which is short noncoding RNA that suppresses gene expression at the posttranscriptional level. A growing line of evidence strongly suggests that miRNAs play a crucial role in craniofacial development and pathogenesis of orofacial clefts (Yan, Dai, et al. 2020). For example, Dicer, a type III ribonuclease, helps the transformation from pre-miRNAs to mature miRNAs. Loss of Dicer results in absence of mature miRNAs, and the cranial neural crest (CNC) cell–specific Dicer1 knockout (Dicer1 F/F ;Wnt1-Cre) mice exhibited severe midfacial deformities due to decreased cell proliferation and increased apoptosis in the developing craniofacial regions (Nie et al. 2011).

The FaceBase Consortium (https://www.facebase.org) has made publicly available gene and miRNA expression data sets in mice developing maxillary prominences. We have previously compared the adjacent development stages independently and reported developmental stage-specific gene regulatory mechanisms in the maxillary prominences (Yan, Jia, et al. 2020). However, the pairwise comparison obstructs the discovery of continuous expression changes over the entire development, making it unclear how genes are expressed and regulated in a spatiotemporal-specific manner. In this study, we applied the natural cubic spline model to study the expression during the entire developmental time course in the maxillary process. Our analysis facilitates understanding of molecular processes during craniofacial development and uncovers the etiology of craniofacial developmental defects.

Methods

Natural Cubic Spline Model

All data sets were downloaded from the FaceBase portal (https://www.facebase.org). Detailed processing steps can be found in our previous publication (Yan, Jia, et al. 2020). We applied the natural cubic splines to detect various gene expression patterns. The model includes piecewise polynomials with natural boundaries in the end. The R package splines (version: 4.1.2) was used to conduct the analysis. The ns() function was used to generate the basis matrix for piecewise-cubic splines and natural boundaries. The raw count matrix was quantile normalized and then fed into the model for each gene separately. The natural cubic spline model is expressed as follows:

Yi=β0X13+β1X12+β2X1+β3X2+ε

where Yi is the normalized expression of gene i, X1 represents the developmental stages, and X2 represents the region. The term ε is residual. Multiple testing correction was controlled by the Benjamini–Hochberg method (Benjamini and Hochberg 1995). For miRNAs, the RNA sequencing count matrix was normalized using the voom() function in the limma package (version: 3.46.0). We then applied a similar spline model but without the region variable.

Clustering Analysis

The clustering analysis is described in the Appendix.

Gene Set Enrichment Analysis

The gene set enrichment analysis is described in the Appendix.

Curation of miRNA-Gene Regulation Pairs and Network Construction

Details are described in the Appendix.

In Vitro Experiments

Detailed information about in vitro experiments is described in the Appendix.

Results and Discussion

To study genes and miRNAs related to lip and palate development, we analyzed the expression data available at the FaceBase (Samuels et al. 2020), which contains gene expression profiles in both distal and proximal regions of 129S6 mouse maxillary processes (accession ID: FB00000804/GSE67985) and miRNA expression profiles of C57BL/6J mouse maxillary processes (accession ID: FB00000663.01–FB00000666.01) at E10.5, E11.5, E12.5, E13.5, and E14.5 (Fig. 1).

Figure 1.

Figure 1.

Open in a new tab

Experimental design. The gene and microRNA expression data covering 5 developmental stages (E10.5 to E14.5) were collected from the FaceBase portal. Using differential expression analysis, we identified genes and microRNAs with significant spatiotemporal expression patterns. Di, distal; E, embryonic day; N, nasal epithelium; O, oral epithelium; Pr, proximal.

Differential Gene Expression Analysis Reveals Spatiotemporal Expression Patterns

To explore gene expression profiles in an unsupervised manner, we performed principal component analysis (PCA) (Fig. 2A). Technical replicates clustered near each other, indicating high data quality and reproducibility. Moreover, samples were ordered by developmental stage, from E10.5 to E14.5, along the first principal component (PC), which explains the largest variance in the data. These results revealed that variations between developmental stages were greater than the differences between regions.

Figure 2.

Figure 2.

Open in a new tab

Differential gene expression analysis revealed spatiotemporal expression patterns. (A) Principal component analysis (PCA) shows the global gene expression profiles in reduced dimensions. Points represent samples colored by developmental stages. Shapes represent regions. (B) Heatmap shows differential expression of genes (rows) across developmental stages and regions (columns). Bars on the top of the heatmap indicate the developmental stage and region of each sample. Bars on the left indicate the groups and clusters of differentially expressed genes. (C) Boxplots with fitted curves show expression levels (y-axis) across 5 developmental stages (x-axis) for 1 representative gene for each cluster. Colors represent regions. (D) Dot plot shows the normalized enrichment ratio (x-axis) of enriched pathways (y-axis) for each cluster. Shapes represent annotation databases (GO BP, GO CC, GO MF, KEGG). Color represents false discovery rate (FDR), from blue (low) to red (high). Node size represents the number of enriched genes in the pathway. BP, biological process; CC, cellular component; GO, gene ontology; KEGG, Kyoto Encyclopedia of Genes and Genomes; MF, molecular function.

To detect genes with differential expression patterns over developmental stages (E10.5 to E14.5) and regions (distal vs. proximal of the maxillary process), we applied natural cubic splines, which fit gene expression values into piecewise polynomials. Comparing to the pairwise comparison between adjacent time points, natural cubic splines include data across all time points in 1 model, which is expected to decrease the false discovery rate (FDR) and allows for the discovery of continuous expression changes, especially nonlinear expression patterns over the entire development. In the model, the dependent variable (Y) was gene expression. Independent variables (X) included the developmental stage and region where the stage was converted to a cubic term and region was a covariate. By restricting P values and coefficients for each covariate, we identified the following 3 groups of differentially expressed genes (DEGs) (see Methods section for details, Appendix Fig. 1, Appendix Table 1): 3,927 temporal DEGs, 314 spatial DEGs, and 494 spatiotemporal DEGs. For each group of DEGs, we further conducted clustering analysis to resolve expression patterns more granularly. In total, we identified 8 clusters of DEGs with distinct spatiotemporal expression patterns (Fig. 2B). For example, expression of Meox2 (cluster 1, temporal) increased with time and showed no significant difference between distal and proximal regions (Fig. 2C). Bmp4 (cluster 4, spatial) showed higher expression in the distal region, consistent with a previous study (Zhang et al. 2002). The expression of Krt5 (cluster 5, spatiotemporal) showed both temporal and spatial expression patterns in our analysis. It increased with time and was consistently higher in the distal region when compared to the proximal region.

To interpret biological processes and pathways underlying each group of DEGs, we performed gene-set enrichment analysis (Fig. 2D). For temporal DEGs, genes in cluster 1 were enriched in 127 GO terms and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways (FDR < 0.01), including extracellular matrix (ECM) structural constituent (adjusted P < 2.2 × 10−6) and ECM-receptor interaction (adjusted P = 8.96 × 10−8). Genes in cluster 2 were found most enriched in RNA splicing (adjusted P = 1.97 × 10−6) and noncoding RNA (ncRNA) metabolic process (adjusted P = 3.75 × 10−5).

We then investigated the related pathways of spatial DEGs. The genes in cluster 3 were enriched in 70 pathways, most of which are synapse-related pathways. Genes in cluster 4 were related to regulation of epithelial cell differentiation (adjusted P = 2.33 × 10−5) and mesenchymal development (adjusted P = 3.82 × 10−6).

Spatiotemporal genes showed temporal expression patterns that differed between distal and proximal regions. Genes in cluster 5 (spatiotemporal group), whose expression was upregulated at late developmental stages (E12.5 to E14.5) and higher in the distal maxillary process, were found to be associated with odontogenesis (adjusted P = 2.85 × 10−4) and ossification (adjusted P = 2.99 × 10−4). In agreement with these gene expression patterns, odontogenesis starts thickening oral epithelium at E12.5, and intramembranous ossification in the maxilla starts at E14.5. Genes in cluster 6, whose expression was sequentially upregulated during development and higher in the proximal region, were involved in synaptic membrane (adjusted P = 1.41 × 10−12). Genes in cluster 7, whose expression was sequentially downregulated during development and higher in the proximal maxillary process, were enriched in cell fate commitment (adjusted P = 5.35 × 10−6). Genes in cluster 8 (spatiotemporal group) presented higher expression at early developmental stages (E10.5 to E11.5) of distal regions, when the maxillary processes start to grow toward the medial and nasal processes and gradually form the upper lip. In line with maxillary development, cluster 8 genes were enriched in morphogenesis of embryonic epithelium (adjusted P = 2.29 × 10−5). Aberrations in cluster 8 genes may result in lip malformations.

Integration of DEGs with Genes Related to Orofacial Clefts

To assess the association between gene expression and phenotypes relevant for orofacial clefts, we integrated data from CleftGeneDB (https://bioinfo.uth.edu/CleftGeneDB), a database of genes manually curated for orofacial clefts (cleft genes) with experimental evidence (Xu et al. 2021). These genes were curated by craniofacial developmental biologists from the literature and databases such as dbGaP (https://www.ncbi.nlm.nih.gov/gap/), TSEA-DB (https://bioinfo.uth.edu/TSEADB/) (Jia et al. 2019), Mouse Genome Informatics (MGI) (http://www.informatics.jax.org/), and FaceBase Portal (https://www.facebase.org/). Out of 591 cleft-related genes, 203 genes (34.3%) showed significant expression patterns (Appendix Fig. 2), including 109 temporal genes, 30 spatial genes, and 64 spatiotemporal genes. Fisher’s exact test suggested these cleft genes were enriched in distal regions (clusters 4, 5, and 8; Appendix Fig. 3A), which is consistent with biology that distal regions of maxillary processes develop into lip and palate.

We further downloaded the curated single-nucleotide polymorphisms (SNPs) associated with craniofacial malformations from CleftGeneDB. These data contained 670 significant SNPs in 127 human cleft genes that were experimentally validated. We converted mouse cleft DEGs to human orthologs and found that 24 cleft genes were common in humans and mice. These human cleft genes and their variants are summarized in Appendix Table 2. For example, Efnb1, a DEG from cluster 1 (Appendix Fig. 3B), showed an increase in expression over time, and its human ortholog EFNB1 harbored 3 SNPs identified in the individuals with cleft lip with/without cleft palate (Wieland et al. 2005). Similar to a previous publication, these common genes can be used to construct a conserved regulatory network in humans and mice (Li et al., 2020). In addition, mouse models of these genes may suggest similar variants and abnormal processes in human craniofacial malformations.

miRNAs Exhibited 2 Distinct Expression Patterns across Developmental Stages

To explore the miRNA expression data in an unsupervised manner, we followed our analysis strategy from the gene expression data and performed PCA (Fig. 3A). Technical replicates clustered near each other, indicating high data quality and reproducibility. After performing differential expression analysis, we identified 177 miRNAs with significant expression patterns across developmental stages (adjusted P < 0.05) (Fig. 3B). The clustering analysis revealed 2 unique clusters of these significant miRNAs, with 80 miRNAs decreasing and 97 miRNAs increasing in expression from E10.5 to E14.5 (Fig. 3C). For example, mmu-miR-5622-5p exhibited a decreasing expression pattern while mmu-miR-24-2-5p was increasing across developmental stages (Fig. 3D).

Figure 3.

Figure 3.

Open in a new tab

MicroRNAs (miRNAs) exhibited distinct expression patterns across developmental stages. (A) Principal component analysis (PCA) of miRNA expression data. Each point represents a sample colored by developmental stage. (B) Volcano plot shows the developmental stage coefficient (x-axis) and adjusted P value (y-axis) for each miRNA. Blue colors highlight statistically significant miRNAs. Dashed horizontal line represents the adjusted P value cutoff. (C) Heatmap shows differential expression of miRNAs (rows) across developmental stages (columns). Bars on the top of the heatmap indicate the developmental stage of each sample. Bars on the left indicate the groups of differentially expressed miRNAs. (D) Boxplots show expression levels (y-axis) across 5 developmental stages (x-axis) for 1 representative gene for each group.

Regulatory Networks

MicroRNA-regulatory networks have been found important in developmental processes and related diseases (Li et al. 2020). However, they have not been systematically applied to study the spatial or temporal features in the course of craniofacial development. To uncover spatiotemporal miRNA-gene regulations, we constructed regulatory networks by integrating genes and miRNAs (see Methods for details).

The cluster-specific networks were assembled using each group of DEGs and miRNAs. We were specifically interested in the network of clusters 4, 5, and 8 as mouse cleft genes were enriched in these 3 clusters. When constructing the network, we further filtered the miRNA-gene pairs to those with a negative correlation to capture the repressive nature of miRNAs, which reflects the true miRNA regulation in the cellular system (Li et al. 2020; Li et al. 2019).

The cluster 4–specific network contained 491 unique nodes, including 138 genes, 353 miRNAs, and 1,099 edges. Among 138 genes, 24 were mouse cleft genes regulated by 186 miRNAs (Fig. 4A). Based on the definition of hubs in the network (Yu et al. 2017), we pinpointed 6 hub genes, namely, Satb2 (degree = 40), Trps1 (degree = 37), Has2 (degree = 25), Tbx22 (degree = 22), Vcan (degree = 21), and Pigv (degree = 21). For example, SATB2-associated syndrome, an autosomal dominant congenital disorder, was characterized by intellectual disability, cleft palate or high-arched palate, microdontia, and micrognathia (Zarate et al. 2018). Mice with deficiency for Satb2 (Satb2Cre/Cre) exhibited cleft palate, incisor agenesis, macroglossia, and micrognathia due to increased apoptosis. Mutations in TBX22 were associated with cleft lip with or without cleft palate (CL/P) with/without hypodontia and ankyloglossia in humans (Kaewkhampa et al. 2012), and mice with Tbx22 deficiency showed submucous cleft palate with ankyloglossia due to decreased osteogenesis.

Figure 4.

Figure 4.

Open in a new tab

Regulatory networks for mouse cleft-related differentially expressed gene (DEG) clusters (clusters 4, 5, and 8) and their microRNAs. (A) Cluster 4–specific network visualized by Cytoscape. Triangles (green) represent microRNAs and circles (pink) represent genes. The mouse cleft DEGs are highlighted in blue. The node size is proportional to the degree in the network. (B) Cluster 5–specific network. (C) Cluster 8–specific network.

The cluster 5–specific network contained a total of 147 edges and 129 unique nodes, 9 of which were mouse cleft-related genes, including Runx2, Fgfr2, Spry1, Acvr1, Edn2, Pitx2, Irf6, Adamts3, and Dhrs3 (Fig. 4B). Notably, 2 hub miRNAs, mmu-miR-325-3p (degree = 29) and miR-384-5p (degree = 18), were identified in the network that regulated cleft genes. Of note, fibroblast growth factors (FGFs) bind to tyrosine kinase FGF receptors (FGFR1–4) and transduce FGF signaling, which regulates cell proliferation, differentiation, and migration during development. For example, mutations in FGFR2 are associated with syndromic craniosynostosis with/without CL/P such as Apert syndrome, Crouzon syndrome, and nonsyndromic CL/P (Azoury et al. 2017). Mice with epithelial-specific deletion of Fgfr2 (K14-Cre;Fgfr2b F/F ) exhibited cleft of the secondary palate (Rice et al. 2004), and mice with a gain-of-function mutation in Fgfr2c(Fgfr2c C342Y/C342Y ) exhibited cleft of the secondary palate and craniosynostosis (Eswarakumar et al. 2004). Mutations in SPRY1 and SPRY2 (de Araujo et al. 2016) have been found in individuals with nonsyndromic CL/P. Spry2-null mice exhibited cleft of the secondary palate, while Spry1-null mice showed no craniofacial deformities but defects in the kidneys (Basson et al. 2005).

There were 156 nodes and 185 edges in the cluster 8–specific network (Fig. 4C). Mouse cleft genes in this network include Piga, Map3k7, Ermp1, Tbx2, Snai1, Bmi1, Fgf10, Lhx6, Pbx2, Prrx2, and Vegfa. The hub nodes include miR-218-5p (degree = 25), miR-24-3p (degree = 24), miR-338-5p (degree = 18), and miR-10b-5p (degree = 18). Transcription factors Tbx2 and Tbx3 were expressed in palatal mesenchyme during palatogenesis. Tbx2-null mice exhibited cleft in the secondary palate due to a failure in the growth of the palatal shelves (Zirzow et al. 2009). CNC cell-specific deletion of Pbx1/2 (Wnt1-Cre;Pbx1 F/F ;Pbx2 F/F ) and triple heterozygous of Pbx1/2/3 (Pbx1+/–;Pbx2+/–;Pbx3+/–) mice exhibited cleft of the secondary palate (Ferretti et al. 2011). Mice with deletion of both Prrx1 and Prrx2 (Prrx1–/–;Prrx2–/–) exhibited additional phenotypes such as cleft in the mandible or abnormal mandibular morphology, lack of or single lower incisors, and polydactyly (Lu et al. 1999).

User-Friendly Web Tool Enables Interactive Visualization

Analysis of these data required substantial computer code. Therefore, we created a public and user-friendly Shiny web tool to facilitate interactive exploration of these results (https://fyan.shinyapps.io/facebase_shiny). Our web tool has multiple levels of functionality, allowing the user to visualize gene and miRNA expression patterns across developmental stages. The user enters a gene or miRNA symbol and obtains a graphical display of the expression pattern across developmental stages (Appendix Fig. 4A, B).

Of note, our web tool enables visualization of the miRNA-gene regulatory network. Users upload interested gene lists and select desired confidence levels of miRNA-gene pairs to show the network (Appendix Fig. 4C). The gene list can be curated genes from publications related to orofacial cleft phenotype, experimentally identified disease-associated genes, or genes with variants and mutations from genome-wide association studies (GWAS). The miRNA-gene network will facilitate the understanding of underlying regulatory mechanisms of the user-interested phenotype.

Experimental Validation

To evaluate the functional significance of miRNAs identified in our analysis, we conducted cell proliferation assays in primary mouse embryonic palatal mesenchymal (MEPM) cells isolated from E13.5 C57BL/6J mouse embryos, as well as O9-1 cells, an established cranial neural crest cell line. The cells were treated with a mimic of miR-325-3p, miR-384-5p (hub miRNAs in clusters 4 and 5), miR-218-5p, miR-338-5p (hub miRNAs in clusters 4 and 8), or negative control. In primary MEPM cells, we observed that miR-338-5p mimic significantly suppressed cell proliferation while its inhibitor activated cell proliferation in MEPM cells (Fig. 5A), suggesting that miR-338-5p is functionally relevant in palate development. To validate the predicted miRNA-gene networks (Fig. 4), we conducted quantitative reverse-transcription polymerase chain reaction (qRT-PCR). We found that the predicted target genes (Bmpr1a, Satb2, Tbx2, Snai1, and Bmi1 for miR-338-5p) were significantly downregulated with the treatment of a mimic and upregulated with the treatment of an inhibitor for miR-338-5p in MEPM cells, respectively (Fig. 5B). Similar results were found in O9-1 cells (Fig. 5C). In agreement with these changes in cell proliferation, the predicted target genes (Satb2 and Snail1 for miR-325-3p [cluster 4]; Bmpr1a and Satb2 for miR-325-3p [cluster 4]; Ednra and Arid5b for miR-384-5p [cluster 4]; Adamts3, Fgfr2, and Runx2 for miR-325-3p [cluster 5]; Edn2 and Acvr1 for miR-384-5p [cluster 5]; Bmi1 for miR-338-3p [cluster 8]) were downregulated with treatment of each miRNA mimic and upregulated with treatment of each miRNA inhibitor, respectively (Fig. 5D–F). Finally, to test potential contribution of these miRNAs to chemical-induced cleft palate, we analyzed the expression of miR-325-3p, 338-5p, and 384-5p in the developing palates of mice treated with all-trans retinoic acid (atRA), a cleft palate mouse model, and found that these miRNAs were significantly upregulated at E13.5 and E14.5 (Fig. 5G). Taken together, these results suggest that predicted miR-gene networks play a role in cleft palate.

Figure 5.

Figure 5.

Open in a new tab

Experimental validation results. (A) Cell proliferation assays in mouse embryonic palatal mesenchymal cells treated with indicated microRNA (miRNA) mimic (left panel) or inhibitor (right panel). (B) Quantitative reverse-transcription polymerase chain reaction (RT-PCR) for genes predicted as the targets of miR-325-5p in cluster 4 (top) and cluster 8 (bottom). (C) Cell proliferation assays in O9-1 cells treated with the indicated miRNA mimic (left panel) or inhibitor (right panel). (D–F) Quantitative RT-PCR for genes predicted as the targets of each miRNA in cluster 4 (D), cluster 5 (E), and cluster 8 (F). (G) MiRNA expression in the developing palate of all-trans retinoic acid (atRA)–induced cleft palate mice at embryonic day (E) 13.5 and E14.5. Blue bars indicate the mice treated with control vehicle. Red bars indicate the mice treated with atRA at E11.5. n = 6 per group. *P < 0.05. **P < 0.01. ***P < 0.001.

Limitations

There were several limitations of this study. First, only temporal miRNA expression data were available, which did not contain spatial information. Future work will be integrating spatial information of miRNA expression once it becomes available to provide high resolution of gene regulation during maxillary development. Second, part of curated miRNA-gene interaction pairs was computationally predicted, which might result in false interaction pairs in the regulatory networks even though a stringent threshold was applied to decrease the false-positive rate. Last, in vitro experiments with MEPM cells and O9-1 cells validated the functional significance of hub miRNAs identified in our analysis and their repression of orofacial cleft–related gene expression. However, O9-1 is a multipotent cranial neural crest cell line from E8.5, while the regulation in our study was from E10.5 to E14.5, when the neural crest cells have differentiated into craniofacial mesenchymal progenitors. Thus, in vivo experiments using orofacial cleft mouse models will provide authentic validation.

Conclusion

In this study, our analysis revealed spatiotemporal gene expression patterns and constructed regulatory networks of orofacial cleft–associated genes. The patterns form a basis for maxillary process development research. These expression profiles and our web tool provide an important resource for generating potential hypotheses and follow-up investigation of the function and roles of miRNAs through additional experimental validation. The regulatory network allows the integration with user-interested gene lists to understand the underlying complex regulatory mechanisms of the disease. When additional cleft expression data become available, we will expand this approach for high-resolution spatiotemporal gene expression in craniofacial development.

Author Contributions

F. Yan, contributed to data acquisition and analysis, drafted and critically revised the manuscript; L.M. Simon, contributed to conception and data analysis, drafted and critically revised the manuscript; A. Suzuki, contributed to data interpretation, drafted and critically revised the manuscript; C. Iwaya, J. Iwata, contributed to data interpretation and experimental validation, drafted and critically revised the manuscript; P. Jia, contributed to conception and design, critically revised the manuscript; Z. Zhao, contributed to conception and design, drafted and critically revised the manuscript. All authors gave final approval and agree to be accountable for all aspects of the work.

Supplemental Material

sj-pdf-1-jdr-10.1177_00220345221105816 – Supplemental material for Spatiotemporal MicroRNA-Gene Expression Network Related to Orofacial Clefts
Click here for additional data file. (1.2MB, pdf)

Supplemental material, sj-pdf-1-jdr-10.1177_00220345221105816 for Spatiotemporal MicroRNA-Gene Expression Network Related to Orofacial Clefts by F. Yan, L.M. Simon, A. Suzuki, C. Iwaya, P. Jia, J. Iwata and Z. Zhao in Journal of Dental Research

Acknowledgments

We thank Hiroki Yoshioka for the technical assistance; Yinying Wang, Aimin Li, and Haodong Xu for cleft gene resources in the lab; and the FaceBase investigators for data sharing.

Footnotes

A supplemental appendix to this article is available online.

Declaration of Conflicting Interests: The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding: The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: F. Yan is a CPRIT Predoctoral Fellow in the Biomedical Informatics, Genomics, and Translational Cancer Research Training Program (BIG-TCR) funded by the Cancer Prevention & Research Institute of Texas (CPRIT RP210045). We thank the technical support from the Cancer Prevention and Research Institute of Texas (CPRIT RP180734). This study was supported by grants from the National Institute of Dental and Craniofacial Research (R03D E028103 to Z. Zhao and P. Jia and R03DE027393 to Z. Zhao and J. Iwata) and partially supported by grants (R03DE027711 to P. Jia; R03DE028340, R03DE026509, and R03DE026208 to J. Iwata; R01DE030122 to Z. Zhao and J. Iwata; and R01LM012806 to Z. Zhao) and UTHealth School of Dentistry faculty funds to J. Iwata. The funders had no role in the study design, data collection, analysis, decision to publish, or preparation of the manuscript.

Data Availability: All data sets were downloaded from the FaceBase portal (https://www.facebase.org).

References

  1. Azoury SC, Reddy S, Shukla V, Deng C-X. 2017. Fibroblast growth factor receptor 2 (FGFR2) mutation related syndromic craniosynostosis. Int J Biol Sci. 13(12):1479–1488. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Basson MA, Akbulut S, Watson-Johnson J, Simon R, Carroll TJ, Shakya R, Gross I, Martin GR, Lufkin T, McMahon AP, et al. 2005. Sprouty1 is a critical regulator of GDNF/RET-mediated kidney induction. Dev Cell. 8(2):229–239. [DOI] [PubMed] [Google Scholar]
  3. Benjamini Y, Hochberg Y. 1995. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc Series B Stat Methodol. 57(1):289–300. [Google Scholar]
  4. Bush JO, Jiang R. 2012. Palatogenesis: morphogenetic and molecular mechanisms of secondary palate development. Development. 139(2):231–243. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. de Araujo TK, Secolin R, Félix TM, de Souza LT, Fontes MÍB, Monlleó IL, de Souza J, Fett-Conte AC, Ribeiro EM, Xavier AC, et al. 2016. A multicentric association study between 39 genes and nonsyndromic cleft lip and palate in a Brazilian population. J Craniomaxillofac Surg. 44(1):16–20. [DOI] [PubMed] [Google Scholar]
  6. Eswarakumar VP, Horowitz MC, Locklin R, Morriss-Kay GM, Lonai P. 2004. A gain-of-function mutation of Fgfr2c demonstrates the roles of this receptor variant in osteogenesis. Proc Natl Acad Sci U S A. 101(34):12555–12560. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Ferretti E, Li B, Zewdu R, Wells V, Hebert JM, Karner C, Anderson MJ, Williams T, Dixon J, Dixon MJ, et al. 2011. A conserved Pbx-Wnt-p63-Irf6 regulatory module controls face morphogenesis by promoting epithelial apoptosis. Dev Cell. 21(4):627–641. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Grubb M, Golden A, Withers A, Vellone D, Young A, McLachlan K. 2021. Screening approaches for identifying fetal alcohol spectrum disorder in children, adolescents, and adults: a systematic review. Alcohol Clin Exp Res. 45(8):1527–1547. [DOI] [PubMed] [Google Scholar]
  9. Jia P, Dai Y, Hu R, Pei G, Manuel AM, Zhao Z. 2019. TSEA-DB: a trait–tissue association map for human complex traits and diseases. Nucleic Acids Res. 48(D1):D1022–D1030. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Kaewkhampa A, Jotikasthira D, Malaivijitnond S, Kantaputra P. 2012. TBX22 mutation associated with cleft lip/palate, hypodontia, and limb anomaly. Cleft Palate Craniofac J. 49(2):240–244. [DOI] [PubMed] [Google Scholar]
  11. Li A, Jia P, Mallik S, Fei R, Yoshioka H, Suzuki A, Iwata J, Zhao Z. 2020. Critical microRNAs and regulatory motifs in cleft palate identified by a conserved miRNA–TF–gene network approach in humans and mice. Brief Bioinform. 21(4):1465–1478. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Li A, Qin G, Suzuki A, Gajera M, Iwata J, Jia P, Zhao Z. 2019. Network-based identification of critical regulators as putative drivers of human cleft lip. BMC Med Genomics. 12(Suppl 1):16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Lu MF, Cheng HT, Kern MJ, Potter SS, Tran B, Diekwisch TG, Martin JF. 1999. prx-1 functions cooperatively with another paired-related homeobox gene, prx-2, to maintain cell fates within the craniofacial mesenchyme. Development. 126(3):495–504. [DOI] [PubMed] [Google Scholar]
  14. Nie X, Wang Q, Jiao K. 2011. Dicer activity in neural crest cells is essential for craniofacial organogenesis and pharyngeal arch artery morphogenesis. Mech Dev. 128(3–4):200–207. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Rice R, Spencer-Dene B, Connor EC, Gritli-Linde A, McMahon AP, Dickson C, Thesleff I, Rice DPC. 2004. Disruption of Fgf10/Fgfr2b-coordinated epithelial-mesenchymal interactions causes cleft palate. J Clin Invest. 113(12):1692–1700. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Samuels BD, Aho R, Brinkley JF, Bugacov A, Feingold E, Fisher S, Gonzalez-Reiche AS, Hacia JG, Hallgrimsson B, Hansen K, et al. 2020. FaceBase 3: analytical tools and FAIR resources for craniofacial and dental research. Development. 147(18):dev191213. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Suzuki A, Abdallah N, Gajera M, Jun G, Jia P, Zhao Z, Iwata J. 2018. Genes and microRNAs associated with mouse cleft palate: a systematic review and bioinformatics analysis. Mech Dev. 150:21–27. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Suzuki A, Sangani DR, Ansari A, Iwata J. 2016. Molecular mechanisms of midfacial developmental defects. Dev Dyn. 245(3):276–293. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Warner DR, Mukhopadhyay P, Brock G, Webb CL, Michele Pisano M, Greene RM. 2014. MicroRNA expression profiling of the developing murine upper lip. Dev Growth Differ. 56(6):434–447. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Wieland I, Reardon W, Jakubiczka S, Franco B, Kress W, Vincent-Delorme C, Thierry P, Edwards M, König R, Rusu C, et al. 2005. Twenty-six novel EFNB1 mutations in familial and sporadic craniofrontonasal syndrome (CFNS). Hum Mutat. 26(2):113–118. [DOI] [PubMed] [Google Scholar]
  21. Xu H, Yan F, Hu R, Suzuki A, Iwaya C, Jia P, Iwata J, Zhao Z. 2021. CleftGeneDB: a resource for annotating genes associated with cleft lip and cleft palate. Sci Bull Fac Agric Kyushu Univ [epub ahead of print 1 Jan 2021]. doi: 10.1016/j.scib.2021.07.008 [DOI] [PubMed] [Google Scholar]
  22. Yan F, Dai Y, Iwata J, Zhao Z, Jia P. 2020. An integrative, genomic, transcriptomic and network-assisted study to identify genes associated with human cleft lip with or without cleft palate. BMC Med Genomics. 13(Suppl 5):39. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Yan F, Jia P, Yoshioka H, Suzuki A, Iwata J, Zhao Z. 2020. A developmental stage-specific network approach for studying dynamic co-regulation of transcription factors and microRNAs during craniofacial development. Development. 147(24):dev192948. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Yu D, Lim J, Wang X, Liang F, Xiao G. 2017. Enhanced construction of gene regulatory networks using hub gene information. BMC Bioinf. 18(1):186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Zarate YA, Smith-Hicks CL, Greene C, Abbott M-A, Siu VM, Calhoun ARUL, Pandya A, Li C, Sellars EA, Kaylor J, et al. 2018. Natural history and genotype-phenotype correlations in 72 individuals with SATB2-associated syndrome. Am J Med Genet A. 176(4):925–935. [DOI] [PubMed] [Google Scholar]
  26. Zhang Z, Song Y, Zhao X, Zhang X, Fermin C, Chen Y. 2002. Rescue of cleft palate in Msx1-deficient mice by transgenic Bmp4 reveals a network of BMP and Shh signaling in the regulation of mammalian palatogenesis. Development. 129(17): 4135–4146. [DOI] [PubMed] [Google Scholar]
  27. Zirzow S, Lüdtke TH-W, Brons JF, Petry M, Christoffels VM, Kispert A. 2009. Expression and requirement of T-box transcription factors Tbx2 and Tbx3 during secondary palate development in the mouse. Dev Biol. 336(2):145–155. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

sj-pdf-1-jdr-10.1177_00220345221105816 – Supplemental material for Spatiotemporal MicroRNA-Gene Expression Network Related to Orofacial Clefts
Click here for additional data file. (1.2MB, pdf)

Supplemental material, sj-pdf-1-jdr-10.1177_00220345221105816 for Spatiotemporal MicroRNA-Gene Expression Network Related to Orofacial Clefts by F. Yan, L.M. Simon, A. Suzuki, C. Iwaya, P. Jia, J. Iwata and Z. Zhao in Journal of Dental Research

Articles from Journal of Dental Research are provided here courtesy of International and American Associations for Dental Research

RESOURCES