Identification of candidate downstream targets of TGFβ signaling during palate development by genome-wide transcript profiling

Abstract

Nonsyndromic orofacial clefts are common birth defects whose etiology is influenced by complex genetic and environmental factors and gene-environment interactions. Although these risk factors are not yet fully elucidated, it is known that alterations in transforming growth factor-beta (TGFβ) signaling can cause craniofacial abnormalities, including cleft palate, in mammals. To elucidate the downstream targets of TGFβ signaling in palatogenesis, we analyzed the gene expression profiles of Tgfbr2(fl/fl) ;Wnt1-Cre mouse embryos with cleft palate and other craniofacial deformities resulting from the targeted inactivation of the Tgfbr2 gene in their cranial neural crest (CNC) cells. Relative to controls, palatal tissues obtained from Tgfbr2(fl/fl) ;Wnt1-Cre mouse embryos at embryonic day 14.5 (E14.5) of gestation have a robust gene expression signature reflective of known defects in CNC-derived mesenchymal cell proliferation. Groups of differentially expressed genes (DEGs) were involved in diverse cellular processes and components associated with orofacial clefting, including the extracellular matrix, cholesterol metabolism, ciliogenesis, and multiple signaling pathways. A subset of the DEGs are known or suspected to be associated with an increased risk of orofacial clefting in humans and/or genetically engineered mice. Based on bioinformatics analyses, we highlight the functional relationships among differentially expressed transcriptional regulators of palatogenesis as well as transcriptional factors not previously associated with this process. We suggest that gene expression profiling studies of mice with TGFβ signaling defects provide a valuable approach for identifying candidate mechanisms by which this pathway controls cell fate during palatogenesis and its role in the etiology of human craniofacial abnormalities.

Fig. 1

Images of palates from Tgfbr2^fl/fl;Wnt1-Cre and Tgfbr2^fl/fl mouse embryos at E14.5. A: SEM images of the palates of wild-type C57BL/6J mice. Boxed area was dissected out from Tgfbr2^fl/fl (Control) and Tgfbr2^fl/fl;Wnt1-Cre mice for gene expression microarray analysis. Pr, primary palate. B: Hematoxylin and eosin staining of E14.5 Tgfbr2^fl/fl (Control) and Tgfbr2^fl/fl;Wnt1-Cre mice. The boxed area was dissected out from Tgfbr2^fl/fl (Control) and Tgfbr2^fl/fl;Wnt1-Cre mice for microarray analysis.

Fig. 2

Hierarchical clustering analysis. A: Hierarchical clustering was performed on gene expression values of the 419 most highly variant (CV>0.1) transcripts of autosomal genes. B: Hierarchical clustering was performed on expression values of 30 probe sets representing the most variably expressed transcripts (CV>0.1) that reside on sex chromosomes. All dendrograms were generated using Euclidean distance and average linkage metrics. Rows and columns provide expression values from individual transcripts and samples, respectively. The color bar indicates differences (based on standard deviations) in the expression value assigned a given probe set from the mean expression score of the same probe set across all samples.

Fig. 3

Confirmatory quantitative PCR analysis of selected DEGs. Bars represent mean transcript expression counts of selected genes within palatal tissue of three samples each from Tgfbr2^fl/fl E14.5 control mice (gray bars) and Tgfbr2^fl/fl;Wnt1-Cre E14.5 mice (white bars). Error bars represent standard deviations. * P <0.05; ** P <0.01 based on a two-tailed Student’s t-test.

Fig. 4

Differentially expressed genes (DEGs) related to the cell cycle. KEGG analysis identified the “cell cycle” pathway as being enriched for DEGs found in our comparisons of palatal tissue from Tgfbr2^fl/fl;Wnt1-Cre and Tgfbr2^fl/fl mouse embryos at E14.5. DEGs with higher (white text in red boxes) and lower (white text in green boxes) expression in Tgfbr2^fl/fl;Wnt1-Cre embryos are highlighted.

Fig. 5

Molecular interactions among all DEGs. Based on IPA software, the top scoring gene network for all DEGs considered together was involved in the following processes: cell cycle, cellular assembly and organization, DNA replication, recombination, and repair. Red and green nodes respectively represent DEGs with higher (red) and lower (green) expression in palatal tissue from Tgfbr2^fl/fl;Wnt1-Cre relative to Tgfbr2^fl/fl mouse embryos at E14.5. Unshaded nodes are inferred by the IPA software. Solid lines between nodes indicate physical interaction between the connected elements, while dashed lines indicate indirect interaction through additional molecular components.

Fig. 6

Molecular interactions among transcriptional regulators. Using the connect tool in IPA software we identified molecular interactions between (i) transcriptional regulators that were differentially expressed in palatal tissue from Tgfbr2^fl/fl;Wnt1-Cre and Tgfbr2^fl/fl mouse embryos and (ii) TFs whose binding site motifs were enriched in the genomic regions 2-kb upstream and downstream of the transcription start site of all DEGs. Transcriptional regulators with higher and lower expression in Tgfbr2^fl/fl;Wnt1-Cre mouse embryos are colored in red and green, respectively. TFs identified based on enrichment for their binding site motifs are shaded in gray. Solid and dashed lines between nodes indicate direct and indirect interactions, respectively.