The Tip60/Ep400 chromatin remodeling complex impacts basic cellular functions in cranial neural crest-derived tissue during early orofacial development

Kat5-dependent alterations of gene expression in a cranial neural crest cell line

To study the role of Kat5 in cells of the cranial neural crest, we first resorted to mouse O9-1 cells as a suitable and well characterized cellular model.18 In these cells, we used CRISPR/Cas9-dependent genome editing to inactivate the Kat5 gene by targeting position 669 of the coding sequence in exon 8 (Fig. 1a). Position 669 is localized in the first common exon of Kat5 immediately in front of the sequences coding for the MYST type HAT domain (Supplementary Fig. 1a). According to Inference of CRISPR Edits (ICE) analysis, we succeeded in obtaining three independent clones that carried insertions or deletions in all four Kat5 alleles of the tetraploid O9-1 cells (Supplementary Fig. 1b). These are predicted to lead to frameshifts and immediately ensuing premature translational termination (Supplementary Fig. 1b). Quantitative RT-PCR confirmed a drastic reduction of Kat5 transcripts in all clones (Fig. 1b). In agreement, Kat5 protein was not detectable in any of the clones by immunocytochemical staining or by Western blotting of whole cell extracts (Fig. 1c, d; Supplementary Fig. 1c, d).

Fig. 1figure 1

CRISPR/Cas9-dependent Kat5 inactivation in O9-1 cells. a Schematic representation of the Kat5 gene with zoom-in on guide sequence and adjacent regions in the targeted exon 8 (red box). b Kat5 transcript levels in wildtype O9-1 cells (WT) and gene-edited clones Kat5/1 – Kat5/3 as determined by RT-PCR (n = 3). Levels in wildtype cells were set to 1 ± SEM. c, d Immunocytochemical detection of Kat5 protein in wildtype O9-1 cells (WT) and gene-edited clones (c), used for quantification (n = 3) of Kat5 levels (d). Nuclei were counterstained with DAPI. Scale bar: 150 µm. eg Summary of the consequences of Kat5 inactivation in O9-1 cells as determined by RNA sequencing of gene-edited clones (n = 3) and wildtype cells (n = 4). Shown are the number of up-regulated (red) and down-regulated (blue) genes upon Kat5 inactivation (log2-fold ≥ 1, P value ≤ 0.05) (e), a bi-clustering heatmap for visualization of the expression profile of the top 30 differentially expressed genes sorted and plotted by their (absolute) log2 transformed expression values in the samples (f), and a gene set enrichment analysis of preranked genes to identify cellular processes and structures affected by Kat5 inactivation (g). Statistical significance was determined by one-way ANOVA and Dunnett’s post test (***P ≤ 0.001)

Using total RNA from the Kat5-deficient clones and O9-1 controls, we performed RNA sequencing studies. Kat5-deficient clones and controls clustered separately in PCA plots (Supplementary Fig. 1e). In total, we obtained 230 significantly deregulated genes (DEGs), of which 37 were upregulated and 193 were downregulated (absolute value of log2-fold ≥ 1, p value ≤ 0.05) in the Kat5 knockout (ko) clones (Fig. 1e). Heatmaps of the 30 most strongly deregulated genes confirmed concordance among the samples in the respective groups (Fig. 1f). The proportion of up- versus downregulated genes is in accord with the proposed role of Kat5 as a general activator. Gene set enrichment analysis (GSEA) revealed that enriched sets were associated primarily with various aspects of the intermediate metabolism (e.g., small molecule biosynthesis/catabolism, organic hydroxyl compound, monocarboxylic acid and lipid metabolism), protein biosynthesis (e.g., ribosome biogenesis, RNA splicing, RNA nuclear export) and cellular proliferation (e.g., cell cycle, DNA replication, mitosis) (Fig. 1g). KEGG clustering of downregulated genes by process confirmed the strong alterations in gene sets associated to carbohydrate, lipid and amino acid metabolism and revealed links to several signaling pathways (Supplementary Fig. 1f).

Ep400-dependent alterations of gene expression in a cranial neural crest cell line

To see how characteristic these alterations are for the Tip60/Ep400 complex in O9-1 cells, we next performed CRISPR/Cas9-dependent genome editing for the Ep400 gene by targeting position 3201 of the coding sequence in exon 15 (Fig. 2a). Position 3201 is localized at the ATP binding site in the front part of the split ATPase domain of Ep400 (Supplementary Fig. 2a). We obtained one clone with insertions or deletions in all four Ep400 alleles that led to frameshifts and premature translation termination (clone 1 in Supplementary Fig. 2b). A second clone contained a frameshift mutation in three alleles and a deletion of four amino acids in the remaining allele (clone 3 in Supplementary Fig. 2b). The third had a frameshift in one allele and a six amino acid deletion in the other three (clone 2 in Supplementary Fig. 2b). Amino acid deletions were all localized in the ATP binding site and should minimally lead to inactivation of Ep400. Quantitative RT-PCR again pointed to a drastic reduction of Ep400 transcripts in all three clones (Fig. 2b). In addition, immunocytochemical stainings failed to detect Ep400 in all clones arguing for an absence of the protein (Fig. 2c, d). Clustering of samples was again evident from the PCA plot (Supplementary Fig. 2c). In case of Ep400, we obtained a total number of 733 DEGs, 492 of them up- and 281 downregulated (absolute value of log2-fold ≥ 1, P value ≤ 0.05) in the knockout clones (Fig. 2e). Heatmaps of the 30 most strongly deregulated genes again confirmed concordance among the samples in the respective groups (Fig. 2f). A comparable number of up- and downregulated genes has been observed in RNA sequencing analyses of other Ep400-containing and Ep400-deficient samples.19,20,21

Fig. 2figure 2

CRISPR/Cas9-dependent Ep400 inactivation in O9-1 cells. a Schematic representation of the Ep400 gene with zoom-in on guide sequence and adjacent regions in the targeted exon 15 (red box). b Ep400 transcript levels in wildtype O9-1 cells (WT) and gene-edited clones Ep4005/1 – Ep4005/3 as determined by RT-PCR (n = 3). Levels in wildtype cells were set to 1 ± SEM. c, d Immunocytochemical detection of Ep400 protein in wildtype O9-1 cells (WT) and gene-edited clones (c), used for quantification (n = 3) of Ep400 levels (d). Nuclei were counterstained with DAPI. Scale bar: 150 µm. eg Summary of the consequences of Ep400 inactivation in O9-1 cells as determined by RNA sequencing of gene-edited clones (n = 3) and wildtype cells (n = 4). Shown are the number of up-regulated (red) and down-regulated (blue) genes upon Ep400 inactivation (log2-fold ≥ 1, P value ≤ 0.05) (e), a bi-clustering heatmap for visualization of the expression profile of the top 30 differentially expressed genes sorted and plotted by their (absolute) log2 transformed expression values in the samples (f), and a gene set enrichment analysis of preranked genes to identify cellular processes and structures affected by Ep400 inactivation (g). Statistical significance was determined by one-way ANOVA and Dunnett’s post test (**P ≤ 0.01; ***P ≤ 0.001)

Inactivation of the chromatin remodeling subunit of the Tip60/Ep400 complex predictably led to deregulation of genes associated with chromatin/chromosome organization in GSEA (Fig. 2g). In addition, and similar to results on Kat5 ko clones, enriched gene sets in Ep400 ko clones were associated with small molecule metabolism (in particular amino acid metabolism) and aspects of protein biosynthesis (e.g., translation initiation, cotranslational ER targeting of proteins, peptide biosynthesis) according to GSEA. KEGG clustering by process confirmed that downregulation of gene expression in the absence of Ep400 affected amino acid metabolism and protein biosynthesis (via aminoacyl-tRNA biosynthesis) and again pointed to alterations in carbohydrate metabolism (glycolysis/gluconeogenesis) (Supplementary Fig. 2d).

Almost two thirds of the DEGs obtained for Kat5 were also deregulated in Ep400 ko clones. To compare DEG profiles in Kat5 and Ep400 ko clones in an unbiased manner, we used the rank–rank hypergeometric overlap (RRHO) algorithm.22 Our analysis revealed a strong and statistically significant overlap between gene-expression signatures that primarily concerned the downregulated genes in Kat5 and Ep400 ko clones (Fig. 3a). Application of the same algorithm to the terms from GSEA shows a similarly high concordance for the terms with low P value (Fig. 3b). KEGG pathway analysis argued that the common DEGs for Kat5 and Ep400 were involved in aspects of amino acid and carbohydrate metabolism as well as pluripotency/stem cell pathways (Fig. 3c). However, none of the DEGs listed in the latter category was a bona fide pluripotency gene. They rather represented genes associated more generally with pathways involved in precursor cell properties such as proliferation and increased protein synthesis.

Fig. 3figure 3

Common consequences of Kat5 and Ep400 inactivation on gene expression in O9-1 cells. a RRHO analysis for comparison of DEGs in Kat5 and Ep400 ko clones ranked by their degree of differential expression. The heatmap shows a strong and statistically significant overlap in the downregulated genes. b RRHO analysis of terms obtained by GSEA of preranked genes from Kat5 and Ep400 ko clones ranked by p value reveals a strong overlap in terms with low P value. c KEGG pathway analysis for the identification of the main cellular processes (gray dots, red terms) and the key components (blue dots, black terms) affected by both Kat5 and Ep400 inactivation. d Quantitative RT-PCR to validate expression changes of DEGs related to carbohydrate (left) and amino acid (middle) metabolism or protein biosynthesis (right) shared between Kat5 and Ep400 ko clones. Normalized transcript levels for each gene in wildtype O9-1 cells (WT) were set to 1, and levels in ko clones (n = 3) expressed relative to it. e, f Chromatin immunoprecipitation to determine the occupancy of H2A.Z (e) and acetylated H4 (H4ac, f) near the transcriptional start site of the Pck2, Aldh18a1, Asns, Aars and Eif4ebp1 genes in wildtype O9-1 cells, Kat5 ko clone Kat5/3 and Ep400 clone Ep400/2. Fasl and Gm5741 genes served as controls. Amounts in precipitates were normalized to input and are presented as enrichment with histone-specific antibodies over IgG controls. Statistical significance was determined by 2-way ANOVA and Dunett’s post test (e, f) (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001)

A list of top downregulated genes categorized by their association with carbohydrate and amino acid metabolism as well as protein biosynthesis is presented in Supplementary Fig. 2e. For select DEGs, down-regulation in Kat5 and Ep400 ko clones was confirmed by quantitative RT-PCR (Fig. 3d). Interestingly, Pck2, Aldh18a1, Asns, Aars, and Eif4ebp1 as representatives of genes associated with carbohydrate and amino acid metabolism or protein biosynthesis exhibited statistically significant increases in H2A.Z accumulation and statistically significant decreases in H4 acetylation near their transcriptional start sites in Kat5 and Ep400 ko clones according to chromatin immunoprecipitation analysis (Fig. 3e, f). In contrast, no comparable alterations were observed for the transcriptional start regions of the control genes Fasl and Gm5741. These data argue that at least some of the DEGs are direct and shared targets for Kat5 and Ep400 action.

Impact of Kat5 and Ep400 on glycolysis, translation, proliferation and cell death in a cranial neural crest cell line

To evaluate the consequences of the observed expression changes on basic cellular functions in Kat5 ko and Ep400 ko clones, we analyzed several physiological parameters and compared them to the original O9-1 line. As a first step, we looked into the rate of ATP generation using the Seahorse XFe analyzer. In line with the RNA sequencing data and the inferred changes in gene expression, we measured lower ATP generation rates by glycolysis in the absence of Kat5 or Ep400 (Fig. 4a). This was consistent between all three clones with reductions ranging from 36% to 44% among Kat5 ko clones and from 34 to 56% among Ep400 ko clones (Supplementary Fig. 3a, b). In contrast, mitochondrial ATP generation by oxidative phosphorylation was not statistically different between Kat5 and Ep400 ko clones on the one and controls on the other side when averaged over all three clones (Supplementary Fig. 3c). At single clone level, one Kat5 ko clone and two Ep400 ko clones exhibited statistically significant reductions in mitochondrial ATP generation (Supplementary Fig. 3d, e). However, the respective levels of reduction were mild.

Fig. 4figure 4

Altered basic functions in O9-1 cells upon Kat5 and Ep400 inactivation. ad Comparison of O9-1 cells before (WT, black bars) and after inactivation of Kat5 (blue bars) or Ep400 (green bars) regarding rate of glycolytic ATP generation (determined in a Seahorse XFe analyzer as pmol per minute) (a), translation (determined as OPP incorporation in nascent transcripts with levels in WT cells set to 100%) (b), apoptosis (determined as percentage of cleaved caspase 3-positive cells among all cells) (c) and cell increase during 48 h (determined by crystal violet staining with staining at 0 h set to 1) (d). Red line in (b) corresponds to OPP incorporation in the presence of cycloheximide. Values were averaged over all three clones and are expressed as mean ± SEM. eh Percentage of cells undergoing mitosis (e, f) or being positive for phosphohistone H3 (pHH3) (g, h) in 30% confluent cultures as additional parameters for proliferation. Quantifications (e, g) and representative phase contrast images with overlaid fluorescent DAPI signal (violet–blue to white; scale bar: 100 µm; f) or immunofluorescent images (pHH3 in red, DAPI in blue; scale bar: 200 µm; h) are shown. Statistical significance was determined by one-way ANOVA and Dunnett’s post test (*P ≤ 0.05; ***P ≤ 0.001)

To measure protein synthesis, incorporation of the puromycin analog OPP was used as a proxy for the rate of translation. Pretreatment of cells with the protein synthesis inhibitor cycloheximide was used to determine background levels (Fig. 4b, red dotted line) and analysis of control O9-1 cells defined the standard level of 100%. Intriguingly, both Kat5- and Ep400-deficient cells exhibited lower protein synthesis with 74% ± 4.8% of control levels averaged over all Kat5 ko clones and 71% ± 3.2% for Ep400 ko clones (Fig. 4b). All Kat5 ko clones and Ep400 ko clones consistently exhibited significantly decreased rates of translation (Supplementary Fig. 3f, g).

Despite the decreased rate of protein synthesis, cell death was comparable between Kat5 ko and Ep400 ko clones on the one side and control O9-1 cells on the other both when averaged over all three clones (Fig. 4c) or analyzed in each single Kat5 ko and Ep400 ko clone (Supplementary Fig. 3h, i). However, proliferation was significantly lower in the absence of Kat5 or Ep400 when determined over a period of 48 h by crystal violet staining (Fig. 4d). Again, effects were consistent between ko clones with a 32%–38% decrease for Kat5 ko clones and a 43%–52% decrease for Ep400 ko clones (Supplementary Fig. 4a–c). This translated into a strongly reduced percentage of mitotic figures in cultures of Kat5- or Ep400-deficient O9-1 cells (Fig. 4e, f) with little variation among the ko clones for each gene (varying between 3.0% and 3.2% for Kat5 ko clones and between 2.4% and 3.7% for Ep400 ko clones as compared to 7.9% for the control O9-1 cells; Supplementary Fig. 4d, e). Qualitatively similar decreases were also detected in phosphohistone H3 stainings (Fig. 3g, h and Supplementary Fig. 4f, g). In summary, these results confirm that the changes in gene expression upon Kat5 or Ep400 deletion indeed result in decreased glycolysis and glycolytic ATP production, lowered protein biosynthesis and substantially reduced proliferation without detectable increases in cell death in vitro.

Impact of Kat5 and Ep400 on the cranial neural crest and facial development in vivo

In the second set of experiments, we studied whether the Kat5- and Ep400-dependent changes are also relevant in cranial neural crest cells in vivo as they migrate into and form the mesenchyme of the first two pharyngeal arches that later give rise to most of the facial structures.8,23 For this purpose, we crossed mice carrying a floxed Kat5 or Ep400 allele with the Wnt1::Cre transgene and a Rosa26-stopflox-YFP reporter to be able to target and identify neural-crest derived cells in the pharyngeal arches (Fig. 5a). Controls contained only the combination of Wnt1::Cre transgene and a Rosa26-stopflox-YFP reporter. Immunohistochemical staining of control tissue revealed that Kat5 and Ep400 proteins are similarly present and predominantly nuclear in the ectodermally-derived YFP-negative epidermis and the neural crest-derived YFP-positive mesenchyme of the pharyngeal arch (Fig. 5b). This argued for a comparable nuclear co-localization of Kat5 and Ep400 proteins in epidermal and neural crest-derived mesenchymal cells, but also underlined the necessity of conditional deletion to specifically study the role of Kat5 and Ep400 in the neural crest-derived cells.

Fig. 5figure 5

Developmental consequences of Kat5 and Ep400 inactivation in the cranial neural crest. a Schematic representation of the floxed alleles, the Wnt1::Cre used to inactivate Kat5 and Ep400 and the Rosa26-stopflox-YFP allele used to visualize successful inactivation in the cranial neural crest. Exons are depicted as black boxes, loxP sites as open triangles. pA polyadenylation sequence, SA adenovirus splice acceptor, PGK-Neo pgk-neomycin, YFP enhanced yellow fluorescent protein, Wnt1 prom Wnt1 promoter, Cre Cre recombinase, Wnt1 enh Wnt1 enhancer. b Immunohistochemical staining of tissue from the first pharyngeal arch of control embryos at E10.5 with YFP (green) and Kat5 (red, upper row) or Ep400 (red, lower row). DAPI was used to counterstain nuclei. Scale bar: 25 µm. c, d Reflected-light microscopic (RL, upper rows) and YFP-autofluorescent (lower rows) images of whole mount control (ctrl) embryos and age-matched heterozygous as well as homozygous embryos with Kat5 (Kat5Δ/+, Kat5Δ/Δ) and Ep400 (Ep400Δ/+, Ep400Δ/Δ) deletions at E9.5 (c) and E10.5 (d)

Inspection of the whole mount embryos by reflected-light microscopy or autofluorescence revealed no major alterations or malformations at embryonic day (E) 9.5. This concerned the overall appearance, shape or size of the pharyngeal arches and the contribution of YFP-positive cranial neural crest cells in embryos lacking one or both copies of the Kat5 (Kat5Δ/+ and Kat5Δ/Δ) or Ep400 (Ep400Δ/+ and Ep400Δ/Δ) gene (Fig. 5c).

By E10.5, a reduced presence of neural crest cells in and contribution to the first two pharyngeal arches in Kat5Δ/Δ and Ep400Δ/Δ embryos was evident from the smaller size and altered morphology of the pharyngeal arches in whole mounts and the dramatically reduced YFP autofluorescence (Fig. 5d). In heterozygously deleted, i.e., Kat5Δ/+ and Ep400Δ/+ embryos, size and autofluorescent signal of pharyngeal arches remained relatively normal at E10.5.

In line with the impression from whole mount inspections, the total number of YFP-labeled neural crest-derived cells in the pharyngeal arch mesenchyme of Kat5Δ/Δ and Ep400Δ/Δ embryos was still fairly normal at E9.5 arguing against a major defect in early neural crest stem cell generation or migration in the embryos (Fig. 6a). By contrast, the number of cranial neural crest cells was strongly decreased in the first pharyngeal arch of Kat5Δ/Δ and Ep400Δ/Δ embryos at E10.5 (Fig. 6b, c). Cells in the pharyngeal arch of Kat5Δ/Δ and Ep400Δ/Δ embryos also displayed a significantly decreased expression of Pck2, Tpi1, Aldh18a1, Gpt2, Aars and Eif3h as genes involved in carbohydrate and amino acid metabolism or protein synthesis (Fig. 6d). This compares well with our previous observation in Kat5- and Ep400-deficient O9-1 cells (Fig. 3d).

Fig. 6figure 6

Consequences of Kat5 and Ep400 inactivation on gene expression, proliferation and survival of the developing cranial neural crest. a, b Overall numbers of YFP-labelled cells in the first pharyngeal arches of control (ctrl) embryos and age-matched heterozygous as well as homozygous embryos (n = 3 for each genotype) with Kat5 (Kat5Δ/+, checkered blue; Kat5Δ/Δ, blue) and Ep400 (Ep400Δ/+, checkered green; Ep400Δ/Δ, green) deletions at E9.5 (a) and E10.5 (b). Absolute numbers ± SEM correspond to cells per first pharyngeal arch. c DAPI stain (left) and YFP-autofluorescence (right) of transverse sections from control, heterozygous and homozygous embryos with Kat5 or Ep400 deletions at E10.5 showing maxillary (X) and mandibular (B) branch of the first pharyngeal arch. Scale bar: 300 µm. d Quantitative RT-PCR to determine transcript levels of Pck2, Tpi1, Aldh18a1, Gpt2, Aars and Eif3h in pharyngeal arches of ctrl, Kat5Δ/Δ and Ep400Δ/Δ embryos at E10.5. Normalized transcript levels for each gene in ctrl embryos were set to 1, and levels in Kat5Δ/Δ and Ep400Δ/Δ embryos (n = 3) were expressed relative to it. eg Percentages of phosphohistone H3- (pHH3-) positive (e, f) and BrdU-labelled (g) neural crest cells in the first pharyngeal arches of control, hetero- and homozygous embryos (n = 3 for each genotype). Percentages are given as mean value ± SEM. h, i Rate of cleaved caspase 3-positive apoptotic cells in the various genotypes at E9.5 (h) and E10.5 (i). Statistical significance was determined by multiple t-test (d) or one-way ANOVA and Sidak’s post test (ei) (*P ≤ 0.05; **P ≤ 0.01; ***P ≤ 0.001)

Proliferation rates were also substantially reduced in embryos with homozygous neural crest-specific deletion of Kat5 or Ep400 as determined by immunohistochemical staining for phosphohistone H3 (Fig. 6e, f; for representative stainings see Supplementary Fig. 5a–c). Intriguingly, a reduction in phosphohistone H3-positive cells was already visible at E9.5 (Fig. 6e; for representative stainings see Supplementary Fig. 5b). The reduced proliferation rates in Kat5Δ/Δ and Ep400Δ/Δ embryos were also confirmed in BrdU incorporation studies at E10.5. One hour after treatment, only 7%–18% of neural crest cells had incorporated BrdU as compared to 44% in controls (Fig. 6g; Supplementary Fig. 5d).

It is noteworthy that even Kat5Δ/+ and Ep400Δ/+ embryos exhibited reduced numbers of phosphohistone H3-positive cells at E9.5 (Fig. 6e; for representative stainings see Supplementary Fig. 5b). This correlated with a lower number of neural crest-derived cells in the pharyngeal arch mesenchyme at E10.5, although the reduction in cell number was not statistically significant. By E10.5, the reduction of phosphohistone H3-positive cells in Kat5Δ/+ and Ep400Δ/+ embryos had disappeared (Fig. 6f; for representative stainings see Supplementary Fig. 5c). Nevertheless, this argues that even heterozygous loss of Kat5 or Ep400 may have mild effects on neural crest proliferation.

In contrast to O9-1 cells in vitro, neural crest cells with homozygously deleted Kat5 or Ep400 additionally exhibited increased rates of apoptosis as determined by immunohistochemical stainings for cleaved caspase 3 at both E9.5 and E10.5 (Fig. 6h, i; Supplementary Fig. 6a, b). Heterozygous deletion in Kat5Δ/+ and Ep400Δ/+ embryos also led to slightly increased cell death rates that did not reach statistical significance.

As a consequence of reduced proliferation and increased apoptosis, neural crest cell numbers were decreased by 70%–75% in the first pharyngeal arch of Kat5Δ/Δ and Ep400Δ/Δ embryos at E10.5. In Kat5Δ/+ and Ep400Δ/+ embryos, neural crest cell numbers in the pharyngeal arches were only slightly decreased.

When analyzed shortly before birth at E18.5, the early cranial neural crest defect had translated into a near absence of facial structures in embryos with homozygous Kat5 deletion (Fig. 7a). Similar gross facial malformations were also observed in age-matched embryos with homozygous Ep400 deletion (Supplementary Fig. 6c), whereas neither mice with heterozygous Kat5 deletion nor mice with heterozygous Ep400 deletion exhibited any obvious facial malformations upon visual inspection (data not shown).

Fig. 7figure 7

Perinatal phenotype of mice with neural crest-specific Kat5 inactivation. a Reflected-light microscopic (RL, upper row) and YFP-autofluorescent (lower row) images of heads from control (ctrl) embryos and age-matched homozygous embryos with Kat5 deletion (Kat5Δ/Δ) at E18.5. b Alcian blue/alizarin red stainings of cartilaginous and osseous head structures in control and Kat5Δ/Δ embryos at E18.5. c Immunohistochemical staining of the remaining facial structures in Kat5Δ/Δ embryos at E18.5 with antibodies directed against YFP (green), Sp7 (red) and Sox6 (white). Nuclei were counterstained with DAPI (blue). For overview of the structure and orientation of the plane of sectioning, see left panel. d Alcian blue/alizarin red stainings of head (top) and mandibles (bottom) from control and Kat5Δ/+ embryos at E18.5. e Determination of mandibular length in control and Kat5Δ/+ embryos at E18.5 (n = 3). f Visualization of the anterior (left) and middle (right) part of the secondary palate of control and Kat5Δ/+ embryos at E18.5 by DAPI and YFP staining of coronal sections. Arrowheads point to clefts in the secondary palate. Scale bars: 200 µm (c, f). g Quantification of secondary palatal thickness (middle region) in control and Kat5Δ/+ embryos at E18.5 (n = 3). Statistical significance was determined by unpaired t-test (*P ≤ 0.05)

A combined alcian blue/alizarin red staining of Kat5Δ/Δ embryos revealed a drastic reduction of cartilage and bone in the facial region (Fig. 7b). However, some osseous and cartilaginous structures remained. Due to the strong malformations, it was impossible to assign the remnants to any particular facial bone or cartilage. However, immunohistochemistry verified that Sp7-positive bone-forming cells and Sox6-positive chondrocytes were present (Fig. 7c). Each group of cells labelled additionally with the YFP lineage marker, attesting to their neural crest origin (Fig. 7c). We therefore conclude that neither osteogenic nor chondrogenic differentiation is completely blocked.

A closer look at the facial skeleton of Kat5Δ/+ embryos revealed no missing structures. However, the mandible was significantly shortened (Fig. 7d, e). Histochemical analyses with DAPI and YFP revealed the presence of unilateral or bilateral clefts in the anterior most regions of the secondary palate in a subset of the embryos (Fig. 7f, left). From the middle to the most posterior region, the secondary palate was fused, but appeared substantially thinner than its counterpart from age-matched controls (Fig. 7f, right and Fig. 7g). These findings argue that a heterozyogous loss of Kat5 in the cranial neural crest is sufficient to induce mild orofacial clefting.

留言 (0)

沒有登入
gif