p73 Is Required for Multiciliogenesis and Regulates the Foxj1-Associated Gene Network

Abstract

We report that p73 is expressed in multiciliated cells (MCCs), is required for MCC differentiation, and directly regulates transcriptional modulators of multiciliogenesis. Loss of ciliary biogenesis provides a unifying mechanism for many phenotypes observed in p73 knockout mice including hydrocephalus; hippocampal dysgenesis; sterility; and chronic inflammation/infection of lung, middle ear, and sinus. Through p73 and p63 ChIP-seq using murine tracheal cells, we identified over 100 putative p73 target genes that regulate MCC differentiation and homeostasis. We validated Foxj1, a transcriptional regulator of multiciliogenesis, and many other cilia-associated genes as direct target genes of p73 and p63. We show p73 and p63 are co-expressed in a subset of basal cells and suggest that p73 marks these cells for MCC differentiation. In summary, p73 is essential for MCC differentiation, functions as a critical regulator of a transcriptome required for MCC differentiation, and, like p63, has an essential role in development of tissues.

Figure 1. Global Ablation of p73 Eliminates MCCs

(A) Representative H&E and IF micrographs of the indicated tissues from p73+/+ and −/− mice. Arrowheads mark cilia and flagella (testis). IF was performed using antibodies recognizing the cilia marker *α-tubulin (green) and p73 (red). (B) Representative micrographs of *α-tubulin in embryonic day 18.5 tracheas and lungs from p53−/−, p63−/− and p73−/− animals. (scale bars = 25 µm). See also Figure S1.

Figure 2. p73−/− Mice Exhibit Severe Airway Phenotypes Including Hyperplasia and Epithelial Loss

(A) Representative H&E images of the terminal airways in 18 month old p73+/+ and p73−/− mice. The p73−/− mice exhibit areas of epithelial loss (arrow) and small nodules of hyperplastic epithelium (+). (B) Immunohistochemistry (IHC) staining of pan-cytokeratin and α-SMA of above mice. Arrows in panel B indicate areas of epithelial loss as well as hypertrophy and hyperplasia of the smooth muscle (scale bar= 50 µm). See also Figure S2.

Figure 3. Loss of p73 and p63 in the Respiratory Epithelium Leads to Significant Changes in Cellular Composition

Respiratory epithelium from murine tracheas and bronchioles with the indicated genotypes were analyzed for cell type as percent of total DAPI nuclear marker. Due to the perinatal lethality of p63−/− mice, E18.5 mice were assessed for this genotype. Both E18.5 and post-pubertal mice were assessed for p73 genotypes. Manual quantification was performed on micrographs from at least eight animals with four fields of view per animal. Representative images can be found in Figure S3. p-values are annotated in the table under the bar graphs in panels A and B as <0.0001 (•••), <0.001 (••), and <0.01 (•). See also Figure S3 and Table S1.

Figure 4. p73 is Expressed in a Subset of Basal Cells in the Murine Trachea as well as in MCCs

Representative IF staining of murine tracheas and bronchioles. (A) p73 (red) is 100% co-localized with Foxj (green) in MCCs in the bronchioles (upper panel). p73 (red) co-localizes with Foxj1 (green) in a subset of cells in the trachea (lower panel). (B) IF staining of p73 (red) and p63 (green) shows co-localization of p63 and p73 in a subset of basal tracheal epithelial cells (scale bars = 25 µm). Tables provide the percentage of cells that have single or dual expression of p73, p63 or Foxj1. Values represent the average of 32 fields of view that were quantified (four views from eight animals). See also Figure S4.

Figure 5. p73 Binds and Regulates Target Genes Found in Cilia-Associated Gene Set

(A) Table showing differentially expressed genes in indicated MTECs after ectopic p73 expression (results shown are relative to control). Cells were infected with a lentivirus containing a TAp73β expression construct or an empty vector control. Cultures were maintained for 5 days in basal growth media and 1 day in differentiation media, followed by RNA harvest and sequencing. Differential gene expression analysis between duplicate p73 overexpression and control samples was performed for each genotype. The number of differentially expressed genes (adjusted p-value <0.1) was quantified for 4 different categories: all protein-coding genes (All Genes), genes with p73 binding sites within 25 kb of their transcriptional start sites (TSS) (p73 Proximal Genes), a Cilia-Associated Gene Set, and the overlap of the previous two categories (p73 Proximal Cilia-Associated) For the three latter categories, the p-value for category-specific enrichment versus all protein-coding genes was calculated using the hypergeometric test. (B) Table listing cilia-associated genes bound and regulated by p73. For each binding site, the q-value significance and distance to the respective TSS are indicated along with notation for whether a corresponding p63 binding site was found at the same location. For each gene, the log2 fold change and adjusted p-value from the differential gene expression analysis performed in A are presented. p73 (**) is included as a reference for expression change and genes are ordered based upon increasing q-value. (C) Integrative Genomics Viewer screenshots for select genes from panel B in which the four tracks show ChIP-seq data normalized to 1× depth of coverage and presented with identical scales. The bottom three tracks represent DNA reads that were obtained after ChIP with the antibodies listed to the left, and the top track is the input sample for comparison. At the bottom of each panel is an annotated exon/intron gene structure displayed on the same scale as the ChIP-seq tracks, and the arrow in the bottom left annotates the gene orientation. See also Figures S5 and S6, as well as Table S2, Table S3, Table S4 and Table S5.

Figure 6. p73 Regulates the Expression of Foxj1

(A) MTECs were infected with a lentivirus containing a TAp73β overexpression construct (Ectopic p73) or an empty vector control (Control) and grown in basal growth media for 3 days, after which IF was performed for the indicated proteins. (B) After growth in basal media, parallel cultures were transferred to differentiation media for 3 days and stained identically. For both panels A and B, DAPI counterstaining (blue) was performed to determine the percentage of cells expressing the indicated proteins. The bottom graph presents quantification of p73- and Foxj1-positive cells averaged from quadruplicate experiments with six fields of view per condition (•• represents p-value <0.001, ••• represents p-value <0.0001) with error bars representing standard deviation.