Inducible knockout of ΔNp63 alters cell polarity and metabolism during pubertal mammary gland development

Abstract

The ΔNp63 isoform of the p53-family transcription factor Trp63 is a key regulator of mammary epithelial stem cells that is involved in breast cancer development. To investigate the role of ΔNp63 at different stages of normal mammary gland development, we generated a ΔNp63-inducible conditional knockout (cKO) mouse model. We demonstrate that the deletion of ΔNp63 at puberty results in depletion of mammary stem cell-enriched basal cells, reduces expression of E-cadherin and β-catenin, and leads to a closed ductal lumen. RNA-sequencing analysis reveals reduced expression of oxidative phosphorylation (OXPHOS)-associated proteins and desmosomal polarity proteins. Functional assays show reduced numbers of mitochondria in the mammary epithelial cells of ΔNp63 cKO compared to wild-type, supporting the reduced OXPHOS phenotype. These findings identify a novel role for ΔNp63 in cellular metabolism and mammary epithelial cell polarity.

The mammary gland is unique in that its initial major development occurs postnatally, and dynamic changes in development continue due to alterations in the reproductive state and age of the female [1]. The extensive tissue remodeling that occurs during adulthood, pregnancy, and involution suggests the involvement of multipotent stem and lineage-restricted progenitor cells [2].

P63 is a member of the larger p53 family of transcription factors and can be expressed as two isoforms [3]. The longer isoform is referred to as the TA isoform, which exhibits tumor suppressor activity. The other isoform is the ΔN isoform, which is predominantly expressed in the basal epithelial layer of mammary glands, skin, esophagus, and prostate [4-8], and plays key roles in tissue homeostasis. Additionally, ΔNp63 can have oncogenic functions in a number of cancers including breast, prostate, and head squamous cell carcinoma [4,9-11]. In our earlier study, we found that ΔNp63 is critically involved in maintaining the stem-like/basal state of the mammary epithelium in normal development and basal breast cancer via activation of Wnt signaling [4]. However, the lack of an inducible conditional knockout model for ΔNp63 has precluded the analysis of ΔNp63 at different postnatal developmental phases. Here, we filled this gap by generating such a system and inducibly knocking out the ΔN isoform of p63 at puberty, adulthood, and during pregnancy. Our data show an indispensable function for ΔNp63 in tissue homeostasis of the mammary gland during puberty.

Formation of the lumen, a functional unit of mammary gland development, is associated with the polarization that maintains the apical and basolateral integrity of mammary epithelial cells [12,13]. This process is also tightly linked to cell metabolism. For example, recent studies show that polarization of hepatocytes is associated with increased metabolism, including a shift from glycolysis to mitochondrial oxidative phosphorylation. Interestingly, many metabolic pathways are emerging as important regulatory mechanisms for programming stem cell fates [14]. However, it is currently unclear how these metabolic pathways are regulated at different stages of mammary gland development. Moreover, the mechanisms driving polarization of mammary epithelial cells in vivo remain elusive and the connection between cell polarity and cell metabolism during mammary gland development, particularly during the high-energy required stages like puberty and lactation, remains poorly understood. Notably, ΔNp63 regulates the expression of genes involved in metabolism in normal skin homeostasis and squamous cell carcinoma [15]. As we have shown that ΔNp63 is critical for maintenance of tissue homeostasis of the mammary gland [4], it raises the intriguing possibility that ΔNp63-dependent effects on high-energy cellular metabolism during pubertal mammary gland development may contribute to this process.

In this study, we demonstrated that ΔNp63 ablation compromises the functional organization (closed lumen) of the mammary epithelial cells during puberty due to depletion of mammary stem cell (MaSC)-enriched basal cells. Decreased mammary stem cells in ΔNp63 conditional knockout (cKO) mammary gland are associated with reduced expression of polarity proteins such as E-cadherin and β-catenin. Additionally, we found that ΔNp63 deletion exerts no significant effects in either adult or lactation stage, suggesting it is dispensable in later stages of mammary gland development. Finally, RNA-seq data revealed that desmogleins and desmocolins, key cell adhesion molecules [16,17], and multiple oxidative phosphorylation pathways required for cell metabolism are downregulated by loss of ΔNp63 at puberty, suggesting an important function for ΔNp63 in both cell polarity and metabolism. Functional studies demonstrate reduced mitochondrial number in the mammary epithelial cells with loss of ΔNp63, supporting decreased OXPHOS mitochondrial genes in RNA-seq data. Thus, our study reveals novel functions for ΔNp63 in mammary gland development, which may also be important for breast cancer development, where similar processes are involved for cancer progression.

Materials and methods

Animal studies and establishment of inducible ΔNp63-knockout mouse model

Animals were housed and treated following Institutional Animal Care and Use Committee (IACUC) of the University of Pennsylvania. The generation of ΔNp63-floxed mouse has been described previously [18]. We used iK 14-CreER (iK14-Cre) [19,20] mouse to cross with ΔNp 63-floxed mouse to acquire iK14; ΔNp63-floxed mice. We carried out DNA preparation from tail biopsies to perform the genotyping. The genotypes of iK14-Cre recombinase and floxed allele for ΔNp63 were determined by PCR using the following primer pairs Cre: 5′-GAGTGATGAGG TTCGCAAGA-3′ and 5′-CTCCACCAGAGACGGAAA TC-3′, amplifying a fragment of 600 bp and ΔNp63-floxed primers are as follows: 5′-TTCACATTCACACAGA CAGCTCC-3′ and 5′-ACAGTCCTCTGCTTTCAGC-3′, amplifying a fragment of 850 bp for the floxed allele and 750 bp for the endogenous ΔNp63 WT allele. iK14-Cre; ΔNp63^f/f mice were treated once with 1.5 mg.kg⁻¹ tamoxifen (Sigma, St. Louis, MO, USA) dissolved in corn oil (Sigma) by intraperitoneal route.

RNA-sequencing analysis

We isolated primary MECs from WT and ΔNp63 cKO for RNA-seq at 7 days or 35 days postinduction of deletion at puberty. As the functional impact of ΔNp63 deletion was equally prominent at both induction time points, we performed RNA-sequencing analysis with both sets of samples. Hence, we obtained four FASTQ files from WT and ΔNp63 cKO at two-time point inductions at four weeks of age. Both cKO samples were very consistent and showed similarity in changes of gene expression by principal component analysis (PCA) of RNA-seq. Both the WT and cKO were pooled, respectively. 100 000 sorted mammary epithelial cells were used for RNA-seq analysis. Raw sequence files (fastq) for 4 samples were mapped using salmon (https://combine-lab.github.io/salmon/) against the mouse transcripts described in gene code (version M21, built on the mouse genome GRCm38.p6, https://www.gencodegenes.org), with an average of 44.6M total input reads for each sample, and an 79.9% average mapping rate. Transcript counts were summarized to the gene level using tximport (https://bioconductor.org/packages/release/bioc/html/tximport.html), and normalized and tested for differential expression using DESeq2 (https://bioconductor.org/packages/release/bioc/html/DESeq2.html). A 2-factor statistical model was used including genotype (WT/cKO) and puberty time point. Gene Set Enrichment Analysis (GSEA, http://software.broadinstitute.org/gsea/index.jsp) was run for the contrast in preranked mode using the DESeq2 statistic of the ranking metric. Mouse gene symbols were mapped to human gene orthologs using Ensembl’s BioMart (http://www.ensembl.org/biomart/martview/). The statistics with the highest absolute value were chosen, where there were redundant mappings.

Mammary gland whole-mount carmine staining

Dissected fourth and fifth mammary glands were spread on a glass slide. Mammary glands were fixed for 6 h in a 10% formalin solution. The fixed mammary glands were hydrated using descending ethanol gradient and stained in carmine alum solution for overnight. The samples were then dehydrated in ascending grades of ethanol and cleared with xylene overnight for imaging. Carmine alum solution was prepared with 1 g carmine (Millipore Sigma, Burlington, MA, USA) and 2.5 g aluminum potassium sulfate in 400 mL distilled water and boiled until solution cleared. The final volume was adjusted to 500 mL with distilled water after filtering through 0.45-micron filters to remove debris and nondis-solved powder. Whole-mount carmine-stained glands were imaged under a Leica dissection microscope.

Flow Cytometry and cell sorting

Primary MECs were isolated, as previously described [21,22]. Briefly, the fourth and fifth mammary glands were harvested, minced, and digested in DMEM : F12 (1 : 1) medium containing hyaluronidase and collagenase for 50–60 min at 37 °C with vigorous shaking every 15 min. After digestion, a 90-s treatment with trypsin and 4-min treatment with dispase and DNase I at 37 °C were performed. The RBCs were lysed with 0.64% NH₄Cl solution for 3 min. The single-cell suspension was filtered through 40-micron strainers and treated with a combination of anti-mouse CD24-PE (1 : 100, BD Biosciences, Franklin Lakes, NJ, USA), anti-mouse CD29-FITC (1 : 100, Bio-Rad, Hercules, CA, USA), and lineage antibodies (CD45, CD31, and Ter119 at 1 : 100 from BD Biosciences) for 30 min at room temperature in the dark. MitoTracker FITC (Thermo Fisher Scientific, Waltham, MA, USA) was used along with other primary antibodies as per the manufacturer’s instructions. Flow cytometry analysis was performed after secondary antibody treatment using the LSRII Flow Cytometer (BD Biosciences), and data were analyzed using FlowJo software (TreeStar, Inc, Ashland, OR, USA). We analyzed and sorted DAPI-negative live cells and lineage negative (Lin⁻) mammary epithelial cells. Mammary basal cells were gated as basal (P4, lin⁻CD24⁺CD29^hi) and luminal cells (P5, lin⁻CD24⁺CD29^low). Sorted mammary epithelial cells (MECs) were used for RNA extraction and RNA-sequencing analyses.

Protein extraction and western blot analysis

We used dissociated mammary gland single cells to make protein extracts using RIPA buffer as previously described [4,21,23]. We used the β-actin antibody (1 : 10 000) to control equal protein loading. We used ΔN63 antibody [4] (1 : 1000) dilution to detect the protein levels of ΔNp63 in mammary gland samples.

H&E, immunohistochemistry, and immunofluorescence

Mammary glands were harvested and fixed in formalin for 4 h at room temperature before embedding in paraffin blocks. Antigen retrieval was performed by boiling slides in sodium citrate solution (10 mM sodium citrate, 0.05% Tween-20, pH 6.0) for 20 min in a microwave oven. For IHC, we add two additional blockings for streptavidin/biotin and H₂O₂ using a Streptavidin/Biotin Blocking Kit (Vector Labs, Burlingame, CA, USA) and 3% H₂O₂. We used 20% goat serum to block nonspecific binding of antibodies. Mouse on Mouse (M.O.M.) block and diluent were used for mouse/rat primary antibodies. The primary antibody dilutions were as follows: rabbit ant-ΔNp63 (1:60 for IHC and 1:20 for IF), rat anti-K8 (1:25 for IF, Cat-TROMA-1s, DSHB), rabbit anti-K14 (1:200 for IF, Cat-PRB155P, BioLegend, San Diego, CA, USA), mouse anti-Ki67 (1:50) for IHC, Cat-Ki67-MM1-L-CE, Leica Microsystems, Wetzlar, Germany), mouse anti-β-catenin (1:100 for IF, Cat-QA21316, Life Tech Corp, Waltham, MA, USA), and mouse anti-E-cadherin (1:100 for IF, Cat-610181, BD Biosciences). We used appropriate secondary antibodies for IHC and IF. Intensity was measured on a scale of 0–3, and abundance was measured on a scale of 01-100. For IHC, the DAB peroxidase (HRP) substrate kit (Vector Labs) was used to detect the signals. The sections were then counter-stained with hematoxylin, dehydrated, and mounted using Permount (Fisher Scientific, Pittsburgh, PA, USA). For IF, the sections were mounted with DAPI containing mounting media (Vector Labs).

Mitochondrial mass was evaluated with 200 nM MitoTracker Green (MTG) (Cat-M7514; ex488, em525, Thermo Fisher Scientific). Cells were seeded at a density of 4 × 10⁴ cells per well of a 4-well chamber slide, and 48 h later, the cells were incubated with MTG for 30 min at 37 °C in the dark, after which time cells were collected and washed in phosphate-buffered saline (PBS). Cells were then suspended in complete medium. All IF and IHC images were taken using a Nikon TiE microscope. Scoring was done by examining multiple fields of view (FOV) per sample. N = 3 sample was analyzed for most studies.

qRT-PCR analysis

Total RNA was isolated from whole mammary glands or sorted cells using the Thermo RNA extraction kit and quantified using a NanoDrop (Thermo). Real-time RT-PCR was performed on the Applied Biosystems StepOne Plus PCR machine (Thermo Fisher Scientific) using SYBR Green Power (Life Technologies, Waltham, MA, USA). The gene-specific primer sets were used at a final concentration of 0.2 μm. The primer sequences for mouse ΔNp63, 5′-TGCCCAGACT CAATTTAGTGA-3′ and 5′-GAGGAGCCGTCCTGAAT CTG-3′, mouse Dsg2 5′-GTGGTCTGCTTGGACTTTG GA-3′ and 5′- GGAACGGTTTGCCTTCATTTC-3′, mouse Dsg3 5′-GCCACGGATGCAGATGAAC-3′ and 5′- CGGA CTTCTCCGGTGTTTC-3′, mouse Sod2 5′-CAGACCTGC CTTACGACTATGG-3′ and 5′-CTCGGTGGCGTTGAGA TTGTT-3′, mouse Ndufs4 5′-CCGTCTGTAGAGTTC CATCCA-3′ and 5′-CTGCATGTTATTGCGAGCAGG-3′, mouse Ndufs1 5′-AGGATATGTTCGCACAACTGG-3′ and 5′-TCATGGTAACAGAATCGAGGGA-3′, mouse Atp13a3 5′-GAGTCAGTCACAACAGATGCG-3′ and 5′-TGCA CAAGATACACCTTCATCC-3′, mouse CoxII 5′-CATCT GAAGACGTCCTCCACTCAT-3′ and 5′-TCGGTTTGAT GTTACTGTTGCTTGAT-3′, and mouse Gapdh 5′- TTCCA CTCTTCCACCTTCGATGC-3′, and 5′-GGGTCTGGGA TGGAAATTGTGAGG-3′. All qRT-PCR assays were performed in duplicate in at least three independent experiments using minimum of three different tissue samples.

Statistical analysis

Results were reported as mean ± SEM (standard error of the mean). The significance of differences was calculated using a two-tailed Student’s t-test for all datasets.

Results

Conditional knockout of ΔNp63 shows an indispensable role in ductal morphogenesis at puberty

To determine the stage-specific function of ΔNp63 in mammary gland development, we generated inducible iK14-Cre; ΔNp63^f/f (cKO) mice by mating ΔNp63-floxed mice [18] with inducible K14-Cre (iK14-Cre) mice [20]. To assess the activity of the inducible K14-Cre recombinase in mammary ducts, we mated our iK14-Cre mice with ROSA26-tdTomato mice to generate iK14-CreER;ROSA26-Tomato mice. iK14-Cre was induced by 1.5 mg.kg⁻1 tamoxifen treatment following an established protocol (Fig. S1A, [2]). This concentration of tamoxifen has been shown to have no deleterious effect on mammary gland development. Flow cytometric data of mammary epithelial cells reveal expression of Tomato only with induction and further demonstrated that Tomato⁺ cells were present in both basal (lin⁻CD24⁺CD29^hi) and luminal (lin⁻CD24⁺CD29^low) cell population (Fig. S1B-D) indicating that iK14-Cre recombinase is active in both basal and luminal cells. Consistent with flow cytometry analysis data, immunofluorescence images show expression of both basal (K14) and luminal (K8) markers significantly colocalized with Tomato in both basal and luminal layers of iK14-CreER; ROSA26-tdTomato mice compared to control (Fig. S1E). Notably, no Tomato expression was observed in the control mice, indicating that iK14-Cre recombinase activity is not leaky (Fig. S1E).

To delineate the functional contribution of ΔNp63 in the pubertal mammary gland, we induced ΔNp63 deletion at 4 weeks and analyzed glands after 7 days in WT and ΔNp63 cKO (Fig. 1A). The deletion of ΔNp63 was confirmed at both the mRNA and protein levels in mammary epithelial cells (Figs 1B-C). As expected, TAp63 was expressed at very low levels in mammary gland epithelial cells, corroborating earlier published data [4] and showed no significant difference between WT and ΔNp63 cKO mammary epithelial cells (data not shown). While there were no gross body changes between genotypes, ΔNp63 cKO mice exhibited some loss of hair compared to WT mice (Fig. 1D). Since ΔNp63 also augments mammary stem cells during normal mammary gland development [4], we next investigated whether conditional knockout of ΔNp63 affected mammary duct morphology. We found a significant reduction in mammary ductal elongation and branching in ΔNp63 cKO mice compared with WT littermates at 7 days postinduction of deletion, corroborating our earlier straight KO data (Fig. 1E and [4]). The altered ductal morphology of the mammary gland was associated with abnormal organization of ductal cells and a closed lumen upon the deletion of ΔNp63 (Fig. 1F and Fig. S2A,B). Furthermore, we observed reduced extracellular matrix (ECM) around the ducts of ΔNp63 cKO mice mammary gland compared to WT (Fig. 1F). Reduced expression of ΔNp63 was further confirmed by IHC (Fig. 1G,H and Fig. S2C). To understand the impact of ΔNp63 deletion on the basal layer, we costained for ΔNp63 and the basal marker (K14) and found a decrease in overall basal cells and ΔNp63⁺ basal cells (Fig. 1I-L).

Fig. 1.

The loss of ΔNp63 at puberty affects ductal morphogenesis. (A) The schematic diagram illustrates the experimental timeline and strategy for the induction of ΔNp63 deletion by tamoxifen treatment at pubertal stage in iK14-Cre; ΔNp63^f/f mice. (B) qRT-PCR data show ΔNp63 mRNA expression in P4 (basal population, lin⁻CD24⁺CD29^hi), P5 (luminal population, lin⁻CD24⁺CD29^low), and P6 (mesenchymal cells) populations of mammary epithelial cells (MECs) of indicated ΔNp63 genotypes. (C) Western blot shows reduced ΔNp63 protein expression in ΔNp63 cKO mammary epithelial cells compared to WT. (D) Representative whole-body mouse images show some loss of hair growth in ΔNp63 cKO compared to WT mice. (E) Representative images of alum-carmine-stained whole-mount mammary outgrowths show reduced ductal growth in ΔNp63 cKO mammary gland compared to WT at indicated ages. Black arrows indicate length of ducts from lymph node. (F) Representative H&E images depict closed ductal lumen in ΔNp63 cKO mammary gland compared to WT mice. (G, H) Representative IHC images show reduced ΔNp63⁺ staining and narrow lumen in ΔNp63 cKO duct compared with WT duct, which is quantified in H. (G) Black arrow indicates ΔNp63⁺ cell in ΔNp63 cKO mammary duct. (I) Representative IF images show reduced ΔNp63 expression in basal cells in ΔNp63 cKO mammary ducts compared to WT. White arrows indicate ΔNp63⁺- and K14⁺-stained cells in WT mammary duct. (J–L) The scatter plots show quantification of ΔNp63⁺ cells, K14⁺ (basal cells), and both positive cells upon costaining of ΔNp63 and K14 as showed in I. Student’s t-test was used to calculate P-values. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, and ***P < 0.001. Scale bars, 2 mm in (E), 40 μm in (F, G), and 20 μm in (I).

The deletion of ΔNp63 in the mammary gland at puberty by a single dose of tamoxifen followed by a long-term follow-up (35 days) showed a similar reduction in ductal morphogenesis and aberrant ductal formation (Fig. S3A-F) as seen at 7 days after induction. The closed lumen was also associated with reduced ΔNp63 expression at this time point (Fig. S3C-F), suggesting that loss of ΔNp63 impacts mammary stem cells, which are long-lived.

Conditional knockout of ΔNp63 shows reduced polarity genes at puberty

Immunostaining with the basal/myoepithelial marker K14 shows a dramatic reduction in the basal/ myoepithelial layer in ΔNp63 cKO mammary glands compared to WT (Fig. 2A). Basal/myoepithelial cells impact cell polarity by maintaining cross-talk between the ECM and luminal cells in the mammary gland [24]. Lack of polarity is often seen with closed lumen [24]. To assess the mechanistic basis for the abnormal polarity of mammary epithelial cells in the absence of ΔNp63, we next examined the expression of β-catenin, which is normally present in apical and basolateral membranes of polarized mammary epithelial cells. Staining for β-catenin shows reduced β-catenin expression in mammary ducts. (Fig. 2B-D), suggesting altered apical–basal polarity of mammary epithelial cells in ΔNp63 cKO mammary glands compared to WT. Next, we analyzed the effects of ΔNp63 deletion on E-cadherin expression, which also mediates apical and basolateral cell polarity in epithelial cells [16,25]. We found decreased E-cadherin levels in ΔNp63 cKO mammary ducts (Fig. 2E-G), suggesting loss of ΔNp63 reduces E-cadherin mediated cell polarity in the ΔNp63 cKO mammary glands. Finally, mammary gland development at puberty is accompanied by cell proliferation, which drives ductal elongation and bifurcation. We found that the cell proliferation marker Ki67 is significantly reduced in ΔNp63 cKO mammary glands compared to WT (Fig. 2H-I). To evaluate the impact of ΔNp63 deletion on mammary stem cell-enriched basal cells, we analyzed this population by flow cytometry. Flow cytometry analysis showed a significant reduction in the mammary stem cell-enriched basal population (Fig. 2J-K) in cKO mice compared to WT mice. Together, these findings suggest that ΔNp63 may be a master regulator of pubertal mammary gland development, regulating both mammary stem cell function and polarity.

Fig. 2.

ΔNp63 is required in maintaining cellular polarity of mammary epithelial cells. (A) Representative IF images show reduced expression of basal marker K14 in ΔNp63 cKO mammary ductal cells compared to WT, confirming depletion of basal layers with the loss of ΔNp63. Insets are magnified to illustrate the expression of K14. (B-D) Representative IF images show reduced β-catenin expression along with basal marker (K14) in ΔNp63 cKO mammary duct, indicating alteration in cell polarity, which is quantified in C and D. White arrows and a bar indicate the fewer β-catenin-expressing cells in ΔNp63 cKO mammary ducts. (E–G) Representative IF images show reduced E-cadherin expression along with basal marker (K14) in ΔNp63 cKO mammary duct, indicating alteration in cellular polarity, which is quantified in F and G. (E) White arrows indicate the fewer K14⁺ basal cells in ΔNp63 cKO mammary ducts. (H) Representative IHC images show reduced number of Ki67⁺ proliferating cells in ΔNp63 cKO mammary ducts compared to WT mammary ducts. Quantification is shown in the I. (J) Flow cytometry analysis profile shows reduced populations of MaSC-enriched basal cells (P4) in Lin⁻ MECs from ΔNp63 cKO mice compared to WT. Lin⁻ epithelial population was obtained following published protocol [22]. P5 population represents luminal population of the mammary gland. (K) Bar graph represents percentage of P4 cells (basal) and P5 cells (luminal) in WT and ΔNp63 cKO MECs after flow cytometry. Student’s t-test was used to calculate p-values. Data are presented as the mean ± SEM. **P < 0.01 and ***p < 0.001. Scale bars, (A) is 40 μm, (B) is 10 μm, and (E and H) is 20 μm.

Conditional knockout of ΔNp63 has dispensable effects on ductal morphogenesis at adulthood

To delineate the functional contribution of ΔNp63 in the adult mammary gland, we induced ΔNp63 deletion at 8 weeks (Fig. 3A) and analyzed glands after 14 days in WT and ΔNp63 cKO mice by whole-mount alum-carmine staining (Fig. 3B). There were no dramatic morphological differences in ΔNp63 cKO compared to WT mammary glands (Fig. 3B-C). Whole-body examination also showed no significant phenotypic changes between virgin ΔNp63 cKO and WT mice (data not shown). Notably, no effects on the size of the ductal lumen or ductal cell organization were observed between mammary glands of adult WT and ΔNp63 cKO mice, as seen by H&E staining (Fig. 3D,E). Consistent with H&E results, IF staining of tissue sections showed no significant change in expression of K14 and K8 markers following the deletion of ΔNp63 (Fig. 3F). Flow cytometry analysis data also revealed the deletion of ΔNp63 has no impact on stem cell-enriched basal or luminal cells in ΔNp63 cKO mice compared to WT mice (Fig. 3G,H). Altogether, our data suggest that loss ΔNp63 is not required for the maintenance of ductal morphogenesis at adulthood.

Fig. 3.

The loss of ΔNp63 at the adult stage shows dispensable function on ductal morphogenesis. (A) The schematic diagram illustrates the experimental timeline and strategy for the induction of ΔNp63 deletion by tamoxifen at the adult stage. (B) Representative alum-carmine-stained whole-mount mammary outgrowths are showing no significant difference between WT and ΔNp63 cKO. (C) Western blot shows reduced ΔNp63 expression in ΔNp63 cKO mammary epithelial cells compared to WT. (D, E) Representative H&E images show no changes in the morphology of ductal lumen between WT and ΔNp63 cKO mammary gland. (F) Representative IF images show the K14⁺ basal and K8⁺ luminal layers in WT and ΔNp63 cKO mammary gland. Insets show magnified zoomed-in images to illustrate no difference in the expression of K14 and K8 in the mammary duct of ΔNp63 cKO mammary gland compared to WT gland. (G) Flow cytometry profile shows different populations of MECs from WT and ΔNp63 cKO mice. No significant difference in any population was observed between the two groups. Numbers within the plots are in percentages. (H) Bar graph represents percentage of P4 (basal) compared to P5 (luminal) cells after flow analysis. Student’s t-test was used to calculate P-values. Data are presented as the mean ± SEM. Scale bars, (B, D) is 2 mm. (E) is 40 μm, and (F) is 20 μm.

Conditional knockout of ΔNp63 exhibits redundant function on alveologenesis

During pregnancy, a transcriptionally induced cascade of complex interactions between stromal and epithelial compartments initiates extensive ductal branching, differentiation, and alveolar proliferation that results in milk production. To determine whether mammary gland development during pregnancy is dependent on ΔNp63, we induced ΔNp63 deletion prior to mating and performed all experiments at lactation day 1 (Fig. 4A). Whole-body images showed no major difference between WT and ΔNp63 cKO mice (Fig. 4B). Conditional deletion of ΔNp63 during pregnancy did not affect the number of surviving litters, suggesting the presence of milk in mammary glands of ΔNp63 cKO mice (Fig. 4C). Whole-mount alum-carmine staining further confirmed no major alterations in alveologenesis in ΔNp63 knockout mammary glands compared to WT mammary glands (Fig. 4D-F), suggesting that ΔNp63 is not required for the maintenance of adult mammary gland during pregnancy. As expected, the mammary stem cell-enriched population dramatically increased during pregnancy compared to virgin stage [26]. Flow cytometry analysis showed no alteration in either basal or luminal populations in ΔNp63 knockout mice compared to WT mice (Fig. 4G,H). Taken together, these results indicate that loss of ΔNp63 during pregnancy has minimal effects on mammary gland development.

Fig. 4.

Loss of ΔNp63 at lactation shows a dispensable function in alveologenesis. (A) The schematic diagram illustrating the experimental timeline and strategy for the induction of ΔNp63 by tamoxifen at lactation. (B) Representative whole-body image shows no effects of ΔNp63 on skin development and hair follicle in ΔNp63 cKO compared to WT mice. (C) The scatterplot represents number of pups in WT and ΔNp63 cKO mice. (D-E) Representative images of alum-carmine-stained whole-mount mammary outgrowths from WT and cKO mice; insets are magnified to illustrate the morphology of ductal alveoli. (F) Representative H&E images show lobules in WT and ΔNp63 cKO mice. (G) Flow cytometry profiles show different populations of MECs from WT and ΔNp63 cKO mice showing no major difference between two groups. Numbers within the plots are in percentages. (H), Bar graph represents percentage of P4 cells (basal) compared to P5 cells (luminal) after flow cytometry. Student’s t-test was used to calculate p-values. Data are presented as the mean ± SEM. Scale bars, 2 mm in (D, E) and 40 μm in (F), respectively.

ΔNp63 regulates oxidative phosphorylation and cell polarity through desmosomal proteins

To understand the functional role of ΔNp63 in the mammary gland at puberty, we sorted mammary epithelial cells from both WT and ΔNp63 cKO mice in which ΔNp63 was deleted at puberty and performed an unbiased RNA-sequencing analysis (Fig. 5A). As expected, RNA-seq analysis showed reduced signatures for several published mammary stem cells, basal cell signatures, and EMT in ΔNp63 cKO cells in comparison with WT cells (Fig. 5B and data not shown). Interestingly, we also found several oxidative phosphorylation pathways to be downregulated in ΔNp63 cKO mammary epithelial cells (Fig. 5C-E). Specifically, genes important for OXPHOS such as cytochrome c oxidase (Cox), NADH ubiquinone oxidoreductase Fe-S protein (Nduf), ATPase type 13A3 (Atp13A3), and superoxide dismutase 2 (Sod2) were significantly downregulated in ΔNp63 cKO mammary epithelial cells compared to WT (Fig. 5F and RNA-sequencing data). To test the functional impact of ΔNp63 deletion on oxidative phosphorylation, we used the MitoTracker probe, which stains active mitochondria. Loss of ΔNp63 significantly decreased the MitoTracker staining in mammary epithelial cells (Fig. 5G,H), suggesting a direct role for ΔNp63 in regulating oxidative phosphorylation by increasing mitochondria mass. Collectively, our findings showed for the first time a plausible connection between ΔNp63 and the increased cell metabolism required for the active proliferation and migration of cells during pubertal mammary gland development.

Fig. 5.

Loss of ΔNp63 is associated with reduced OXPHOS and polarity-related desmosomal proteins. (A) A schematic diagram shows strategy to sort mammary epithelial cells from both WT and ΔNp63 cKO pubertal mice followed by high-throughput RNA-sequencing. (B-E) Gene set enrichment analyses (GSEA) demonstrate increased mammary stem cells (B) and oxidative phosphorylation (C-E) signatures in WT MECs, compared to ΔNp63 cKO MECs. (F) qRT-PCR analysis of oxidative phosphorylation-related mitochondrial genes shows a decrease in mRNA levels of these genes upon the deletion of ΔNp63 compared to WT at puberty. n = 2 samples were used to perform qRT-PCR analysis, each sample has technical duplicates. (G, H) Representative IF images (G) and flow cytometry (H) analysis show fewer uptake of Mito Tracker probe in ΔNp63 cKO MECs compared with WT. (I) List of genes was made using cutoff log2 fold change from RNA-seq data, showing an increase in desmogleins (Dsg3, Dsg1b, Dsg1a, Dsg1c, and Dsg2) and desmocolins (Dsc2) in WT MECs compared to ΔNp63 cKO MECs. (J) qPCR analysis of desmogleins shows a decrease in mRNA levels upon the deletion of ΔNp63 compared with WT at puberty. n = 2 samples were used to perform qPCR analysis, each sample has technical duplicates. (K) The working model shows unique role of ΔNp63 as a master regulator of polarity and cell metabolism, and the deletion of ΔNp63 leads to closed lumen phenotype along with significant reduction of total K14⁺ basal cells. ΔNp63 controls the expression of OXPHOS genes, which are essential to produce high energy for ΔNp63⁺ mammary stem cells. Besides, ΔNp63 upregulates the expression of cell adhesion and polarity-related genes such as β-catenin, E-cadherin, and desmogleins. The deletion of ΔNp63 leads to the decreased expression of cell metabolism and polarity-related genes, and collectively that affects the mammary stem cell function and leads to the closed lumens of mammary ducts. Student’s t-test was used to calculate p-values. Data are presented as the mean ± SEM. *P < 0.05 and **P < 0.01. Scale bar, 40 μm in (G).

Several studies [27-29] also suggest that epithelial cells require a high-energy metabolic output to maintain their polarization through the involvement of several cell adhesion molecules. The presence of a closed lumen is often associated with abnormal polarity of mammary epithelial cells [30]. Desmosomal proteins are important for mammary epithelial cell polarity in vitro [13]. Indeed, our unbiased RNA-seq data showed significant downregulation of several desmosomal proteins including Dsg2, Dsg3 (desmogleins), and Dsc2 (desmocolin2) in ΔNp63 knockout mammary epithelial cells (Fig. 5I,J). As desmosomal proteins are important for polarization of mammary epithelial cells forming lumen formation [24], our findings support the importance of ΔNp63 in both cellular metabolism and maintenance of polarity at the pubertal stage.

Discussion

Transcription factors are essential for mammary gland development [31,32], but the lack of inducible conditional knockouts has hindered analyses exploring the role of these factors at specific developmental stages. Here, we have assessed the role of the ΔN isoform of p63 (ΔNp63) using a tamoxifen-inducible ΔNp63-knockout mouse model and identified a novel role for ΔNp63 in maintaining postnatal growth and the morphological phenotype of the mammary gland specifically at puberty. Our study demonstrates that loss of ΔNp63 at puberty alters cell metabolism and polarity of ductal epithelial cells, leading to reduced ductal morphogenesis and closed lumens. We also found that ΔNp63 is not required in the later stages of development, suggesting that ΔNp63 is dispensable for mammary gland maintenance and development during adulthood and pregnancy, respectively. This observation differs from earlier studies showing a function for p63 in lactation [33], which could be due to different models (straight knockouts of p63 vs. isoform-specific deletion of ΔNp63) and strains of mice used.

Metabolic alterations can play a role in determining cellular phenotypes and interactions [34]. Several oxidative phosphorylation pathways are reduced in ΔNp63 cKO mammary glands compared with wild-type glands at puberty. Furthermore, mitochondria biogenesis is affected by loss of ΔNp63 in mammary epithelial cells. Earlier studies have linked the function of ΔNp63 to cell metabolic pathways such as glycolysis and oxidative phosphorylation in keratinocytes and squamous cell carcinoma (SCC) [15]. However, similar studies have not yet been undertaken in primary mammary epithelial cells. Using our novel model, we identified a potential role for ΔNp63 in regulating oxidative phosphorylation (OXPHOS)-related mitochondrial genes. As stem cells have high-energy demands, it is possible that ΔNp63⁺ mammary stem cells achieve higher energy required for the massive proliferation and migration of the cells at puberty through increased OXPHOS.

Adhesion molecules are required for adhesion between epithelial cells and the induction of cellular polarity [29,35-37]. The precise regulation of cell-cell interactions is an essential feature of the development and function of the mammary gland lumen, which later plays a crucial role in milk production. IF and RNA-sequencing data in this study revealed that several adhesion molecules, including E-cadherin, β-catenin, and the desmosomal proteins Dsg2/3 and Dsc2, were reduced in ΔNp63 cKO mammary epithelial cells. A complex interplay between ECM proteins and epithelial cells maintains cell polarity [38], and our studies suggest that ΔNp63⁺ mammary stem cell-enriched basal cells in the mammary gland may be responsible for interacting with ECM proteins to maintain polarity and lumen formation of luminal cells. Specifically, in the absence of ΔNp63, reduced numbers of K14⁺ basal cells are associated with reduced polarization and closed lumen formation. Desmogleins are part of desmosomal junction that holds neighboring cells together with its adhesive function. Moreover, desmogleins such as Dsg2/3 aid in lumen formation in vitro culture of mammary epithelial cells. We speculate that ΔNp63⁺ basal cells are held together in the basement membrane through these desmosomal proteins. Additionally, some of these desmosomal proteins also aid in apical–basal cell polarity to maintain an open lumen [13,24,39]. In the absence of ΔNp63 during puberty, expression of these desmosomal proteins is decreased, reducing basal cell adhesion and leading to their depletion. Reduced levels of these desmosomal proteins also affect the polarity of mammary epithelial cells, leading to a closed lumen in the absence of ΔNp63 during puberty. On the contrary, as polarization leading to lumen formation has already occurred in adults, loss of ΔNp63 at this stage does not affect morphology. Future studies will delineate how ΔNp63 mediated desmosomal proteins, as well as E-cadherin and β-catenin regulate polarity of mammary epithelial cells.

In conclusion, in addition to the well-known function of ΔNp63 in regulating mammary stem cells and EMT [4], we identified for the first time a potential role for ΔNp63 in the maintenance of cellular metabolism (by regulating OXPHOS and mitochondria mass) and polarity at puberty (Fig. 5K), a time at which the mammary gland is undergoing major modeling and needs high energy. Similar mechanisms may exist in breast cancers or other cancers where ΔNp63 is important, but will require further careful evaluation. Thus, our model will be of paramount importance to delineate the function of ΔNp63-mediated cellular metabolism and polarity—two processes required for normal tissue homeostasis and dysregulated in multiple models of cancer.

Supplementary Material

Suppl files

Fig. S1. iK14-CreER recombinase-based lineage tracing suggests that iK14-CreER expression is present in both basal and luminal cells of the mammary epithelium.

Fig. S2. The deletion of ΔNp63 at puberty dramatically influences mammary gland development.

Fig. S3. Deletion of ΔNp63 at puberty influences long-lived mammary cells.

Abbreviations

ECM

extracellular matrix

FOV

fields of view

GSEA

gene set enrichment analyses

MaSC

mammary stem cell

MTG

MitoTracker Green

OXPHOS

oxidative phosphorylation

PBS

phosphate-buffered saline

PCA

principal component analysis

SCC

squamous cell carcinoma