Hoxa2^+/− mice were maintained by backcrossing to C57BL/6J wild-type mice and the heterozygous mice were intercrossed to generate Hoxa2^+/+ and Hoxa2^−/− embryos for analysis in this study (Smith et al., 2009). Pregnant mice were sacrificed by isoflurane inhalation followed by cervical dislocation. Embryos were staged according to Kaufman (1992) and were considered E0 on the day the vaginal plug was found. Genotypes were confirmed by PCR analyses as described in Gendron-Maguire et al. (1993). The protocol for the use of animals was approved by the University of Saskatchewan's Animal Research Ethics Board and adhered to Canadian Council on Animal Care guidelines for humane animal use.

Embryos were harvested from timed-pregnant mice and the heads were fixed with 4% paraformaldehyde in phosphate buffered saline, pH 7.4 (PBS) for 24 h. Embryo heads were placed in 30% sucrose in PBS for at least 24 h, followed by embedding in optimal cutting temperature compound (OCT; Tissue-Tek®) and serially sectioned at 10 μm thickness. Histological sections taken anterior or posterior to the first molar tooth bud were designated as anterior and posterior palate sections, respectively, and sections taken at the plane of the first molar tooth bud were designated as middle palate sections (Welsh and O'Brien, 2009; Smith et al., 2013). Tissue sections were placed on 0.5% gelatin-coated glass slides, air-dried at room temperature for at least 2 h, rehydrated for 30 min in PBS, and blocked for 30 min at room temperature in PBS containing 4% skim milk and 0.1% Triton X-100. Sections were then incubated overnight at 4°C with primary antibody diluted in PBS. The antibodies used were: Six2 rabbit polyclonal antibody (Proteintech®; 1:500 dilution), E-cadherin rat monoclonal antibody (Sigma; 1:200 dilution) and Ki-67 rat monoclonal antibody (Affymetrix eBioscience®; 1:100 dilution). Sections were rinsed twice for 5 min in PBS followed by incubation with secondary antibody for 1 h at room temperature (anti-IgG Alexa fluor 488 antibody, 1:200 dilution or anti-IgG Alexa fluor 594, 1:400 dilution; Molecular Probes®). Finally, sections were rinsed twice in PBS and mounted in ProLong® Gold antifade reagent with DAPI (Molecular Probes®).

Total RNA was isolated from excised palatal shelves of wild-type and Hoxa2^−/− embryos at stages E12.5, E13.5, E14.5, and E15.5 using Aurum Total RNA mini kit (BioRad®) as per the manufacturer's protocol. First-strand cDNA synthesis was performed using SuperScript reverse transcriptase (Invitrogen®) with 1 μg of total RNA as per the manufacturer's protocol (Smith et al., 2009).

Gene expression analysis was performed on palatal shelf cDNA samples as described in our previous study (Thangaraj et al., 2017). Briefly, a TaqMan® gene expression assay was used for qRT-PCR analysis of relative Six2 mRNA expression levels. All qRT-PCR reactions were performed using 25 ng of template cDNA, TaqMan Universal Master Mix, FAM-labeled TaqMan Gene Expression assay Mm03003557_S1 for Six2 (Applied Biosystems®), and VIC-labeled endogenous control TaqMan assay for β-actin (Applied Biosystems® assay 4352341E). Cyclin D1 expression was quantified by SYBR Green assay using forward primer 5′-ACCCTGACACCAATCTCCTC-3′ and reverse primer 5′-AAGCGGTCCAGGTAGTTCAT-3′. All reactions were run in biological replicates of 5. Thermocycling parameters were: 2 min at 50°C, 10 min at 95°C, followed by 40 cycles of 95°C for 15 s and 60°C for 70 s. The C_T values obtained were analyzed using the 2^−ΔΔCT method to determine the relative expression of target genes in wild-type and Hoxa2 null samples.

To independently confirm the results of our qRT-PCR analyses, we also performed ddPCR gene expression analyses on palatal shelf cDNA samples, following established protocols (Hindson et al., 2013). Briefly, oil-emulsified PCR reaction mixtures containing palatal shelf cDNA were amplified in 96-well plates on a Bio-Rad Tetrad 2 Peltier Thermal Cycler under the following conditions: 95°C for 10 min then 40 cycles of 95°C for 15 s and 60°C for 1 min (2.5°C/s ramp rate) with a final 10 min hold at 98°C. After amplification, the plates were transferred to a Bio-Rad QX 100 Droplet Reader, which aspirated oil-emulsified PCR products from each well and counted numbers of FAM-positive and VIC-positive droplets, sampling at 100 kHz. Discrimination between droplets containing amplified target (positives) from those which did not (negatives) was achieved by applying a global fluorescence amplitude threshold. Gene transcript concentrations for each palatal RNA sample were calculated using dedicated ddPCR Poisson distribution computational modeling software (Bio-Rad®). We employed the same Six2 and β-actin TaqMan assays for our ddPCR analyses as described in our qRT-PCR protocol.

Palatal shelves were dissected from wild-type or Hoxa2^−/− embryos and homogenized in RIPA Buffer (150 mM NaCl, 10 mM Tris, 0.1%-SDS, 1% Triton X-100, 1% deoxycholate, 5 mM EDTA) supplemented with a protease inhibitor cocktail (Sigma®) as previously described (Brown and Nazarali, 2010). Sample aliquots containing equal total protein were loaded onto 10% polyacrylamide-SDS gels. Following electrophoresis, the proteins were transferred to Immunoblot PVDF membranes (Bio-Rad®). The membranes were blocked overnight at 4°C in PBS containing 4% skim milk, followed by incubation for 1 h at room temperature with rabbit polyclonal Six2 primary antibody (Proteintech®; diluted 1:2,000 in PBS containing 4% skim milk). This was followed by four 15 min washes in PBST (PBS containing 0.08% Tween-20) and incubation for 1 h at room temperature with horseradish peroxidase-conjugated anti-rabbit secondary antibody (Bio-Rad®; diluted 1:3,000 in PBS containing 4% skim milk). After four 15 min washes in PBST, the membranes were incubated with Clarity® Western ECL substrate (Bio-Rad®) and the signal detected using a SYNGENE® image analyzer. As a control for equal protein loading, the membranes were subsequently washed overnight at 4°C in PBS, followed by incubation with anti-β-tubulin (Developmental Studies Hybridoma Bank) at 1:2,000 dilution for 1 h at room temperature. Subsequently, membranes were washed with PBST, incubated with anti-mouse IgG horseradish peroxidase conjugate (Bio-Rad®; 1:2,000 dilution), and processed for chemiluminescence protein detection as described above. After imaging, semi-quantitative densitometry was performed using AlphaView® software to generate an integrated density value for each Six2 protein band, which was normalized to the β-tubulin density value from the same sample. Four separate Western blots, each having both wild-type and Hoxa2^−/− samples from E12.5, E13.5, E14.5, E15.5 palatal shelves, were performed. For each blot, the normalized Six2 expression values in the various samples (wild-type and Hoxa2^−/− palates at E12.5, E13.5, E14.5, E15.5 stages) were compared to the value of the E12.5 wild-type sample on the same blot, which was arbitrarily assigned a relative expression level of 1. A total of four blots (n = 4) were analyzed using two-way ANOVA to compare Six2 protein expression in wild-type and Hoxa2^−/− samples.

Primary cultures of MEPM cells were established as previously described (Iwata et al., 2012; Iyyanar and Nazarali, 2017). Briefly, palatal shelves from E13.5 embryos were aseptically micro-dissected and placed in Hank's balanced salt solution. The palatal shelves were then incubated with 0.25% trypsin/EDTA in Ca⁺/Mg⁺⁺-free PBS for 20 min at 37°C, briefly triturated, and passed through a 70 μm cell strainer (BD Falcon®) to generate a dissociated mesenchymal cell suspension. Trypsin action was terminated by adding complete DMEM/F-12 medium (Dulbecco's Modified Eagle's Medium/Ham's medium F12 [1:1] supplemented with 10% fetal bovine serum, 1% antibiotic/antimycotic solution [Sigma®], and L-glutamate) to the cell suspension. The cells were plated on poly-D-lysine coated plates and cultured at 37°C in 5% CO₂ incubator.

A pre-designed Six2 siRNA (Invitrogen® Silencer Select assay s73795) and a negative control siRNA (Invitrogen®) were utilized. MEPM cells (5 × 10³) were plated in a 96-well plate until they reached 60–80% confluency. Aliquots of 50 nM siRNA were mixed with Lipofectamine 3000 (Invitrogen®) according to the manufacturer's protocol. The siRNA-Lipofectamine complex (10 μl) was added to each well of MEPM cells containing 100 μl of serum free medium. Following incubation at 37°C for 12 h, the transfection medium was replaced with fresh medium. MEPM cells were analyzed for cellular DNA content after 48 h from the time of transfection using the CyQUANT NF cell proliferation assay kit (Life Technologies®) following the manufacturer's protocol. Briefly, the transfected MEPM cells were incubated with 1X dye binding solution at 37°C for 45 min in the dark. Fluorescence was measured with a BioTek® microplate reader at 485 nm excitation and 530 nm emission wavelengths.

Following IHC staining of histological sections, the number of Ki-67 positive mesenchymal cells in the palatal shelf were counted manually using Image J software and this number was divided by the total cross-sectional area (mm²) of the palatal mesenchyme. Additionally, the numbers of Ki-67 positive mesenchyme cells in the nasal vs. oral halves of the sectioned palatal shelves were counted and divided by their respective areas (mm²); n = 4.

All statistical analyses and graph construction were performed using GraphPad® Prism 5.0 software. The Western blot, gene expression, and cell count data were evaluated using two-way analysis of variance (ANOVA) followed by Bonferroni post-hoc tests.

Our qRT-PCR analyses revealed that Six2 mRNA is expressed in the developing palatal shelves of wild-type mouse embryos from E12.5 to E15.5, with highest expression at E12.5 and E13.5 (Figure 1A). The palatal shelves of Hoxa2^−/− null embryos showed a significant upregulation of Six2 mRNA levels relative to wild-type palates at stages E12.5 to E14.5 (Figure 1A). These trends were confirmed on independent biological samples using the ddPCR technique as an alternate method to quantify Six2 gene transcript levels (Figure 1B). Consistent with the Six2 mRNA expression profiles, Western blot analysis revealed that Six2 protein is present in the developing palatal shelves from E12.5 to E15.5, with peak expression at E13.5 in wild-type embryos (Figures 1C,D). At stages E12.5 to E14.5, Six2 protein levels were significantly higher in Hoxa2^−/− palatal shelves compared to wild-type palatal shelves (Figure 1D). These results demonstrate that Six2 is expressed intrinsically in the developing palate and is negatively regulated by Hoxa2 during palatogenesis.

We next examined the spatial distribution of Six2 protein in the developing SP. IHC analyses of mid-coronal sections of heads from wild-type mouse embryos revealed abundant Six2 protein expression in the mesenchyme of the palatal shelves from stages E12.5 to E15.5 (Figures 2A–D). The intensity of Six2 immunostaining in the palatal shelf mesenchyme of wild-type embryos appeared higher at earlier stages of palatogenesis (E12.5–E13.5) (Figures 2A,B) compared to later stages (E14.5–E15.5) (Figures 2C,D). At stages E12.5 and E13.5, prior to palatal shelf elevation, Six2 protein was observed throughout the palatal mesenchyme in both the prospective “nasal half” of the palatal shelf (located nearest the tongue at these pre-elevation stages) as well as in the “oral half” of the palatal shelf (located furthest from the tongue) (Figures 2A,B). However, by E14.5 to E15.5, after the wild-type palatal shelves have reoriented to a horizontal position above the tongue, there was a conspicuous loss of Six2 immunostaining within a layer of palatal mesenchyme located in the nasal half of the palatal shelf, immediately subjacent to the surface palatal epithelium (Figures 2C,D). In contrast, Six2 protein expression persisted throughout the palatal mesenchyme in the oral half of the palatal shelf in wild type embryos (Figures 2C,D).

The palatal shelves of Hoxa2 null embryos, unlike those of wild-type embryos, fail to elevate and instead remain oriented vertically downward on either side of the developing tongue (Figures 2E–H). Within these Hoxa2^−/− palatal shelves, Six2 protein expression persisted through stage E15.5 in palatal mesenchyme cells of both the nasal half of the palatal shelf (positioned nearest the tongue) as well as the oral half of the palatal shelf (located furthest from the tongue) (Figures 2E–H). Therefore, in comparison to wild-type embryos, the loss of Hoxa2 function expands the spatial domain of Six2 expression within the palatal mesenchyme at stages E14.5–E15.5 (compare Figures 2A–D to Figures 2E–H).

The outer epithelial cell layer that coats the nasal and oral surfaces of the palatal shelf displayed a strikingly different pattern of Six2 protein distribution. In wild-type palatal shelves, Six2 immunostaining was prominent in the surface epithelium located on the nasal side of the palatal shelf at both pre-elevation and post-elevation stages (Figures 2A–D; Figures 3A,B). Conversely, Six2 protein was undetectable in the surface epithelium located on the oral side of the palatal shelf (Figures 2A–D, 3A,D). The pattern of Six2 expression within the palatal epithelium was largely unaffected by loss of Hoxa2 function. In Hoxa2^−/− embryos, like wild-type embryos, Six2 expression was prominent in epithelial cells located on the nasal side of the palatal shelf (facing the tongue), with little or no Six2 protein detectable in the epithelium covering the oral side of the palatal shelf (located furthest from the tongue) (Figures 2E–H, 4A,B,D). These findings were confirmed by co-staining palatal sections from wild-type (Figure 3) and Hoxa2^−/− embryos (Figure 4) with an antibody for E-cadherin, a characteristic epithelial marker (Figures 3E–L, 4E–L). Examination of the palatal sections at high magnification revealed that the surface epithelium on the nasal side of the palatal shelf co-expressed both Six2 and E-cadherin proteins in wild-type embryos (Figures 3B,F,J) as well as in Hoxa2^−/− mutants (Figures 4B,F,J). By contrast, epithelial cells on the oral side of the palatal shelf were positive for E-cadherin alone in both wild-type (Figures 3D,H,L) and Hoxa2^−/− embryos (Figures 4D,H,L).

Interestingly, in the palatal shelves of E14.5 wild-type embryos, we also observed prominent expression of Six2 together with E-cadherin within cells of the MES, the point of contact/adhesion between the apical tips of the two horizontally elevated palatal shelves (Figures 3C,G,K). However, Six2 immunostaining was only faintly visible in medial edge epithelium (MEE) cells located at the apical tips of the Hoxa2^−/− palatal shelves, which fail to elevate and make midline contact (Figures 4C,G,K).

To explore the relationships between Six2, Hoxa2 and cell proliferation during SP development, we performed double immunofluorescence staining on histological sections from the anterior, middle, and posterior regions of E13.5 wild-type and Hoxa2 null palatal shelves using Six2 antibody in combination with an antibody for Ki-67, a nuclear protein expressed exclusively in proliferating cells (Figure 5).

We observed that the mesenchyme of both wild-type and Hoxa2^−/− palatal shelves contained large numbers of proliferating, Ki-67 positive cells (Figures 5C,D). Cell counts performed on the immunostained sections revealed that the number of Ki-67 positive palatal mesenchyme cells per unit area was significantly higher in Hoxa2^−/− palatal shelves compared to wild-type in both the anterior and posterior regions of the palate (Figure 6A). This trend was also observed in the middle region of the palate, although the difference there was not statistically significant. Interestingly, in all three regions along the A-P axis of the palate (anterior, middle, and posterior), the density of proliferating Ki-67 positive mesenchyme cells was significantly higher in the nasal half of the palatal shelf compared to its oral half (Figure 6B). This was the case for both wild-type as well as Hoxa2 null embryos (Figure 6B).

Many Six2-expressing palatal mesenchyme cells in both wild-type and Hoxa2 null embryos were actively proliferating, as evidenced by Ki-67 co-expression (Figures 5E, F). Cell counts revealed that the numbers of these Six2/Ki-67 double-positive palatal mesenchyme cells per unit area were significantly higher in Hoxa2^−/− mice compared to wild-type in all three regions along the A-P axis of the palate (anterior, middle, and posterior) (Figure 6C).

We next investigated whether Six2 is a positive regulator of cell proliferation in the palatal mesenchyme and whether upregulation of Six2 is responsible for the increased cell proliferation in Hoxa2^−/− palatal mesenchyme. Primary cultures of MEPM cells were transfected with siRNA targeting Six2 mRNA to determine the effects of Six2 knockdown on proliferation of palatal mesenchyme cells in vitro. As shown in Figure 7A, treatment with Six2 siRNA resulted in ~90% reduction in Six2 mRNA expression in MEPM cultures when compared to cultures administered either siRNA delivery vehicle alone (mock treatment) or a non-targeting negative control siRNA. Importantly, siRNA-mediated Six2 knockdown decreased cell proliferation in both wild-type and Hoxa2^−/− MEPM cultures (Figure 7B) and also reduced mRNA levels of Cyclin D1, a cell cycle regulator (Figure 7C). In the Hoxa2^−/− MEPM cultures, Six2 knockdown restored both cell proliferation and Cyclin D1 expression down to levels approximating those of wild-type control MEPM cultures treated with either siRNA delivery vehicle alone or the negative control siRNA (Figures 7B,C).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.