tp53-dependent and independent signaling underlies the pathogenesis and possible prevention of Acrofacial Dysostosis–Cincinnati type

Abstract

Ribosome biogenesis is a global process required for growth and proliferation in all cells, but disruptions in this process surprisingly lead to tissue-specific phenotypic disorders termed ribosomopathies. Pathogenic variants in the RNA Polymerase (Pol) I subunit POLR1A cause Acrofacial Dysostosis–Cincinnati type, which is characterized by craniofacial and limb anomalies. In a zebrafish model of Acrofacial Dysostosis–Cincinnati type, we demonstrate that polr1a^–/– mutants exhibit deficient 47S rRNA transcription, reduced monosomes and polysomes and, consequently, defects in protein translation. This results in Tp53-dependent neuroepithelial apoptosis, diminished neural crest cell proliferation and cranioskeletal anomalies. This indicates that POLR1A is critical for rRNA transcription, which is considered a rate limiting step in ribosome biogenesis, underpinning its requirement for neuroepithelial cell and neural crest cell proliferation and survival. To understand the contribution of the Tp53 pathway to the pathogenesis of Acrofacial Dysostosis–Cincinnati type, we genetically inhibited tp53 in polr1a^–/– mutant embryos. Tp53 inhibition suppresses neuroepithelial apoptosis and partially ameliorates the polr1a mutant phenotype. However, complete rescue of cartilage development is not observed due to the failure to improve rDNA transcription and neural crest cell proliferation. Altogether, these data reveal specific functions for both Tp53-dependent and independent signaling downstream of polr1a in ribosome biogenesis during neural crest cell and craniofacial development, in the pathogenesis of Acrofacial Dysostosis–Cincinnati type. Furthermore, our work sets the stage for identifying Tp53-independent therapies to potentially prevent Acrofacial dysostosis–Cincinnati type and other similar ribosomopathies.

Introduction

Ribosome biogenesis is the process by which a cell makes ribosomes, the ribonucleoprotein complexes required for translation of mRNA into protein. The process of ribosome biogenesis begins with transcription of rRNA by RNA Polymerases (Pol) I and III. Pol I transcribes the 47S rRNA within the nucleolus and is considered a rate-limiting step of ribosome biogenesis (1). The 47S rRNA is subsequently processed into the 18S, 5.8S and 28S rRNAs. In contrast, Pol III transcribes 5S rRNA within the nucleus. Each of these rRNAs is modified and then associate with numerous ribosomal proteins for ribosome assembly. Perturbation of any step in this process can result in congenital disorders termed ribosomopathies, which are characterized by defects in growth and development leading to tissue-specific phenotypes, such as craniofacial malformations, skeletal anomalies and/or defects in hematopoiesis (2–4).

We recently described a novel ribosomopathy characterized by facial dysmorphism and variable limb anomalies, which we termed Acrofacial Dysostosis–Cincinnati type (AFDCIN; MIM 616462) (5). We identified and described three individuals with phenotypes ranging from mild malar hypoplasia with dysplastic ears, micrognathia and short, broad fingers, to severe hypoplasia of maxillary and zygomatic bones with severe micrognathia and bowed femurs (5). Each affected individual was genetically diagnosed with a unique heterozygous pathogenic variant in POLR1A, which encodes the largest subunit of Pol I and comprises part of the catalytic core of the enzyme (6).

Craniofacial anomalies often arise from perturbation of the development of neural crest cells, a multipotent progenitor cell population born during early embryogenesis (7). Neural crest cells are induced in the dorsal neuroepithelium, undergo an epithelial to mesenchymal transformation, delaminate and migrate throughout the embryo, ultimately differentiating into a wide variety of cell types and tissues including cartilage and bone of the craniofacial skeleton as well as the neurons and glia of the peripheral nervous system (8,9). Disruptions in neural crest cell development can lead to a variety of conditions and disorders known as neurocristopathies (10).

Our previous work in zebrafish revealed that loss-of-function mutations in polr1a result in a deficiency of neural crest cell precursors, which underlies the pathogenesis of cranioskeletal defects, characterizing AFDCIN as a neurocristopathy as well as a ribosomopathy (5). The deficiency in neural crest cell precursors is a result of elevated levels of Tp53-dependent neuroepithelial cell death in association with reduced rDNA transcription and rRNA synthesis (5). rDNA transcription is considered a rate-limiting step in ribosome biogenesis, and the activation of Tp53 is likely due to perturbed ribosome biogenesis.

We recently discovered that Tp53 plays a role in the pathogenesis of AFDCIN (5), but the actual contribution of the Tp53 response to the characteristic tissue-specific phenotype was not determined and remains unknown. In the present study, we examined the role of Tp53 in a polr1a mutant zebrafish model of AFDCIN, with the goal of preventing this congenital disorder. We show that genetic inhibition of tp53 prevents neuroepithelial cell death and improves the phenotype of polr1a mutants during early stages of embryonic development, demonstrating a critical role for tp53 in the pathogenesis of AFDCIN. However, we also discovered that tp53 inhibition does not improve global ribosome biogenesis or neural crest cell proliferation in late-stage polr1a mutant embryos. The persistence of reduced neural crest cell proliferation, without alterations in neural crest cell differentiation, contributes to hypoplasia of cranioskeletal elements in a Tp53-independent manner. Collectively, our results demonstrate that Tp53-dependent and Tp53-independent mechanisms play critical roles during early and late stages of embryogenesis, respectively, in the pathogenesis of AFDCIN.

Results

AFDCIN is characterized by varying degrees of mandibulofacial dysostosis and other craniofacial malformations, together with limb anomalies (5). In support of POLR1A dysfunction causing these phenotypes, in vivo studies in zebrafish demonstrated polr1a loss-of-function mutants exhibit cranioskeletal and pectoral fin defects characteristic of AFDCIN (5). Moreover, structures that are affected in humans and zebrafish with AFDCIN correlate with the tissue-specific domains of enriched polr1a expression during embryogenesis. Furthermore, polr1a loss-of-function was shown to compromise rDNA transcription, resulting in the induction of Tp53-dependent neuroepithelial cell death. Collectively, this leads to generation of a diminished population of migrating neural crest cells, with reduced proliferation capacity, and consequently, hypoplasia of the craniofacial skeleton (5). However, the actual extent of the contribution of the Tp53 response to pathogenesis of AFDCIN phenotype was not determined and remains unknown.

rDNA transcription and rRNA synthesis are considered rate limiting steps in ribosome biogenesis. Therefore, to test our hypothesis that the reduction in rRNA synthesis observed in polr1a^{hi3639Tg/hi3639Tg} (hereafter referred to as polr1a^–/–) mutant zebrafish (5) would result in deficient ribosome biogenesis, leading to induction of Tp53-dependent neuroepithelial cell death, we performed polysome profiling on wild-type and polr1a^–/– zebrafish. We identified significant reductions in the 40S, 60S and 80S monosome peaks, as well as smaller and fewer polysome peaks in polr1a^–/– mutant zebrafish compared with controls (Fig. 1A). Quantification of the polysome profiles revealed the 40S and 60S ribosomes were reduced in polr1a^–/– embryos to about 60% of controls, whereas the 80S and polysomes were diminished to about 30% of controls (Table 1). Reduced abundance of polysomes is suggestive of perturbed ribosome loading of mRNA, decreased numbers of ribosomes translating a given mRNA and diminished translation initiation. Consistent with these ideas, we confirmed through ³⁵S-Met incorporation that there was indeed a significant 75% reduction in protein synthesis in polr1a^–/– mutant zebrafish compared with controls (Fig. 1B), which correlated with reductions in their rDNA transcription and polysome profiles. Thus, polr1a^–/– mutant zebrafish exhibit a global decrease in ribosome biogenesis and consequently translation.

Table 1.

Quantification of polysome profiles

	40S Subunit	60S Subunit	80S Monosome	Polysomes
Control	1.00 ± 0.03	1.00 ± 0.01	1.01 ± 0.12	1.02 ± 0.18
polr1a–/–	0.61 ± 0.04	0.65 ± 0.05	0.33 ± 0.07	0.28 ± 0.02
P-value	0.0002*	0.0004*	0.0010*	0.0021*

n = 3 for control and polr1a^–/– samples. *Significant at P < 0.05, Student’s t-test.

Figure 1.

polr1a mutant embryos display reduced ribosome biogenesis. Polysome profiling shows a reduction in the 40S, 60S, 80S monosomes, and polysomes in polr1a^–/– embryos indicating an overall decrease in ribosome biogenesis (A). polr1a^–/– embryos exhibit an approximately 75% reduction in the rate of protein synthesis as determined by ³⁵S-Met incorporation normalized to protein concentration (cpm/ng) (B). *P < 0.05, Student’s t-test. Error bars represent 95% confidence intervals.

Given that alterations in ribosome biogenesis can activate Tp53 (11–14), and also that Tp53 plays a major role in the pathogenesis of multiple ribosomopathies (15–17), we sought to understand the complete role of Tp53-dependent neuroepithelial cell death in polr1a^–/– embryos and in the pathogenesis of AFDCIN. We hypothesized that inhibition of tp53 would prevent neuroepithelial cell death, leading to increased production and survival of neural crest progenitors, and perhaps prevention of the cranioskeletal anomalies characteristic of AFDCIN.

To analyze the role of Tp53 in the pathogenesis of AFDCIN, we crossed polr1a^–/– zebrafish to tp53^M214K zebrafish. The tp53^M214K allele, hereafter referred to as tp53^–/–, contains a mutation in the DNA binding domain resulting in production of non-functional Tp53 protein (18). At 24 hours post fertilization (hpf), polr1a^–/– embryos are characterized by severe hypoplasia of the head and eyes (Supplementary Material, Figs S1 and S2) in association with extensive cell and tissue death (5). In contrast, 24 hpf polr1a^–/–; tp53^–/– double homozygous mutant zebrafish are phenotypically indistinguishable from their wild-type siblings (Supplementary Material, Figs S1 and S2). This indicates that Tp53 inhibition can rescue the gross morphology of polr1a^–/– embryos during early development. However, by 48 hpf, polr1a^–/–; tp53^–/– embryos exhibit hypoplasia of the head and eyes relative to control embryos (compare Fig. 2A versus C and D versus F; see also Supplementary Material, Fig. S2). This phenotype is not as severe as that observed in polr1a^–/– mutants, which exhibit more pronounced differences in embryo size and craniofacial tissues, together with some edema in the hindbrain region relative to controls (Fig. 2B and E). At 3 days post fertilization (dpf), these comparative differences remain, with polr1a^–/– embryos displaying the most severe phenotype, whereas polr1a^–/–; tp53^–/– embryos exhibit some improvement in eye, head and overall embryo size (Fig. 2G–L). This pattern continues through 5 dpf, and although both polr1a^–/– and polr1a^–/–; tp53^–/– embryos display hypoplastic craniofacial structures, small pectoral fins and cardiac edema relative to controls, the morphological phenotype of polr1a^–/–; tp53^–/– embryos remains intermediate between polr1a^–/– and wild-type embryos (Fig. 2M–R). Interestingly, tp53 inhibition lengthened the lifespan of a portion of polr1a^–/– embryos from 5 to 8 dpf, but these polr1a^–/–; tp53^–/– embryos still become edemic (Supplementary Material, Fig. S3). Nonetheless, this indicates that tp53 inhibition can rescue the phenotype of polr1a^–/– embryos during early embryogenesis, until about 24–36 hpf, and improve their overall lifespan from 5 to 8 dpf, but cannot rescue the characteristic craniofacial and fin phenotypes throughout later stages of embryogenesis.

Figure 2.

tp53 inhibition improves the early phenotype of polr1a^–/– embryos. Live images of control, polr1a^–/–, and polr1a^–/–; tp53^–/– embryos reveal overall improvements in the size and morphology of polr1a^–/–; tp53^–/– embryos at 2 dpf (A–F), together with reduced microphthalmia of the eye (e), and suppression of edema around the hindbrain (white arrows, compare A, C, to edema in B). At 3 dpf (G–L), polr1a^–/–; tp53^–/– embryos remain smaller than controls, but show improved morphology relative to polr1a^–/– embryos, particularly with respect to eye development (e). By 5 dpf (M–R), polr1a^–/– and polr1a^–/–; tp53^–/– embryos display a similar overall phenotype with pericardial edema (heart, h) small pectoral fins, and reduced lifespan. Scale bar = 200 µm.

To determine if genetic inhibition of tp53 reduced cell death in polr1a^–/–; tp53^–/– embryos, we performed TUNEL staining on mutant and wild-type embryos ranging from 24 to 60 hpf (Fig. 3), which covers the period of neural crest cell migration into the pharyngeal arches and their subsequent differentiation. The start of this time course also preceded the visible onset of a gross morphological phenotype in polr1a^–/–; tp53^–/– embryos. At 24 hpf, polr1a^–/– embryos display high levels of cell death throughout the head, particularly in the neuroepithelium (Fig. 3), in agreement with our previous report (5). Quantitative reverse transcriptase-polymerase chain reaction (qPCR) confirmed the upregulation of tp53 in concert with the loss of polr1a, in polr1a^–/– embryos (Supplementary Material, Fig. S4). In addition, we also determined that p21/cdkn1a, a downstream target of Tp53 which is known to promote cell cycle arrest and apoptosis (19,20) was significantly upregulated in polr1a^–/– embryos (Supplementary Material, Fig. S4B). In contrast, polr1a^–/–; tp53^–/– embryos exhibit very low levels of cell death, which is consistent with a phenotype indistinguishable from controls at this stage (Fig. 3). Furthermore, qPCR revealed the expected downregulation of tp53 in polr1a^–/–; tp53^–/– embryos, together with a concomitant reduction of p21/cdkn1a (Supplementary Material, Fig. S4). The same trend involving high levels of cell death in polr1a^–/– embryos and relatively little cell death in polr1a^–/–; tp53^–/– embryos compared with controls continued to be observed at 36 and 48 hpf. Levels of apoptosis in the polr1a^–/– embryos appear slightly lower and less localized at 60 hpf compared with polr1a^–/– embryos at 48 hpf. Although tp53 inhibition was able to suppress apoptosis in polr1a^–/– embryos for at least two days of development, TUNEL positive cells were observed at 60 hpf in polr1a^–/–; tp53^–/– embryos, at levels which were comparable to polr1a^–/– embryos (Fig. 3). This suggested that although Tp53-dependent cell death played an important role during early embryogenesis in pathogenesis of the AFDCIN phenotype, there may also be a Tp53-independent role during late embryogenesis.

Figure 3.

Cell death is reduced in polr1a^–/–; tp53^–/– mutant embryos. Increased cell death revealed by TUNEL staining is apparent in polr1a^–/– embryos at all stages relative to controls. Inhibition of tp53 reduced cell death in 24–48 hpf polr1a^–/–; tp53^–/– embryos to control levels. Scale bar = 100 µm.

Given the reduction in cell death in polr1a^–/–; tp53^–/– embryos during early development, we hypothesized that this would promote the survival of neural crest cells and their capacity to generate craniofacial cartilage. Therefore, we performed Alcian blue and Alizarin red staining for craniofacial cartilage and bone, respectively, in wild-type and mutant embryos beginning at 3 dpf. polr1a^–/– embryos exhibit severe hypoplasia of craniofacial cartilage relative to controls and an absence in particular of cartilage in the viscerocranium (Fig. 4A, B, D and E). By comparison, 3 dpf polr1a^–/–; tp53^–/– embryos also display craniofacial cartilage hypoplasia relative to controls, but they clearly have more developed trabecular cartilage elements compared with polr1a^–/– embryos (Fig. 4A–F). By 5 dpf, all the cartilage elements of the larval head skeleton are well differentiated and easily identified in control embryos (Fig. 4G and J). In contrast, both polr1a^–/– and polr1a^–/–; tp53^–/– embryos again display varying degrees of hypoplasia of the majority of skeletal elements (Fig. 4H, I, K and L). However, an improvement is evident in the formation of Meckel’s cartilage and the pectoral fin in polr1a^–/–; tp53^–/– embryos compared with polr1a^–/– embryos (Fig. 4, white arrows). Although skeletal elements derived from pharyngeal arches 2–7, such as the ceratohyal and ceratobranchial cartilages, appear to be absent at this stage in polr1a^–/– mutants (Fig. 4J–L), portions of these elements are present in polr1a^–/–; tp53^–/– embryos that survive until 8 dpf with the greatest improvement seen in cartilage elements derived from arches 1 and 2 (Supplementary Material, Fig. S3). Thus, reduced global cell death and more specifically reduced neuroepithelial cell death, led to an improved phenotype in 24 hpf polr1a^–/–; tp53^–/– embryos. However, the majority of neural crest cell-derived cartilage elements still failed to form properly by 5–8 dpf.

Figure 4.

polr1a^–/–; tp53^–/– embryos show slight amelioration of skeletal development. Alcian blue staining of cartilage and Alizarin red staining of bone reveals severe skeletal hypoplasia in both polr1a^–/– and polr1a^–/–; tp53^–/– embryos at 3 dpf (A–F) and 5 dpf (G–L). At 3 dpf, mutant embryos lack cartilage development in the viscerocranium (brackets). polr1a^–/–; tp53^–/– embryos show improved formation of trabeculae (black arrowhead C, F) relative to polr1a^–/– embryos (black arrowhead B, E), but remain hypoplastic relative to controls (black arrowhead A, D). By 5 dpf, some partial cartilage elements are present in mutant embryos such as Meckel’s cartilage (white arrowheads G–I), the ethmoid plate (ep, J, K), and small pectoral fins (pf, J, K). Ceratobranchial cartilage elements are missing in both polr1a^–/– and polr1a^–/–; tp53^–/– embryos (cb 1–5, J–L). Overall, polr1a^–/–; tp53^–/– embryos (L) have slightly more cartilage than polr1a^–/– embryos (K). Schematics of the 5 dpf skeletal elements are drawn in M–O with the neurocranium in light blue, Meckel’s cartilage in red, other first and second arch cartilage elements in dark blue, and the ceratobranchial cartilages in yellow. Scale bar = 200 µm.

Cell differentiation is typically associated with decreases in ribosome biogenesis (21,22). Given that mutations in polr1a result in reduced ribosome biogenesis, we explored whether this triggered premature differentiation in polr1a^–/– embryos. To determine if the perturbation of cartilage formation in polr1a^–/– embryos was a consequence of a defect in neural crest cell differentiation, we examined the expression of markers of cartilage and bone development. In situ hybridization for sox9a, which is expressed in neural crest cell derived cartilage progenitors and is also a critical driver of cartilage differentiation (23), revealed the presence of cartilage progenitor cells in polr1a^–/– and polr1a^–/–; tp53^–/– embryos at 36, 48 and 60 hpf (Fig. 5A–R). However, the sox9a positive progenitor population in polr1a^–/– embryos is severely reduced compared with controls. In contrast to polr1a^–/– embryos, the domain of sox9a expression in polr1a^–/–; tp53^–/– embryos is broader and similar to wild-type controls, but some reduction in the posterior pharyngeal arches persists (Fig. 5G–I, black arrows). This is in accordance with perturbed cartilage differentiation in these arches (Fig. 4, brackets). Similarly, sox9a expression is absent from the pectoral fin progenitor domain (Fig. 5G–L, open white arrowheads) in both polr1a^–/– and polr1a^–/–; tp53^–/– embryos at 36 and 48 hpf. Although a small domain of sox9a expression is detected within the pectoral fin region in 60 hpf polr1a^–/–; tp53^–/– embryos (Fig. 5M–R, open white arrowheads), the diminished domains of sox9a expression are consistent with the severe pectoral fin cartilage hypoplasia observed in both mutants. Analysis of sox9a expression by qPCR at 48 hpf confirmed these observations (Supplementary Material, Fig. S5), with polr1a^–/– embryos showing significantly less sox9a expression than both control and polr1a^–/–; tp53^–/– embryos. Although there was not a significant difference in the relative levels of expression between control and polr1a^–/–; tp53^–/– embryos by qPCR, changes in the domains of expression are evident by in situ hybridization (Fig. 5) and are consistent with the observed skeletal phenotype (Fig. 4).

Figure 5.

tp53 inhibition improves the size of the cartilage precursor population in polr1a^–/– mutants. In situ hybridization labeling for cartilage precursors using sox9a (A–R), and for developing cartilage using col2a1a (S–AJ), in polr1a^–/– and polr1a^–/–; tp53^–/– embryos. At 48 hpf, sox9a expression is absent from the posterior pharyngeal arches in polr1a^–/– embryos (E), and is severely reduced in polr1a^–/–; tp53^–/– embryos (F, black arrowheads). This population gives rise to cartilage elements of the viscerocranium which are absent in polr1a^–/– and polr1a^–/–; tp53^–/– embryos at 3 dpf. sox9a expression is absent within the pectoral fin in 48 hpf (white open arrows) polr1a^–/– and polr1a^–/–; tp53^–/– embryos (E and F versus D), although by 60 hpf, a small population of sox9a+ cells is visible in the polr1a^–/–; tp53^–/– embryos. Additionally, col2a1a is expressed in a smaller domain in polr1a^–/– and polr1a^–/–; tp53^–/– embryos from 36 hpf (S–X), and a clear reduction in the population extending to become the trabecular cartilage elements is evident at 48 hpf (Y–AD, white arrowheads). By 60 hpf, the size of the ethmoid plate is smaller in polr1a^–/– and polr1a^–/–; tp53^–/– embryos, and expression in the pectoral fin is absent (AE–AJ, black open arrowheads). Scale bar = 200 µm.

Sox9 is a transcription factor which activates genes involved in cartilage differentiation, including type II collagen, Col2a1 (24). In zebrafish, col2a1a is expressed in the developing cartilage of the pharyngeal arches, cranial mesoderm and otic capsule, as well as the notochord (25). In polr1a^–/– and polr1a^–/–; tp53^–/– embryos, the domains of col2a1a expression were diminished in size relative to controls, in a manner likely proportional to the size of the mutant embryos. The intensity of expression, however, appeared similar to controls (Fig. 5S–AJ) and, consistent with this, qPCR analysis at 48 hpf did not reveal any significant changes in the relative expression levels of col2a1a expression (Supplementary Material, Fig. S5). This indicates that the cartilage differentiation program is occurring at the right time in 3 dpf polr1a^–/– and polr1a^–/–; tp53^–/– embryos, but as revealed via sox9a expression, there is a reduced number of cartilage progenitor cells undergoing differentiation. We detected col2a1a expression in regions of the neurocranium and otic capsule in all embryos (Fig. 5S–AJ). However, in contrast to control embryos, col2a1a was not detected in the pectoral fin of 60 hpf polr1a^–/– and polr1a^–/–; tp53^–/– embryos (Fig. 5AE–AJ, open black arrowheads). Additionally, both polr1a^–/– and polr1a^–/–; tp53^–/– embryos did not express col2a1a within the posterior pharyngeal arches, which correlates with their failure to form the pharyngeal arch-derived cartilage elements at later stages of embryogenesis. Altogether, our data reveal that polr1a may not be specifically required for cartilage differentiation, but instead may play a role in controlling the size of the neural crest cell-derived cartilage progenitor cell population through proliferation.

Tp53-dependent cell death was suppressed in polr1a^–/–; tp53^–/– embryos (Fig. 3); however, this was not sufficient to completely rescue the cranioskeletal phenotype. We therefore hypothesized that tp53 inhibition did not prevent cell cycle arrest or other defects in proliferation. Consistent with this idea, although no change was detected in the level of p21 in polr1a^–/–; tp53^–/– embryos at 24 hpf relative to controls, qPCR analysis revealed an approximately 2-fold increase at 48 hpf (Supplementary Material, Fig. S4), indicating a role for Tp53-independent regulation of the cell cycle at late stages of embryogenesis. Immunostaining for phospho-Histone H3 (pHH3), a marker of cells in the G2-M phase of the cell cycle (26), was therefore used to evaluate cell proliferation in wild-type and mutant embryos (Fig. 6). At 24 hpf, although polr1a^–/– embryos appear to be smaller and have a reduced number of pHH3+ cells relative to controls, there was no difference between polr1a^–/–; tp53^–/– and control embryos (Fig. 6A–C). To determine the level of global proliferation, IMARIS software was used to measure the volume of the embryo based on DAPI nuclei staining and to count the number of pHH3+ cells within that volume. Quantification revealed that control embryos had a significantly higher number of proliferating cells (5.06 × 10⁻⁵ µm^–3; n = 7) compared with polr1a^–/– embryos (2.87 × 10⁻⁵ µm^–3; n = 4). In contrast, polr1a^–/–; tp53^–/– embryos displayed a similar level of global cell proliferation to controls (4.55 × 10⁻⁵ cells/µm³; n = 7; Fig. 6G). This indicates that in polr1a^–/–; tp53^–/– embryos, there is no significant difference in global proliferation relative to controls prior to the onset of a visible gross morphological phenotype. Furthermore, it suggests that the differences observed in polr1a^–/– embryos compared with controls may be due, at least in part, to Tp53-dependent regulation of cell proliferation.

Figure 6.

Global proliferation is improved in polr1a^–/–; tp53^–/– embryos. Antibody staining for phospho-Histone H3 (pHH3, red) reveals a significant reduction in global proliferation in polr1a^–/– embryos at 24 hpf (B) relative to controls (A), whereas polr1a^–/–; tp53^–/– embryos (C) are not significantly different. Quantification in G. However, by 36 hpf, polr1a^–/– (E) and polr1a^–/–; tp53^–/– (F) embryos both show a significant reduction in proliferation relative to controls (D). Consistent with their improved phenotype, polr1a^–/–; tp53^–/– embryos show an improvement in global proliferation relative to polr1a^–/– embryos. Quantification in H. Scale bar = 100 µm. Error bars represent 95% confidence intervals and *P < 0.05 (one-way ANOVA, Tukey test).

At 36 hpf, a similar trend was observed with polr1a^–/– embryos exhibiting less proliferation than controls and polr1a^–/–; tp53^–/– embryos displaying more proliferation than polr1a^–/– embryos (Fig. 6D–F). Quantification of these differences revealed significantly less global proliferation in polr1a^–/– embryos (2.05 × 10⁻⁵ cells/µm³, n = 6) compared with controls (2.89 × 10⁻⁵ cells/µm³, n = 6). In contrast, a significant increase in proliferation was observed in polr1a^–/–; tp53^–/– embryos (2.54 × 10⁻⁵ cells/µm³, n = 6) compared with polr1a^–/– embryos. The levels remained slightly lower than in wild-type controls, however, this difference was not statistically significant (Fig. 6H, Tukey test). This indicates that proliferation in 36 hpf polr1a^–/–; tp53^–/– embryos is increased relative to polr1a^–/– embryos, which is consistent with their improved phenotype and indicative of a tp53-dependent role in cell proliferation.

Given the suppression of apoptosis and restoration of proliferation in polr1a^–/–; tp53^–/– embryos during early embryogenesis, the subsequent reduction in the neural crest cell-derived cartilage progenitor population (Fig. 5) and the severity of the cartilage defects at 3 dpf (Fig. 4) was surprising. We therefore hypothesized that proliferation may be specifically affected within the neural crest cell population. To test this idea, we bred sox10:gfp transgenic zebrafish, in which neural crest cells were labeled with GFP, into the background of control, polr1a^–/–, and polr1a^–/–; tp53^–/– embryos. At 24 hpf, pharyngeal arches 1 and 2 (outlined in white) are populated by neural crest cells (Fig. 7A–C) and by 36 hpf the same pharyngeal arches are larger and more clearly defined (Fig. 7D–F). Neural crest cells proliferate and differentiate into the cartilage elements derived from pharyngeal aches 1 and 2 including Meckel’s cartilage, palatoquadrate, hyosymplectic and ceratohyal elements (27). We initially examined the neural crest cell population in pharyngeal arches 1 and 2 via confocal imaging and determined their volume using IMARIS software. At 24 hpf, polr1a^–/– embryos have a significantly smaller arch volume (1.12 × 10⁻⁵ µm³; n = 4) than controls (1.88 × 10⁻⁵ µm³; n = 7). In contrast, the arch volume in polr1a^–/–; tp53^–/– embryos (1.72 × 10⁻⁵ µm³; n = 7) was very similar to controls (Fig. 7G). The lack of a significant difference in arch volume between control and polr1a^–/–; tp53^–/– embryos suggests that at 24 hpf, the sizes of the neural crest cell populations were very similar. This is consistent with the observed inhibition of neuroepithelial apoptosis (Fig. 3) and improved global cell proliferation in polr1a^–/–; tp53^–/– embryos (Fig. 6). Thus, neural crest cells in polr1a^–/–; tp53^–/– embryos formed and migrated properly.

Figure 7.

NCC proliferation is significantly reduced in polr1a^–/– and polr1a^–/–; tp53^–/– mutant embryos. Analysis of the NCC population using sox10: gfp transgenic embryos at 24 hpf (A–C) revealed the volume of pharyngeal arches 1 and 2 (white outline) is significantly smaller in polr1a^–/– but not polr1a^–/–; tp53^–/– embryos relative to controls. Quantification in G. By 36 hpf (D–F), the volume of pharyngeal arches 1 and 2 is significantly reduced in both polr1a^–/– and polr1a^–/–; tp53^–/– embryos. Quantification in H. To account for this difference in size, the proliferation within pharyngeal arch 1 and 2 was analyzed. At 24 hpf, polr1a^–/– and polr1a^–/–; tp53^–/– embryos tended to show reduced proliferation, although the difference was not statistically significant (quantification in I). At 36 hpf polr1a^–/– and polr1a^–/–; tp53^–/– embryos both showed an approximately 50% reduction in NCC proliferation (quantification in J). Scale bar = 100 µm. Error bars represent 95% confidence intervals and *P < 0.05 (one-way ANOVA, Tukey test).

However, by 36 hpf, pharyngeal arches 1 and 2 were significantly smaller in both polr1a^–/– (1.54 × 10⁻⁵ µm³, n = 5) and polr1a^–/–; tp53^–/– (2.09 × 10⁻⁵ µm³, n = 6) embryos relative to controls (3.20 × 10⁻⁵ µm³, n = 5) (Fig. 7H). Even though the early stages of neural crest cell development improved in polr1a^–/–; tp53^–/– embryos relative to polr1a^–/– embryos (Fig. 7F versus E), the pharyngeal arches remained at least 35% smaller than controls. To account for the size difference that occurs between 24 and 36 hpf in polr1a^–/–; tp53^–/– embryos, we examined proliferation specifically within the neural crest cell population of pharyngeal arches 1 and 2. We counted the number of sox10: gfp+ and pHH3+ cells within the pharyngeal arches of 24 and 36 hpf zebrafish embryos to determine the percentage of neural crest cells that were proliferating (Fig. 7D–F, quantification in I, J). At 24 hpf, 9.04% of neural crest cells in control embryos (n = 7) were proliferating. In contrast, only 4.71% of neural crest cells in polr1a^–/– embryos (n = 4), and 5.28% of neural crest cells in polr1a^–/–; tp53^–/– embryos (n = 7) were proliferating (Fig. 7I). Although the differences did not reach statistical significance, there was a clear trend in both 24 hpf polr1a^–/– and polr1a^–/–; tp53^–/– embryos that only about half the neural crest cells were proliferating relative to controls.

Quantification of neural crest cell proliferation at 36 hpf revealed the same trend, however, now with statistical significance (Fig. 7J). The percentage of proliferating neural crest cells in 36 hpf control embryos was 13.24% (n = 5), although in polr1a^–/– (n = 6) and polr1a^–/–; tp53^–/– (n = 6) embryos it was 6.00% and 6.38%, respectively. Thus, there is a more than 50% reduction in proliferating neural crest cells during early development, which results in a reduced population of neural crest cells within the pharyngeal arches as evidenced by the diminished domains of sox9a and col2a1a expression (Fig. 5). Consequently, this results in severe hypoplasia of the craniofacial cartilage during later development (Fig. 4), irrespective of tp53 status.

The persistence of deficient neural crest cell proliferation and cartilage defects in polr1a^–/–; tp53^–/– embryos led us to hypothesize that this was mechanistically due to the inability of tp53 inhibition to restore rRNA synthesis and ribosome biogenesis. To test this idea, we used qPCR to estimate the levels of 47S transcription, which as a rate limiting step of ribosome biogenesis, is a very sensitive measure of the status of ribosome biogenesis. More specifically, we used primers to amplify the spacer regions of the 47S rRNA transcript including the 5′ externally transcribed spacer (ETS), internally transcribed spacer (ITS) 1 and ITS2. At 24 hpf, which is prior to the onset of a visible gross morphological phenotype in polr1a^–/–; tp53^–/– embryos, we observed a significant reduction in the relative expression levels of all the spacer regions in polr1a^–/– and polr1a^–/–; tp53^–/– embryos (Fig. 8A). The amount of 18S was not significantly different at this stage, but this is not surprising given the stability and longevity of fully processed maternal 18S rRNA, which lasts for at least 3 days within the embryo (28). However, by 48 hpf, there were significant reductions in expression of all spacer regions examined, as well as the 18S rRNA. Reductions in transcription ranged from 60 to 80% in comparisons between polr1a^–/– and polr1a^–/–; tp53^–/– embryos relative to controls (Fig. 8B). Diminished production of rRNA is associated with decreased ribosome biogenesis and protein synthesis (as shown in Fig. 1), leading to a failure of cell growth and proliferation. Therefore, polr1a loss-of-function disrupts ribosome biogenesis and Tp53-independent regulation of cell proliferation particularly within the neural crest cell population. This in turn diminishes the pool of cartilage progenitors and underpins the tissue-specificity of the cranioskeletal phenotype in AFDCIN.

Figure 8.

rRNA synthesis is not improved in polr1a^–/–; tp53^–/– mutant embryos. Quantitative RT-PCR was used to examine the 47S rRNA transcript including, the 5′ETS, ITS1, and ITS2 spacer regions, to estimate relative transcription of 47S rRNA. polr1a^–/– and polr1a^–/–; tp53^–/– embryos show significant reductions in the relative amount of 47S rRNA at 24 hpf (A), prior to the onset of a visible phenotype in polr1a^–/–; tp53^–/– embryos. A significant difference was not observed for 18S rRNA at 24 hpf. At 48 hpf, the amount of 47S rRNA remains significantly reduced and a reduction in 18S rRNA is also observed (B). Error bars represent 95% confidence intervals and *P < 0.05 (one-way ANOVA).

Discussion

AFDCIN is characterized by varying severity of craniofacial malformations together with limb anomalies (5). We previously established polr1a mutant zebrafish as a model of this neurocristopathy and ribosomopathy disorder and showed that polr1a loss-of-function compromises rDNA transcription, resulting in Tp53-dependent neuroepithelial cell death. This in turn diminishes the population of migrating neural crest cells and reduces their proliferation capacity, resulting in hypoplasia of the craniofacial skeleton (5). However, the extent of the Tp53 contribution to the pathogenesis of AFDCIN had not been previously explored or tested.

In this study, we discovered that tp53-dependent and tp53-independent signaling contribute to the AFDCIN phenotype in polr1a^–/– zebrafish. polr1a^–/– embryos display a high level of neuroepithelial cell death and severe craniofacial hypoplasia as early as 12 hpf (5). In contrast, neuroepithelial cell death was suppressed in polr1a^–/–; tp53^–/– embryos (Fig. 3) and consequently, at least up until about 24–36 hpf, polr1a^–/–; tp53^–/– embryos appear morphologically normal. This suggests that tp53-dependent signaling plays a role in the pathogenesis of the polr1a^–/– mutant phenotype during early embryogenesis, particularly in regulating neuroepithelial cell survival. Quantification of the neural crest cell population within pharyngeal arches 1 and 2 revealed a significant restoration in pharyngeal arch size at 24 hpf in polr1a^–/–; tp53^–/– embryos relative to polr1a^–/– mutant embryos, such that the volume of the arches in polr1a^–/–; tp53^–/– embryos was not significantly different to controls (Fig. 7). This indicates that tp53 inhibition prevented the loss of neural crest cells during early embryogenesis, which resulted in partial restoration of the normal complement of cartilage precursors (Fig. 5).

Surprisingly, despite these improvements in early neural crest cell development, cartilage development was still perturbed at 3 dpf and thereafter (Fig. 4). Because neural crest cells are able to form, migrate and express markers of cartilage differentiation in polr1a^–/–; tp53^–/– embryos (Fig. 5), we hypothesized that the failure to completely rescue the craniofacial phenotype was due to deficient neural crest cell proliferation within the pharyngeal arches (Fig. 7). Subsequently, we found that polr1a^–/–; tp53^–/– embryos exhibited an improvement in global cell proliferation relative to polr1a^–/– mutants (Fig. 6), indicating that there is some phenotypic amelioration upon tp53 inhibition. However, proliferation specifically within the pharyngeal arch neural crest cell population was not improved (Fig. 7). This suggests that the regulation of cell proliferation specifically within the pharyngeal arch neural crest cell population in the pathogenesis of AFDCIN, is tp53-independent. Thus, the limited rescue of cranioskeletal development in 5 dpf polr1a^–/–; tp53^–/– embryos is likely due to the failure to sufficiently restore ribosome biogenesis and proliferation specifically in neural crest cells, which illustrates that tp53-dependent and tp53-independent signaling contributes to the pathogenesis of AFDCIN (Fig. 9). Consistent with these observations, studies in human cell lines have demonstrated that knockdown of POLR1A also reduced proliferation in both a Tp53-dependent and Tp53-independent manner (12,29).

Figure 9.

Tp53-dependent cell death and Tp53-independent regulation of cell proliferation function in the AFDCIN phenotype. In polr1a^–/– mutant embryos, reduced rRNA synthesis results in diminished ribosome biogenesis, an increase in cell death and diminished NCC proliferation which collectively underlie the cranioskeletal anomalies characteristic of AFDCIN. Upon inhibition of tp53, cell death is suppressed, however, this is not sufficient to prevent pathogenesis of the AFDCIN phenotype. This is due to Tp53-independent regulation of NCC proliferation which limits amelioration of malformations of the craniofacial skeleton.

Ribosome biogenesis is a highly regulated process, essential to all cells. Hence, it is not surprising that multiple pathways converge to regulate rRNA transcription and ribosome biogenesis, including Tp53, c-Myc and ARF (30–35). One particularly well-studied pathway is the ribosomal protein (RP)-Mdm2-Tp53 pathway (12,36–39). During normal cell growth and proliferation, Mdm2 ubiquitinates Tp53, targeting it for proteasomal degradation (40,41). In contrast, perturbations in the balance between rRNA synthesis and RP synthesis as well as other alterations in ribosome biogenesis, can cause nucleolar stress, which disrupts nucleolar morphology (38) and impairs Mdm2-Tp53 interactions (11). Under such conditions, certain ribosomal proteins including Rpl5/uL18 and Rpl11/uL5, together with 5S RNA, bind to Mdm2 and inhibit its ubiquitin ligase activity, which leads to Tp53 stabilization (42–44). However, perturbation of ribosome biogenesis may also trigger a Tp53 response without any alteration of nucleolar morphology (12,36). Nonetheless, Tp53 then functions in multiple ways to halt cell cycle progression. Firstly, by activating downstream targets such as p21 which promote cell cycle arrest and apoptosis (19). This is consistent with our observations of p21 activation in polr1a^–/– embryos, and its suppression in response to tp53 inhibition. Secondly, Tp53 can disrupt the interaction between SL1 and UBF, two factors that are required for initiating Pol I transcription, thus resulting in diminished rRNA synthesis and consequently reduced ribosome biogenesis (45). Furthermore, in vitro studies have revealed that the relative level of ribosome biogenesis prior to inhibiting rRNA transcription is important in determining whether cells undergo Tp53-dependent cell cycle arrest or apoptosis (14). Inhibiting rRNA synthesis in cells with high levels of ribosome biogenesis tended to trigger the apoptotic pathway, while cells with a relatively low rate of ribosome biogenesis underwent cell cycle arrest (14). This distinction in the Tp53 response depending on the relative levels of ribosome biogenesis may contribute to the tissue-specificity of phenotypes observed in ribosomopathies.

Tp53 stabilization and accumulation has been shown to play a key role in the pathologies of several ribosomopathies including Treacher Collins syndrome (16,46), Bent Bone Dysplasia syndrome (MIM 614592) (47), 5q– syndrome (MIM 153550) (48), Shwachman–Diamond syndrome (SDS; MIM 260400) (49) and Diamond Blackfan anemia (DBA; MIM 612561) (15–17,37,50–52). Interestingly, Treacher Collins syndrome, which is caused by pathogenic variants in TCOF1 (TCS1; MIM 154500), POLR1C (TCS3; MIM 248390) and POLR1D (TCS2; MIM 613717), is a mandibulofacial dysostosis syndrome that is phenotypically similar to AFDCIN type (53,54). However, a role for Tp53 in syndrome pathogenesis is not limited solely to ribosomopathies characterized by bone and cartilage defects. Several other ribosomopathies, including DBA, SDS and 5q– syndrome, are associated with myelodysplastic syndrome (a failure of normal blood cell development) and increased risk of developing certain cancers. DBA is associated with pathogenic variants in one of several ribosomal proteins (e.g. RPS19, RPS7, RPL5, RPL11) and presents with defects primarily in hematopoiesis (55), although pathogenic variants in SBDS, a gene which functions in 60S ribosome assembly, account for the majority of individuals with SDS (56). The pathogenic variants identified in affected individuals with either Treacher Collins syndrome or AFDCIN have not been linked to increases in malignant transformation or anemia, as occurs in these other ribosomopathies. This is not unexpected as malignant transformation often requires an upregulation of rRNA synthesis, which is impaired in these syndromes. Despite the distinct phenotypes characteristic of these syndromes, Tp53-dependent signaling plays a central role in the cellular response to various ribosomal stressors including perturbation of rDNA transcription, rRNA synthesis and modification, or ribosomal protein production. We observed Tp53 activation and stabilization in response to perturbed rRNA synthesis in polr1a^–/– embryos (5). Consistent with these findings, manipulation of POLR1A in human cancer cell lines revealed that Tp53 activation occurs in response to a disruption in the balance between rRNA transcription and ribosomal protein production (12).

Given the demonstrated importance of Tp53-dependent signaling in the pathogenesis of multiple ribosomopathies, the limited phenotypic rescue of polr1a^–/–; tp53^–/– embryos, in contrast to other Pol I subunit mutants, such as polr1c^–/–; tp53^–/– and polr1d^–/–; tp53^–/–, which have been used in the study of Treacher Collins syndrome (46), was surprising. Mutations in Tcof1 in mice, and in polr1c and polr1d in zebrafish, each result in diminished rRNA synthesis, ribosome biogenesis and translation, which leads to perturbed craniofacial cartilage development (46,57). Tp53 inhibition in Tcof1^+/– mice is able to completely restore normal craniofacial development and rescue neonatal lethality through to adulthood (16). tp53 inhibition in the background of polr1c or polr1d mutant zebrafish embryos is also capable of restoring normal cartilage development (46), but does not rescue long term embryo survival. Although Tp53 does seem to have a consistent role in the pathogenesis of Treacher Collins syndrome irrespective of the gene mutated, the unique responses observed in each of these models to tp53 inhibition could be due to: (i) distinct functions of Tcof1 versus Polr1c and Polr1d; (ii) haploinsufficiency of the Tcof1 mouse model versus embryonic (but not maternal) nulls for polr1c and polr1d zebrafish models or (iii) species-specific differences in the tp53 response.

Tcof1, for example, has known roles in stimulating rRNA transcription (39), rRNA processing (58) and oxidative stress-induced DNA damage repair (59–61). In a Tcof1^+/– mouse model, Tp53-dependent cell death was found to be the primary cause of the phenotype. Despite no apparent global improvement in ribosome biogenesis in response to Tp53 inhibition, rescue of the Tcof1^+/– phenotype was still quite complete (59). Polr1a, Polr1c and Polr1d, however, are subunits of the core Pol I enzyme itself and even when the Tp53 pathway was blocked, there was no improvement in rRNA transcription (Fig. 8). Although there was also no improvement in the Tcof1^+/– model, this would suggest that the level of rRNA transcription and ribosome biogenesis in polr1a, polr1c and polr1d mutant embryos fell below the threshold required for proper embryogenesis and survival to adulthood. The failure of Tp53 inhibition to significantly improve cartilage formation and patterning in polr1a mutants, in contrast to what we previously observed in polr1c and polr1d mutants, may be due to the stronger effect of polr1a loss-of-function on rRNA transcription and ribosome biogenesis in addition to Tp53-independent regulation of the ribosomal stress response. It is important to note that POLR1A forms part of the catalytic core of Pol I and may therefore be more critical to basal levels of activity of Pol I than POLR1C and POLR1D, which exist as a heterodimer at the structural periphery of Pol I (6,62).

Similar to our polr1a^–/– zebrafish model of AFDCIN, other ribosomopathies also do not exhibit a complete rescue or prevention of their disease phenotype upon Tp53 inhibition. This highlights a potential role for Tp53-independent signaling in ribosomopathy pathologies (13,63–66). For example, although Tp53 inhibition can rescue craniofacial anomalies and morphological defects in zebrafish models of DBA, it does not rescue the defects in hematopoiesis (13,66,67). Instead, supplementation with l-leucine, which stimulates the mTOR pathway (68,69), significantly rescued both morphological and hematopoietic defects in the DBA models (64–66). However, these alternative methods to stimulate Pol I transcription are unlikely to be successful in ameliorating the phenotypes caused by mutations in polr1a, polr1c and polr1d, because the function of Pol I itself is disrupted. Other zebrafish models with deficient ribosome biogenesis have demonstrated functions for Tp53-indpendent pathways including AKT (67) and autophagy (70). Furthermore, studies in Tp53-null human cell lines have also revealed Tp53-independent responses to alterations in ribosome biogenesis. For example, when Tp53 was inhibited in cell lines lacking POLR1A, a decrease in proliferation was observed (29), which is consistent with our in vivo observations. This decrease in proliferation was mediated through the transcription factor E2F-1, which lies downstream of Retinoblastoma protein (pRb). Interestingly, inhibiting Tp53 and pRb together restored E2F-1 expression and proliferation in POLR1A deficient cells (29).

Collectively, these studies provide substantial evidence of a significant role for both Tp53-dependent and Tp53-independent signaling in response to alterations in rDNA transcription and ribosome biogenesis. Distinct perturbations in ribosome biogenesis elicit different effects, especially when it comes to Tp53-independent responses. Given the variety of ribosomopathy phenotypes, there must be some specificity in both how different cells and tissues regulate ribosome biogenesis as well as how they respond to ribosomal stress. This may result from the fact that ribosome biogenesis is under the regulation of multiple signaling pathways, and that ribosomal proteins play extra-ribosomal functions (71,72). However, although these ideas have long been appreciated, the potential for tissue-specific regulation of rRNA transcription has been largely ignored. Cell-type specific regulation of rRNA transcription has been observed in the cornea and lens epithelia of the adult mouse eye and, in general, higher levels of rRNA transcription correlated with higher levels of proliferation (73). More recently, the revelation that maternal-specific and somatic-specific mature rRNA species (18S, 5.8S, 28S) have distinct structures in zebrafish, has added another layer of regulation complexity to the translation of specific mRNAs by the ribosome (28). Together, these studies indicate that rRNA may play an underappreciated tissue-specific role in the regulation of translation and proliferation.

Reduced rRNA transcription and ribosomal biogenesis in polr1a^–/– mutants may therefore elicit distinct Tp53-dependent and Tp53-independent responses in different tissues at different times during embryonic development. Neural crest cell cells appear especially sensitive to alterations in rRNA transcription as demonstrated by their apoptosis and decreased proliferation in connection with polr1a loss-of-function. Inhibiting Tp53-dependent cell death is not sufficient to completely prevent the pathogenesis of AFDCIN in our zebrafish model because Tp53-independent signaling regulates the proliferation of neural crest cells. In the future, it will be important to understand mechanistically, how and why, distinct cells and tissues respond differently to the same global perturbations in rRNA synthesis and ribosome biogenesis, particularly with respect to Tp53-independent responses. Furthermore, the type or location of any pathogenic variant in a gene may play a role in the tissue-specificity of these responses, as mutations in POLR1A and POLR1C are also associated with leukodystrophy, a disorder of the white matter in the central nervous system (74,75). The polr1a^–/–, polr1c^–/– and polr1d^–/– zebrafish models, together with the identification of a broader range of pathogenic variants, will be instrumental in addressing these issues and for developing targeted therapies to ameliorate and prevent AFDCIN and other similar ribosomopathies.

Materials and Methods

Animals

Adult zebrafish (Danio rerio) were housed and maintained in the Stowers Institute for Medical Research Zebrafish Facility. Zebrafish embryos were raised at 28.5°C and staged according to Kimmel et al. (76). Heterozygous polr1a^hi3639Tg fish were identified as previously described (5) and homozygous mutant embryos were identified by morphology and then confirmed by PCR. 1-Phenyl-2-thiourea (0.003%) was added to the embryo media to prevent pigment development. polr1a heterozygous mutant lines were crossed with reporter lines including Tg(7.2kb-sox10: gfp) (77), referred to as sox10:gfp, as well as the tp53^M214K line (18).

Live imaging and skeletal stains

Embryos were prepared for imaging via anesthesia with MS-222 and mounting on a depression slide in 2% methyl cellulose while submerged in E2 media. Embryos were imaged using a Leica MZ16 microscope equipped with a Nikon DS-Ri1 camera and NIS Elements BR 3.2 imaging software. When appropriate, manual Z stacks were taken, and the images were assembled using Helicon Focus software. Alcian blue and Alizarin red staining was performed as previously described (78). Embryos were cleared in glycerol and imaged on the system described earlier.

In situ hybridization, immunostaining, TUNEL

In situ hybridization was performed according to standard protocols as previously described (46). Probes against sox9a and col2a1a were used to examine cartilage development. Signal was detected using NBT/BCIP and embryos were post-fixed with 4% paraformaldehyde, washed with phosphate-buffered saline (PBS), cleared in glycerol and imaged on the system described earlier. Whole-mount immunostaining was performed as previously described (Westerfield, 2000, The Zebrafish Book) using primary antibodies against GFP (1:500, Invitrogen) and pHH3 (1:2000, Millipore). Fluorescent secondary antibodies, either Alexa-488 or Alexa-546 (1:500, Invitrogen) were used for detection. TUNEL was performed to assess for apoptosis (46). Embryos were imaged using a Zeiss upright 700 confocal microscope and images were captured and processed using Zen software.

Image quantification and processing

IMARIS software was used to quantify regions of interest from confocal images. To quantify global proliferation, embryo volume and cell counts were generated using the automated settings for each image. Surfaces were generated manually for pharyngeal arches 1 and 2 and then the volume was quantified using the automated settings. Within the surface generated for arches 1 and 2, the spots tool was used to generate cell counts. For cell counts per volume, the surfaces were generated with smoothing set to 0.725 µm and a threshold of 25, whereas the diameter in the spots tool was set to 4 µm with the quality set to 5. Images were assembled in Adobe Photoshop and adjustments of brightness and contrast were made uniformly.

Polysome profiling and protein synthesis assays

For polysome profiling, embryos were sorted by phenotype at 3 dpf and 150 embryos were used for each biological replicate. Polysome profiling was conducted as previously described (46). Briefly, embryos were deyolked and rinsed with ice cold PBS and then dissociated in ice-cold lysis buffer. Homogenized zebrafish were centrifuged at 15 000g at 4°C for 10 min and the supernatant was kept for analysis. Lysates were loaded onto a 10–50% sucrose gradient and ultra-centrifuged at 4°C at 40 000 rpm for 2 h in an SW-41 Ti rotor (Beckman). Each gradient was passed through a UA-6 absorbance reader system (Teledyne ISCO) using a syringe pump (Brandel) to evaluate UV absorbance profiles. Absorbance at 254 nm was recorded using WinDaq data acquisition software (DATAQ INSTRUMENTS) and the profiles were plotted in Microsoft Excel. Three replicates were performed for control and mutant samples. Polysome profiles were quantified as previously described (47). For ³⁵S-Met incorporation assay, embryos were identified by phenotype at 24 hpf and 50 embryos were collected per sample. A minimum of three biological replicates were used in each experiment. Embryos were deyolked and washed with PBS and then dissociated by incubation in TrypLE (Gibco) at 28°C for 30 min. Any remaining tissue was dissociated by pipet. Dissociated cells were washed in PBS and incubated in DME medium without methionine (Sigma-Aldrich) at 28°C for 30 min. Cells were then incubated in DME containing 20 µCi of ³⁵S-methionine (PerkinElmer) at 28°C for 1 hour. Cells were centrifuged and the pellet was washed with ice-cold PBS three times to remove excess ³⁵S-Met. Cells were lysed in RIPA buffer (50 mm Tris pH 7.5, 150 mm NaCl, 0.1% SDS, 1% Triton X-100, Protease Inhibitor (1:1000, Sigma)) for 30 min on ice and vortexed for 5 s every 10 min. Protein concentration was determined using the Qubit Fluorimeter and the remaining extract was counted in Ecoscint to determine ³⁵S radioactivity in a liquid scintillation spectrometer (Beckman LS6500).

qPCR

RNA was collected from zebrafish embryos using the Qiagen miRNeasy Mini Kit and tested for quality and concentration on an Agilent 2100 Bioanalyzer. Equal amounts of RNA were used to synthesize cDNA for qPCR with the Superscript III kit (Invitrogen) using random hexamer primers. The primers sequences and methods used for polr1a, tp53, 5′ETS, ITS1, ITS2 and 18S rRNA were identical to those previously described (5,15,46). Primer sequences for p21 and sox9a were obtained from (79) and (80), respectively. Primers sequences for col2a1a were 5′-GGATTCCACTTTAGCTATGC-3′ forward and 5′-GTAGGTGATGGTCTGAGTG-3′ reverse. β-Actin, canx and ef1 were used as controls. PerfeCTa (Quanta Biosciences) reaction mix and the ABI 7900HT real time PCR cycler were used to measure cDNA amplification. The QuantStudio7 real time PCR cycler (Thermo Fisher) was used to measure cDNA amplification from sox9a, col2a1a, tp53, and p21 primer sets. Three to five biological replicates were run in technical triplicate for each experiment. No reverse-transcriptase and no template samples were included as negative controls. Biogazelle qbase+ software was used to analyze data and perform statistical analyses.

Statistics

A two-tailed Student’s t-test was used to determine significance between the means from the quantification of polysome profiles and ³⁵S-Met experiment. qPCR data were analyzed using Biogazelle qbase+ software to perform one-way ANOVA and post-tests. For image quantification analyses, one-way ANOVA was performed to determine if there were significant differences between groups. If the ANOVA results were significant, the Tukey test was used to determine which differences were significant, or the Tukey–Kramer was used when groups had a differing number of samples.

Study approval

The zebrafish used in this study were approved under Protocol # 2015–0138 of the Stowers Institute for Medical Research Institution Animal Care and Use Committee.

Supplementary Material

Supplementary Material is available at HMG online.

Acknowledgements

The authors thank members of the Trainor laboratory for their discussion and input during the completion of this study. We also thank the Stowers Institute for Medical Research Aquatics Facility for zebrafish care and maintenance, Tom Schilling for the sox10: gfp zebrafish, and Tatjana Piotrowski for the tp53^M214K zebrafish and the col2a1a plasmid.

Conflict of Interest statement. None declared.

Funding

National Institute for Dental and Craniofacial Research (F31DE023017; K.E.N.W.), Stowers Institute for Medical Research (PAT), National Institute for Child Health and Development (T32 HD060549; C.L.N.), National Institute of Health (DE025222; A.E.M.) and March of Dimes (#6-FY15–233; A.E.M.). Original data underlying this manuscript can be accessed from the Stowers Original Data Repository at http://www.stowers.org/research/publications/libpb-1238.