Identification of adult mineralized tissue zebrafish mutants

Abstract

Zebrafish craniofacial, skeletal, and tooth development closely resembles that of higher vertebrates. Our goal is to identify viable adult zebrafish mutants that can be used as models for human mineralized craniofacial, dental, and skeletal system disorders. We utilized a large-scale forward-genetic chemical N-ethyl-nitroso-urea (ENU) mutagenesis screen to identify 17 early lethal homozygous recessive mutants with defects in craniofacial cartilage elements, and 7 adult homozygous recessive mutants with mineralized tissue phenotypes including craniofacial shape defects, fused sutures, dysmorphic or missing skeletal elements, scoliosis, and neural arch defects. One mutant displayed both an early lethal homozygous phenotype and an adult heterozygous phenotype. These results extend the utility of the zebrafish model beyond the embryo, to study human bone and cartilage disorders.

Skeletal malformations, including craniofacial and dental abnormalities, are among the most common birth defects. Orofacial clefting, primarily consisting of cleft lip and/or cleft palate, occurs in about 1 in 700 births, while craniosynostosis, premature fusion of the cranial sutures, occurs in 1 out of 2500 births (Coussens et al., 2007; Juriloff and Harris, 2008). The estimated number of congenital vertebral malformations is approximately 0.5–1 per 1000 births, and skeletal dysplasia, disorders characterized by abnormalities of cartilage and bone growth resulting in abnormal shape and size of the skeleton, occurs in approximately 1 in 4,000 to 5,000 births (Raggio et al., 2009). Normal craniofacial and skeletal development requires precise timing and coordination of proper growth factor signaling pathways, transcription factor activities, and tissue interactions. In the past two decades, several genes implicated in craniofacial and skeletal development have been identified. However, we are still far from fully understanding the complex process of mineralized tissue development, size and shape determination, and homeostasis. Although skeletal malformations can be a part of a genetic syndrome, the majority of cases are non-syndromic, making them even more difficult to study.

The zebrafish has proved to be not only an excellent model to study vertebrate development but also a useful model for human diseases. There are many advantages of using zebrafish as a model, including ease of handling and economical maintenance, short reproductive cycle, external fertilization and development, production of large numbers of synchronous and rapidly developing embryos per mating, and the optical transparency of zebrafish embryos. These advantages have facilitated large-scale forward genetic mutagenesis screens that have yielded the discovery of numerous mutants, and the identification of genes important for the development of many different organs and tissues, including craniofacial and tooth development (Neuhauss et al., 1996; Piotrowski et al., 1996; Schilling et al., 1996). These elegant studies demonstrated that specification and development of craniofacial skeletal elements are governed by similar signaling pathways in humans and zebrafish. However, almost all known mutagenesis screens had several significant limitations. For one, mutants were usually selected based on visible defects, which limited the detection to the most severe phenotypes. Also, the mutants were screened at early embryonic stages (5–6 days post fertilization (dpf)), which limited the screens to the identification of genes important to the early development of predominantly cartilaginous skeletal elements, while the genes controlling later stages of bone development and homeostasis were not studied.

We have used a forward genetic chemical mutagenesis screen to identify novel adult craniofacial, skeletal and tooth mutants by screening for mineralized tissue defects in viable adult F3 zebrafish. The mutagenesis of adult male zebrafish was performed as previously described (Solnica-Krezel et al., 1994). Ninety two, adult three month old zebrafish males were subjected to three 1 hour treatments with N-ethyl-nitro-urea (ENU). Mutagenized F0 males were crossed to wild type females to generate 1,084 F1 individuals, which were raised to sexual maturity and outcrossed to wild type adults, generating 131 F2 families. Sexually mature F2 family members were then incrossed in pairwise fashion to generate 12–15 F3 clutches for analysis. The F3 embryos were initially screened daily to 5–6 dpf to identify embryonic lethal (EL) phenotypes. When EL phenotypes were detected and confirmed, embryos were fixed at 4–5 dpf depending on severity of the phenotype, and stained with Alcian Blue (Ab) for cartilage and Alizarin Red (AR) for mineralized structures. Viable siblings from EL phenotype clutches were raised to 16–21 mm in body length, stained with Alizarin red (AR), and screened for adult heterozygous phenotypes. When no EL phenotype was observed, all of the embryos of the clutch were grown to juvenile stage (16–21 mm in body length at ~ 6–8 weeks), stained with AR, and analyzed for mineralized tissue defects.

Out of the 49 screened F2 families, we isolated 22 mutants which displayed an EL phenotype (Table 1). Two mutants had EL phenotypes but did not display any visible craniofacial defects (83N and 91N, Table1 and data not shown). Three EL mutants (8N, 30N and 85N) exhibited a small head, small eye embryonic phenotype. Ab/AR staining revealed that these mutants did not develop any detectable craniofacial structures (Table1, data not shown). One of the mutants, 93N, appeared to be anemic, and exhibited reduced pigmentation and a curved tail. Interestingly, cartilaginous craniofacial elements were present in the 93N mutant, but stained only faintly with Ab suggesting defects in extracellular matrix formation and cartilage differentiation (Table1, Figure 1). The largest group of EL mutants had several common features, including smaller heads, some with curved bodies, and heart edema (7N, 9N, 12N, 69N, 78N, 100N, 131N, Table1). Ab/AR staining revealed similar skeletal defects, such as reduced or defective anterior arches, as well as absent posterior arches (Figure 1, Figure 3 and data not shown). One mutant, 69N lyon (lyo), displayed a distinctive phenotype (Figure 1) consisting of a shortened Meckel’s cartilage, rostrocaudal elongated and split ethmoid plate resembling cleft palate, and shortened and perhaps duplicated ceratohyal cartilage elements (Figure 1). The surviving putative heterozygous and wild type siblings of all EL phenotype families were also analyzed for adult phenotypes, with the expectation that 2/3 would be heterozygous, and the remaining 1/3 would be wild type. Out of the 22 EL families that were analyzed, we found one family, 78N/knjaz(knz), which also displayed adult skeletal defects. AR staining revealed upper jaw shape defects in knjaz mutants, and defects in tail structures (Figure 2). Pleurostyle and hypural bones were not properly formed and articulated in the tails of knjaz mutants, as compared to wild type siblings (Figure 2).

Table 1.

Summary of early lethal phenotypic ENU families.

Family	Name	Phenotype		Embryonic phenotype	Skeletal phenotype
		Onset	Lethal
3N		2dpf	3dpf	Heart edema/curved body	Not determined
7N		2dpf	7dpf	Small head/curved body	Small pharyngeal arches
8N		3dpf	7dpf	Small head/ curved body	No defined craniofacial cartilage
9N	Fast track (ftt)	3dpf	8dpf	Small head/ enlarged gut	Reduced/ missing arches/reduced jaw
11N		3dpf	5dpf	Small head/curved body	Reduced cartilage formation
12N		3dpf	6dpf	Small head/ heart edema	Absent posterior arches / defective ceratohyal
18N		1dpf	3dpf	Heart edema/ small eye	Not determined
30N		3dpf	5dpf	Small head/ small eye	No defined craniofacial cartilage
32N		2dpf	4dpf	Severe deformity	Variable phenotype
48N		2dpf	6dpf	Small head/ small eye/ curved body	Not determined
56N		2dpf	5dpf	Curved body	Variable phenotype
60N		3dpf	6dpf	Curved body	Small pharyngeal arches, reduced low jaw
69N	Lyon (lyo)	3dpf	6dpf	Small head/ small eye/ curved body	Inverted ceratohyal/reduced or missing posterior arches
75N		3dpf	6dpf	Small head/curved body/ edema	Reduced cartilage formation
78N	Knjaz (knz)	2dpf	5dpf	Small low jaw/ curved body/ heart edema	Reduced neurocranium/ reduced anterior arches/ absent posterior arches
83N		3dpf	6dpf	Curved body	No defects found
85N		2dpf	4dpf	Small head/ small eye/ curved body	No defined craniofacial cartilage
91N		2dpf	4dpf	Small tail/ large heart/ edema	No defects found
93N		4dpf	6dpf	Curved tail/ reduced pigment/ anemic	Reduced cartilage staining
100N	Vader (vdr)	3dpf	5dpf	Small head/small eye/ curved body/heart edema	Reduced anterior arches/ absent posterior arches
130N		3dpf	5dpf	Curved body	Not determined
131N		3dpf	5dpf	Small head/ small eye/ curved body	Reduced or absent anterior arches/ absent posterior arches

Early lethal phenotypes were apparent 1–3 days post-fertilization (dpf), and were characterized by distinct features, including reduced pharyngeal arch cartilages, as described.

Figure 1. Early lethal craniofacial phenotypes.

Examples of EL phenotypes are presented. For each, the left hand panels show lateral views of living embryos, middle panels show lateral views of Alcian blue and Alizarin Red (Ab/AR) stained larvae, and right hand panels show ventral views of Ab/AR stained larvae. Small arrows point to pharyngeal teeth (pt). For the 69N Lyon mutant, the black arrow indicates a smaller lower jaw of 69N lyon (lyo), the red arrow indicates a rostrocaudally extended ethmoid plate, a n d the white arrow indicates a shortened Meckel’s cartilage. The black arrowheads point to rudimentary branchial arch cartilages of 69N lyo embryo, and an asterisk indicates the clefted ethmoid plate. The black arrow in the 93N mutant panel indicates aberrant shape of the lower jaw. The black arrows in the 100N mutant panel indicates a reduced lower jaw, Meckel’s cartilage, and ceratohyal cartilage element. Abbreviations: ch, ceratohyal; eth, ethmoid plate; e, eye; M, Meckel’s cartilage; pt, pharyngeal teeth; wt, wild type; ch, 1–5, branchial arches.

Figure 3. Adult craniofacial phenotypes of the 38N belka (bka) and 102N dushka (duk) mutants.

(a–h) Alizarin Red (AR) stained adult zebrafish. (a,b,e,f) Lateral views of heads of wt (a,e), and 38N bka (b) and 102N duk (f) mutant siblings. The arrows in (b) and (f) point to reduced upper jaw in both mutants. The arrowhead in (f) points to bent parasphenoid in 102N duk mutant. (c,d) Dorsal view of upper jaw in wt (c) and 38N bka mutant (d) siblings. The asterisks indicate patent suture in wt maxilla (c), and fused suture in 38N bka mutant maxilla (d). Arrows in (d) indicate small maxillary bones in 38N bka mutants. (g,h) Lateral view of axial skeleton in wt (g) and 102N duk mutant (h) siblings. Arrows indicate small and misshapen neural arch cartilages in 102N duk mutants.

Figure 2. Homozygous and heterozygous phenotypes of the 78N knjaz (knz) mutant.

(a, b) Lateral view of live 4 dpf wt (a) and 78N knjaz (knz) mutant (b) larvae. (c–f) Lateral (c, d) and (e, f) views of Ab/AR stained 4 dpf wt (c, e) and 78N knjaz (knz) mutant (e, f) embryos. Arrows in (d) and (f) indicate Meckel’s cartilage, and arrowheads in (d) indicate missing branchial arche cartilages. (g–j) Lateral view of AR stained of adult wt (g, i) and 78N knjaz (knz) mutant (h, j) zebrafish. Arrow in (h) points to altered upper jaw morphology in knz mutants. Arrow in (j) points to defects in pleurostyle and hypural bones of knz mutant. Abbreviations: ch, ceratohyal; eth, ethmoid plate; e, eye; M, Meckel’s cartalige; wt, wild type; 1–5, branchial arches.

Among the 27 F2 families that were not embryonically lethal, we identified 4 families with adult homozygous recessive craniofacial phenotypes (Table 2). Several mutants exhibited varying degrees of midfacial hypoplasia. AR staining for mineralized tissues revealed that the 38N belka (bka) mutant exhibited a severe reduction in size of the upper jaw (Figure3). Belka mutants also exhibited missing kinethmoid and fused maxilla (Figure 3). In addition, approximately 50% of belka mutants exhibited scoliosis in the tail region. Similarly to belka, 102N or dushka (duk) mutants exhibited severely reduced upper jaws (Figure 3), however their maxillas were not fused. Dushka mutants also exhibited bent parasphenoid bones and branched neural arches, which are normally straight in wild type zebrafish (Figure3). The least affected mutant, 72N, exhibited only slight midfacial hypoplasia, reduced kinethmoid, and asymmetric maxilla (data not shown). The 72N mutant also exhibited defects in tail structures similar to that observed in knjaz heterozygous mutants (data not shown).The only craniofacial defect we could detect in the fourth mutant, 17N, was a bent basihyal in the lower jaw, as compared to wild type siblings (data not shown). Among the F2 families that did not exhibit EL phenotypes, we found 3 mutants exhibiting axial skeletal defects, but which did not display detectable craniofacial abnormalities (Table 2). In two of these families, 92N/droog (dro) and 99N, we observed defects in neural arch elements, however this phenotype was more dramatic in dro mutants (Figure 4). Mutants in 74N families displayed scoliosis in the tail region (Figure 4). The ratio of mutants to wild type siblings in F3s generated from several pairs of F2 74Ns ranged from 33% to 67%, suggesting a possible dominant phenotype.

Table 2.

Summary of adult mineralized tissue phenotypic ENU families.

Family	Name	Craniofacial phenotype	Axial skeleton phenotype
17N		Bent basihyal	Not detected
38N	Belka (bka)	Reduced upper jaw, mused maxilla, absent kinethmoid	50% scoliosis in tail region
72N		Reduced kinethmoid, asymmetric maxilla	Not detected
74N		Not detected	Scoliosis in tail region
92N	Droog (dro)	Not detected	Missing neural arches
99N		Not detected	Missing neural arches
102N	Dushka (duk)	Reduced upper jaw, bent parasphenoid bone	Curved neural arches

Adult phenotypes were identified in 16–21 mm fish (6–8 weeks post fertilization). Craniofacial and skeletal defects were revealed by Alizarin Red stain for mineralized tissues.

Figure 4. Adult skeletal phenotypes of 74N, 92N droog (dro), and 99N mutants.

Lateral views of wt AR stained axial skeleton (a). Lateral view of AR stained axial skeletons of 92N dro (b–d) and 99N (e–f) mutants. Arrows in (b–f) point to missing neural arches. (g) Lateral view of AR stained wt (g), and 74N mutant (h,i) tails. Arrows point to misshaped tail structures in 74N mutant. Dorsal view of AR stained wt (j) and 74N mutant (k, l) tails. Arrows point to scoliosis in 74N mutant tails.

Overall, our screen represents one of the few systematic genetic screens focusing on viable adult skeletal defects. We are currently working to map the mutations with assigned names (See Tables 1 and 2) to identify and test candidate genes. We anticipate that these studies will contribute to the generation of a more comprehensive collection of mineralized tissue zebrafish mutants relevant to human skeletal, craniofacial, and dental disease phenotypes, and which will become valuable tools for elucidating novel genes and signaling pathways regulating mineralized tissue development, remodeling, and homeostasis.

Materials and Methods

Zebrafish husbandry

Wild type (AB) zebrafish were bred and raised at 28.5°C at Tufts Zebrafish Facility, in a controlled environment with 14/10 h light/dark cycle, as previously described (Westerfield, 1995).

ENU mutagenesis and breeding

ENU mutagenesis was performed as previously described (Solnica-Krezel and Driever, 1994). Briefly, young adult male zebrafish (F0) were mutagenized with ENU and out-crossed to wild type females. The resulting F1 fish were raised to sexual maturity and outcrossed to wild type fish, to generate F2 families. Adult F2 family members were then incrossed in pair-wise fashion to generate F3 fish, which were raised to 16–21 mm (6–8 wpf), and analyzed.

Alcian Blue (Ab) and Alizarin Red (AR) staining

Four to seven dpf zebrafish were stained for cartilage and bone with Ab/AR, as previously described (Yelick and Connolly, 2010). Juvenile stage zebrafish, ~16–21 mm in length, were stained with AR for mineralized tissues using modified versions of previously published methods (Connolly and Yelick, 2010).