Abstract

Schizophrenia may arise from subtle abnormalities in brain development due to alterations in the functions of candidate susceptibility genes such as Disrupted-in-schizophrenia 1 (DISC1) and Neuregulin 1 (NRG1) . To provide novel insights into the functions of DISC1 in brain development, we mapped the expression of zebrafish disc1 and set out to characterize its role in early embryonic development using morpholino antisense methods. These studies revealed a critical requirement for disc1 in oligodendrocyte development by promoting specification of olig2 -positive cells in the hindbrain and other brain regions. Since NRG1 has well-documented roles in myelination, we also analyzed the roles of nrg1 and ErbB signalling in zebrafish brain development and we observed strikingly similar defects to those seen in disc1 morphant embryos. In addition to their effects on oligodendrocyte development, knock-down of disc1 or nrg1 caused near total loss of olig2 -positive cerebellar neurones, but caused no apparent loss of spinal motor neurones. These findings suggest that disc1 and nrg1 function in common or related pathways controlling development of oligodendrocytes and neurones from olig2 -expressing precursor cells. Like DISC1 and NRG1 , OLIG2 and ERBB4 are promising candidate susceptibility genes for schizophrenia. Hence our findings in the zebrafish embryo suggest that hitherto unappreciated neurodevelopmental connections may exist between key human schizophrenia susceptibility genes. These connections could be investigated in Disc1 and Nrg1 mouse models and in genetically defined groups of patients in order to determine whether they are relevant to the pathobiology of schizophrenia. GenBank accession number for Danio rerio disc1 : EU273350.

INTRODUCTION

Schizophrenia is a devastating, clinically heterogeneous psychiatric disorder affecting 0.5–1% of the worldwide population. The available evidence suggests that schizophrenia may have its origins in early brain development. Histopathological studies reveal the presence of mis-localized neurones beneath the prefrontal cortex in brains from people with schizophrenia ( 1 ), which is suggestive of a defect in neuronal production, migration, guidance and/or integration into target tissue. Other studies have indicated the presence of abnormalities in myelination in the brains of people with schizophrenia ( 2 ), while diffusion tensor imaging suggests white matter disorganization in specific brain areas ( 3 , 4 ), raising the possibility that myelinating oligodendrocyte functions may also be compromised in this disorder. Taken together, these findings suggest that schizophrenia and related psychiatric disorders are likely to be the result of phenotypically complex developmental abnormalities arising from aberrant interactions between multiple cell types within the brain ( 5 , 6 ).

Linkage and association studies have implicated genetic risk factors in schizophrenia aetiology, including Disrupted-in-schizophrenia 1 (DISC1) and Neuregulin 1 (NRG1). DISC1 was identified through the striking segregation of a chromosomal translocation that disrupts the DISC1 gene with schizophrenia and other major mental illness in a large Scottish pedigree ( 7 ). Mouse models with mutations in Disc1 have behavioural and pathological alterations reminiscent of schizophrenia ( 8–11 ). Studies utilizing siRNA and retroviral vectors carrying shRNA constructs targeted against Disc1 mRNA revealed important in vivo functions for Disc1 in regulation of neuronal migration, neurite outgrowth, synaptic transmission and control over functional integration of neurons into the developing and adult brain ( 12 , 13 ). While DISC1 is clearly implicated in the control of neuronal development, there is a relative paucity of data regarding the molecular function of Disc1 during early brain development, and it is currently unknown whether DISC1 functions in glial as well as in neuronal development.

NRG1 encodes a transmembrane signalling protein that can be proteolytically cleaved to release a diffusible intercellular signal. Both membrane-tethered and soluble forms of Neuregulin 1 signal to target cells via membrane-spanning receptor tyrosine kinases of the ErbB family. The NRG1 gene has been implicated in the aetiology of schizophrenia on the basis of a high-resolution genetic mapping study ( 14 ), and subsequent association studies of schizophrenia in Caucasian and Asian populations have supported this observation ( 15 ). Neuregulin 1 has well-established roles in controlling neuronal migration, axon guidance, synaptic transmission and oligodendroglial maturation ( 16–18 ), but in-depth analysis of its neurodevelopmental functions has been complicated by the fact that mutation of murine Nrg1 causes embryonic lethality at around embryonic day 10.5 due to a cardiac defect, precluding later analysis of CNS phenotypes (see 19 for review).

The zebrafish exhibits a powerful combination of attributes which makes this organism particularly well-suited to the molecular genetic analysis of vertebrate neurodevelopmental mechanisms. A battery of techniques for both forward and reverse genetic analysis makes it possible to manipulate the functions of specific genes of interest. In addition, the rapid external development of the zebrafish embryo, together with its optical transparency, facilitate detailed in vivo visualization of molecular probes for analyzing cell type, cell behaviour and subcellular structures in both living and fixed specimens. To provide new insights into the function of disc1 in brain development, we set out to characterize the role of zebrafish disc1 in the development of the early embryo. Our studies reveal a critical requirement for disc1 in promoting specification of oligodendrocytes. In view of the known requirements for NRG1 in myelination and its status as a genetic risk factor for schizophrenia, we then compared oligodendrocyte specification in disc1 and nrg1 morphant zebrafish embryos. Strikingly, disc1 and nrg1 morphant embryos exhibited very similar defects of oligodendrocyte development. In addition, we found that both disc1 and nrg1 were required for specification of olig2 -positive cerebellar neurones. Several recent gene expression profiling studies have implicated the oligodendrocyte lineage in the pathogenesis of schizophrenia, supporting previous neuropathological studies, but our results provide the first direct demonstration of a function for Disc1 in oligodendrocyte development. Furthermore, our findings suggest that disc1 and nrg1 function in common or related pathways controlling development of oligodendrocytes and neurones from olig2 -expressing precursor cells. These neurodevelopmental studies in the zebrafish embryo suggest new lines of investigation that could be undertaken in Disc1 and Nrg1 mouse models, and in human schizophrenia patients.

RESULTS

Expression of disc1 in the zebrafish embryo

The zebrafish disc1 orthologue is partially annotated on zebrafish chromosome 13 as gene Q8AV88_DANRE (Ensembl Gene ID ENSDARG00000021895). This region of the zebrafish genome shows synteny with human chromosome 1; TSNAX / tsnax and EGLN1 / egln1 being immediately upstream of DISC1 / disc1 in both genomes. We used bioinformatic methods and RT–PCR to amplify a 3190 bp cDNA encoding the full coding DNA sequence of disc1 . The cDNA encoded a 998 amino acid protein, the N-terminal half of which is poorly conserved, whereas the C-terminal portion shares 35% identity and 53% similarity over 543 residues with human DISC1, and COILS analysis ( 20 ) demonstrated that this C-terminal portion also has a high probability of forming coiled-coil regions, as does the human protein.

Whole-mount in situ hybridization of a disc1 antisense riboprobe to 26 h post-fertilization (hpf) embryos revealed disc1 expression in scattered cells located superficially in dorsal regions of the embryo (Fig.  1 A). Transverse sectioning through the trunk revealed expression in likely delaminating neural crest (Fig.  1 B). Expression was also prominent in streams of cells surrounding the optic tectum, optic vesicles and at the midbrain-hindbrain boundary, as well as in ventral parts of the otic vesicles (Fig.  1 A–C). By 48–52 hpf, disc1 expression was detectable weakly throughout the brain, consistent with its known roles in neuronal morphogenesis and differentiation (Fig.  1 D and E). In addition, d isc1 expression was particularly prominent in the cranial neural crest-derived cartilages of the lower jaw and also in the pectoral fin buds and sensory patches of the inner ear (Fig.  1 D–F). Strongest expression of disc1 was observed in the Meckel's and ceratohyal cartilages of the lower jaw, which are derivatives of the anterior cranial neural crest. No labelling was observed in embryos hybridized with a control sense probe (data not shown).

Figure 1.

Analysis of disc1 expression in the developing zebrafish embryo. Whole-mount in situ hybridization of disc1 ( AH ) and olig2 ( I ) riboprobes to zebrafish embryos at 26 hpf (A–C) and 50 hpf (D–I). Lateral views of whole embryos at 26 (A) and 50 hpf (E). (A) Strong expression of disc1 is visible in cells surrounding the optic vesicles (arrowhead) and in two ventral patches within each otic vesicle (arrow). (B) Transverse section through the trunk at the level of the red line shown in (A) reveals disc1 expression in delaminating neural crest (arrow). (C) Dorsal view of the head shows disc1 expression between the optic vesicles and tectum and across the midbrain-hindbrain boundary. (D) Lateral view of the head at 50 hpf showing strong disc1 expression in lower jaw cartilages (arrow) and a uniform, lower level of expression in the brain. (E) Lateral view of 50 hpf embryo showing that abundant expression of disc1 is restricted to the brain, craniofacial structures and pectoral fin. (F) Ventral view showing strong disc1 expression in anterior neural crest-derived Meckel's and ceratohyal cartilages at 50 hpf. (G) Dorsal view of hindbrain showing strong disc1 expression in midline cells (arrowhead). (H and I) Transverse sections through the hindbrain at the level of r5 at 50 hpf, showing strong co-localization of disc1 (H) and olig2 (I) transcripts in a restricted sub-population of midline cells. Note that ventro-laterally located olig2 -expressing cells that have migrated away from the midline do not express detectable levels of disc1 .

Figure 1.

Analysis of disc1 expression in the developing zebrafish embryo. Whole-mount in situ hybridization of disc1 ( AH ) and olig2 ( I ) riboprobes to zebrafish embryos at 26 hpf (A–C) and 50 hpf (D–I). Lateral views of whole embryos at 26 (A) and 50 hpf (E). (A) Strong expression of disc1 is visible in cells surrounding the optic vesicles (arrowhead) and in two ventral patches within each otic vesicle (arrow). (B) Transverse section through the trunk at the level of the red line shown in (A) reveals disc1 expression in delaminating neural crest (arrow). (C) Dorsal view of the head shows disc1 expression between the optic vesicles and tectum and across the midbrain-hindbrain boundary. (D) Lateral view of the head at 50 hpf showing strong disc1 expression in lower jaw cartilages (arrow) and a uniform, lower level of expression in the brain. (E) Lateral view of 50 hpf embryo showing that abundant expression of disc1 is restricted to the brain, craniofacial structures and pectoral fin. (F) Ventral view showing strong disc1 expression in anterior neural crest-derived Meckel's and ceratohyal cartilages at 50 hpf. (G) Dorsal view of hindbrain showing strong disc1 expression in midline cells (arrowhead). (H and I) Transverse sections through the hindbrain at the level of r5 at 50 hpf, showing strong co-localization of disc1 (H) and olig2 (I) transcripts in a restricted sub-population of midline cells. Note that ventro-laterally located olig2 -expressing cells that have migrated away from the midline do not express detectable levels of disc1 .

Within the developing brain, an elevated level of disc1 expression was detected in the midline of the hindbrain at around 50 hpf (Fig.  1 G). Transverse sectioning through rhombomere 5 (r5) of the hindbrain revealed that disc1 was abundantly and selectively expressed in a restricted population of midline neural progenitors located half-way along the dorso-ventral axis of the hindbrain (Fig.  1 H), a region where olig2 expression is also prominent (Fig.  1 I). Previous studies of olig2 expression in zebrafish and other species demonstrated that olig2 is an early marker of oligodendrocyte precursor cells (OPC) and is required for oligodendrocyte specification ( 21 ), raising the possibility that disc1 also plays a role in the specification of oligodendrocyte precursors.

disc1 is required for development of cranial neural crest-derived viscerocranial cartilages in zebrafish

To determine whether reduced levels of disc1 activity affected the development of structures in which disc1 was expressed, a morpholino antisense oligonucleotide designed to specifically inhibit translation of disc1 mRNA (disc1aug) was injected into embryos at the 1-cell stage. While control morpholino-injected embryos showed no abnormalities over the period of development studied (Fig.  2 A), by 4–5 days post-fertilization (dpf), disc1 morpholino-injected embryos failed to inflate the swim bladder, they had small, mis-shapen eyes, a grossly mis-shapen head, abnormal pectoral fins, pericardial oedema, a curved trunk and tail and they lacked a morphologically distinct, properly articulated lower jaw (Fig.  2 B). A recent study demonstrated that some morpholinos can elicit off-target developmental abnormalities upon microinjection into zebrafish embryos, through the activation of a p53-dependent cell death programme ( 22 ). To ascertain whether any aspects of the morphant phenotype induced by disc1 morpholinos were due to such off-target effects, the phenotypes produced by co-injecting a 1.5-fold molar excess of a p53-specific morpholino along with either the disc1aug morpholino or the control morpholino were compared. Figure  2 shows that both the overall morphological phenotype and the craniofacial defects that are induced by disc1aug morpholino are completely independent of a requirement for p53. Alcian blue staining of the extracellular matrix associated with cartilage elements further demonstrated that knock-down of disc1 dramatically inhibited specification and morphogenesis of cranial neural crest-derived viscerocranial cartilages (Fig.  2 F; Table  1 ). Although the ethmoid plate and trabeculae of the cranium were relatively normal in disc1 morphant embryos (Fig.  2 D), they were significantly shorter than in control morpholino-injected embryos (Fig.  2 C). However, the ceratobranchial, basihyal and basibranchial cartilages were absent in disc1aug morpholino-injected embryos, and the Meckel's and ceratohyal cartilages of disc1 morphants (Fig.  2 F) were much smaller than in control morpholino-injected embryos (Fig.  2 E).

Figure 2.

Morphological and craniofacial defects induced by disc1 knock-down. ( A and B ) Low power bright field images of live 5 dpf embryos co-injected with either control and p53 morpholinos (A), or a disc1 translation-blocking morpholino (disc1aug) and a p53 morpholino (B). Injection of the disc1 morpholino specifically induces a p53-independent phenotype that includes small, mis-shapen eyes, a grossly mis-shapen head, pericardial oedema (black arrowhead), lack of a morphologically recognizable lower jaw (red arrow), abnormal pectoral fins, a curved trunk and tail, and failure of the swim bladder to inflate (asterisk). Morpholino-mediated knock-down of disc1 inhibits formation of the viscerocranial cartilages ( CF ). Cartilage elements were stained using Alcian Blue in disc1 and p53 (D and F) and control and p53 morpholino-injected (C and E) embryos. Ventral views at the level of the ethmoid plate and trabeculae (C and D) and the viscerocranial cartilages (E and F) from the same embryos are shown. In disc1 morphants the ethmoid plate (red arrowhead) and trabeculae (red arrow) are smaller but correctly shaped [compare (D) with (C)], whereas the ceratobranchial, basihyal and basibranchial cartilages were absent, and the Meckel's (green arrowhead) and ceratohyal (green arrow) cartilages were much smaller than in control morpholino-injected embryos [compare (F) with (E)].

Figure 2.

Morphological and craniofacial defects induced by disc1 knock-down. ( A and B ) Low power bright field images of live 5 dpf embryos co-injected with either control and p53 morpholinos (A), or a disc1 translation-blocking morpholino (disc1aug) and a p53 morpholino (B). Injection of the disc1 morpholino specifically induces a p53-independent phenotype that includes small, mis-shapen eyes, a grossly mis-shapen head, pericardial oedema (black arrowhead), lack of a morphologically recognizable lower jaw (red arrow), abnormal pectoral fins, a curved trunk and tail, and failure of the swim bladder to inflate (asterisk). Morpholino-mediated knock-down of disc1 inhibits formation of the viscerocranial cartilages ( CF ). Cartilage elements were stained using Alcian Blue in disc1 and p53 (D and F) and control and p53 morpholino-injected (C and E) embryos. Ventral views at the level of the ethmoid plate and trabeculae (C and D) and the viscerocranial cartilages (E and F) from the same embryos are shown. In disc1 morphants the ethmoid plate (red arrowhead) and trabeculae (red arrow) are smaller but correctly shaped [compare (D) with (C)], whereas the ceratobranchial, basihyal and basibranchial cartilages were absent, and the Meckel's (green arrowhead) and ceratohyal (green arrow) cartilages were much smaller than in control morpholino-injected embryos [compare (F) with (E)].

Table 1.

Quantification of viscerocranial cartilage defects

Treatment n Normal staining (%) Severe loss of viscerocranial cartilages (%) 
Uninjected 89 99 
CoMO+p53MO 18 100 
Disc1aug 56 38 62 
Disc1aug+p53MO 39 49 51 
Treatment n Normal staining (%) Severe loss of viscerocranial cartilages (%) 
Uninjected 89 99 
CoMO+p53MO 18 100 
Disc1aug 56 38 62 
Disc1aug+p53MO 39 49 51 

disc1 is required for development of oligodendrocytes in the zebrafish hindbrain

Although disc1 is expressed throughout the brain of wild-type embryos at 2 dpf, disc1 transcripts were most abundant in the same region as olig2 -expressing hindbrain ventricular zone midline progenitors (Fig.  1 H and I). Differentiated zebrafish oligodendrocytes are a highly migratory population of cells which first appear in the midline of hindbrain rhombomeres 5 and 6 at around 3–4 dpf, before dispersing laterally into the mantle region ( 23 ). Oligodendrocytes develop from olig2- expressing OPC, which also express sox10 prior to the onset of terminal differentiation and the synthesis of myelin components encoded by genes such as mbp and plp1b . To determine whether disc1 is required for production and/or differentiation of oligodendrocytes, disc1aug morphants and control morpholino-injected embryos were analyzed for expression of the OPC markers olig2 and sox10 , as well as for the expression of plp1b and mbp , markers of terminally differentiated oligodendrocytes. At 32 hpf, expression of olig2 in the hindbrain was similar in uninjected, control morpholino- and disc1aug morpholino-injected embryos, and limited to a small group of midline cells of the ventricular zone spanning rhombomeres 5 and 6 (Fig.  3 A and B). By 50 hpf, the hindbrain olig2 expression domain in both uninjected embryos and embryos microinjected with 0.6 pmol control morpholino had expanded along the midline from r1 to r7 and had become contiguous with the domain of olig2 expression in the spinal cord (Fig.  3 C, D and G). In contrast, expression of olig2 in embryos microinjected with 0.6 pmol disc1aug morpholino was sharply restricted to the ventricular zone midline of rhombomeres 5 and 6 (Fig.  3 F), demonstrating that disc1 is required for expansion of the pool of olig2- expressing cells in the hindbrain (Fig.  3 F; Table  2 ). Moreover, the effect of the disc1aug morpholino on olig2 expression was dose-dependent, as embryos injected with 0.3 pmol of the disc1aug morpholino showed a milder reduction in the number of olig2 -expressing cells in the hindbrain, and some of these cells were located outside rhombomeres 5 and 6 (Fig.  3 E). To confirm that the reduced number of olig2 -expressing cells was due to specific inhibition of disc1 expression, a second morpholino, disc1e2i2, was designed to interfere with splicing of disc1 primary transcripts at the exon 2-intron 2 splice junction (Fig.  4 ). When the disc1e2i2 morpholino was microinjected into embryos, only a few olig2 -expressing cells were specified and they were sharply confined to the midline territory within rhombomeres 5 and 6 (Fig.  4 D), just as was observed with the disc1aug morpholino. These results also contrasted to the situation in control morpholino-injected embryos, where an extensive population of olig2 -expressing cells developed in the hindbrain midline along its full rostro-caudal extent (Fig.  4 A and B). Moreover, as was observed with the disc1aug morpholino, the effects of the disc1e2i2 morpholino on olig2 expression were dose-dependent (Fig.  4 C). RT–PCR analysis with primers spanning exons 2–4 of disc1 demonstrated that the higher (1.2 pmol) dose of disc1e2i2 morpholino used caused near total loss of correctly processed disc1 mRNA in the morphant embryos ( Supplementary Material, Fig. S1 ). Figure  3 H further shows that expansion of the population of hindbrain olig2 -positive cells was specifically inhibited by the disc1aug morpholino in a fully p53-independent manner. Thus, given that both disc1aug and disc1e2i2 morpholinos caused very similar dose-dependent reductions in the number of olig2 -expressing cells, and that the phenotypic effects of the disc1aug morpholino were p53-independent, we concluded that the consistent effects of these two disc1 morpholinos were due to specific inhibition of disc1 expression and were not a consequence of off-target effects causing p53-dependent cell death.

Figure 3.

disc1 is required for expansion of the population of olig2 -positive cells in the hindbrain. Whole-mount in situ hybridization with an olig2 riboprobe was performed on embryos fixed at 32 hpf ( A and B ), 50 hpf ( CF ) and 51 hpf ( G and H ). Expression of olig2 is initiated in hindbrain rhombomeres 5 and 6 midline cells of uninjected (A) and disc1aug morpholino-injected (B) embryos by 32 hpf. The location of hindbrain rhombomeres 5 and 6 is indicated by the brackets. Expression of olig2 is detected along the entire length of the hindbrain midline of uninjected (C) and control morpholino-injected (D) embryos at 50 hpf, whereas in contrast, expression of olig2 remained restricted to the midline of rhombomeres 5 and 6 in embryos injected with a high dose (0.6 pmol) of the disc1aug morpholino (F) at 50 hpf. Embryos injected with a lower dose (0.3 pmol) of the disc1aug morpholino showed a reduction in the number of olig2 -expressing cells in the hindbrain at 50 hpf (E). Similarly, in a separate experiment, expression of olig2 was detected along the entire length of the hindbrain midline of control morpholino- plus p53 morpholino-injected (G) embryos at 51hpf, whereas in disc1aug morpholino- plus p53 morpholino-injected embryos (H), expression of olig2 remained restricted to the midline of rhombomeres 5 and 6.

Figure 3.

disc1 is required for expansion of the population of olig2 -positive cells in the hindbrain. Whole-mount in situ hybridization with an olig2 riboprobe was performed on embryos fixed at 32 hpf ( A and B ), 50 hpf ( CF ) and 51 hpf ( G and H ). Expression of olig2 is initiated in hindbrain rhombomeres 5 and 6 midline cells of uninjected (A) and disc1aug morpholino-injected (B) embryos by 32 hpf. The location of hindbrain rhombomeres 5 and 6 is indicated by the brackets. Expression of olig2 is detected along the entire length of the hindbrain midline of uninjected (C) and control morpholino-injected (D) embryos at 50 hpf, whereas in contrast, expression of olig2 remained restricted to the midline of rhombomeres 5 and 6 in embryos injected with a high dose (0.6 pmol) of the disc1aug morpholino (F) at 50 hpf. Embryos injected with a lower dose (0.3 pmol) of the disc1aug morpholino showed a reduction in the number of olig2 -expressing cells in the hindbrain at 50 hpf (E). Similarly, in a separate experiment, expression of olig2 was detected along the entire length of the hindbrain midline of control morpholino- plus p53 morpholino-injected (G) embryos at 51hpf, whereas in disc1aug morpholino- plus p53 morpholino-injected embryos (H), expression of olig2 remained restricted to the midline of rhombomeres 5 and 6.

Figure 4.

A disc1 splice-blocking morpholino (disc1e2i2) causes similar dose-dependent effects on olig2 expression in the hindbrain to the disc1 translation-blocking morpholino. Uninjected ( A ), 1.2 pmol control morpholino-injected ( B ), 0.6 pmol disc1e2i2- ( C ) and 1.2 pmol disc1e2i2-injected ( D ) embryos at 51 hpf are shown.

Figure 4.

A disc1 splice-blocking morpholino (disc1e2i2) causes similar dose-dependent effects on olig2 expression in the hindbrain to the disc1 translation-blocking morpholino. Uninjected ( A ), 1.2 pmol control morpholino-injected ( B ), 0.6 pmol disc1e2i2- ( C ) and 1.2 pmol disc1e2i2-injected ( D ) embryos at 51 hpf are shown.

Table 2.

Quantification of oligodendrocyte defects

Age Probe Treatment n Expression in r5/r6 (%) No expression in r5/r6 (%) Non-specific abnormalities (%) 
30–32 hpf olig2 Uninjected 68 99 
  CoMO 27 100 
  CoMO+p53MO 89 11 
  Disc1aug 41 80 15 
  Disc1aug+p53MO 20 100 
  Disc1e2i2 16 94 

 
Age Probe Treatment n Widespread distribution (%) Cells restricted to r5/r6 (%) Non-specific abnormalities (%) 

 
50–51 hpf olig2 Uninjected 303 99 
  CoMO 95 99 
  CoMO+p53MO 31 100 
  0.3 pmol disc1aug 72 97 
  0.6 pmol disc1aug 71 17 83 
  Disc1aug+p53MO 43 96 
  0.6 pmol disc1e2i2 63 98 
  1.2 pmol disc1e2i2 41 10 90 
  Nrg1i3e4 29 97 
  DMSO 31 90 10 
  ErbB inhibitor 30 17 83 

 
Age Probe Treatment n Widespread distribution (%) Cells restricted to r5/r6 (%) Non-specific abnormalities (%) 

 
50–53 hpf sox10 Uninjected 35 91 
  CoMO 53 92 
  Disc1aug 68 94 
  Nrg1i3e4 40 97 

 
Age Probe Treatment n Wild-type pattern (%) Reduced/absent expression (%) Non-specific abnormalities (%) 

 
4.5–5 dpf plp1b Uninjected 169 97 
  CoMO 69 95 
  CoMO+p53MO 54 93 
  Disc1aug 41 98 
  Disc1aug+p53MO 55 20 80 
  Nrg1i3e4 71 11 89 
4.5–5 dpf mbp Uninjected 188 94 
  CoMO 40 97 
  CoMO+p53MO 38 100 
  Disc1aug 122 20 79 
  Disc1aug+p53MO 39 15 85 
  Nrg1i3e4 96 13 86 
Age Probe Treatment n Expression in r5/r6 (%) No expression in r5/r6 (%) Non-specific abnormalities (%) 
30–32 hpf olig2 Uninjected 68 99 
  CoMO 27 100 
  CoMO+p53MO 89 11 
  Disc1aug 41 80 15 
  Disc1aug+p53MO 20 100 
  Disc1e2i2 16 94 

 
Age Probe Treatment n Widespread distribution (%) Cells restricted to r5/r6 (%) Non-specific abnormalities (%) 

 
50–51 hpf olig2 Uninjected 303 99 
  CoMO 95 99 
  CoMO+p53MO 31 100 
  0.3 pmol disc1aug 72 97 
  0.6 pmol disc1aug 71 17 83 
  Disc1aug+p53MO 43 96 
  0.6 pmol disc1e2i2 63 98 
  1.2 pmol disc1e2i2 41 10 90 
  Nrg1i3e4 29 97 
  DMSO 31 90 10 
  ErbB inhibitor 30 17 83 

 
Age Probe Treatment n Widespread distribution (%) Cells restricted to r5/r6 (%) Non-specific abnormalities (%) 

 
50–53 hpf sox10 Uninjected 35 91 
  CoMO 53 92 
  Disc1aug 68 94 
  Nrg1i3e4 40 97 

 
Age Probe Treatment n Wild-type pattern (%) Reduced/absent expression (%) Non-specific abnormalities (%) 

 
4.5–5 dpf plp1b Uninjected 169 97 
  CoMO 69 95 
  CoMO+p53MO 54 93 
  Disc1aug 41 98 
  Disc1aug+p53MO 55 20 80 
  Nrg1i3e4 71 11 89 
4.5–5 dpf mbp Uninjected 188 94 
  CoMO 40 97 
  CoMO+p53MO 38 100 
  Disc1aug 122 20 79 
  Disc1aug+p53MO 39 15 85 
  Nrg1i3e4 96 13 86 

Given that olig2 is an early marker for OPC in the zebrafish hindbrain, we investigated the effect of disc1 knock-down on the expression of other markers of the oligodendrocyte lineage during hindbrain development. Analysis of sox10 expression in disc1aug morphant embryos revealed that sox10 -expressing cells were restricted to hindbrain rhombomeres 5 and 6 at 50 hpf (Fig.  5 B) in a similar fashion to olig2 -expressing cells, whereas in control morpholino-injected embryos, sox10 -expressing cells were distributed more extensively along the antero-posterior axis of the hindbrain (Fig.  5 A). In contrast, expression of sox10 elsewhere, such as in the otic vesicles, was unaffected by disc1 knock-down (Fig.  5 A and B). Furthermore, analysis of 5 dpf embryos for expression of the oligodendrocyte differentiation markers plp1b and mbp revealed a profound deficit of differentiated oligodendrocytes in disc1aug morpholino-injected embryos (Fig.  5 D and H), by comparison with control morpholino-injected embryos (Fig.  5 C and G). As was observed for olig2 expression, the effect of disc1 knock-down on expression of plp1b and mbp was completely unaffected by co-injection of the p53 morpholino (compare plp1b expression in Fig.  5 E and F with Fig.  5 C and D, and mbp expression in Fig.  5 I and J with Fig.  5 G and H).

Figure 5.

disc1 is required for development of the oligodendrocyte lineage in the zebrafish hindbrain. Whole-mount in situ hybridization with sox10 ( A and B ), plp1b ( CF ) and mbp ( GJ ) riboprobes on embryos microinjected with control (A, C, E, G and I) and disc1aug (B, D, F, H and J) morpholinos, then fixed at 50 hpf (A and B) and 5 dpf (C–J). Consistent with the olig2 expression patterns described in Figure  3 , the distribution of sox10 -expressing cells at 50 hpf also remained restricted to hindbrain rhombomeres 5 and 6 in disc1aug-injected embryos (B), whereas sox10 -expressing cells were distributed along the entire length of the hindbrain midline in control morpholino-injected embryos (A). In contrast, sox10 expression in the otic vesicle was unaffected by injection of disc1aug morpholino, compare (A) and (B). Expression of the oligodendrocyte differentiation markers plp1b and mbp in the hindbrain was greatly reduced in disc1aug-injected embryos at 5 dpf (D and H), as compared with their widespread expression in the hindbrain of control morpholino-injected sibling embryos (C and G), revealing a profound deficit of differentiated oligodendrocytes in disc1aug-injected embryos. These defects in oligodendrocyte development were p53-independent: Expression of plp1b and mbp in the hindbrain is widespread in a dispersed population of cells at 5 dpf in the hindbrain of embryos injected with a combination of control and p53 morpholinos (E and I), whereas the populations of plp1b and mbp -expressing cells are eliminated in embryos injected with a combination of disc1aug and p53 morpholinos (F and J).

Figure 5.

disc1 is required for development of the oligodendrocyte lineage in the zebrafish hindbrain. Whole-mount in situ hybridization with sox10 ( A and B ), plp1b ( CF ) and mbp ( GJ ) riboprobes on embryos microinjected with control (A, C, E, G and I) and disc1aug (B, D, F, H and J) morpholinos, then fixed at 50 hpf (A and B) and 5 dpf (C–J). Consistent with the olig2 expression patterns described in Figure  3 , the distribution of sox10 -expressing cells at 50 hpf also remained restricted to hindbrain rhombomeres 5 and 6 in disc1aug-injected embryos (B), whereas sox10 -expressing cells were distributed along the entire length of the hindbrain midline in control morpholino-injected embryos (A). In contrast, sox10 expression in the otic vesicle was unaffected by injection of disc1aug morpholino, compare (A) and (B). Expression of the oligodendrocyte differentiation markers plp1b and mbp in the hindbrain was greatly reduced in disc1aug-injected embryos at 5 dpf (D and H), as compared with their widespread expression in the hindbrain of control morpholino-injected sibling embryos (C and G), revealing a profound deficit of differentiated oligodendrocytes in disc1aug-injected embryos. These defects in oligodendrocyte development were p53-independent: Expression of plp1b and mbp in the hindbrain is widespread in a dispersed population of cells at 5 dpf in the hindbrain of embryos injected with a combination of control and p53 morpholinos (E and I), whereas the populations of plp1b and mbp -expressing cells are eliminated in embryos injected with a combination of disc1aug and p53 morpholinos (F and J).

The Neuregulin-ErbB signalling pathway is required for oligodendrocyte development in the zebrafish hindbrain

The dramatic oligodendrocyte specification defect of disc1 morphants is particularly interesting in light of the evidence suggesting that a deficiency of oligodendrocytes accompanies, and indeed may contribute to, the aetiology of schizophrenia ( 24 , 25 ). On this basis, we hypothesized that the zebrafish orthologue of NRG1 might show similar requirements to disc1 for the specification of oligodendrocytes in the developing brain. Analysis of nrg1 expression by whole-mount in situ hybridization using a probe directed against the cytoplasmic tail region found in type I, II and some type III Neuregulin isoforms at 30 hpf showed that nrg1 was expressed throughout the embryonic brain with strongest expression in the hindbrain (Fig.  6 A). At 50 hpf, nrg1 expression was strong in ventral territories of the rhombomeres (Fig.  6 B and C). Transverse sections through the hindbrain at the level of r5 at 50 hpf revealed that the expression domain of nrg1 partially overlapped with that of olig2 in a different way to that of disc1 . Figure  1 I shows that at 50 hpf, olig2 is strongly expressed both in the midline of the hindbrain and in a population of ventrally located cells that extends medio-laterally across the floor of the hindbrain. While disc1 is strongly expressed in the midline of the hindbrain (Fig.  1 H), nrg1 has no detectable expression in midline cells, but is, like olig2 and unlike disc1 , strongly expressed in cells that lie at the base of the hindbrain (Fig.  6 D). In addition, nrg1 is expressed in a column of cells that extends dorsally from the lateral margins of the ventral population of nrg1 -expressing cells on each side of the hindbrain (Fig.  6 D).

Figure 6.

Analysis of embryonic nrg1 expression. Whole-mount in situ hybridization of an nrg1 riboprobe to zebrafish embryos at 30 hpf ( A ) and 50 hpf ( BD ) stages. Lateral views of the head show that nrg1 is expressed uniformly throughout the brain at 30 hpf (A) with strongest expression in patches of cells in the hindbrain. At 50 hpf (B), nrg1 expression is strongest in ventral territories of the hindbrain (arrow). Dorsal view of the hindbrain further illustrating discrete patches of nrg1 expressing cells at 50 hpf, including two arcs of expression extending rostrally through rhombomeres 4 and 5 (C). Transverse sections through the hindbrain at the level of r5 at 50 hpf (D) demonstrate that nrg1 is prominently expressed in ventrally located cells that extend medio-laterally across the floor of the hindbrain (arrow), as well as in a column of cells that extends dorsally from the lateral margins of the ventral population of nrg1 -expressing cells on each side of the hindbrain.

Figure 6.

Analysis of embryonic nrg1 expression. Whole-mount in situ hybridization of an nrg1 riboprobe to zebrafish embryos at 30 hpf ( A ) and 50 hpf ( BD ) stages. Lateral views of the head show that nrg1 is expressed uniformly throughout the brain at 30 hpf (A) with strongest expression in patches of cells in the hindbrain. At 50 hpf (B), nrg1 expression is strongest in ventral territories of the hindbrain (arrow). Dorsal view of the hindbrain further illustrating discrete patches of nrg1 expressing cells at 50 hpf, including two arcs of expression extending rostrally through rhombomeres 4 and 5 (C). Transverse sections through the hindbrain at the level of r5 at 50 hpf (D) demonstrate that nrg1 is prominently expressed in ventrally located cells that extend medio-laterally across the floor of the hindbrain (arrow), as well as in a column of cells that extends dorsally from the lateral margins of the ventral population of nrg1 -expressing cells on each side of the hindbrain.

A previously characterized splice-blocking morpholino that targets the EGF-like domain of all Neuregulin isoforms ( 26 ), and which was recently used to investigate the role of nrg1 in formation of zebrafish dorsal root ganglion neurones ( 27 ), was microinjected into one-cell embryos. Morphant embryos were analyzed by in situ hybridization with probes for olig2 , sox10 , plp1b and mbp in order to determine the effect of reduced nrg1 activity on oligodendrocyte development (Fig.  7 ). Most intriguingly, the effect of nrg1 knock-down on the specification and patterning of olig2 -positive cells at 51 hpf was strikingly similar to the effect of disc1 knock-down. When compared with controls (Fig.  7 A), there was a substantial deficit of olig2 -positive cells in the hindbrain of nrg1 morphants and they were largely restricted to rhombomeres 5 and 6 (Fig.  7 B). Thus, nrg1 promotes oligodendrocyte specification in the zebrafish hindbrain and may perform a function closely related to that of disc1 in this process. Both secreted and membrane-bound forms of Neuregulin 1 elicit their effects by binding to members of the ErbB family of receptor tyrosine kinases and thus activating the ErbB intracellular signalling pathway. In order to test whether the effects on specification and patterning of olig2 -positive cells were due to inhibition of Neuregulin 1 signalling via members of the ErbB receptor tyrosine kinase family, uninjected wild-type embryos were exposed continuously to a chemical ErbB inhibitor at a concentration of 25 µM from 32–52 hpf. Whole-mount RNA in situ hybridization of drug-treated embryos and control DMSO-treated siblings revealed that inhibition of ErbB signalling elicited a considerable reduction in the number of olig2 -positive cells in the hindbrain (Fig.  7 D), as was observed in response to knock-down of nrg1 or disc1 , in comparison to the situation in control embryos that were exposed to DMSO alone (Fig.  7 C). Analysis with a second marker for oligodendrocyte precursors ( sox10 ), demonstrated that sox10 -expressing cells were restricted to hindbrain rhombomeres 5 and 6 in nrg1 morphants at 50 hpf (Fig.  7 F), whereas in control morpholino-injected embryos, they were distributed more extensively along the antero-posterior axis of the hindbrain (Fig.  7 E). In situ hybridization analysis of nrg1 and control morphants for plp1b and mbp expression at 5 dpf revealed a near total loss of plp1b - and mbp -positive cells in the hindbrain (Fig.  7 G–J), just as was seen in disc1 morphants (Fig.  5 ). RT–PCR analysis with primers spanning exons 2–6 of nrg1 demonstrated that the dose of nrg1i3e4 morpholino used caused near total loss of correctly processed nrg1 mRNA in the morphant embryos ( Supplementary Material, Fig. S1 ).

Figure 7.

The Neuregulin-ErbB signalling pathway is required for oligodendrocyte development in the zebrafish hindbrain. Whole-mount in situ hybridization with olig2 ( AD ), sox10 ( E and F ), plp1b ( G and H ) and mbp ( I and J ) riboprobes on uninjected (A), DMSO-treated (C), control morpholino- (E, G and I), nrg1i3e4 morpholino-injected (B, F, H and J) and ErbB inhibitor treated (D) embryos fixed at 51 hpf (A and B), 52 hpf (C and D), 50 hpf (E and F) and 5 dpf (G–J). Dorsal views show that expression of olig2 in hindbrain of nrg1 morphants (B) is reduced and restricted to midline cells compared to uninjected controls (A). Constant exposure of wild-type embryos to 25 µM N -(4-((3-Chloro-4-fluorophenyl)amino)pyrido[3,4-d]pyrimidin-6-yl)2-butynamide (an inhibitor of ErbB receptor tyrosine kinases) in 2% DMSO, from 30–52 hpf (D) reduced the number of hindbrain olig2 -expressing cells and restricted their distribution to the midline, as was observed when either disc1 or nrg1 was knocked down, and these effects were not observed in control embryos exposed to DMSO alone (C). The distribution of sox10 -expressing cells was restricted to hindbrain rhombomeres 5 and 6 in nrg1i3e4-injected embryos at 50 hpf (F), whereas sox10 -expressing cells were distributed along the entire length of the hindbrain midline in control morpholino-injected embryos (E). Expression of plp1b (H) and mbp (J) was also substantially reduced in the hindbrain of nrg1 morphants, as compared with control morpholino-injected embryos (G and I), indicating a severe impairment of oligodendrocyte development.

Figure 7.

The Neuregulin-ErbB signalling pathway is required for oligodendrocyte development in the zebrafish hindbrain. Whole-mount in situ hybridization with olig2 ( AD ), sox10 ( E and F ), plp1b ( G and H ) and mbp ( I and J ) riboprobes on uninjected (A), DMSO-treated (C), control morpholino- (E, G and I), nrg1i3e4 morpholino-injected (B, F, H and J) and ErbB inhibitor treated (D) embryos fixed at 51 hpf (A and B), 52 hpf (C and D), 50 hpf (E and F) and 5 dpf (G–J). Dorsal views show that expression of olig2 in hindbrain of nrg1 morphants (B) is reduced and restricted to midline cells compared to uninjected controls (A). Constant exposure of wild-type embryos to 25 µM N -(4-((3-Chloro-4-fluorophenyl)amino)pyrido[3,4-d]pyrimidin-6-yl)2-butynamide (an inhibitor of ErbB receptor tyrosine kinases) in 2% DMSO, from 30–52 hpf (D) reduced the number of hindbrain olig2 -expressing cells and restricted their distribution to the midline, as was observed when either disc1 or nrg1 was knocked down, and these effects were not observed in control embryos exposed to DMSO alone (C). The distribution of sox10 -expressing cells was restricted to hindbrain rhombomeres 5 and 6 in nrg1i3e4-injected embryos at 50 hpf (F), whereas sox10 -expressing cells were distributed along the entire length of the hindbrain midline in control morpholino-injected embryos (E). Expression of plp1b (H) and mbp (J) was also substantially reduced in the hindbrain of nrg1 morphants, as compared with control morpholino-injected embryos (G and I), indicating a severe impairment of oligodendrocyte development.

The striking similarity in the effect of disc1 and nrg1 knock-down on olig2 expression suggests that they may function in the same or related pathways controlling oligodendrocyte development. In order to begin to address this we investigated the combined effect of disc1 and nrg1 morpholinos on development. However, co-injection of either high doses (0.6 pmol disc1aug + 0.2 pmol nrg1e3i4) or low doses (0.3 pmol disc1aug + 0.1 pmol nrg1e3i4) of disc1 and nrg1 morpholinos caused a severe impairment of embryonic development by 24 hpf and embryos failed to survive to 48 hpf, when oligodendrocytes can be investigated by in situ hybridization (not shown), so we were unable to draw conclusions as to their combined effect on oligodendrocyte development.

Specification of olig2 -expressing cerebellar neurones requires both disc1 and nrg1

It has recently been shown that olig2 marks a population of cerebellar neurones with features of eurydendroid neurones in the zebrafish brain ( 28 ). While analyzing olig2 expression in disc1 and nrg1 morphants it became apparent that morphants for either gene exhibited a dramatic loss of olig2 expression in the developing cerebellum (results not shown). In order to investigate these effects further, disc1aug and nrg1i3e4 morpholinos were microinjected into Tg ( olig2:egfp ) embryos and EGFP fluorescence at 50 hpf was analyzed using confocal microscopy. While two semicircular arcs of EGFP-fluorescent cells were clearly seen in the cerebellum of control morpholino-injected embryos (Fig.  8 A), these cells were almost completely eliminated in both disc1aug- (Fig.  8 C) and nrg1i3e4-injected (Fig.  8 E) embryos. Consistent with the results shown in Figures  3 and 7 , Tg ( olig2:egfp ) - expressing cells were confined to the midline region of rhombomeres 5 and 6 of the hindbrain in disc1 (Fig.  8 C) and nrg1 (Fig.  8 E) morphants, whereas in control embryos, Tg ( olig2:egfp ) - expressing cells were found throughout the midline of the hindbrain, both anterior and posterior to rhombomeres 5 and 6 and these cells were also dispersed more laterally (Fig.  8 A). In the midbrain of disc1 and nrg1 morphant embryos, Tg ( olig2:egfp ) - expressing cells were much more restricted in their distribution than in control morpholino-injected embryos, but there is a paucity of data concerning the fate and function of these cells.

Figure 8.

Disc1 and nrg1 are required for correct specification of olig2 -positive cerebellar neurones but not spinal motor neurones. Confocal imaging of GFP fluorescence in olig2:gfp transgenic embryos at 50 hpf illustrates a near total loss of GFP-expressing cells in the cerebellum, which are a subset of neurones reported to have similarities to eurydendroid neurones ( 28 ). These cells are seen as two semicircular arcs in control morpholino-injected embryos (arrows in A ), but are almost completely absent in disc1aug- ( C ) and nrg1i3e4-injected ( E ) embryos. Consistent with our findings using whole-mount in situ hybridization for olig2 , GFP expression in the posterior hindbrain (bracketed) was mostly confined to midline cells in rhombomeres 5 and 6 in disc1 (C) and nrg1 (E) morphants, compared with control injected embryos (A). Conversely, knock-down of disc1 ( D ) or nrg1 ( F ) had minimal effects on specification of olig2 -positive cells in the spinal cord, although mild defects in spinal motor neurone axon outgrowth are apparent, more so in nrg1 morphants (F) than disc1 morphants (D).

Figure 8.

Disc1 and nrg1 are required for correct specification of olig2 -positive cerebellar neurones but not spinal motor neurones. Confocal imaging of GFP fluorescence in olig2:gfp transgenic embryos at 50 hpf illustrates a near total loss of GFP-expressing cells in the cerebellum, which are a subset of neurones reported to have similarities to eurydendroid neurones ( 28 ). These cells are seen as two semicircular arcs in control morpholino-injected embryos (arrows in A ), but are almost completely absent in disc1aug- ( C ) and nrg1i3e4-injected ( E ) embryos. Consistent with our findings using whole-mount in situ hybridization for olig2 , GFP expression in the posterior hindbrain (bracketed) was mostly confined to midline cells in rhombomeres 5 and 6 in disc1 (C) and nrg1 (E) morphants, compared with control injected embryos (A). Conversely, knock-down of disc1 ( D ) or nrg1 ( F ) had minimal effects on specification of olig2 -positive cells in the spinal cord, although mild defects in spinal motor neurone axon outgrowth are apparent, more so in nrg1 morphants (F) than disc1 morphants (D).

Interestingly, while disc1 and nrg1 appear to have multiple roles in the specification of olig2 -positive cells throughout the brain, there was no apparent loss of EGFP-fluorescent cells in the spinal cord of Tg ( olig2:egfp ) embryos injected with disc1 (Fig.  8 D) or nrg1 (Fig.  8 F) morpholinos compared with controls (Fig.  8 B), where olig2 expression is a marker of both oligodendrocyte precursors and motor neurones. However, subtle defects in motor neurone axon outgrowth, which included shorter axons and minor pathfinding errors, were observed in both disc1 and nrg1 morphant embryos (Fig.  8 B, D and F).

DISCUSSION

Our studies on the neurodevelopmental functions of the orthologues of key schizophrenia susceptibility genes in the zebrafish embryo have revealed several important novel findings. We show that disc1 is required for development of oligodendrocytes and cerebellar neurones in the hindbrain by promoting specification of olig2 -positive precursors. In addition, we find that nrg1 function and ErbB signalling are also required for development of hindbrain oligodendrocytes through a similar mechanism involving expansion of the pool of olig2 -positive precursors. Hence, we conclude that disc1 and ErbB-mediated Nrg1 signalling are essential for the development of multiple neural cell types that are derived from olig2- expressing precursor cells.

RT–PCR analysis demonstrated that the strong phenotypes observed in embryos microinjected with either disc1e2i2 or nrg1 morpholinos result from near total elimination of disc1 and nrg1 mRNAs in the embryo. Importantly, partial disc1 knock-down, achieved by microinjecting low doses of either disc1aug or disc1e2i2 morpholinos, caused less severe defects in the development of olig2 -expressing cells, demonstrating close relationships between the levels of mature disc1 and nrg1 mRNAs and the number of olig2 -positive cells in the hindbrain. Recent studies in mice demonstrate that in addition to oligodendrocytes and motor neurones, olig2 -positive progenitors can give rise to astrocytes and GABAergic, cholinergic and glutamatergic neurones ( 29 ). Therefore, if the neurodevelopmental functions of disc1 and nrg1 that we describe are conserved in higher vertebrates, then even subtle functional genetic variations in DISC1 and NRG1 might be predicted to influence the development of several different neural cell types within the CNS, with complex phenotypic effects. Additionally, development of neural cell types that are not derived from olig2 lineages will most likely also be affected by functional variations in DISC1 and NRG1 .

Our results show that knock-down of disc1 does not prevent establishment of the initial domain of olig2 expression in the midline of rhombomeres 5 and 6, but disc1 knock-down does inhibit expansion of this domain both along the antero-posterior axis of the hindbrain midline and laterally as oligodendrocytes become dispersed within the brain. These observations suggest that disc1 is required to maintain production of some of the olig2- expressing cells in the developing brain. Our initial studies have focused on hindbrain oligodendrocyte development since we observed prominent disc1 expression in a pattern consistent with that of oligodendrocyte precursors in the hindbrain, but there is also a paucity of data regarding the development of oligodendrocytes in more anterior regions of the zebrafish brain. However, it is apparent that reduced disc1 activity has a profound effect on the distribution of olig2 -positive cells in the forebrain and midbrain. Therefore, it will be important to investigate both the specification of oligodendrocytes in the zebrafish midbrain and forebrain, and to elucidate the requirements for disc1 in development of these brain regions.

The expression of Disc1 in the developing mouse brain has been studied using both in situ hybridization and immunocytochemistry ( 30 , 31 ). In these studies, Disc1 expression during development was shown to be most prominent in the hippocampus, the zebrafish equivalent of which has not been definitively identified, and the thalamus. In the zebrafsh, disc1 is strongly expressed in the midbrain at 26 hpf (Fig.  1 A and C). There are no reports of Disc1 expression in the mouse posterior hindbrain, so it remains to be seen whether the role we describe for disc1 in oligodendrocyte development is conserved in rodents. However, roles for Disc1 in both embryonic ( 30 , 32 ) and adult ( 13 ) neurogenesis have been postulated. We observed diffuse disc1 expression throughout the developing zebrafish brain at 26–50 hpf and consistent with a role in neurogenesis, reduced expression of the neurogenesis marker ash1b was observed in the brain of disc1 morphants ( Supplementary Material, Fig. S2 ). Hence disc1 is likely to have other roles in neural development in addition to the role we describe in the specification of olig2 -positive cells.

A growing body of evidence indicates that schizophrenia is a neurodevelopmental disorder with multiple subtle histopathological and anatomical brain abnormalities. Compelling molecular genetic evidence implicates DISC1 in the pathogenesis of this disorder, and several recent functional studies in mice have strengthened this view ( 8–13 ). The known interaction partners of DISC1 implicate this protein in neuronal morphogenesis and function ( 33–37 ), and inhibition of Disc1 during development using RNAi severely impairs the development and functions of cortical neurons in mice ( 12 ). While abnormalities of neuronal development and function characterize the brains of both schizophrenia patients and Disc1 mutant mice, substantial evidence indicates that myelinating glial cell functions are also compromised in schizophrenia ( 2 , 24 , 25 , 38–41 ). Our studies of disc1 in the developing zebrafish embryo indicate that disc1 is both abundantly expressed in regions of the hindbrain where olig2 -positive cells are born and required for the development of oligodendrocytes in the brain, raising the possibility that in people with schizophrenia, compromised DISC1 function disrupts oligodendrocyte development, leading to the myelination defects that have been previously documented. It will now be particularly interesting to investigate whether white matter abnormalities are a feature of the Scottish DISC1 pedigree ( 42 ).

While disc1 is expressed particularly strongly in midline oligodendrocyte precursors of the hindbrain prior to their terminal differentiation in both medial and lateral locations, we also found that disc1 is expressed at a low level throughout the brain, consistent with its known functions in neurones, and our results indicate that disc1 is required for specification of a population of olig2 -positive cerebellar neurones ( 28 ). However, disc1 is most abundantly expressed in the cranial neural crest-derived viscerocranial cartilages, and disc1 knock-down severely disrupted their development. This may be of particular interest given that a high prevalence of minor facial abnormalities is associated with schizophrenia ( 43 , 44 ), but it must be borne in mind that such abnormalities have not been reported in the Scottish DISC1 family ( 42 ).

Of the many candidate susceptibility genes that have been linked to schizophrenia, NRG1 is one of the most promising based on both genetic ( 14 , 15 ) and neurobiological ( 17 ) evidence. Similarly, there is some evidence that the gene for one of the Neuregulin 1 receptors, ERBB4 , is a candidate susceptibility gene for schizophrenia ( 45 ), and a more recent study suggests that multiple interactions between genes in the Neuregulin-ErbB signalling network impart increased susceptibility to schizophrenia ( 46 ). Our results demonstrate that both genetic and chemical inhibition of Neuregulin-ErbB signalling had strikingly similar effects to that elicited by disc1 knock-down on the number and distribution of olig2 -expressing cells in the brain. These results raise the possibility that disc1 and nrg1 may function in the same or related pathways controlling the proliferation of olig2 -positive neural precursors in the brain. There is evidence that OLIG2 is a schizophrenia-susceptibility gene, both alone and in combination with other genes related to oligodendroglial development, namely CNP and ERBB4 ( 41 ). However, no interaction effect on disease risk was found between OLIG2 and NRG1 . It is unclear as to why ERBB4 , a receptor for NRG1 , should show genetic interaction with OLIG2 whereas NRG1 itself did not, while association between OLIG2 and DISC1 has not been tested. Therefore, further genetic association studies could be undertaken to explore whether interactions between DISC1 , OLIG2 , NRG1, ERBB4 and downstream signalling molecules such as AKT1 confer increased susceptibility to schizophrenia. Future studies will continue to exploit the potential of the zebrafish to elucidate further the roles of Disc1 and Neuregulin-ErbB signalling pathways in early brain development and to generate new experimental systems in which neurodevelopmental mechanisms and genetic pathways of relevance to schizophrenia can be dissected.

MATERIALS AND METHODS

Zebrafish maintenance

Wild-type AB zebrafish and olig2:egfp ( 47 ) transgenic embryos (kindly provided by Dr B. Appel, Vanderbilt University) were used for this study and embryos were obtained by natural mating.

Zebrafish disc1 cloning and sequencing

The zebrafish disc1 orthologue is partially annotated on zebrafish chromosome 13 as gene Q8AV88_DANRE (Ensembl Gene ID ENSDARG00000021895). The annotated gene lacked the 5′ end of the sequence and BLAST searches were used to identify EST clone CR928719, encoding the 5′ end of the gene. RT–PCR was performed using Herculase enhanced DNA polymerase (Stratagene) to amplify a 3190 bp cDNA encoding the full coding DNA sequence of disc1 . The sequences of the forward and reverse primers used were 5′-gcgtcgcgtttggagttg-3′ and 5′-cttgtgcgcatcagttttcat-3′, respectively. The 3190 bp product was cloned into pCRII-TOPO and sequenced at the Genetics Core Facility, School of Medicine and Biomedical Sciences, University of Sheffield.

Microinjection of morpholinos

Morpholino antisense oligonucleotides (MO) against disc1 were designed by Gene Tools, LLC (Philomath, OR, USA). MO against the initiating AUG and exon 2-intron 2 splice donor site of disc1 were used. A characterized morpholino that targets the EGF-like domain essential for function of all neuregulin isoforms ( 26 ) was purchased from Open Biosystems (Huntsville, AL, USA). A standard control morpholino (CoMO) that targets human β-globin pre-mRNA and a zebrafish p53 MO which suppresses apoptosis caused by some MO ( 22 ) were also purchased from Gene Tools, LLC. Morpholino stocks (1–2 m m in water) were diluted to 0.1 m m (nrg1e3i4), 0.15 or 0.3 m m (disc1aug) and 0.3 or 0.6 m m (disc1e2i2) respectively with water then 2 nl microinjected with phenol red into one-cell embryos. Embryos were maintained at 28.5°C in E3 medium and dechorionated with fine forceps prior to fixation. The sequences of the MO were as follows:

disc1aug: 5′-CCTGACCATTCCTGCGAACATCATG-3′

disc1e2i2: 5′-TTGTTGTTTAGTTCTCACCTTGGTG-3′

nrg1e3i4: 5′- TGCTGGTGGCTGCTGCACAGAGGAA-3′

CoMO: 5′-CCTCTTACCTCAGTTACAATTTATA 3′

p53MO: 5′-GCGCCATTGCTTTGCAAGAATTG-3′.

RNA in situ hybridization

Digoxigenin-labelled probes were prepared using SP6 and T7 polymerases as recommended by the manufacturer (Roche). Sense and antisense disc1 riboprobes were generated from the pCRII-TOPO-DISC1-FL clone as described earlier, nrg1 and mbp riboprobes were generated from IMAGE clones 8339257 and 4955400, respectively, and previously described probes included plp1b ( 23 ), sox10 ( 48 ) and olig2 ( 21 ). Whole-mount in situ hybridization was performed using standard procedures ( 49 ).

Histological analysis

Cartilages were stained with Alcian blue, a dye that stains the extracellular matrix associated with chondrocytes, and preparations were made as previously described ( 50 ).

ErbB inhibitor treatment

A 1 m m stock solution of the EGFR/ErbB2/ErbB4 inhibitor N -(4-((3-Chloro-4-fluorophenyl)amino)pyrido[3,4-d]pyrimidin-6-yl)2-butynamide (Calbiochem 324840) was prepared in DMSO. The drug in DMSO and DMSO alone were added directly to E3 medium containing embryos at 30–32 hpf and embryos exposed to the drug continuously prior to fixation at 50–53 hpf.

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online .

FUNDING

Funding from the MRC to support the Centre for Developmental and Biomedical Genetics (Grants G0400100 and G0700091 to P.I.) is gratefully acknowledged. The core BMS-MBB Light Microscopy Facility was funded by a grant from The Wellcome Trust. C.A.R. is supported by a NARSAD Distinguished Investigator award. Research in the laboratory of V.T.C. is supported by the Wellcome Trust and the Medical Research Council.

ACKNOWLEDGEMENTS

We thank Professor Philip Ingham for support and encouragement. We are also grateful to Dr Tanya Whitfield, Dr Henry Roehl, Dr Andrew Grierson and Professor Philip Ingham for sharing fish, experimental materials and other reagents, and to Lisa Gleadall, Susanne Surfleet and Matthew Green for fish care.

Conflict of Interest statement . None declared.

REFERENCES

1
Akbarian
S.
Kim
J.J.
Potkin
S.G.
Hetrick
W.P.
Bunney
W.E.
Jr
Jones
E.G.
Maldistribution of interstitial neurons in prefrontal white matter of the brains of schizophrenic patients
Arch. Gen. Psychiatry
 , 
1996
, vol. 
53
 (pg. 
425
-
436
)
2
Flynn
S.W.
Lang
D.J.
Mackay
A.L.
Goghari
V.
Vavasour
I.M.
Whittall
K.P.
Smith
G.N.
Arango
V.
Mann
J.J.
Dwork
A.J.
, et al.  . 
Abnormalities of myelination in schizophrenia detected in vivo with MRI, and post-mortem with analysis of oligodendrocyte proteins
Mol. Psychiatry
 , 
2003
, vol. 
8
 (pg. 
811
-
820
)
3
Kanaan
R.A.
Kim
J.S.
Kaufmann
W.E.
Pearlson
G.D.
Barker
G.J.
McGuire
P.K.
Diffusion tensor imaging in schizophrenia
Biol. Psychiatry
 , 
2005
, vol. 
58
 (pg. 
921
-
929
)
4
Kubicki
M.
Westin
C.F.
McCarley
R.W.
Shenton
M.E.
The application of DTI to investigate white matter abnormalities in schizophrenia
Ann. N. Y. Acad. Sci.
 , 
2005
, vol. 
1064
 (pg. 
134
-
148
)
5
Arnold
S.E.
Talbot
K.
Hahn
C.G.
Neurodevelopment, neuroplasticity, and new genes for schizophrenia
Prog. Brain Res.
 , 
2005
, vol. 
147
 (pg. 
319
-
345
)
6
Rapoport
J.L.
Addington
A.M.
Frangou
S.
Psych
M.R.
The neurodevelopmental model of schizophrenia: update 2005
Mol. Psychiatry
 , 
2005
, vol. 
10
 (pg. 
434
-
449
)
7
Millar
J.K.
Wilson-Annan
J.C.
Anderson
S.
Christie
S.
Taylor
M.S.
Semple
C.A.
Devon
R.S.
Clair
D.M.
Muir
W.J.
Blackwood
D.H.
, et al.  . 
Disruption of two novel genes by a translocation co-segregating with schizophrenia
Hum. Mol. Genet.
 , 
2000
, vol. 
9
 (pg. 
1415
-
1423
)
8
Clapcote
S.J.
Lipina
T.V.
Millar
J.K.
Mackie
S.
Christie
S.
Ogawa
F.
Lerch
J.P.
Trimble
K.
Uchiyama
M.
Sakuraba
Y.
, et al.  . 
Behavioral phenotypes of Disc1 missense mutations in mice
Neuron
 , 
2007
, vol. 
54
 (pg. 
387
-
402
)
9
Hikida
T.
Jaaro-Peled
H.
Seshadri
S.
Oishi
K.
Hookway
C.
Kong
S.
Wu
D.
Xue
R.
Andrade
M.
Tankou
S.
, et al.  . 
Dominant-negative DISC1 transgenic mice display schizophrenia-associated phenotypes detected by measures translatable to humans
Proc. Natl. Acad. Sci. USA
 , 
2007
, vol. 
104
 (pg. 
14501
-
14506
)
10
Li
W.
Zhou
Y.
Jentsch
J.D.
Brown
R.A.
Tian
X.
Ehninger
D.
Hennah
W.
Peltonen
L.
Lonnqvist
J.
Huttunen
M.O.
, et al.  . 
Specific developmental disruption of disrupted-in-schizophrenia-1 function results in schizophrenia-related phenotypes in mice
Proc. Natl. Acad. Sci. USA
 , 
2007
, vol. 
104
 (pg. 
18280
-
18285
)
11
Pletnikov
M.V.
Ayhan
Y.
Nikolskaia
O.
Xu
Y.
Ovanesov
M.V.
Huang
H.
Mori
S.
Moran
T.H.
Ross
C.A.
Inducible expression of mutant human DISC1 in mice is associated with brain and behavioral abnormalities reminiscent of schizophrenia
Mol. Psychiatry
 , 
2008
, vol. 
13
 (pg. 
173
-
186
)
12
Kamiya
A.
Kubo
K.
Tomoda
T.
Takaki
M.
Youn
R.
Ozeki
Y.
Sawamura
N.
Park
U.
Kudo
C.
Okawa
M.
, et al.  . 
A schizophrenia-associated mutation of DISC1 perturbs cerebral cortex development
Nat. Cell Biol.
 , 
2005
, vol. 
7
 (pg. 
1167
-
1178
)
13
Duan
X.
Chang
J.H.
Ge
S.
Faulkner
R.L.
Kim
J.Y.
Kitabatake
Y.
Liu
X.B.
Yang
C.H.
Jordan
J.D.
Ma
D.K.
, et al.  . 
Disrupted-in-schizophrenia 1 regulates integration of newly generated neurons in the adult brain
Cell
 , 
2007
, vol. 
130
 (pg. 
1146
-
1158
)
14
Stefansson
H.
Sigurdsson
E.
Steinthorsdottir
V.
Bjornsdottir
S.
Sigmundsson
T.
Ghosh
S.
Brynjolfsson
J.
Gunnarsdottir
S.
Ivarsson
O.
Chou
T.T.
, et al.  . 
Neuregulin 1 and susceptibility to schizophrenia
Am. J. Hum. Genet.
 , 
2002
, vol. 
71
 (pg. 
877
-
892
)
15
Li
D.
Collier
D.A.
He
L.
Meta-analysis shows strong positive association of the neuregulin 1 (NRG1) gene with schizophrenia
Hum. Mol. Genet.
 , 
2006
, vol. 
15
 (pg. 
1995
-
2002
)
16
Buonanno
A.
Fischbach
G.D.
Neuregulin and ErbB receptor signaling pathways in the nervous system
Curr. Opin. Neurobiol.
 , 
2001
, vol. 
11
 (pg. 
287
-
296
)
17
Fischbach
G.D.
NRG1 and synaptic function in the CNS
Neuron
 , 
2007
, vol. 
54
 (pg. 
495
-
497
)
18
Vartanian
T.
Fischbach
G.
Miller
R.
Failure of spinal cord oligodendrocyte development in mice lacking neuregulin
Proc. Natl. Acad. Sci. USA
 , 
1999
, vol. 
96
 (pg. 
731
-
735
)
19
Falls
D.L.
Neuregulins: functions, forms, and signaling strategies
Exp. Cell Res.
 , 
2003
, vol. 
284
 (pg. 
14
-
30
)
20
Lupas
A.
Van Dyke
M.
Stock
J.
Predicting coiled coils from protein sequences
Science
 , 
1991
, vol. 
252
 (pg. 
1162
-
1164
)
21
Cunliffe
V.T.
Casaccia-Bonnefil
P.
Histone deacetylase 1 is essential for oligodendrocyte specification in the zebrafish CNS
Mech. Dev.
 , 
2006
, vol. 
123
 (pg. 
24
-
30
)
22
Robu
M.E.
Larson
J.D.
Nasevicius
A.
Beiraghi
S.
Brenner
C.
Farber
S.A.
Ekker
S.C.
p53 activation by knockdown technologies
PLoS Genet.
 , 
2007
, vol. 
3
 pg. 
e78
 
23
Brosamle
C.
Halpern
M.E.
Characterization of myelination in the developing zebrafish
Glia
 , 
2002
, vol. 
39
 (pg. 
47
-
57
)
24
Hof
P.R.
Haroutunian
V.
Copland
C.
Davis
K.L.
Buxbaum
J.D.
Molecular and cellular evidence for an oligodendrocyte abnormality in schizophrenia
Neurochem. Res.
 , 
2002
, vol. 
27
 (pg. 
1193
-
2000
)
25
Uranova
N.
Orlovskaya
D.
Vikhreva
O.
Zimina
I.
Kolomeets
N.
Vostrikov
V.
Rachmanova
V.
Electron microscopy of oligodendroglia in severe mental illness
Brain Res. Bull.
 , 
2001
, vol. 
55
 (pg. 
597
-
610
)
26
Milan
D.J.
Giokas
A.C.
Serluca
F.C.
Peterson
R.T.
MacRae
C.A.
Notch1b and neuregulin are required for specification of central cardiac conduction tissue
Development
 , 
2006
, vol. 
133
 (pg. 
1125
-
1132
)
27
Honjo
Y.
Kniss
J.
Eisen
J.S.
Neuregulin-mediated ErbB3 signaling is required for formation of zebrafish dorsal root ganglion neurons
Development
 , 
2008
, vol. 
135
 (pg. 
2615
-
2625
)
28
McFarland
K.A.
Topczewska
J.M.
Weidinger
G.
Dorsky
R.I.
Appel
B.
Hh and Wnt signaling regulate formation of olig2(+) neurons in the zebrafish cerebellum
Dev. Biol.
 , 
2008
, vol. 
318
 (pg. 
162
-
171
)
29
Ono
K.
Takebayashi
H.
Ikeda
K.
Furusho
M.
Nishizawa
T.
Watanabe
K.
Ikenaka
K.
Regional- and temporal-dependent changes in the differentiation of Olig2 progenitors in the forebrain, and the impact on astrocyte development in the dorsal pallium
Dev. Biol.
 , 
2008
, vol. 
320
 (pg. 
456
-
468
)
30
Austin
C.P.
Ky
B.
Ma
L.
Morris
J.A.
Shughrue
P.J.
Expression of Disrupted-in-schizophrenia-1, a schizophrenia-associated gene, is prominent in the mouse hippocampus throughout brain development
Neuroscience
 , 
2004
, vol. 
124
 (pg. 
3
-
10
)
31
Meyer
K.D.
Morris
J.A.
Immunohistochemical analysis of Disc1 expression in the developing and adult hippocampus
Gene Expr. Patterns
 , 
2008
, vol. 
8
 (pg. 
494
-
501
)
32
Schurov
I.L.
Handford
E.J.
Brandon
N.J.
Whiting
P.J.
Expression of disrupted in schizophrenia 1 (DISC1) protein in the adult and developing mouse brain indicates its role in neurodevelopment
Mol. Psychiatry
 , 
2004
, vol. 
9
 (pg. 
1100
-
1110
)
33
Brandon
N.J.
Handford
E.J.
Schurov
I.
Rain
J.C.
Pelling
M.
Duran-Jimeniz
B.
Camargo
L.M.
Oliver
K.R.
Beher
D.
Shearman
M.S.
, et al.  . 
Disrupted in schizophrenia 1 and Nudel form a neurodevelopmentally regulated protein complex: implications for schizophrenia and other major neurological disorders
Mol. Cell. Neurosci.
 , 
2004
, vol. 
25
 (pg. 
42
-
55
)
34
Millar
J.K.
Pickard
B.S.
Mackie
S.
James
R.
Christie
S.
Buchanan
S.R.
Malloy
M.P.
Chubb
J.E.
Huston
E.
Baillie
G.S.
, et al.  . 
DISC1 and PDE4B are interacting genetic factors in schizophrenia that regulate cAMP signaling
Science
 , 
2005
, vol. 
310
 (pg. 
1187
-
1191
)
35
Miyoshi
K.
Honda
A.
Baba
K.
Taniguchi
M.
Oono
K.
Fujita
T.
Kuroda
S.
Katayama
T.
Tohyama
M.
Disrupted-in-schizophrenia 1, a candidate gene for schizophrenia, participates in neurite outgrowth
Mol. Psychiatry
 , 
2003
, vol. 
8
 (pg. 
685
-
694
)
36
Morris
J.A.
Kandpal
G.
Ma
L.
Austin
C.P.
DISC1 (Disrupted-in-schizophrenia 1) is a centrosome-associated protein that interacts with MAP1A, MIPT3, ATF4/5 and NUDEL: regulation and loss of interaction with mutation
Hum. Mol. Genet.
 , 
2003
, vol. 
12
 (pg. 
1591
-
1608
)
37
Ozeki
Y.
Tomoda
T.
Kleiderlein
J.
Kamiya
A.
Bord
L.
Fujii
K.
Okawa
M.
Yamada
N.
Hatten
M.E.
Snyder
S.H.
, et al.  . 
Disrupted-in-schizophrenia-1 (DISC-1): mutant truncation prevents binding to NudE-like (NUDEL) and inhibits neurite outgrowth
Proc. Natl. Acad. Sci. USA
 , 
2003
, vol. 
100
 (pg. 
289
-
294
)
38
Hakak
Y.
Walker
J.R.
Li
C.
Wong
W.H.
Davis
K.L.
Buxbaum
J.D.
Haroutunian
V.
Fienberg
A.A.
Genome-wide expression analysis reveals dysregulation of myelination-related genes in chronic schizophrenia
Proc. Natl. Acad. Sci. USA
 , 
2001
, vol. 
98
 (pg. 
4746
-
4751
)
39
Tkachev
D.
Mimmack
M.L.
Ryan
M.M.
Wayland
M.
Freeman
T.
Jones
P.B.
Starkey
M.
Webster
M.J.
Yolken
R.H.
Bahn
S.
Oligodendrocyte dysfunction in schizophrenia and bipolar disorder
Lancet
 , 
2003
, vol. 
362
 (pg. 
798
-
805
)
40
Roy
K.
Murtie
J.C.
El-Khodor
B.F.
Edgar
N.
Sardi
S.P.
Hooks
B.M.
Benoit-Marand
M.
Chen
C.
Moore
H.
O'Donnell
P.
, et al.  . 
Loss of erbB signaling in oligodendrocytes alters myelin and dopaminergic function, a potential mechanism for neuropsychiatric disorders
Proc. Natl. Acad. Sci. USA
 , 
2007
, vol. 
104
 (pg. 
8131
-
8136
)
41
Georgieva
L.
Moskvina
V.
Peirce
T.
Norton
N.
Bray
N.J.
Jones
L.
Holmans
P.
Macgregor
S.
Zammit
S.
Wilkinson
J.
, et al.  . 
Convergent evidence that oligodendrocyte lineage transcription factor 2 (OLIG2) and interacting genes influence susceptibility to schizophrenia
Proc. Natl. Acad. Sci. USA
 , 
2006
, vol. 
103
 (pg. 
12469
-
12474
)
42
Blackwood
D.H.
Fordyce
A.
Walker
M.T.
St Clair
D.M.
Porteous
D.J.
Muir
W.J.
Schizophrenia and affective disorders–cosegregation with a translocation at chromosome 1q42 that directly disrupts brain-expressed genes: clinical and P300 findings in a family
Am. J. Hum. Genet.
 , 
2001
, vol. 
69
 (pg. 
428
-
433
)
43
Scutt
L.E.
Chow
E.W.
Weksberg
R.
Honer
W.G.
Bassett
A.S.
Patterns of dysmorphic features in schizophrenia
Am. J. Med. Genet.
 , 
2001
, vol. 
105
 (pg. 
713
-
723
)
44
Sivkov
S.T.
Akabaliev
V.H.
Minor physical anomalies in schizophrenic patients and normal controls
Psychiatry
 , 
2003
, vol. 
66
 (pg. 
222
-
233
)
45
Norton
N.
Moskvina
V.
Morris
D.W.
Bray
N.J.
Zammit
S.
Williams
N.M.
Williams
H.J.
Preece
A.C.
Dwyer
S.
Wilkinson
J.C.
, et al.  . 
Evidence that interaction between neuregulin 1 and its receptor erbB4 increases susceptibility to schizophrenia
Am. J. Med. Genet. B Neuropsychiatr. Genet.
 , 
2006
, vol. 
141
 (pg. 
96
-
101
)
46
Benzel
I.
Bansal
A.
Browning
B.L.
Galwey
N.W.
Maycox
P.R.
McGinnis
R.
Smart
D.
St Clair
D.
Yates
P.
Purvis
I.
Interactions among genes in the ErbB-Neuregulin signalling network are associated with increased susceptibility to schizophrenia
Behav. Brain Funct.
 , 
2007
, vol. 
3
 pg. 
31
 
47
Shin
J.
Park
H.C.
Topczewska
J.M.
Mawdsley
D.J.
Appel
B.
Neural cell fate analysis in zebrafish using olig2 BAC transgenics
Methods Cell Sci.
 , 
2003
, vol. 
25
 (pg. 
7
-
14
)
48
Dutton
K.A.
Pauliny
A.
Lopes
S.S.
Elworthy
S.
Carney
T.J.
Rauch
J.
Geisler
R.
Haffter
P.
Kelsh
R.N.
Zebrafish colourless encodes sox10 and specifies non-ectomesenchymal neural crest fates
Development
 , 
2001
, vol. 
128
 (pg. 
4113
-
4125
)
49
Westerfield
M.
The Zebrafish Book. A Guide for the Laboratory Use of Zebrafish (Danio rerio)
2000
4th edn
Eugene
Univ. of Oregon Press
50
Schilling
T.F.
Piotrowski
T.
Grandel
H.
Brand
M.
Heisenberg
C.P.
Jiang
Y.J.
Beuchle
D.
Hammerschmidt
M.
Kane
D.A.
Mullins
M.C.
, et al.  . 
Jaw and branchial arch mutants in zebrafish I: branchial arches
Development
 , 
1996
, vol. 
123
 (pg. 
329
-
344
)