Embryos

Japanese flounder (Paralichthys olivaceus) embryos at the two cell stage were collected from a hatchery tank at the National Research Institute of Aquaculture, Mie, Japan. They were kept in an incubator set at 20°C.

Disulfiram and α, α′-dipyridyl treatment

A stock solution of disulfiram (5 mM) was prepared in DMSO and added to seawater to give the final concentrations used in the experiments. α, α′-dipyridyl was directly dissolved in seawater. Fifty embryos were cultured in 20 ml of filtered seawater supplemented with the drugs.

Cartilage staining

Cartilages were visualized with 0.1% Alcian blue 8GX dissolved in 1% HCl/70% ethanol. Some of the stained embryos were bleached in 2% H₂O₂ in 1% KOH to remove pigments. Samples were mounted in 80% glycerol.

In situ hybridization

Embryos were fixed in 4% paraformaldehyde in phosphate-buffered saline, pH 7.0, overnight, and kept in methanol at −20°C. The preparation of antisense riboprobes of flounder Hoxd-4 and shh and whole-mount in situ hybridization were conducted as described (Suzuki et al., 1988, 1989).

Effects of disulfiram on Hoxd-4 and shh expressions

The development of pharyngeal cartilage and the expression pattern of Hoxd-4 and shh in flounder embryos are briefly summarized as follows. At the prim-5 stage, 60 hours post-fertilization (hpf) in flounder, cartilage precursor cells have finished their migration from the hindbrain into the pharyngeal arches, and the Hoxd-4 expression domain shows clear anterior borders in the hindbrain and pharyngeal area (Fig. 1 A, B). After this stage, cartilage precursor cells actively divide for 36 h, during which period shh expression begins as spots and widely expands on the pharyngeal endoderm (Fig. 1H). Then, the cartilage matrix begins to be synthesized, and cartilage elements form in the mandibular, hyoid, and 1st–4th gill arches by the open-mouth stage, 130 hpf (Figs. 2A and 3A, B).

Fig. 1

Hoxd-4 and shh expression in normal and disulfiram treated flounder embryos. Embryos in A–G were hybridized with Hoxd-4 antisense riboprobe at 60 hpf when normal embryos reached the prim-5 stage. The anterior borders of the Hoxd-4 expression domain in the hindbrain (h) and pharyngeal area (p) are marked by arrows and arrowheads, respectively. The expression in the pectoral fin (f) is marked by asterisks. A, B: normal embryos. C, D–E: embryos incubated at 22 - 24 hpf (shield stage) with 5 and 50 μM disulfiram, respectively. F, G: embryos incubated with 1 μM disulfiram at 22–60 hpf. The anterior border of the Hoxd 4 expression domain is unclear only in embryos in D, E. O: otic vesicle. Embryos in H and I are hybridized with the shh probe (96 hpf). H: normal embryo.I: embryo incubated with 1 μM disulfiram at 60–96 hpf. shh is expressed in the pharyngeal endoderm (arrows) and pectoral fin buds (arrowheads) at the same level in H and I. Bar, 500 μm in A, C, and E; 500 μm in D and F; 250 μm in B and G; 200 μm in H and I.

Fig. 2

Normal, disulfiram-, and α, α′-dipyridyl-treated embryos after cartilage staining. Embryos were stained at 130 hpf when normal embryos reached the open-mouth stage. A: normal embryo. B, C, D: embryos incubated with 1 μM disulfiram at 22–60, 60–130, and 22–130 hpf, respectively. 22, 60, and 130 hfp correspond with the shield, prim-5, and open-mouth stage, respectively. E: embryo incubated with 5 μM α, α′-dipyridyl at 60–130 hpf. Anomaly of the cartilaginous skeleton is obvious in C, D, and E. Anomaly of the noto-chord is obvious in B and D. Bar, 500 μm

Fig. 3

Cartilaginous skeletons of normal, disulfiram-, and α, α′-dipyridyl-treated embryos. The embryos shown in Fig. 2 A–C and E were bleached, and their craniofacial area was viewed from the lateral (left figures) and ventral (right figures) direction. A, B: normal embryo. C, D: embryo incubated with 1 μM disulfiram at 22–60 hpf. E, F: embryo incubated with 1 μM disulfiram at 60 −130 hpf. G, H: embryo incubated with 5 μM α, α′-dipyridyl at 60–130 hpf. In E–H, bending of cartilage elements is obvious (arrowheads). bb, basibrancial; bh, basihyal; cb, ceratobranchial; ch, ceratohyal; hs, hyosymplectic; m, Meckel's cartilage; pf, pectoral fin plate; pq, palatoquadrate; sc, sclerotic cartilage; t, trabeculae cranii. Bar, 170 μm.

Pulse incubation with RA for 1 hr at the shield stage, 22 hpf, shifts the Hoxd-4 expression border anteriorly, indicating sensitivity of the Hoxd-4 gene to RA at this stage (Suzuki et al., 1998). The inhibitory effect of disulfiram on RA synthesis in vivo appears at a concentration of 1 μM in zebrafish (Marsh-Armstrong et al., 1995). Based on the data, we first incubated flounder embryos with 5 μM disulfiram for 2 hr at the shield stage and examined the Hoxd-4 expression during the hatching period. The pulse exposure did not affect the position of the anterior borders of the Hoxd-4 expression domain in the hindbrain and pharyngeal area, even though the trunk was slightly shortened by the drug (Fig. 1C). When the disulfiram-treated embryos were further developed in normal seawater until the open-mouth stage, the pharyngeal cartilages developed normally (data not shown). Pulse exposure to a high dose (50 μM) of disulfiram strongly truncated the trunk (Fig. 1D, E). In these embryos, the anterior border of the Hoxd-4 expression domain was unclear, but the expression in the pharyngeal area and pectoral fin bud could still be recognized. Since the embryos treated with this concentration of the drug did not survive later than the prim-5 stage, the phenotype of cartilage anomaly could not be observed.

Treatment with 1 μM disulfiram from the 2–3 somite stage to the prim-22 stage induces truncation of the trunk and undulation of the notochord in zebrafish embryos (Marsh-Armstrong et al., 1995). Then, to reveal the effect of continuous exposure to disulfiram during the period from the initiation to the establishment of the Hoxd-4 expression domain, we incubated flounder embryos with 1 μM disulfiram from the shield (22 hpf) to prim-5 stage (60 hpf). As reported in zebrafish, the trunk of the treated embryos was truncated, and the noto-chord was undulated (Fig. 4B). The anterior borders of the Hoxd-4 expression domain were normally established both in the hindbrain and pharyngeal area (Fig. 1F, G). Hoxd-4 expression at the pectoral fin buds also began normally. Thus, Hoxd-4 expression in flounder embryos was insensitive to disulfiram at a concentration which causes anomaly in the notochord. Some of the treated embryos were transferred to normal seawater and developed further, and the structure of cartilages was observed at the open-mouth stage (130 hpf). The trunk was kept truncated, but cartilage elements formed in the mandibular, hyoid, and 1st - 4th gill arches, as in normal embryos (Figs. 2B and 3C, D). We could not detect structural anomalies in the cartilage elements. The pectoral fin cartilages also formed normally. Thus, it was concluded that disulfiram does not affect the establishment of the Hoxd-4 expression domain or cartilage development when given before the prim-5 stage, even though it affects notochord development.

Fig. 4

Effects of disulfiram and α, α′-dipyridyl on the notochord and trunk. All embryos were fixed at 60 hpf when normal embryos reached the prim-5 stage. A: normal embryo. B, C: embryos incubated with 1 μM disulfiram and 5 μM α, α′-dipyridyl, respectively, from the shield stage (22 hpf). In both embryos, the notochord (n) is undulated (arrows), and the trunk is truncated and ventrally curved. Bar, 500 μm.

To observe the effect of disulfiram on shh expression in the pharyngeal area and pectoral fin buds, embryos were incubated with 1 μM disulfiram for 36 hr from the prim-5 stage. In the treated embryos, the shh expression domain in the pharyngeal endoderm was established normally, and the shh expression at the pectoral fin also started normally (Fig. 1I).

Effects of disulfiram and α, α′-dipyridyl on embryogenesis

Because data could not be obtained to support the idea that disulfiram suppresses Hoxd-4 and shh expression in the pharyngeal area, we decided to examine whether disulfiram affects the mechanical stability of cartilage.

In the embryos incubated with 1 μM disulfiram from the prim-5 to the open mouth stage, cartilage elements in the pharyngeal arches were obviously bent (Fig. 2C), as previously reported (Suzuki et al., 2000). Detailed observation showed that cartilage elements existed in the mandibular, hyoid, and four gill bars, as in normal embryos (Fig. 3 E, F). Large cartilaginous elements were all severely bent, including the Meckel's (m) and palatoquadrate (pq) cartilages in the mandibular arches, the ceratohyal (ch) and hyosymplectic (hs) cartilages in the hyoid arches, and the ceratobrancial (cb) and basibranchial (bb) cartilages in the gill arches. Small cartilage elements, such as the basihyal (bh) in the hyoid arch, were hard to identify due to severe malformation. Other than the pharyngeal cartilages, the pectoral fin plate (pf) cartilages were deeply undulated. The trabeculae cranii (t) of the neurocranium was also bent, and the sclerotic cartilage(sc) around the eyes was undulated. Thus, all cartilage elements in the embryos, whether of neural or mesodermal origin, were affected by disulfiram. In spite of these strong anomalies, major cartilage elements of the pharyngeal arches could be identified. The notochord was not undulated by disulfiram at this stage (Fig. 3C). When embryos were treated with disulfiram from the shield to the open-mouth stage, both pharyngeal cartilages and the notochord were malformed, showing a combination of truncation of the trunk and bending of cartilages (Fig. 3D).

In the embryos treated with 5 μM α, α′-dipyridyl from the prim-5 to the open-mouth stage, large cartilage elements formed in the mandibular to the 4th gill arches were severely bent (Figs. 2E and 3G, H). Anomalies were also observed in the pectoral fin plates, trabeculae, and sclerotic cartilage. In general, the degree of anomaly induced by α, α′-dipyridyl was more severe than that by disulfiram. The phenotype of pectoral fin malformation induced by the two drugs was similar. Thus, the bending of cartilage elements was a common anomaly induced by disulfiram and α, α′-dipyridyl.

Treatment with 5 μM α, α′-dipyridyl after the prim-5 stage did not affect the notochord (Fig. 2E), but treatment at the shield to the prim-5 stage caused severe truncation of the trunk and undulation of the notochord (Fig. 4C). Thus, it was indicated that disulfiram and α, α′-dipyridyl exert similar stage-specific teratogenic effects on the notochord at the early embryonic stage and on the cartilaginous skeleton at the organogenesis stage.

Unsusceptibility of Hoxd-4 and shh expression to disulfiram treatment

Deficiency of vitamin A, a precursor of RA, in rat embryos causes loss of Hoxd-4 expression in the hindbrain and reduces the posterior rhombomeric segments of hindbrain and related posterior pharyngeal arches (White et al., 2000). The phenotype of the pharyngeal skeletal anomaly induced by RA-deficiency is, however, unclear. Since disulfiram decreases the RA content of embryos (Vallari and Pietruszko 1982; Marsh-Armstrong et al., 1995), we expected disulfiram to suppress Hoxd-4 and shh expression in the pharyngeal area and to induce skeletal anomalies via loss of the expression. Contrary to our expectations, loss of gene expression was not achieved by disulfiram at a concentration which caused anomalies in the notochord. An extremely high dose of disulfiram was needed to affect Hoxd-4 expression. However, disulfiram exhibits various side effects, particularly on the central nervous system (Peachey et al., 1981), and it is ambiguous whether the effect on Hoxd-4 expression is caused by suppression of RALDH. Even when embryos were exposed to disulfiram throughout early embryogenesis until the prim-5 stage, cartilages formed normally when the embryos were transferred to normal seawater. This supports the idea that the processes of hindbrain and pharyngeal arch development at the early embryonic stage are unsusceptible to the effect of disulfiram.

The RA content of embryos and localization of RALDH mRNA were recently analyzed in detail in zebrafish and chick. The level of RALDH-2 mRNA expression is high in trunk somites and low in the hindbrain (Swindell et al., 1999). Accordingly, the level of RA in the hindbrain region is particularly low (Marsh-Armstrong et al., 1995; Maden et al., 1998). Thus, RA is synthesized de novo in embryos to control levels regionally. However, since RA is contained in 8-cell-stage zebrafish embryos at the same level as in the shield to the prim-5 stage (Costaridis et al., 1996), RA seems to exist also as a maternal factor in the yolk of teleosts. Therefore, to explain why disulfiram did not suppress the Hoxd-4 and shh expressions, we speculate that RA is present in yolk at a concentration that is high enough to activate RA-responsive genes and that complete RA-deficiency could not be reproduced by disulfiram. When a relatively high dose of disulfiram was administered locally to the chick limb buds, loss of shh expression at the zone of polarizing activity was successfully induced (Stratford et al., 1996). On the other hand, when whole embryos were exposed to the drug, as in our experiments, strong side effects seemed to appear in various tissues without suppressing the expression of RA-responsive genes.

Effects of disulfiram on cartilages

Disulfiram induced severe bending in all cartilaginous skeletons irrespective of their origin, including pharyngeal arches, neurocranium, and pectoral fin. Mechanically unstable type II collagen is known to cause dyschondroplasia in craniofacial and limb skeletons (Metsäranta et al., 1992). Examining the effect of α, α′-dipyridyl, which reduces the mechanical stability of collagen (Fiedler-Nagy et al., 1981), on embryo-genesis, we found that severe bending of cartilaginous skeletons is also induced by this drug. From the similarity of phenotype, we infer that disulfiram induces bent cartilages by affecting the mechanical stability of collagen, as does α, α′dipyridyl. Our hypothesis on the mechanism of the teratogenic effect is that disulfiram inhibits ascorbate oxidase by Cu²⁺-chelating activity, and the resultant disruption of ascorbic acid metabolism causes suppression of the hydroxylation of collagen.

Zebrafish mutant Brk is characterized by bending of the pharyngeal cartilaginous skeletons, so the loss of mechanical stability of the cartilages is attributed to mutation (Neuhauss et al., 1996). Another character of Brk is the reduction of melanocytes. Disulfiram inhibits the pigmentation of skin by inhibiting a Cu²⁺-dependent enzyme, tyrosinase (Marsh-Armstong, 1995). Reduction of melanocytes could also be realized in flounder embryos incubated with disulfiram (Fig. 1 B, G). Thus, the phenotype of anomaly in Brk and disulfiram-treated embryos is closely related with both cartilage and melanocyte development. Therefore, a gene coding for a Cu^2+-dependent enzyme which metabolizes tyrosine and ascorbic acid (Goldstein et al., 1964; Menniti et al., 1986) might be mutated in Brk.

Undulation of the notochord is a remarkable anomaly of zebrafish embryos treated with disulfiram (Costaridis et al., 1996; Marsh Armstrong et al., 1995). A similar anomaly was reproduced in flounder embryos by not only disulfiram but also α, α′-dipyridyl. Although the notochord is a non-chondrogenic tissue, cells of the notochord express type II collagen mRNA at the early embryonic stage (Bieker and Yazdani-Buicky, 1992; Yan et al., 1995). Though the undulation is attribut to the ability of disulfiram to suppress RA synthesis (Costaridis et al., 1996; Marsh-Armstrong et al., 1995), it is also possible that undulation of the notochord is another phenotype of the anomaly caused by the inhibition of collagen maturation.