Experimental animals

Embryos of the brook lamprey, Lampetra reissneri, and the swordtail, Xiphophorus helleri, were used. Sexually mature lampreys were kept in tank water at 20°C and embryos were obtained after natural spawning or artificial fertilization. Lamprey embryos were staged according to the criteria of Tahara (1988) and were used at stages 22–30. Swordtails were kept in an aquarium at 25°C and pregnant females were sacrificed at appropriate intervals. Their embryos were used at stages 8–23 (Tavolga, 1949). The numbers of somites at the various developmental stages of the swordtail and lamprey embryos examined in this study are listed in Table 1.

Table 1

Somite number in developmental stages of lamprey and swordtail embryos

Histology

Embryos were fixed with Bouin's fixative. After dehydration in an ethanol-toluene series, they were embedded in Paraplast (Oxford Labware, St. Louis, MO, USA). Transverse sections were serially cut at a thickness of 6 to 10 μm. Sections were pre-stained with alcian blue (pH 2.5) for 2 hr to visualize the sclerotomal region and poststained by Masson-Goldner's method or hematoxylin and eosin. They were mounted with Entellan (Merck, Darmstadt, Germany).

Immunohistochemistry

Embryos were fixed overnight in a 4% solution of paraformaldehyde in phosphate-buffered saline (PBS, pH 7.4) at 4°C and then washed several times with PBS. The fixed embryos were immersed in 10% to 30% sucrose in PBS and embedded in OCT compound (Miles, Elkhart, IN, USA). Serial 10 –15-μm sections were attached to albumin-coated glass slides. Immunohistochemistry was performed according to methods described previously (Hirata et al., 1997). The following primary antibodies were used: (1) HNK-1 (Becton Dickinson, San Jose, CA, USA; Lot 40557), (2) mouse monoclonal antibody against native chondroitin sulfate, CS-56 (Seikagaku Corp., Tokyo), (3) rabbit antiserum to human fibronectin (anti-FN; Cappel, Durham, NC, USA) for the identification of swordtail fibronectin, and (4) rabbit polyclonal antibody against human fibronectin (anti -FN; Calbiochem, Sandiego, CA, USA) which identifies lamprey fibronectin (Wright, 1986). The following affinity-purified secondary antibodies were used: (1) FITC-conjugated goat anti-mouse immunoglobulins (IgG, IgA, IgM) (Cappel), (2) rhodamine-conjugated goat anti-mouse immunoglobulins (Cappel), and (3) FITC-conjugated goat anti-rabbit IgG (Cappel). Double-stainings were performed with HNK-1 and anti-FN. HNK-1 staining always preceded that of anti-FN. Antibodies were used at the following dilutions: HNK-1, 1:40; CS-56, 1:500; anti-FN, 1:500; and secondary antibodies, 1:40. Primary antibodies except for CS-56 were applied to frozen sections overnight at 4°C. CS-56 was applied for 30 min at room temperature. In order to confirm that CS-56-positive materials were chondroitin sulfate, sections were treated with 0.1 U/ml protease free chondroitinase ABC (Seikagaku Corp.), which digests chondroitin sulfate chains specifically, for 3 hr at 37°C in a moist chamber (Nishikawa and Sasaki, 1993). Secondary antibodies were used for 1 hr at room temperature. Some sections of lamprey embryos were nuclear-stained with a 0.4 μg/ml solution of DAPI (4′,6-diamino-2-phenylindole; Sigma, St. Louis, MO, USA) in PBS to make cell identification easier.

Estimation of volume of the somite and sclerotome

The length of somitic (myomere + sclerotome) and sclerotomal regions was measured along the dorso-ventral (D-V), proximo-distal (P-D) and antero-posterior (A-P) axes. The A-P length was estimated based on the number of serial transverse sections (each 10 μm thick) covering one somite or one sclerotome. The maximal D-V and P-D length were measured in these transverse sections by using an ocular micrometer. Since the somite and sclerotome are roughly regarded as being cubic, the volume of these structures was calculated as the multiplication of D-V, P-D and A-P values and the proportion of the sclerotome per somite was determined.

Relationship between sclerotomal development and neural crest cell migration in the swordtail

Fibronectin and chondroitin sulfate synthesized by sclerotomal cells play important roles in the formation of neural crest cell migratory routes in amniotes (Le Douarin, 1982). Therefore, we have examined temporal changes of the distribution of these extracellular matrix molecules in the sclerotome during the period of neural crest cell migration in the swordtail. In the premigratory stages of neural crest cells (Fig. 2A–D), morphologically detectable sclerotome was observed in the ventro-medial portion of the somite (Fig. 2D). The area surrounding the notochord, that contains the sclerotome, was CS-56-immunoreactive. Intense staining was seen in the boundary between the neural tube and the notochord (Fig. 2C). No fibronectin-like immunoreactivity was detected (Fig. 2B). In early migratory stages (Fig. 2E–H), HNK-1-positive cells (Fig. 2E) were present in the sclerotome that spread dorsally and reached the level of the neural tube (Fig. 2H). This observation shows that the ventral pathway is formed in the sclerotome, as in amniotes. The sclerotome contained FN and CS-56-positive materials (Fig. 2F, G). In intermediate-late migratory stages (Fig. 2I–L), the sclerotome reached the dorsal-most level of the neural tube (Fig. 2L) and contained HNK-1-positive cells (Fig. 2I). FN and CS-56-positive materials were observed in the sclerotome (Fig. 2J, K). CS-56-positive materials disappeared in sections treated with chondroitinase ABC (data not shown). These results indicate that the swordtail sclerotome contains FN and CS-56-positive chondroitin sulfate (CS), and that their expression in the sclerotome is correlated with HNK-1-positive neural crest cell migration along the ventral pathway. Furthermore, sclerotomal development along the antero-posterior axis progressed concomitantly with neural crest cell migration (Fig. 3). HNK-1-immunoreactive cells in the ventral pathway were observed ventrally beyond the notochord throughout the swordtail trunk.

Fig. 2

Sclerotomal development during neural crest cell migration in swordtail embryos. (A-D) Premigratory stages of neural crest cells in the ventral pathway. Transverse sections at 21-somite levels of stage 13 embryos. (E-H) Early migratory stages. Transverse sections at 17-somite levels of stage 12 embryos. (I-L) Intermediate-late migratory stages. Transverse sections at 10-somite levels of stage 12 embryos. Sections were stained with HNK-1 (A, E, I), antifibronectin (B, F, J) and CS-56 (C, G, K). D, H and L show sections stained histologically. The distribution of neural crest cells (NC) and fibronectin (FN) was examined in sections double-stained with HNK-1 and anti-fibronectin. NO, notochord; NT, neural tube; Scl, sclerotome. Scale bars: 20 μm.

Fig. 3

Spatial and temporal changes in sclerotome formation during neural crest cell migration along the ventral pathway in swordtail embryos. Striped area displays somitic levels where neural crest cells migrate along ventral pathways at stages 12 (A), 13 (B) and 15 (C). Hatched area represents somitic levels where the sclerotome is formed at the corresponding stages. Numerals show somite numbers. Thick vertical lines correspond to the last somitic levels at respective stages. A, anterior; P, posterior; D, dorsal; V, ventral.

Regional differences in neural crest cell distribution in the lamprey trunk

HNK-1-positive neural crest cells appeared in the lamprey trunk at stage 23 (Hirata et al., 1997). The distribution of these cells in the ventral pathway shifted caudally from stage 23 to stage 27. There were regional differences in the distribution of HNK-1-immunoreactive cells in the trunk. In the branchial region, the distribution of the cells in the ventral pathway expanded ventrally beyond the notochord (Fig. 4A). Careful observations of serial transverse sections indicated that HNK-1-positive cells at the notochord level in this region are distributed at intervals of about 60 μm which corresponds to the length of one somite along antero-posterior axis. This means that their distribution in the ventral pathway is restricted to the intersomitic space at the notochord level. In the post-branchial trunk, no HNK-1-positive cells were observed ventrally beyond the boundary between the neural tube and notochord (Fig. 4B). These regional differences along the antero-posterior axis are compatible with the observation of the distribution of melanophores which are a typical derivative of neural crest cells even in the lamprey (Newth, 1956). We observed melanophores in the area surrounding the notochord in the branchial region (data not shown) but not in the space between the notochord and somite in the post-branchial region (Hirata et al., 1997).

Fig. 4

Lamprey neural crest cells in the ventral pathway (red arrowheads). (A) Transverse section at 3-somite level in the branchial region of a stage 23 embryo. (B) Transverse section at 35-somite level in the post-branchial trunk of a stage 24 embryo. Sections were stained with HNK-1. NO, notochord; NT, neural tube. Scale bar: 20 μm.

Relationship between sclerotomal development and neural crest cell migration in the lamprey

Since sclerotomal development in the swordtail is closely correlated with neural crest cell migration along the ventral pathway, we have explored their correlation in the lamprey. In the branchial region where HNK-1-positive cells are distributed ventrally beyond the notochord, cell masses associated with CS-56-positive materials were observed at the boundary between the neural tube and notochord and near the ventrolateral sides of the notochord (Fig. 5A–D). These cell masses appeared intersegmentally (Fig. 5E). This intersegmental distribution was similar to that of HNK-1-positive cells at the notochord level in the ventral pathway (Fig. 4A). Since CS-56 immunoreactivity disappeared after digestion with chondroitinase ABC (data not shown), CS-56-positive materials represented lamprey CS. The CS-containing region extended caudally up to the 20-somite level with advancing developmental stages (Fig. 7), but CS-containing regions were not found in the post-branchial trunk posterior to the 20-somite level during the period in which HNK-1-positive cells were present in the ventral pathway (Fig. 5F–H; Fig. 7). In the branchial region, loose mesenchymal cells were histologically observed in similar portions to the CS-containing regions (Fig. 6A). No mesenchymal cells were found beside the notochord in the post-branchial region (Fig. 6B). Moreover, cartilage, a main derivative of the sclerotome, was formed in the branchial region almost corresponding to the CS-containing region after stage 30 (Fig. 6C). No cartilage was observed in the post-branchial trunk during the developmental stages examined (Fig. 6D). These observations strongly suggest that the CS-containing region located intersegmentally in the branchial region is the lamprey sclerotome. In Lampetra fluviatilis embryos, Damas (1944) histologically found the sclerotome in a region corresponding to the CS-containing region detected in the present observation. Taken together, these results suggest that in lamprey embryos, sclerotomal development progresses concurrently with the neural crest cell migration only in the branchial regions, and in the post-branchial regions neural crest cell migration occurs without overt development of the sclerotome (Fig. 7).

Fig. 5

CS-56-positive cell masses in the lamprey trunk. (A-D) Differential interference microscope image (A), nuclear staining (B) and fluorescence (C, D) micrograph of a transverse section at 6-somite level in the branchial region of stage 23 embryo. (E) Horizontal section at 4-6-somite levels (4S-6S) of a stage 23 embryo. Rostral is to the left. (F-H) Differential interference microscope image (F), nuclear staining (G) and fluorescence (H) micrograph of a transverse section at 30-somite level in the post-branchial region of stage 24 embryo. All sections were stained with antibody against native chondroitin sulfate, CS-56. NO, notochord; NT, neural tube. Scale bars: 20 μm.

Fig. 6

(A, B) Histological sections of a stage 27 lamprey embryo. (A) Transverse section at 5-somite level in the branchial region. Mesenchymal cells (small arrowheads) are found in the space between the notochord and the myomere. (B) Transverse section at 30-somite level in the post-branchial region. (C, D) Histological sections of a stage 30 lamprey larva. (C) Transverse section at 5-somite level. Cartilage (large arrowhead) is observed beside the ventro-lateral portion of the notochord. (D) Transverse section at 30-somite level. The dorsal root ganglia (DRG), which are formed in the intersomitic space, are observed in all sections. NO, notochord; NT, neural tube. Scale bars: 50 μm.

Fig. 7

Spatial and temporal changes of sclerotome formation during neural crest cell migration along the ventral pathway in lamprey embryos. Striped area displays somitic levels where neural crest cells migrate along the ventral pathway at stages 23 (A), 24 (B), 25 (C) and 27 (D). Hatched area represents somitic levels where the sclerotome is formed at the corresponding stages. Numerals show somite numbers. Vertical thick lines correspond to the last somitic levels at respective stages. A, anterior; P, posterior; D, dorsal; V, ventral.

Volume of the sclerotome in lamprey and swordtail embryos

Since histological and immunohistochemical observations suggested that the size of the CS-containing sclerotomal region in the lamprey is much smaller than that in the swordtail, the volume of the sclerotome was measured in both embryos. Whereas the swordtail sclerotome occupied approximately 10% of the somite (sclerotome + myomere) at the respective somitic levels during the period in which HNK-1-positive neural crest cells were observed in the ventral pathway (Fig. 8), the proportion of the lamprey sclerotome was approximately 1% at somitic levels where the CS-containing region was observed (Fig. 8).

Fig. 8

Sclerotomal size at respective somitic levels during neural crest cell migration in stage (St.) 13 swordtail (A) and stage 24 lamprey (B) embryos. Percentage of the sclerotome per somite (Scl/SO) was calculated as described in Materials and Methods. It is noteworthy that the value of ordinate in A is 10-fold greater than that in B. The percentage was similar in stage 12, 13 and 15 swordtail embryos and in stage 23, 24 and 25 lamprey embryos.