Gastrulation was examined in Xenopus embryos injected with various polysaccharides into the blastocoel cavity. The progression of gastrulation was assessed by observing pigmentation and yolk plug size in vegetative view embryos. In heparin-or dextran-sulphate-injected embryos, gastrulation was significantly retarded. This was further confirmed in tissue sections of embryos. In contrast, no such retardation was found in embryos injected with hyaluronic acid or chondroitin sulphate. A quantitative analysis showed that the extent of retardation in heparin-or dextran-sulphate-injected embryos was dose-dependent and that, after the initial retardation of up to 2–3 h, gastrulation progressed at a similar rate to controls.

At the time when untreated sibling embryos hatched, embryos injected with heparin or dextran sulphate showed abnormalities in their external appearance and swimming behavior in a dose-dependent manner. When these embryos were examined histologically or immuno-histochemically using tissue-specific monoclonal antibodies, it was found that central nervous system (CNS), especially the brain and eye structures, were most severely damaged. The extent of damage was again dosedependent. In contrast, neural-crest-derived melano-phores were abundant even in aneural larvae. No such change was found in embryos injected with hyaluronic acid or chondroitin sulphate.

Gastrulation is a dynamic cell migration process, which plays a central role in the early development of a wide variety of animal species (Balinsky, 1976; Alberts et al., 1983). The amphibian nervous system is, for instance, known to be derived from the ectoderm which is underlain by the axial mesoderm as a result of gastrulation (Spemann, 1938; Slack, 1983; Gurdon, 1987).

In recent years, fibronectin, a common component of the extracellular matrix (ECM), was shown to be indispensable for cell movement during gastrulation in amphibian and insect embryos using anti-fibronectin serum (Boucaut et al. 1984a) and fibronectin-derived oligopeptides (Boucaut et al. 19846; Naidet et al. 1987). On the other hand, in various cell culture systems, glycosaminoglycans and related substances, which are also common ECM components (Sakashita et al. 1980; Hay, 1981; Hynes and Yamada, 1982), have been shown to have facilitating (Klebe et al. 1986) or inhibitory (Newgreen et al. 1982) effects on cell migration. Since the presence of glycosaminoglycans, such as heparin and chondroitin sulphate, has been reported in the amphibian gastrula (Kosher and Searls, 1973) and heparin is known to bind fibronectin and laminin in ECM (Hay, 1981; Yamada, 1983), I wished to study the effects of various polysaccharides, including heparin, on the gastrulation process in Xenopus embryos. It was shown that heparin and dextran sulphate retarded the gastrulation process up to an initial 2 –3 h period but hyaluronic acid and chondroitin sulphate did not.

In addition, when the later development in the heparin-or dextran-sulphate-injected embryos was examined, I found that CNS differentiation especially was defective. Possible mechanisms underlying the inhibitory action of these polysaccharides on the gastrulation process and neural development are discussed. A preliminary note on this work has been reported in abstract form (Mitani, 1988).

Animal care and generation of hybridomas

These were done in essentially the same way as described in the previous paper (Mitani and Okamoto, 1988).

Chemicals

Heparin was purchased from Wako Pure Chemicals, dextran sulphate (Mr, 500 000) from Sigma, chondroitin sulphate A and hyaluronic acid from Seikagaku Kogyo. They were dissolved in Steinberg’s balanced salt solution (SBSS), sterilized by filtration through cellulose acetate filter (pore size, 0.2 μm; loss of polysaccharide during filtration was negligible).

Microinjection

Blastula embryos (stage 8 or 9) were dejellied with 2% cysteine (pH 7.9), washed with MMR (modified modified frog ringer; Newport and Kirschner, 1982) and kept on ice until they were injected with solutions. Embryos were placed on microwells of Terasaki plates and impaled with a glass micropipette (tip diameter, 10 μm) under a dissecting microscope. Solutions were injected into blastocoel cavities by pressure of nitrogen gas (0.4 kg cm−2) using a solenoid valve. The injected volume was 30 nl except for the gastrulation index assessment (100 nl). Since injection of 200nl of SBSS had no significant effect on gastrulation and subsequent development, it can be concluded that the effects of polysaccharides were attributable to their chemical characteristics. If we assume that the volume of blastocoel is about 300 nl (Slack et al. 1973), the final concentration of the chemicals in blastocoel is about one eleventh and one fourth of original concentrations for injected volumes of 30nl and 100 nl, respectively.

Macroscopic observation of gastrulation (gastrulation index) and subsequent development

After embryos were injected with various solutions, they were allowed to develop in a 10% Holtfreter solution. The degree of gastrulation (gastrulation index) was determined by the extent of pigmentation and by the size of yolk plug in the vegetative view of embryos (external criteria according to Nieuwkoop and Faber, 1967) up to the period when control embryos developed to the neural plate stage. For instance, when an embryo that had been injected with heparin at stage 9 (age 7 h) and cultured further for 6h showed an equivalent extent of gastrulation to stage 10.25 (age 10h) of intact embryo judged from the external criteria, then its gastrulation index is 10 h even though the embryo has an actual age of (7+6) h.

The morphology of the larvae was examined after 2 days of incubation, a time when control embryos had hatched. The swimming behavior evoked by the gentle touch of a hair loop was also examined.

Histology

Control embryos and their polysaccharide-injected siblings were fixed with 2 % glutaraldehyde and 10 % formalin in PBS at 4°C overnight, and washed with PBS. They were incubated in PBS containing 10% gelatin overnight at 37 °C. After gelatin was solidified on ice, the gelatin block was fixed as described above, trimmed into a convenient size, and sliced into sections of 100 μm in thickness, and mounted on the gelatin-coated glass slides. This procedure facilitated the correct orientation of gastrulae, which is otherwise very difficult. The slides were air-dried and the tissue sections were stained with thionine.

To examine the histology of neurula- and hatching-stage larvae, they were fixed and washed as above, dehydrated with graded series of alcohol and xylene, embedded in paraffin, cut serially in 10 μm thickness, mounted and stained with thionine.

Immunohistochemistry and monoclonal antibodies

The procedures are essentially same as those in previous papers (Mitani and Okamoto, 1988, 1989). Briefly, control embryos at the hatching stage (stage 37/38) and their polysaccharide-injected siblings were fixed with 0.5% paraformaldehyde for 16 h at 0°C, and washed with PBS. They were embedded in polyacrylamide, and frozen-sectioned with a cryostat (16μm thick). To reveal the structure of nervous and muscular tissues, the sections were stained with tissue-specific monoclonal antibodies (Mabs) described in the previous papers (Mitani and Okamoto, 1988, 1989) and below, using indirect immunofluorescence technique. They were observed under an epi-illuminescent fluorescence microscope (Nikon).

Tissue- and stage-specificities of monoclonal antibodies (Mabs) Mui, NM1, and Nl have been described in previous papers (Mitani and Okamoto, 1988, 1989). Briefly, Mab Mui is specific for myotomal muscle cell. The antigen becomes detectable at stage 20. Mab NM1 is specific for axons, ependymal processes and myotomal muscle cells. The anti-gen(s) is first detected at stage 24 for myotome and stage 29/30 for neural tissue. Mab Nl is specific for dendrite-associated structures in the central nervous system. The antigen is first detected at stage 33/34 at the mid-trunk level, while the distribution of antigen shows a clear rostrocaudal gradient; the antigen is found only at the rostral CNS and not at the caudal level at the hatching stage, but later, it is also found at the tail CNS.

Mab Mcl is specific for both notochord and myocomma (intermyotomal septum). The antigen is first localized at the extracellular surface of notochords (stage 22), then myocomma, suggesting it is secreted from notochord cells and/or muscle cells (Mitani and Okamoto, unpublished). Mab Mel is specific for neural-crest-derived melanophores and retinal pigment cells. The antigen for Mel became detectable at stage 29/30 (Mitani and Okamoto, unpublished).

Effects of polysaccharide on gastrulation

I examined the effects of heparin, dextran sulphate, chondroitin sulphate and hyaluronic acid on cell migration during the period of gastrulation in Xenopus embryos. The polysaccharide solutions or SBSS (carrier solution) were injected into the blastocoel of blastula embryos (stage 8, age 5h or stage 9, age 7h). The external appearance of these embryos was similar to that of intact embryos until the latter reached stage 10 (age 9h), suggesting that the injected embryos underwent normal cleavages. This was confirmed by examining those embryos in tissue sections (data not shown). In normal embryos at stage 10, the pigment concentration appears at the future dorsal lip region in the vegetal hemisphere; this is the first external sign of gastrulation (Nieuwkoop and Faber, 1967). A similar pigment concentration was observed in embryos injected with chondroitin sulphate, hyaluronic acid or SBSS (up to 200 nl and up to 10 mg ml−1 polysaccharide concentration), but it was not observed at age 9h in those injected with heparin or dextran sulphate at doses higher than 1 mg ml−1. Gastrulation of the former group of embryos seemed to proceed normally as judged from external criteria; the pigmentation extended ventrally to make a large circle delineating yolk plug (stage 10.5, age 11 h) and the yolk plug became smaller as invagination of mesoderm proceeded (stage 12.5, age 14.25h). In Fig. 1, the external appearance (vegetative view) of SBSS-(B) or chondroitin sulphate-(C) injected embryos is shown and compared to that of an intact embryo at the same age (stage 12 of control, A). The stage of gastrulation appears similar in these embryos. The hyaluronic acid-injected embryos gave the same results (data not shown). In contrast, gastrulation of heparin-injected embryos at the same age was markedly retarded from the external criteria in a dosedependent manner (Fig. 1DF). This retardation was slight in embryos injected with 0.1mg ml−1 heparin (Fig. ID), but in the embryos injected with heparin with higher concentrations than 1mg ml−1, retardation of gastrulation was apparent (Fig. IE, F); their yolk plugs were larger than control embryos and heavy pigmentation, which had already invaginated in the control embryos, was still visible. The embryos injected with dextran sulphate showed similar abnormalities (data not shown). These were consistently observed throughout five series of experiments.

Fig. 1.

Effects of polysaccharide on gastrulation rate; examples of actual cases showing vegetative view of polysaccharide-injected embryos. (A) Non-injected control embryo; (B) SBSS-injected embryo; (C) chondroitin-sulphate-injected embryo; (D–F) heparin-injected embryos (0.1, 1 and 10mg ml−1, respectively). Bar=1 mm.

Fig. 1.

Effects of polysaccharide on gastrulation rate; examples of actual cases showing vegetative view of polysaccharide-injected embryos. (A) Non-injected control embryo; (B) SBSS-injected embryo; (C) chondroitin-sulphate-injected embryo; (D–F) heparin-injected embryos (0.1, 1 and 10mg ml−1, respectively). Bar=1 mm.

The retarded gastrulation in heparin-or dextran-sulphate-injected embryos was confirmed from the internal criteria (Nieuwkoop and Faber, 1967). Fig. 2 (A–F) shows examples of sagittal sections of the same series of embryos as in Fig. 1 (A–F), except embryos were collected when control embryos reached stage 11. SBSS-(Fig. 2B) or chondroitin sulphate-(Fig. 2C) injected embryos showed normal gastrulation (compare to Fig. 2A, intact embryo) as judged by the extent of the depth of involution from the blastopore toward the animal pole (indicated by a pair of arrows). On the other hand, the gastrulation process in heparin-injected embryos was clearly inhibited in a dose-dependent manner (Fig. 2D–E, arrows). It is noteworthy that the extent of retardation of gastrulation judged from external criteria fairly closely parallels the extent judged from internal criteria in embryos injected with increasing amount of heparin (Fig. 1D–F and Fig. 2D–F). Similar results were obtained with dextran sulphate-injected embryos (data not shown).

Fig. 2.

Assessment of gastrulation process in earlier phase by sagittal sections. (A) Non-injected control embryo; (B) SBSS-injected embryo; (C) chondroitin sulphate-injected embryo; (D–F) heparin-injected embryos (0.1, 1 and 10 mgml−1, respectively). Bar=lmm.

Fig. 2.

Assessment of gastrulation process in earlier phase by sagittal sections. (A) Non-injected control embryo; (B) SBSS-injected embryo; (C) chondroitin sulphate-injected embryo; (D–F) heparin-injected embryos (0.1, 1 and 10 mgml−1, respectively). Bar=lmm.

To examine further the inhibitory’ effect of heparin and dextran sulphate on gastrulation, the time course of gastrulation in embryos injected with increasing amounts of polysaccharide was analyzed in a quantitative way in a series of experiments. The extent of gastrulation (gastrulation index; see Materials and methods) in injected embryos was evaluated on the basis of the external criteria of gastrulation in normal embryos. In Fig. 3A–C the abscissa is the actual age of embryo while the ordinate is the gastrulation index in polysaccharide-injected embryos. Twenty embryos were injected with the same solution and the results were averaged and presented in this figure. The embryos injected with chondroitin sulphate gastrulated with the same time course as control embryos at any concentration (Fig. 3A). In contrast, the inhibitory effects of heparin and dextran sulphate were prominent. The embryos injected with either of these polysaccharides gastrulated more slowly than controls in a dose-dependent manner (Fig. 3B, C). The process of gastrulation appeared markedly suppressed especially during early phases of gastrulation (up to age 12.5 h).

Fig. 3.

Assessment of gastrulation process in polysaccharide-inj ected embryos by vegetative view. (A–C), Time course (mean) of gastrulation process in embryos injected with various concentrations (○, 0.03; and, △, 0.1; □, 0.3; •, 1; ▲, 3; ■, 10 (mg ml−1 )) of chondroitin sulphate, heparin and dextran sulphate, respectively. Abscissa, actual time; ordinate, gastrulation index (for definition, see text). (D) Dose–gastrulation rate relations (mean±s.E.) of chondroitin sulphate-(⦾), heparin-(⟁) and dextran sulphate-(⧈) injected embryos.

Fig. 3.

Assessment of gastrulation process in polysaccharide-inj ected embryos by vegetative view. (A–C), Time course (mean) of gastrulation process in embryos injected with various concentrations (○, 0.03; and, △, 0.1; □, 0.3; •, 1; ▲, 3; ■, 10 (mg ml−1 )) of chondroitin sulphate, heparin and dextran sulphate, respectively. Abscissa, actual time; ordinate, gastrulation index (for definition, see text). (D) Dose–gastrulation rate relations (mean±s.E.) of chondroitin sulphate-(⦾), heparin-(⟁) and dextran sulphate-(⧈) injected embryos.

To examine the dose-response relation, the advancement of gastrulation in polysaccharide-injected embryos was assessed in Fig. 3A-C at the time when the suppressive effect of the polysaccharide was most prominent (around age 12.5 h). Since gastrulation in normal embryos starts at around stage 10 (age 9 h) (Nieuwkoop and Faber, 1967), I calculated a relative gastrulation rate as (gastrulation index–9)/(12.5–9). The gastrulation rate is plotted against the concentration of injected polysaccharide in Fig. 3D. The threshold of the heparin and dextran sulphate were 0.3 and 0.03 mg ml−1, respectively. The effect seemed to saturate at 10 mg ml−1 of both polysaccharide.

It is interesting that after initial suppression of gastrulation until around age 12.5 h, gastrulation appears to regain a similar rate of progression to controls even with higher doses of polysaccharides (Fig. 3B, C). The recovery of gastrulation movement from initial retardation was confirmed in tissue sections of later stage embryos (neurula) that had been injected with 30 nl of 0.1 to 10 mg ml−1 heparin (Fig. 4B, C). In these embryos, we could identify invaginated mesodermal cells differentiating into myotome, when compared to control embryo (Fig. 4A). Cellular differentiation of myotome muscle in heparin-injected embryos was also confirmed using a monoclonal antibody specific for myotomal muscle (see later section).

Fig. 4.

Assessment of gastrulation process in later phase by sagittal sections. (A) Non-injected control embryo; (B–D) heparin-injected embryos (0.1, 1 and 10mg ml−1, respectively). In A, C and D, rostral is to the right, and in B, rostral is to the left. CNS, central nervous system; myo, presumptive myotome; en, endodermal mass. Bar=0.5mm.

Fig. 4.

Assessment of gastrulation process in later phase by sagittal sections. (A) Non-injected control embryo; (B–D) heparin-injected embryos (0.1, 1 and 10mg ml−1, respectively). In A, C and D, rostral is to the right, and in B, rostral is to the left. CNS, central nervous system; myo, presumptive myotome; en, endodermal mass. Bar=0.5mm.

In’ summary, heparin and dextran sulphate, when injected into the blastocoel cavity, markedly retarded blastomere migration during gastrulation but after a few h of suppression these polysaccharides became ineffective. Disappearance of the inhibitory effect could be either due to a loss of the polysaccharide in blastocoel cavity by degradation or sequestration, or due to an enhancement of migrating force of mesodermal cells which antagonizes the inhibitory action by heparin.

Further development of heparin-injected embryos

To examine the effects of polysaccharide on further development, injected embryos were cultured until control embryos reached the hatching stage (stage 37/38, age 2 d 5.5 h). Larvae that had been injected with 30nl of SBSS (Fig. 5B), those injected with 30nl of 10 mg ml−1 chondroitin sulphate (Fig. 5C) and those injected with 30 nl of 10 mg ml−1 hyaluronic acid (data not shown) developed normally, when compared to control larvae (Fig. 5A). On the other hand, larvae that had been injected with heparin (Fig. 5D–F) or those injected with dextran sulphate (data not shown) showed apparent anomalies, although they survived and hatched. The larvae injected with 30nl of 0.1mg ml heparin (Fig. 5D) had eyes, an apparent normal axial structure, except that they were a little shorter than control larvae. Typically the larvae injected with 30 nl of 1 mg ml−1 heparin (Fig. 5E) had an axial structure such as the notochord and the caudal portions of CNS but no eyes or head (about half). However, the other larvae showed milder (like 10 mg ml−1) or severer (like 10 mg ml−1) phenotypes. The majority of larvae injected with 30 nl of 10 mg ml−1 heparin (Fig. 5F) appeared to have no axial structures.

Fig. 5.

Actual examples of macroscopic outlook of control and heparin-injected larvae. (A) Non-injected control larva; (B) SBSS-injected larva; (C) chondroitin sulphate (10 mg ml−1-injected larva. (D–F) Heparin-injected larvae (0.1, 1 and 10 mg ml−1, respectively). Bar=lmm.

Fig. 5.

Actual examples of macroscopic outlook of control and heparin-injected larvae. (A) Non-injected control larva; (B) SBSS-injected larva; (C) chondroitin sulphate (10 mg ml−1-injected larva. (D–F) Heparin-injected larvae (0.1, 1 and 10 mg ml−1, respectively). Bar=lmm.

The swimming behavior was also impaired in parallel with the external appearance. Larvae injected with 30 nl of 0.1mg ml−1 heparin swam as well as control larvae, while ones injected with 30 nl of 1mg ml−1 heparin simply repeated bending and extending movements but could not swim. The most severely affected embryos (injected with 30 nl of 10 mg ml−1 of heparin) did not move at all even when they were agitated by a hair loop.

Development of eyes, the central nervous system and the notochord

In this and following sections, I examine tissue morphogenesis and cellular differentiation in the heparin-injected larvae histologically using serial paraffin sections or immunohistochemically with the aid of tissuespecific Mabs.

Larvae injected with 30 nl of 0.1 mg ml−1 heparin

Brain, eyes and spinal cord structures were formed in serial sections of four larvae. However, the eyes were fused together at the ventromedian region forming a butterfly-like shape in two of these (Fig. 6B1). The brain- and medulla-like structures extended laterally and seemed larger than those of control larvae when compared at various corresponding rostrocaudal levels. For instance, the area of brain-like structure in the cross-section shown (marked ‘B’ in Fig. 6B1) was 10.4 × 10−2mm2, whereas the brain area in control (‘B’ in Fig. 6A1) was 6.6×10−2 mm2 (58% increase of control). The area of the medulla-like and medulla structures in these larvae (‘Me’ in Fig. 6B2 and A2) were 7.2 × 10−2mm2 and 4.7×10−2mm2, respectively (55 % increase). The spinal cord (Fig. 6B3, arrowhead) structure in injected embryos was, in contrast, not very different from control (Fig. 6A3). Thus, these larvae superficially appeared macrocephalic. However, since the length of both brain and spinal cord was shorter, as expected from the outer appearance, it appeared that the apparent macrocephaly was due not to the excessive formation of the rostral nervous system but due to its deformation. The notochord of these larvae appeared normal (Fig. 6B, N).

Fig. 6.

Development of central nervous system in heparin-injected larvae. (A1-3) Non-injected control larva; (B1-3–D1-3) heparin-injected larvae (30 nl of 0.1, 1 and 10 mg ml−1, respectively). B, brain; Me, Medulla or medulla-like structure; N, notochord; myo, myotome. Bar=500; μm.

Fig. 6.

Development of central nervous system in heparin-injected larvae. (A1-3) Non-injected control larva; (B1-3–D1-3) heparin-injected larvae (30 nl of 0.1, 1 and 10 mg ml−1, respectively). B, brain; Me, Medulla or medulla-like structure; N, notochord; myo, myotome. Bar=500; μm.

The white matter of injected larvae appeared normal throughout the CNS, suggesting heparin at this concentration did not inhibit neurite extension. This was confirmed by the Mabs NM1 and Nl, which have a high affinity for neurites in the white matter (Fig. 7C, D) (Mitani and Okamoto, 1988). From these staining patterns and the normal swimming behavior of the animals, I concluded that neuronal differentiation in these embryos is for the most part normal.

Fig. 7.

Development of central nervous system in heparin-injected larvae studied with tissue-specific Mabs. NM1 (A, C, E, G) and NI (B, D, F, H). (A, B) Non-injected control larva. (C/D, E/F and G/H) heparin-injected larvae (30nl of 0.1, 1 and 10mg ml−1, respectively), (a–h) Schematic illustration of larval structure of corresponding staining in upper cases. Bars=100μm.

Fig. 7.

Development of central nervous system in heparin-injected larvae studied with tissue-specific Mabs. NM1 (A, C, E, G) and NI (B, D, F, H). (A, B) Non-injected control larva. (C/D, E/F and G/H) heparin-injected larvae (30nl of 0.1, 1 and 10mg ml−1, respectively), (a–h) Schematic illustration of larval structure of corresponding staining in upper cases. Bars=100μm.

Larvae injected with 30nl of 1.0 mg ml−1 heparin

The CNS of these larvae was more anomalous than those injected with 0.1 mg ml−1 heparin. In the serial sections of the seven whole larvae that showed typical outer appearance, I found neither brain nor eyes. In the region where medulla should normally form, only epidermis without CNS was found dorsafly to the notochord (Fig. 6C1, arrows). However, at the midtrunk level, a spinal cord was found, though it was hypomorphic and flat rather than round as in control larvae (Fig. 6C2, arrowheads). More caudally, a normal-looking spinal cord is found (Fig. 6C3).

The observation with NM1 and N1 was consistent with the above results. NM1 antigen was detected at the caudal level of CNS (Fig. 7E) but not at the rostral level where the CNS was histologically absent (data not shown). On the other hand, N1 antigen was detected neither at the caudal nor at the rostral level (Fig. 7F; neighboring sections to that in Fig. 7E). N1 antigen is known to be localized in the rostral CNS at the hatching stage (Mitani and Okamoto, 1988). These observations suggest that only the caudal CNS was formed. The notochord of these larvae appeared normal (Fig. 6C).

Larvae injected with 30nl of 10mg ml−1 heparin

Neither CNS nor eyes were discerned in serial sections of three whole larvae (examples of sections are shown in Fig. 6D1–D3). Neither NM1 (Fig. 7G) nor N1 (Fig. 7H) antigens were detected in the embryos, except for NM1 antigen in muscle. This is compatible with the absence of CNS in these larvae studied histologically. On the other hand, notochords appeared to be normally differentiated (Fig. 6D2, N).

I conclude that administration of heparin into the blastocoel inhibits differentiation of eyes and the CNS. The extent of inhibition is highly correlated with the gastrulation retardation, which is dependent on the dosage of heparin. A failure in differentiation was severer rostrally than caudally.

Development of neural-crest-derived melanophores

It is interesting to examine how neural crest cells, which are also induced by mesodermal cells as suggested by the grafted-double axes embryos (Spemann, 1938), are induced to differentiate in these heparin-injected embryos.

I stained frozen sections of controls and larvae that had been injected with 30 nl polysaccharide with Mab Mel, which is specific for neural-crest-derived melanophores and retinal pigment cells. The former was easily distinguished from retinal pigment cells, because eyes had round shapes and localized at the brain level, whereas melanophores were scattered around the more caudal CNS and under the epidermis. The development of melanophores thus identified with the aid of Mab Mel was expected to represent some aspect of neural crest differentiation.

In the control non-injected (Fig. 8A, a), SBSS-injected (data not shown) and 10 mg ml−1 chondroitin sulphate-injected (data not shown) larvae, melanophores differentiated and migrated normally (4 cells counted in the section shown). They were normally localized in the larvae. In the larvae injected with 0.lmg ml−1 heparin, melanophores differentiated as control larvae, as expected from the almost normal differentiation of the CNS (Fig. 8B, b, 4 cells).

Fig. 8.

Development of neural-crest-derived melanophores in heparin-injected larvae as examined with tissue-specific Mab on larval frozen sections. (A, a) Non-injected control larva. (B, b–D, d), Heparin-injected larvae (30nl of 0.1, 1 and 10 mg ml−1, respectively), sc, Spinal cord; n, notochord; m, myotome; en, endodermal mass. Bar=500μm.

Fig. 8.

Development of neural-crest-derived melanophores in heparin-injected larvae as examined with tissue-specific Mab on larval frozen sections. (A, a) Non-injected control larva. (B, b–D, d), Heparin-injected larvae (30nl of 0.1, 1 and 10 mg ml−1, respectively), sc, Spinal cord; n, notochord; m, myotome; en, endodermal mass. Bar=500μm.

To my surprise, however, in the larvae injected with 1 mg ml−1 heparin (Fig. 8C, c) or with dextran sulphate (data not shown), many more melanophores differentiated than in the control larvae (34 cells in the section shown). In addition to the excess differentiation of melanophores, they appeared to migrate a certain distance as indicated by the dispersed distribution of the cells in the sections. In these larvae, the CNS differentiation was greatly impaired as described previously. In the larvae that were injected with 10mg ml−1 heparin (Fig. 8D, d) or dextran sulphate (data not shown), excess differentiation of melanophores was also found (33 cells in the section shown).

The differentiation of melanophores in the heparin-or dextran sulphate-injected larvae was observed at all rostrocaudal levels and showed little difference at the various levels. The total number of melanophores in an animal was estimated to be greater in the heparin-injected larvae than in control, even after the fact that the body length is shorter in the injected larvae than in control is taken into account.

Development of myotomes

In non-injected larvae (Fig. 9A) or larvae injected with SBSS, 10 mg ml−1 hyaluronic acid or chondroitin sulphate (data not shown), all myotomal muscle cells aligned longitudinally by the hatching stage. On the other hand, in larvae injected with heparin (1.0 mg ml−1, 30 nl), groups of myotomal muscle cells were directed transversely or obliquely (Fig. 9D, arrow), though others appeared to be aligned normally. The population of abnormally aligned myotomal muscle cells increased in a dose-dependent manner (data not shown).

Fig. 9.

Development of myotome in heparin-injected larvae. (A–C) Non-injected control larva; (D–F) heparin-injected larva (1 mgml−1’, 30nl). (A, D) Cross-sectional view. (B/E and C/F), Detection of Mui and Mcl antigens, respectively, on sagittal frozen sections. Bars=100μm. sc, spinal cord; n, notochord.

Fig. 9.

Development of myotome in heparin-injected larvae. (A–C) Non-injected control larva; (D–F) heparin-injected larva (1 mgml−1’, 30nl). (A, D) Cross-sectional view. (B/E and C/F), Detection of Mui and Mcl antigens, respectively, on sagittal frozen sections. Bars=100μm. sc, spinal cord; n, notochord.

The cellular differentiation of the myotome in these larvae was confirmed by Mabs Mui (specific for the myotomal muscle cells; Mitani and Okamoto, 1989) and Mcl (specific for myocomma and notochord). The Mui and Mcl antigens were detected in frozen sections of the larvae injected with any of the polysaccharide used in the present study (up to 10 mg ml−1, 30 nl). However, it was obvious again that the arrangement of myotomal cell and myocomma was disorganized in heparin-(Fig. 9E, F) and dextran-sulphate-injected embryos (data not shown).

Possible targets of heparin in retarding gastrulation process

In the present study, I demonstrate that the polysaccharides heparin and dextran sulphate retard the gastrulation process when injected into blastocoel of Xeno-pus blastula, but two other polysaccharides, hyaluronic acid or chondroitin sulphate, do not. Relations have been proposed between galactose-binding lectin on the gastrula cell surface of Xenopus laevis and gastrulation movement (Harris and Zalik, 1985). However, since heparin, dextran sulphate and hyaluronic acid do not have galactose chains and chondroitin sulphate does, this hypothesis does not apply to our situation. Neither is it likely that the effect of heparin and dextran sulphate is correlated with sulphate residues which may chelate the calcium ions in blastocoel, since chondroitin sulphate, which has one third of sulphate residues of heparin in its chain, did not have any inhibitory effect at concentrations more than tenfold higher than heparin or more than 100-fold higher than dextran sulphate (see also Slack et al. 1987).

The presence of heparin and chondroitin sulphate has been reported in the amphibian gastrula cell surface (Kosher and Searls, 1973). These polysaccharides have been suggested to play some role in the cell migration process collaborating with ECM protein such as fibronectin (Hay, 1981; Hynes and Yamada, 1982). It has been indeed demonstrated that fibronectin plays a critical role in gastrulation, using antifibronectin serum (Boucaut et al. 1984a) and fibronectin-related oligopeptides (Boucaut et al. 1984b; Naidet et al. 1987). Moreover, binding of heparin was shown to change the conformation of fibronectin (1980; Richter et al. 1985; Ankel et al. 1986) and strengthens the affinity of fibronectin for other ECM components such as collagen (Perkins et al. 1979). It is also known that heparin binds to laminin (Sakashita et al. 1980; Yamada, 1983), which is also present on amphibian gastrula cell surface (Nakatsuji et al. 1985; Darribere et al. 1986). It is thus possible that heparin interacts with fibronectin and/or other ECM components, and changes the affinity among them. As a result of these changes, heparin may retard the gastrulation process.

To the best of our knowledge, there has been no observation showing that dextran sulphate binds to fibronectin. However, it is noteworthy that dextran sulphate mimics the action of heparin in the inhibition of blood coagulation. Since dextran sulphate was more potent than heparin in retarding gastrulation, the mode of action of this polysaccharide should be further explored.

Differentiation of CNS and neural-crest-derived melanophores in heparin-injected embryos

In the present study, it was shown that injection of heparin or dextran sulphate resulted in the development of an incomplete set of neural structures (rostro-caudal progression of CNS deficiencies) accompanied by retardation in gastrulation. Similar CNS deficiencies were observed in embryos that were treated in completely different ways, that is, in embryos exposed to ultraviolet irradiation, cold or pressure at the critical stage before first cleavage (Malacinski et al. 1977; Sharf and Gerhart, 1983; Cooke, 1985). In their experiments, all these physically treated embryos also showed abnormal and partial gastrulation. It has been shown that the period over which the ectoderm cells can respond to the signal from the mesoderm cells is limited. Accordingly, the defect in gastrulation may inhibit the mesoderm from underlying undifferentiated ectoderm at the proper timing and affect subsequent neural development (Slack, 1983; Gurdon, 1987). Thus, the effect of heparin and dextran sulphate on neural development might be attributed, if not entirely, to its action on gastrulation.

Another obvious possibility is that heparin interferes directly with the neural induction process. Recently heparin was shown to inhibit mesoderm induction by binding to molecules of the fibroblast growth factor group (Slack et al. 1987; Kimelman et al. 1988). When examined in in vitro culture with the technique as described in the previous reports (Okamoto and Mitani, 1987; Mitani and Okamoto, 1989), heparin was ineffective in blocking the neural differentiation from ectoderm cells up to the concentration corresponding to 30nl injection of 1.0mg ml−1 heparin in vivo; with this treatment, the rostral part of the CNS and eyes were lost. However, we also found that at a concentration corresponding 30 nl injection of 10 mg ml−1 heparin, which resulted in complete loss of the CNS in vivo, neural differentiation was inhibited to a certain extent, though the epidermal differentiation from ectoderm cells was also inhibited. It is thus suggested that the defects in the CNS as shown in Fig. 5D and E, or 6B and C are secondary ones due to an inhibition of gastrulation, but the defects in the extreme case, such as shown in Figs 5F or 6D, are the result of both indirect and direct effects of heparin (Mitani and Okamoto, unpublished observations).

In the present study, melanophores were shown to differentiate even in the larvae whose CNS was severely restricted by the injection of heparin or dextran sulphate. Melanophores originate from the neural crest in normal development (Balinsky, 1976). Neural crest is thought to be induced from ectoderm by the action of the underlying mesoderm as a ridge of neural plate. The findings in this study strongly suggest that there are differences between the mechanisms of ‘neural induction’ and ‘neural crest induction’.

I thank Professor Kunitaro Takahashi, Professor Yoshiaki Kidokoro, Dr Martin Chalfie, Dr Shigeto Sasaki and Dr Harumasa Okamoto for critical reading of the manuscript and helpful suggestions. This work was supported by grants-in-aid for scientific research from the Ministry of Education, Science and Culture, Japan.

Alberts
,
B.
,
Bray
,
D.
,
Lewis
,
J.
,
Raff
,
M.
,
Roberts
,
K.
and
Watson
,
J. D.
(
1983
).
Molecular Biology of the Cell
.
Garland
,
New York
.
Ankel
,
E. G.
,
Homandberg
,
G.
,
Tooney
,
N. M.
and
Lai
,
C.
(
1986
).
Heparin modulated conformational states of plasma fibronectin: An electron spin resonance label approach
.
Archs Biochem. biophysics
244
,
50
56
.
Balinsky
,
B. I.
(
1976
).
An Introduction to Embryology. 4th
ed.
Saunders
,
Philadelphia
.
Boucaut
,
J. C.
,
Darribere
,
T.
,
Boulekbbache
,
H.
and
Thierry
,
J. P.
(
1984a
).
Prevention of gastrulation but not neurulation by antibodies to fibronectin in amphibian embryos
.
Nature, Lond
.
307
,
364
367
.
Boucaut
,
J. C.
,
Darribere
,
T.
,
Poole
,
T. J.
,
Aoyama
,
H.
,
Yamada
,
K. M.
and
Thierry
,
J. P.
(
1984b
).
Biologically active synthetic peptides as probes of embryonic development: A competitive peptide inhibitor of fibronectin function inhibits gastrulation in amphibian embryos and neural crest cell migration avian embryos
.
J. Cell Biol
.
99
,
1822
1830
.
Cooke
,
J.
(
1985
).
Dynamics of the control of body pattern in the development of the egg and after excision of presumptive head endo-mesoderm
.
J. Embryol. exp. Morph
.
88
,
135
150
.
Darribere
,
T.
,
Riou
,
J. F.
,
,
D. L.
,
Delarue
,
M.
and
Boucaut
,
J. C.
(
1986
).
Synthesis and distribution of laminin-related polypeptides in early amphibian embryos
.
Cell Tissue Res
.
246
,
45
51
.
Gurdon
,
J. B.
(
1987
).
Embryonic induction-molecular prospects
.
Development
99
,
285
306
.
Harris
,
H.
and
Zalik
,
S. E.
(
1985
).
Studies on the endogenous galactose-binding lectin during early development of the embryo of Xenopus laevis
.
J. Cell Sci
.
79
,
105
117
.
Hay
,
E. D.
(
1981
).
Extracellular Matrix
.
J. Cell Biol
.
91
,
205s
223s
.
Hynes
,
R. O.
and
Yamada
,
K. M.
(
1982
).
Fibronectins: multifunctional modular glycoproteins
.
J. Cell Biol
.
95
,
369
377
.
Kimelman
,
D.
,
Abraham
,
J. A.
,
Haaparanta
,
T.
,
Palisi
,
T. M.
and
Kirschner
,
M. W.
(
1988
).
The presence of fibroblast growth factor in the frog egg: Its role as a natural mesoderm inducer
.
Science
242
,
1053
1056
.
Klebe
,
R. J.
,
Escobedo
,
L. V.
,
Bently
,
K. L.
and
Thompson
,
L. K.
(
1986
).
Regulation of the cell motility, morphology, and growth by sulfated glycasaminoglycans
.
Cell Motil Cytoskel.
6
,
273
281
.
Kosher
,
R. A.
and
Searls
,
R. L.
(
1973
).
Sulfated mucopolysaccharide synthesis during the development of Rana pipiens
.
Devi Biol
.
32
,
50
68
.
Malacinski
,
G. M.
,
Brothers
,
A. J.
and
Chung
,
H-M.
(
1977
).
Destruction of components of the neural induction system of the amphibian egg with ultraviolet irradiation
.
Devi Biol
.
56
,
24
39
.
Mitani
,
S.
(
1988
).
Gastrulation and subsequent morphogenesis are inhibited by injection of certain polysaccharide into blastocoels of Xenopus embryos
.
Neurosa. Res. Suppl
.
7
,
s55
.
Mitani
,
S.
and
Okamoto
,
H.
(
1988
).
Monoclonal antibodies against larval nervous system of Xenopus laevis; their specificities and application to the analysis of neural development
.
Neurosci
.
25
,
291
305
.
Mitani
,
S.
and
Okamoto
,
H.
(
1989
).
Embryonic development of Xenopus studied in a cell culture system with tissue-specific monoclonal antibodies
.
Development
105
,
53
59
.
Naidet
,
C.
,
Semeriva
,
M.
,
Yamada
,
K. M.
and
Thierry
,
J. P.
(
1987
).
Peptides containing the cell-attachment recognition signal Arg-Gly-Asp prevent gastrulation m Drosophila embryos
.
Nature, Lond
.
353
,
348
350
.
Nakatsuh
,
N.
,
Hashimoto
,
K.
and
Hayashi
,
M.
(
1985
).
Laminin fibrils in newt gastrulae visualized by the immunofluorescent staining
.
Develop. Growth and Differ
.
27
,
639
643
.
Newgreen
,
D. F.
,
Gibbins
,
I. L.
,
Sauter
,
J.
,
Wallenfels
,
B.
and
Wutz
,
R.
(
1982
).
Ultrastructural and tissue-culture studies on the role of fibronectin, collagen and glycosaminoglycans in the migration of neural crest-cells in the fowl embryo
.
Cell Tissue Res
.
221
,
521
549
.
Newport
,
J.
and
Kirschner
,
M.
(
1982
).
A major developmental transcription in early Xenopus embryos: I. Characterization and timing of cellular changes at the midblastula stage
.
Cell
30
,
675
686
.
Nieuwkoop
,
P. D.
and
Faber
,
J.
(
1967
).
Normal Table of Xenopus laevis (Daudin)
.
North-Holland Publishing Co
.,
Amsterdam
.
Okamoto
,
H.
and
Mitani
,
S.
(
1987
).
Reconstitution of the neural induction process in vitro
.
Neurosci. Res. Suppl
.
5
,
s61
.
Perkins
,
M. E.
,
Ji
,
T. H.
and
Hynes
,
R. O.
(
1979
).
Cross-linking of fibronectin to sulfated proteoglycans at the cell surface
.
Cell
16
,
941
952
.
Richter
,
H.
,
Went
,
C.
and
Hormann
,
H.
(
1985
).
Aggregation and fibril formation of plasma fibronectin by heparin
.
Biol. Chem. Hoppe-Seyler
336
,
509
514
.
Sakashita
,
S.
,
Engvall
,
E.
and
Ruoslahti
,
E.
(
1980
).
Basement membrane glycoprotein laminin binds to heparin
.
FEBS Lett
.
116
,
243
246
.
Share
,
S. R.
and
Gerhart
,
J. C.
(
1983
).
Axis determination in eggs of Xenopus laevis’. A critical period before first cleavage, identified by the common effects of cold, pressure and ultraviolet irradiation
.
Devi Biol
.
99
,
75
87
.
Slack
,
C.
,
Warner
,
A. E.
and
Warren
,
R. L.
(
1973
).
The distribution of sodium and potassium in amphibian embryos during early development
.
J. Physiol
.
232
,
297
312
.
Slack
,
J. M. W.
(
1983
).
From Egg to Embryo-Determinative Events in Early Development
.
Cambridge University Press
,
Cambridge
.
Slack
,
J. M. W.
,
Darlingon
,
B. G.
,
Heath
,
J. K.
and
Godsave
,
S. F.
(
1987
).
Mesoderm induction in early Xenopus embryos by heparin-binding growth factors
.
Nature, Lond
.
326
,
197
200
.
Spemann
,
H.
(
1938
).
Embryonic Development and Induction
.
Yale University Press
.
Yamada
,
K. M.
(
1983
).
Cell surface interactions with extracellular materials
.
A. Rev. Biochem
.
52
,
761
799
.