DEVELOPMENTAL
BIOLOGY
144, 177-188 (1991)
The B30 Ganglioside Is a Cell Surface Marker for Neural Crest-Derived Neurons in the Developing Mouse DIDIER Department
of Cellular
Y. R. STAINIER,*
H. BILDER, AND WALTER GILBERT
DAVID
and Developnwatal Biology and *Department of Biochemistry Massachusetts 02138 Harvard University, Cambridge,
and Molecular
Biology,
Accepted November 27, 1990
We have previously reported the isolation of a monoclonal antibody, mAb B30, that recognizes two minor gangliosides specifically expressed in a small subset of neurons in the developing mouse central nervous system (Stainier and Gilbert, 1989).B30 labels mesencephalic trigeminal neurons shortly after differentiation until about 2 weeks after birth. Postnatally, it also labels two specific monolayers of cerebellar neurons. In this study, we have characterized the B30 immunoreactivity in the developing peripheral nervous system of the mouse. We report that B30 is a marker for neural crest-derived neurons and have used it to follow the neuronal differentiation of neural crest cells in a serum-free chemically defined culture system. Within hours after plating, neural crest cells migrate away from the neural tube explant on a fibronectin or laminin substrate and by 24 hr, up to 15% of them have differentiated into morphologically identifiable neurons. In vitro as in vivo, undifferentiated mouse neural crest cells express the GD3 ganglioside which is recognized by mAb B33, and neural crest-derived neurons can be labeled by mAbs B33, B30, and also E1.9, a specific neuronal cytoskeletal marker. We also show the unique biochemical specificity of mAb B30 and provide experimental evidence for the role of the B30 ganglioside in the cellular adhesion process. o 1x11 Academic press. he. INTRODUCTION
In a search for cell surface molecules involved in cellcell and cell-substratum interactions in the embryonic nervous system, we isolated a monoclonal antibody, mAb B30, which identifies specific subsets of neurons in the developing mouse central nervous system (CNS). B30 defines neurons in the mesencephalic trigeminal nucleus (MesV), appearing at Embryonic Day 9.5 (E9.5), shortly after differentiation. In the postnatal cerebellum, B30 stains the postmitotic premigratory granule cells very transiently, until they start migrating through the molecular layer; it also stains mature Purkinje cells (Stainier and Gilbert, 1989). The distribution of the B30 antigen in the developing mouse peripheral nervous system (PNS) reveals its specific association with neural crest-derived neurons as well as with placode-derived neurons of the sensory ganglia. We tested this specificity in explant cultures of embryonic neural tubes in a serum-free chemically defined medium, and we show that in such cultures B30 is a specific marker for neural crest-derived neurons. Because neural crest cells give rise to a wide variety of cell types, neuronal and others (reviewed by Le Douarin, 1982; see also Anderson, 1989), there has been a long standing interest in establishing an in vitro differentiation system for these cells. Recently, Ziller et al. (1983) developed a defined culture medium which allows the differentiation of avian neural crest cells into derivatives with different phenotypes. This fully defined me177
dium has also been used by Boisseau and Simonneau (1989) to monitor the neuronal differentiation of mouse neural crest cells. The isolation of cell surface markers for undifferentiated mouse neural crest cells (mAb B33) as well as for crest-derived neurons (mAb B30) is instrumental for studying the diversification of crest derivatives. Furthermore, the demonstration that B30 is a marker for neural crest-derived neurons supports the presumed neural crest origin of MesV neurons (Narayanan and Narayanan, 1978). MATERIALS
AND METHODS
Mono&ma1 antibodies. mAb B30 has been previously described (Stainier and Gilbert, 1989). mAb El.9 recognizes a cytoplasmic epitope in the primary sensory and motor neurons during axonal outgrowth, appearing at E8.5 and disappearing by El2 in the mouse (Stainier and Gilbert, 1990). In neural crest cultures, it labels all neural crest-derived neurons as soon as they differentiate. mAbs B30 and El.9 were deposited with the Developmental Studies Hybridoma Bank and are available upon request. mAb B33 is a mouse IgM that recognizes the GD3 ganglioside as shown by thin-layer chromatography (TLC) blot analysis of acidic glycolipids (unpublished). It labels neural crest and placode cells as well as crest- and placode-derived neurons in vivo and in vitro (Stainier and Gilbert, submitted for publication). mAb lC6 is a mouse IgG that binds a surface antigen expressed on all PC12 cells. It was isolated from a fusion in 0012-1606/91 $3.00 Copyright All rights
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
178
DEVELOPMENTAL BIOLOGY
which the immunogen consisted of a preparation of phosphatidylinositol-specific phospholipase C clipped proteins and other secreted molecules from PC12 cells (Stainier, 1990). mAb 4C1, a mouse IgG, also binds a cell surface antigen on all PC12 cells and was isolated from a fusion similar to the one that contained the lC6 hybridoma (Stainier, 1990). These two antibodies were used as control antibodies for the adhesion studies on PC12GlO cells. Immunohistochemistry. Pregnant CD1 mice were obtained from Charles River under a specific breeding schedule: mice were bred for 3 hr from 12 noon to 3 PM on EO and our embryonic day ran from noon to noon. Animals were sacrificed by cervical dislocation or with halothane. Embryos (E8-E12) staged according to Theiler (1989) were slit longitudinally along the forming neuropore and processed as half-embryo wholemounts. Brains of later embryos and postnatal animals were cut into lOO- to 120-pm-thick vibratome sections and stained as previously described (Stainier and Gilbert, 1989). All reactions were done at room temperature (with gentle shaking) in polystyrene culture dishes by sequential incubations in the following: 10% normal goat serum (NGS) in phosphate-buffered saline (PBS) for 1 hr; hybridoma culture supernatant overnight (1218 hr); after three lo-min washes with 10% NGS/PBS, fluoresceinated goat anti-mouse IgM (Southern Biotechnology Associates, Birmingham, AL), diluted l:lOO, for 2 hr. After three more lo-min washes in 10% NGS/PBS, the wholemounts and thick sections were mounted in fresh mounting medium between two coverslips separated by silicone grease. For El.9 and BlO staining, the tissue was fixed with 1% paraformaldehyde for 10 min, rinsed several times with PBS, and incubated for 2 hr in 10% NGS. We used a Lasersharp MRC-500 confocal microscope mounted with fluorescence optics to observe our specimens. Briefly, serial sections were collected and combined by projection. The resulting images were stored in a WORM optical disk (Maxtor Corp., San Jose, CA) and printed using a video printer. Explant cultures and staining. Trunk fragments (two to three somites long) were dissected out of the lower spinal cord of E9 (13- to 20-somite-stage) mouse embryos. The fragments were placed in a 0.5 mg/ml solution of dispase (Boehringer-Mannheim) for 8 min and then washed in PBS with 10% fetal bovine serum for 2 min. The neural tube was carefully dissected away from surrounding somites, membranes, and notochord. The tube was washed in Ham’s F12 medium, plated on a 35-mm collagen-coated culture dish, incubated for 4 hr with 10 pg/ml fibronectin or 20 pg/ml laminin, and cultured at 37°C with 5% CO, in fresh Basic Brazeau Medium (BBM) (Ziller et al, 1983). This medium consists of a 6:3:1 mixture of Ham’s F12 nutrient mixture, Dul-
VOLUME 144, 1991
becco’s modified Eagle medium, and BGjb Fitton-Jackson modification; supplemented with 2 g/liter bovine serum albumin, 2.38 g/liter Hepes buffer, 100 IU/ml penicillin/streptomycin, 100 pg/liter hydrocortisone, 1 pg/liter insulin, 0.4 pg/liter triiodothyronine, 10 ng/ liter glucagon, 0.2 rg/liter parathyroid hormone, 10 mg/ liter transferrin, 0.1 pg/liter epithelial growth factor, and 0.2 pg/liter fibroblast growth factor. Cultures were stained as described above. The primary antibody used was B30 or B33 culture supernatant. For double-labeling analysis, B30 was used with 15 pg/ml of concentrated R24 (a mouse IgG3 that also identifies the GD3 ganglioside; the generous gift of Dr. K. Lloyd). Secondary antibodies were used at 1:200 dilution for 20 min. Gang&o&de preparation and analysis. Freshly dissected postnatal mouse tissue or cultured PC12GlO cells were homogenized in chloroform/methanol (2:1, vol/ vol) and extracted with 0.12 1MKC1 (Suzuki, 1965). The extracts were dried by rotary evaporation, taken up in water, desalted through Bond-Elut columns (Analytithem International), and lyophilized for storage at -20°C. TLC analysis of gangliosides was done on aluminiumbacked TLC plates (silica gel 60, 0.2-mm-thick, E. Merck). The plates were developed in chloroform/methanol/0.25% aqueous CaCl, (60:40:10). Ganglioside standards were obtained from Calbiochem and were detected by resorcinol stain. After drying, the plates were dipped in 0.05% poly(isobutylmethacrylate) (Polysciences) in hexane for 90 sec. The dried chromatograms were sprayed with PBS before soaking in a blocking solution (PBS, 3% BSA, and 5% NGS) for 30 min (all the incubations were done at room temperature). The chromatograms were overlaid with the first antibody (diluted hybridoma supernatant) for 3 hr. After repeated washes in PBS, the chromatograms were overlaid with alkaline phosphatase-conjugated second antibody (diluted 1:lOOO)for 1 hr. After repeated washes in PBS, BCIP and NBT (BRL) were used at a 1:l molar ratio in 0.1 M Tris-HCl (pH 9.5), 0.1 M NaCl, 50 mM MgCl, for the color reaction. Upon completion, the reaction was stopped in water. Enzyme treatments and chromatography on DEAESephadex A-25 (Pharmacia) were done as described by Schwarting et al. (1987). Adhesion assays. PClZGlO cells, from a 98% B30 positive PC12 subclone isolated by sequential dilution cloning, were assayed for adhesion on different substrates: collagen (Vitrogen loo), fibronectin (10 pg/ml), and laminin (20 pg/ml). (PC12 cells are rat pheochromocytoma cells). First, PClZGlO cells, labeled with B30 in suspension, were assayed for reattachment by visual inspection. Second, adherent PClBGlO cells were exposed to growth medium containing mAb B30 or mAb lC6 and
STAINIER, BILDER, AND GILBERT
examined for morphological changes. (Another control antibody, mAb 4C1, that binds to the surface of all PClZGlO cells was also tested in these assays and had no effect on PClZGlO cellular morphology or reattachment.) Cell viability was measured by trypan blue exclusion and was not affected by either treatment. PClZGlO cells are refractory to NGF induction. RESULTS
B30 is a Marker for Neural
Crest-Derived
Neurons
During the development of the mouse CNS, there is a specific and transient expression of the B30 antigen on the surface of MesV neurons, starting shortly after their differentiation until about 2 weeks after birth (Stainier and Gilbert, 1989). MesV neurons are the only primary sensory neurons in the CNS and they are thought to derive from the neural crest (Narayanan and Narayanan, 1978). In the periphery, primary sensory neurons (neurons of the cranial nerves and of the dorsal root ganglia (DRG)) exhibit B30 immunoreactivity on both their central and their peripheral projections. Figure 1 shows El0 mouse embryos stained as wholemounts with mAb B30 and viewed at the level of the thoracic spinal cord. Individual neural crest cells differentiate into B30 immunoreactive DRG neurons (Figs. 1A and lB), which segregate into cohesive and segmentally repeated ganglia (Figs. 1C and 1D). Other neural crestderived neurons such as sympathetic and parasympathetic neurons are also B30 immunoreactive (data not shown). In the trigeminal ganglion, neurons derive from both the neural crest and the epidermal placodes, and all are B30 immunoreactive (Stainier and Gilbert, 1989). Neurons of the VIIth and VIIIth ganglia which are largely placode derived (Verwoerd and van Oostrom, 1979; Altman and Bayer, 1982) are also B30 positive (see the B30-labeled vestibulocerebellar fibers in Fig. 5, Stainier and Gilbert, 1989). This distribution indicates that, in the mouse, mAb B30 stains the neural crest-derived neurons as well as the placode-derived neurons of the sensory ganglia; other placode-derived structures, such as the olfactory epithelium, do not display B30 immunoreactivity. The characteristics and time course of the B30 immunoreactivity in the developing PNS follow the pattern described for the trigeminal system (Stainier and Gilbert, 1989, 1990): immunoreactivity appears about 12 hr after neuronal differentiation and disappears around 2 weeks after birth. Neural Crest Diferentiation
In vitro
Neural tube explant culture. To study the specificity of mAb B30 for neural crest-derived neurons, we set up a serum-free culture system in which to follow the neuro-
179
Mouse Neural Crest Surface Markers
nal differentiation of neural crest cells. We dissected lower trunk fragments (two to three somites long) of E9 mouse embryos (13-20 somites), freed the neural tube from surrounding tissues after a mild dispase treatment, and cultured it in a serum-free chemically defined medium containing hormones, growth factors, and transferrin (BBM) (Ziller et al, 1983). Fibronectin and laminin gave the best adhesion and migration of mouse neural crest cells, whereas collagen alone, as a film or a gel, allowed little or no migration. Figures 2A and 2B show such E9 neural tube explants cultured on fibronectin-coated dishes. Both cellular migration and axonal outgrowth from the explant are extensive. Undifferentiated neural crest cells, flat and polygonal, express the GD3 ganglioside (as revealed by B33 or Ii24 staining, data not shown). The long axons extending from the explant (Fig. 2B, arrows) are the El.9 positive, B30 negative projections of the primary motor neurons of the neural tube. Characterization of the neuronal phenotype. By 18 hr in culture, numerous cells have migrated away from the explant and have acquired a neuronal morphology with round, refringent perikaria and long, thin unbranched processes (Figs. 2C-2E). Staining such cultures with mAb El.9 confirmed the neuronal identity of these cells (Fig. 2F). As shown in Fig. 3, l-day-old explanted neural tube cultures contain B30 positive (arrows) as well as B30 negative (arrowheads) neurons. All the B30 positive cells are El.9 positive, confirming their neuronal identity and the B30 positive and negative neurons express the GD3 ganglioside, which, in these cultures, supports their neural crest origin (data not shown). Biochemical and Functional Ganglioside
Analysis
of the B30
Biochemical analysis. As shown in Stainier and Gilbert (1989; Fig. 12), the B30 immunostaining of a TLC plate of acidic glycolipids reveals two discrete gangliosides expressed in the postnatal midbrain and brainstem, dissected to include the trigeminal sensory components, and in the postnatal cerebellum. One of the bands migrates slightly above GM1 and the other migrates between GDl, and GD3 in a chloroform/methanol/ aqueous CaCl, solvent system. This migration pattern is similar to that of the D1.l antigen (Cheresh et al, 1984) and the Jones antigen (Schlosshauer et al, 1988). These two antibodies were originally described in the developing rat CNS as staining germinal cells (D1.l) (Cheresh et al., 1984) and as associated with neural cell and process migration (Jones) (Mendez-Otero et al., 1988). Now they have been shown to have the same specificity for Oacetylated-GD3 and other acetylated gangliosides: base treatment of brain gangliosides converts O-acetylated-
180
DEVELOPMENTALBIOLOGYV0~~~~144,1991
STAINIER, BILDER, AND GILBERT
GD3 into GD3 as revealed by mAb R24 staining, and this treatment causes the loss of the D1.l and Jones immunoreactivity (Cheresh et aZ.,1984; Blum and Barnstable, 1987). We base treated a TLC plate of P12 acidic glycolipids and analyzed it by B30 (Fig. 4, lanes a and c) and R24 (Fig. 4, lane b) immunostaining. Figure 4 shows that the B30 epitope was not affected by base treatment (compare lane a to the control lane c) and that the B30 ganglioside was not converted into GD3 (lane b). However, base treatment revealed a third B30 immunoreactive band between the two original bands (lane a). On the basis of this biochemical analysis of the B30 epitope and of the different distribution of the D1.l (and Jones) antigen with the B30 antigen, we conclude that mAb B30 has a unique specificity. Functional analysis. We searched for a B30 immunoreactive cell line that would enable us to investigate the function of the B30 ganglioside. We found that 1 in lo4 PC12 cells was B30 positive and that the B30 expressing cells were always clustered. We isolated several B30 positive subclones by sequential dilution cloning and isolated their acidic glycolipids for analysis. Figure 5 shows the B30 immunostaining of acidic glycolipids from one of these subclones, PClZGlO. Only the faster migrating B30 ganglioside is expressed and it migrates as a doublet (lane b). Recent reports have provided evidence that cell surface gangliosides may be involved in cellular interactions and adhesion (Cheresh et ak, 1986; Sariola et al., 1988). Specifically, gangliosides are thought to redistribute actively into discrete areas of the cell, and they may act synergistically with cell surface receptors by creating an appropriate electrostatic environment (Cheresh and Klier, 1986; Cheresh et ah, 1986; Burns et al, 1988). In light of these studies, we investigated the effect of B30 on the adhesion of PClZGlO cells. In the first of a series of qualitative experiments, we observed that the B30 immunostaining of PC12GlO cells in suspension inhibited their subsequent reattachment to fibronectin or laminin. Then, to determine whether cells preattached to these substrates could be displaced by B30, PClZGlO cells were allowed to settle and bind on the plate for a minimum of 18 hr, after which they were overlaid with growth medium containing 50 pg/ml of mAb B30 or 50 pg/ml of mAb lC6, which binds to an antigen present on the surface of all PC12 cells (Stainier, 1990). Figure 6 shows that the addition of mAb B30 to these flattened
MmLseNeural Crest Surface Markers
181
cells caused a rapid and significant cellular retraction and rounding up. The speed of this effect increased with both concentration of antibody and temperature (complete release from the substrate in 15 min at 37°C). Figure ‘7 shows as a parallel control that cells treated with mAb lC6 exhibit no difference in their morphology or their adherence. Figure 8 shows the immunostaining pattern of the cells shown in Fig. 7. The mAb B30 caused the PClZGlO cells that express the B30 antigen to round up (arrow) and also to detach from the plate, while mAb lC6 stained all the cells but had no visible effect on their morphology. This effect was observed on a variety of substrates, including laminin, fibronectin, and collagen. In primary cultures, mouse trigeminal neurons show an even distribution of the B30 ganglioside on the cell body and axonal projections, including the growth cones (Stainier and Gilbert, submitted for publication). Addition of growth medium containing 50 pg/ml of mAb B30 on trigeminal neurons, sparsely plated on a laminin substrate, caused the cell bodies to detach although the neurites remained adherent (data not shown). These experiments suggest a differential adhesion along the extended trigeminal neuron. Also, when we used the confocal microscope to examine developing sensory ganglia, we observed a higher concentration of the B30 antigen at interneuronal junctions (data not shown). Similarly, GD2 and GD3 gangliosides have been reported to be heavily concentrated at the interface of cultured melanoma cells and their substrate (Cheresh et ah, 1984). These observations are consistent with the presumed role of gangliosides in cellular interactions and adhesion. DISCUSSION
Neural crest cells represent a unique system in which to study how uncommitted embryonic cells acquire particular developmental fates. These progenitor cells give rise to a wide variety of cell types, ranging from sensory and autonomic ganglia neurons, and Schwann cells, to facial bone, cartilaginous structures, and melanocytes (reviewed by Le Douarin, 1982). Recent experimental evidence, in vivo and in vitro, indicates that some neural crest cells are multipotent at the time of their initial migration (Baroffio et ah, 1988; Bronner-Fraser and Fraser, 1988). These progenitors proliferate as they migrate in the periphery and make further differentiation
FIG. 1. (A, B) mAb B30 stains developing DRG neurons. El0 (32~somite-stage) mouse embrvo stained with mAb B30 and viewed at the level of the thoraeic spinal cord (parasagittal view). Two newly differentiated and B30 immunoreactive DRG neurons project centrally where their axons run rostrally and caudally for a few segments. Their peripheral projections are not yet visible. Scale bars, (A) 50 pm; (B) 10 pm. (C, D) E10.5 (35- and 3%somite-stage) mouse embryos stained with mAb B30 and viewed at the level of the thoracic spinal cord (parasagittal view). B30 immunoreactive DRG neurons segregate into cohesive and segmentally repeated structures or ganglia. The arrows point to somitic boundaries. Scale bars, 50 Km.
182
DEVELOPMENTALBIOLOGY
VOLlJME144,1%1
FIG. 2. Neuronal differentiation of neural crest cells in culture. (A) Explanted neural tube after 12 hr in culture. Flat, polygonal neural crest cells migrate away from the neural tube on a fibronectin-coated dish. fp, floor plate. Scale bar, 300 pm. (B) Explanted neural tube after 24 hr in culture. Motor neurons project axons (arrows) from the neural tube while the neural crest cells are migrating. These projections are El.9 positive but B30 negative (data not shown). NT, neural tube. Scale bar, 50 pm. (C) Neurons appear by 18 hr of culture. Away from the explanted neural tube, differentiated neurons (arrows) mix with undifferentiated crest cells (arrowhead). Scale bar, 50 hrn. (D, E) Up to 15% of the neural crest cells differentiate into neurons after 24 hr in culture. Away from the explant, numerous neurons are clearly visible and extend long and thin processes. Scale bars, 50 pm. (F) Neural crest-derived neurons are El.9 positive. El.9 labels cells, and their processes, which have migrated away from the explanted neural tube. Scale bar, 50 pm.
STAINIER,BILDER,ANDGILBERT
Mouse Neural Crest Surface Markers
183
FIG. 3. (A-F) B30 immunostaining of neural crest-derived neurons. Explanted neural tube cultures were stained with mAb B30 24 hr after dissection. The cultures contain, away from the neural tube, B30 positive (arrows) and B30 negative (arrowheads) neurons. Scale bars, 50 pm.
184
DEVELOPMENTALBIOLOGY
a
b
c
FIG. 4. Biochemical analysis of the B30 epitope: base treatment of mouse acidic glycolipids reveals a third B30 immunoreactive band. P12 cerebellar acidic glycolipids containing 0.5-l pg sialic acid were applied to the TLC plate and developed with chloroform/methanol/ 0.25% aqueous CaCl, (60:40:10). The plate was then exposed to ammonium hydroxide vapor for 3 hr (or water vapor as a control, lane c) and then immunostained with mAb B30 (lanes a and c) or mAb R24 (lane b). (a) base-treated acidic glycolipids stained with B30. (b) basetreated acidic glycolipids stained with R24. (c) control acidic glycolipids stained with B30. Ganglioside standards from Calbiochem were run in parallel and stained with resorcinol. The arrowheads point to the two original B30 immunoreactive gangliosides.
choices in response to factors in their local environment (reviewed by Anderson, 1989). The isolation and characterization of surface markers for mouse crest cells (such as mAb B33) and some of their derivatives (such as mAb B30) is critical for the further understanding of the cellular mechanisms involved in lineage segregation in this system. Indeed, the
vOLUME144,1991
use of cell surface antibodies to distinguish and manipulate specific cell types and their precursors in culture has led to some elegant work on glial cell determination in the rat optic nerve (reviewed by Raff, 1989). Raff and co-workers were able to show that both cell-cell interactions and intrinsic cellular programs contribute to the diversification of glial cells; they also identified the signaling molecules that mediate some of the cell-cell interactions. Approaches such as cell sorting, immunoablation, and quantitative clonal analysis are being applied to analyze the differentiation of chick neural crest cells (Maxwell et al, 1988; Vogel and Weston, 1988). But, as a variety of molecular genetic strategies are now available to manipulate the mouse genome in both somatic and germ cells (reviewed by Rossant, 1990), there is considerable advantage in establishing a culture system for mouse neural crest differentiation. Using the serum-free chemically defined medium first described by Ziller et al. (1983), we were able to show neuronal differentiation within 18 hr of culture. The observation that, by 24 hr, some neurons are B30 positive while others are B30 negative may reflect the stage of differentiation of these neurons or may indicate the phenotypic diversity of the neuronal derivatives in vitro. Careful immunoablation experiments should resolve this question. The B30 staining pattern in viva indicates that the B30 antigen appears within 12 hr after neuronal differentiation and that all neural crest-derived neurons are B30 immunoreactive. In the CNS, MesV neurons exhibit a characteristic B30 staining pattern: like the neural crest-derived neurons of the PNS, they become B30 positive shortly after differentiation until about 2 weeks after birth. This specific B30 staining pattern of MesV neurons supports the notion that they are neural crest derived. Indeed, celltracing studies in quail-chick chimaeras provided the first line of evidence for the neural crest origin of the centrally located MesV neurons (Narayanan and Narayanan, 1978), although other experiments questioned these results (Straznicky and Rush, 1985). Functional
a
b
FIG. 5. B30 immunostaining of PC12GlO acidic glycolipids. Acidic glycolipids containing 0.5-l pg sialic acid were applied to the TLC plate and developed with chloroform/methanol/0.25% aqueous CaCl, (60:40:10). The plate was then immunostained with mAb B30. (a) acidic glycolipids from P15 mouse cerebellum. (b) acidic glycolipids from PClZGlO cells. Ganglioside standards were run in parallel.
Sign&ance
of the B30 Ganglioside
The B30 epitope is resistant to base treatment and is thus different from the D1.l (or Jones) epitope. It is carried by two minor gangliosides expressed on mouse neural crest-derived neurons (and certain cerebellar neurons) (Stainier and Gilbert, 1989). This restricted distribution raises obvious questions about the potential role and function of these rare gangliosides. While gangliosides have been implicated in a wide range of phenomena from neurotrophism to regeneration (reviewed by Hannun and Bell, 1989), there is good evidence for the involvement of cell surface gangliosides in cellular interactions and adhesion. The adhesion of
STAINIER, BILDER, AND GILBERT
Mcruse Neural
Crest Surface Markers
185
FIG. 6. mAb B30 causes PClZGlO cells to round up and detach from a laminin substrate. PClZGlO cells were allowed to adhere on a laminin substrate for at least 18 hr, after which they were overlaid with growth medium containing 50 rg/ml of B30 and kept at room temperature. Addition of the antibody caused the flattened cells to retract (arrow), round up, and sometimes detach from the substrate. The time course of the experiment is shown in minutes in the upper right hand corner of each micrograph. The micrographs at the bottom of the panel show a Hoffman phase view of the PClZGlO cells followed in this experiment and the distribution of B30 immunoreactivity on these cells.
186
DEVJ&OPMENTAL BIOLOGY
VOLUME 144.1991
FIG. 7. mAb B30 causes PClZGlO cells to round up (arrow), whereas mAb lC6 has no effect on cellular morphology. PClZGlO cells growing on laminin were overlaid with growth medium containing 50 pg/ml of mAb B30 (left) or 50 pg/ml of mAb lC6 (right) and kept at room temperature. The time course is in minutes.
postnatal rat cerebellar cells to fibronectin, inhibited by Arg-Gly-Asp peptides that block the recognition of fibronectin to its receptor, is partially inhibited by antibodies to the fibronectin receptor and is slowed by antibodies to the D1.l ganglioside (Stallcup et al., 1989). These studies implicate both the fibronectin receptor and the D1.l ganglioside in the adhesion of cerebellar cells to fibronectin. The B30 gangliosides are expressed on specific sets of neurons during axon outgrowth and synapse formation. Using PClZGlO cells, a B30 positive subclone of the neural crest-derived PC12 cells, we showed that B30 labeling inhibited the subsequent reattachment of immunoreactive cells, and that the addition of mAb B30 to the culture medium caused flattened cells to retract, round up, and even detach from various substrates. These experiments suggest a role for the B30 gangliosides (at least the faster migrating one) in the cellular adhesion process; however, the experiments on trigeminal neurons point to additional adhesive interactions for the
neuronal processes. The B30 gangliosides may thus be involved in maintaining the cohesion of specific groups of neurons. In the cerebellum, it may be involved in maintaining the cohesion of the postmitotic granule cell layer before migration through the molecular layer; in the periphery, it may be involved in holding together crest-derived neuronal structures, such as the sensory ganglia. We thank Dr. K. Lloyd for mAb R24, Dr. Simonneau for interesting discussions, and C. Alvarez and M. Grether for critical reading of the manuscript.
REFERENCES ALTMAN, J., and BAYER, S. A. (1982). Development of the cranial nerve ganglia and related nuclei in the rat. Adv. Anat. Embryo.!. Cell. Biol 74, l-89. ANDERSON, D. J. (1989). The neural crest cell lineage problem: Neuropoiesis? Neuron 3, l-12. BARROFIO, A., DUPIN, E., and LE DOUARIN, N. N. (1988). Clone-forming
STAINIER, BILDER, AND GILBERT
Mouse Neural Crest Surface Markers
18’7
FIG. 8. Distribution of B30 and lC6 immunoreactivity on PC12GlO cells. The PClZGlO cells shown in Fig. 7 were incubated with fluoresceinated secondary antibody for 15 min. The left panels show the distribution of B30 immunoreactivity. Around the cells that have retracted, B30 labels substrate-attached adhesion plaques. By this stage many cells have detached from the plate. The arrow points to a group of cells that rounded up but did not detach (see Fig. 7). The right panels show the distribution of lC6 immunoreactivity. The arrow points to a flat cell that showed no visible morphological changes throughout the experiment.
ability and differentiation potential of migratory neural crest cells. Proc. Natl. Acad. Sci. USA 85,5325-5329.
BLUM, A. S., and BARNSTABLE,C. J. (1987). O-acetylation of a cell-surface carbohydrate creates discrete molecular patterns during neural development. Proc. Natl. Acad. Sci. USA 84,8716-8720. BOISSEAU,S., and SIMONNEAU,M. (1989). Mammalian neuronal differentiation: Early expression of a neuronal phenotype from mouse neural crest cells in a chemically defined culture medium. Development 106,665-674. BRONNER-FRASER, M., and FRASER,S. E. (1988). Cell lineage analysis reveals multipotency of some avian neural crest cells. Nature 335, 161-164. BURNS,G. F., LUCAS,C. M., KRISSANSEN,G. W., WERKMEISTER,J. A., SCANLON,D. B., SIMPSON,R. J., and VADAS, M. A. (1988). Synergism between membrane gangliosides and Arg-Gly-Asp-directed glycoprotein receptors in attachment to matrix proteins by melanoma cells. J. Cell Biol. 107, 1225-1230. CHERESH, D. A., HARPER, J. R., SCHULZ, G., and REISFELD, R. A. (1984). Localization of the gangliosides GD2 and GD3 in adhesion plaques and on the surface of human melanoma cells. Proc. Natl. Acad. Sci. USA 81,5767-5771.
CHERESH,D. A., and KLIER, F. G. (1986). Disialoganglioside GD2 distributes into substrate-associated microprocesses on human melanoma cells during their attachment to fibronectin. J. Cell Biol 102, 1887-1897. CHERESH,D. A., PIERSCHBACHER, M. D., HERZIG,M. A., and MUJOO,K. (1986). Disialogangliosides GD2 and GD3 are involved in the attachment of human melanoma and neuroblastoma cells to extracellular matrix proteins. J. Cell Biol. 102.688-696. CHERESH,D. A., REISFELD,R. A., and VARKI, A. P. (1984). O-acetylation of disialoganglioside GD3 by human melanoma cells creates a unique antigenic determinant. Science 225,844-846. HANNUN, Y. A., and BELL, R. M. (1989). Function of sphingolipids and sphingolipid breakdown products in cellular regulation. Science 243, 500-507.
LE DOUARIN,N. M. (1982). “The Neural Crest,” pp. 60-67. Cambridge Univ. Press. Cambridge. MAXWELL, G. D., FORBES,M. E., and CHRISTIE,D. S. (1988). Analysis of the development of cellular subsets present in the neural crest using cell sorting and cell cultures. Neuron 1,557-568. MENDEZ-OTERO,R., SCHLOSSHAUER,B., BARNSTABLE,C., and CONSTANTINE-PATON,M. (1988). A developmentally regulated antigen
188
DEVELOPMENTALBIOLOGY
associated with neural cell and process migration. J. Neurosc~ 8, 564-579. NARAYANAN, C. H., and NARAYANAN, Y. (1978). Determination of the embryonic origin of the mesencephalic nucleus of the trigeminal nerve in birds. J. EmbyoL Exp. MwphoL 43,85-105. PUKEL, C. S., LLOYD,K. O., TRAVASSOS,L. R., DIPPOLD,W. G., OETPJXNGkN, H. F., and OLD, L. J. (1982). GD3, a prominent ganglioside of human melanoma: Detection and characterization by a mouse monoclonal antibody. J. Ezp. Med. 155,1133-114’7. RAFF, M. C. (1989). Glial cell diversification in the rat optic nerve. Science 243,1450-1455. ROSSANT, J. (1990). Manipulating the mouse genome: Implications for neurobiology. Neuron 2,323-334. SARIOLA, H., AUFDERHEIDE, E., BERNHARD, H., HENKE-FAHLE, S., DIPPOLD,W., and EKBLOM,P. (1988). Antibodies to cell surface ganglioside GD3 perturb inductive epithelial-mesenchymal interactions. Cell 54, 235-245. SCHLOSSHAUER,B., BLUM, A. S., MENDEZ-OTERO,R., BARNSTABLE, C. J., and CONSTANTINE-PATON,M. (1988). Developmental regulation of ganglioside antigens recognized by the Jones antibody. J. Neurosci. 8.580-592. SCHWARTING,G. A., JIJNGALWALA,F. B., CHOU,D. K., BOYER,A. M., and YAMAMOTO,M. (1987). Sulfated glucuronic acid-containing glycoconjugates are temporally and spatially regulated antigens in the developing mammalian nervous system. Dev. BioL 120,65-76. SCOTT,S. A. (1987). The development of skin sensory innervation patterns. Trends Neurosei. 10,468-473. STAINIER, D. Y. R. (1990). Neuronal differentiation and maturation in the mouse trigeminal sensory system. PhD thesis, Harvard University, Cambridge, MA.
VOLUME 144,1991
STAINIER,D. Y., and GILBERT,W. (1989). The monoclonal antibody B30 recognizes a specific neuronal cell surface antigen in the developing mesencephalic trigeminal nucleus of the mouse. J. Neuroaci 9,24682485. STAINIER, D. Y. R., and GILBERT, W. (1990). Pioneer neurons in the mouse trigeminal sensory system. Proc. NatL Ad. Sci. USA 87, 923-927. STALLCUP,W. B. (1988). Involvement of gangliosides and glycoprotein fibronectin receptors in cellular adhesion to fibronectin. Exp. Cell Res. 177,90-102. STALLCUP,W. B., PYTELA, R., and RUOSLAHTI,E. (1989). A neuroectoderm-associated ganglioside participates in fibronectin receptormediated adhesion of germinal cells to fibronectin. Dev. BioL 132, 212-229. STRAZNICKY,C., and RUSH, R. A. (1985). The effect of nerve growth factor on the developing primary sensory neurons of the trigeminal nerve in chick embryos. Anat. EmbyoL 171,91-95. SUZUKI, K. (1965). The pattern of mammalian brain gangliosides. III. Regional and developmental differences. J. Neurochem. 12,969-979. THEILER, K. (1989). “The House Mouse,” 2nd ed., Springer-Verlag, Berlin. VERWOERD,C. D. A., and VAN OOSTROM,C. G. (1979). Cephalic neural crest and placodes. EmbrgoL Cell BioL 58, l-75. VOGEL, K. S., and WESTON,J. A. (1988). A subpopulation of cultured avian neural crest cells has transient neurogenic potential. Neuron 1,569-577. ZILLER, C., DUPIN, E., BRAZEAU, P., PAIJLIN, D., and LE DOUARIN,N. (1983). Early segregation of a neuronal precursor cell line in the neural crest as revealed by culture in a chemically defined medium. Cell 32, 627-638.
BIOLOGY
144, 177-188 (1991)
The B30 Ganglioside Is a Cell Surface Marker for Neural Crest-Derived Neurons in the Developing Mouse DIDIER Department
of Cellular
Y. R. STAINIER,*
H. BILDER, AND WALTER GILBERT
DAVID
and Developnwatal Biology and *Department of Biochemistry Massachusetts 02138 Harvard University, Cambridge,
and Molecular
Biology,
Accepted November 27, 1990
We have previously reported the isolation of a monoclonal antibody, mAb B30, that recognizes two minor gangliosides specifically expressed in a small subset of neurons in the developing mouse central nervous system (Stainier and Gilbert, 1989).B30 labels mesencephalic trigeminal neurons shortly after differentiation until about 2 weeks after birth. Postnatally, it also labels two specific monolayers of cerebellar neurons. In this study, we have characterized the B30 immunoreactivity in the developing peripheral nervous system of the mouse. We report that B30 is a marker for neural crest-derived neurons and have used it to follow the neuronal differentiation of neural crest cells in a serum-free chemically defined culture system. Within hours after plating, neural crest cells migrate away from the neural tube explant on a fibronectin or laminin substrate and by 24 hr, up to 15% of them have differentiated into morphologically identifiable neurons. In vitro as in vivo, undifferentiated mouse neural crest cells express the GD3 ganglioside which is recognized by mAb B33, and neural crest-derived neurons can be labeled by mAbs B33, B30, and also E1.9, a specific neuronal cytoskeletal marker. We also show the unique biochemical specificity of mAb B30 and provide experimental evidence for the role of the B30 ganglioside in the cellular adhesion process. o 1x11 Academic press. he. INTRODUCTION
In a search for cell surface molecules involved in cellcell and cell-substratum interactions in the embryonic nervous system, we isolated a monoclonal antibody, mAb B30, which identifies specific subsets of neurons in the developing mouse central nervous system (CNS). B30 defines neurons in the mesencephalic trigeminal nucleus (MesV), appearing at Embryonic Day 9.5 (E9.5), shortly after differentiation. In the postnatal cerebellum, B30 stains the postmitotic premigratory granule cells very transiently, until they start migrating through the molecular layer; it also stains mature Purkinje cells (Stainier and Gilbert, 1989). The distribution of the B30 antigen in the developing mouse peripheral nervous system (PNS) reveals its specific association with neural crest-derived neurons as well as with placode-derived neurons of the sensory ganglia. We tested this specificity in explant cultures of embryonic neural tubes in a serum-free chemically defined medium, and we show that in such cultures B30 is a specific marker for neural crest-derived neurons. Because neural crest cells give rise to a wide variety of cell types, neuronal and others (reviewed by Le Douarin, 1982; see also Anderson, 1989), there has been a long standing interest in establishing an in vitro differentiation system for these cells. Recently, Ziller et al. (1983) developed a defined culture medium which allows the differentiation of avian neural crest cells into derivatives with different phenotypes. This fully defined me177
dium has also been used by Boisseau and Simonneau (1989) to monitor the neuronal differentiation of mouse neural crest cells. The isolation of cell surface markers for undifferentiated mouse neural crest cells (mAb B33) as well as for crest-derived neurons (mAb B30) is instrumental for studying the diversification of crest derivatives. Furthermore, the demonstration that B30 is a marker for neural crest-derived neurons supports the presumed neural crest origin of MesV neurons (Narayanan and Narayanan, 1978). MATERIALS
AND METHODS
Mono&ma1 antibodies. mAb B30 has been previously described (Stainier and Gilbert, 1989). mAb El.9 recognizes a cytoplasmic epitope in the primary sensory and motor neurons during axonal outgrowth, appearing at E8.5 and disappearing by El2 in the mouse (Stainier and Gilbert, 1990). In neural crest cultures, it labels all neural crest-derived neurons as soon as they differentiate. mAbs B30 and El.9 were deposited with the Developmental Studies Hybridoma Bank and are available upon request. mAb B33 is a mouse IgM that recognizes the GD3 ganglioside as shown by thin-layer chromatography (TLC) blot analysis of acidic glycolipids (unpublished). It labels neural crest and placode cells as well as crest- and placode-derived neurons in vivo and in vitro (Stainier and Gilbert, submitted for publication). mAb lC6 is a mouse IgG that binds a surface antigen expressed on all PC12 cells. It was isolated from a fusion in 0012-1606/91 $3.00 Copyright All rights
0 1991 by Academic Press, Inc. of reproduction in any form reserved.
178
DEVELOPMENTAL BIOLOGY
which the immunogen consisted of a preparation of phosphatidylinositol-specific phospholipase C clipped proteins and other secreted molecules from PC12 cells (Stainier, 1990). mAb 4C1, a mouse IgG, also binds a cell surface antigen on all PC12 cells and was isolated from a fusion similar to the one that contained the lC6 hybridoma (Stainier, 1990). These two antibodies were used as control antibodies for the adhesion studies on PC12GlO cells. Immunohistochemistry. Pregnant CD1 mice were obtained from Charles River under a specific breeding schedule: mice were bred for 3 hr from 12 noon to 3 PM on EO and our embryonic day ran from noon to noon. Animals were sacrificed by cervical dislocation or with halothane. Embryos (E8-E12) staged according to Theiler (1989) were slit longitudinally along the forming neuropore and processed as half-embryo wholemounts. Brains of later embryos and postnatal animals were cut into lOO- to 120-pm-thick vibratome sections and stained as previously described (Stainier and Gilbert, 1989). All reactions were done at room temperature (with gentle shaking) in polystyrene culture dishes by sequential incubations in the following: 10% normal goat serum (NGS) in phosphate-buffered saline (PBS) for 1 hr; hybridoma culture supernatant overnight (1218 hr); after three lo-min washes with 10% NGS/PBS, fluoresceinated goat anti-mouse IgM (Southern Biotechnology Associates, Birmingham, AL), diluted l:lOO, for 2 hr. After three more lo-min washes in 10% NGS/PBS, the wholemounts and thick sections were mounted in fresh mounting medium between two coverslips separated by silicone grease. For El.9 and BlO staining, the tissue was fixed with 1% paraformaldehyde for 10 min, rinsed several times with PBS, and incubated for 2 hr in 10% NGS. We used a Lasersharp MRC-500 confocal microscope mounted with fluorescence optics to observe our specimens. Briefly, serial sections were collected and combined by projection. The resulting images were stored in a WORM optical disk (Maxtor Corp., San Jose, CA) and printed using a video printer. Explant cultures and staining. Trunk fragments (two to three somites long) were dissected out of the lower spinal cord of E9 (13- to 20-somite-stage) mouse embryos. The fragments were placed in a 0.5 mg/ml solution of dispase (Boehringer-Mannheim) for 8 min and then washed in PBS with 10% fetal bovine serum for 2 min. The neural tube was carefully dissected away from surrounding somites, membranes, and notochord. The tube was washed in Ham’s F12 medium, plated on a 35-mm collagen-coated culture dish, incubated for 4 hr with 10 pg/ml fibronectin or 20 pg/ml laminin, and cultured at 37°C with 5% CO, in fresh Basic Brazeau Medium (BBM) (Ziller et al, 1983). This medium consists of a 6:3:1 mixture of Ham’s F12 nutrient mixture, Dul-
VOLUME 144, 1991
becco’s modified Eagle medium, and BGjb Fitton-Jackson modification; supplemented with 2 g/liter bovine serum albumin, 2.38 g/liter Hepes buffer, 100 IU/ml penicillin/streptomycin, 100 pg/liter hydrocortisone, 1 pg/liter insulin, 0.4 pg/liter triiodothyronine, 10 ng/ liter glucagon, 0.2 rg/liter parathyroid hormone, 10 mg/ liter transferrin, 0.1 pg/liter epithelial growth factor, and 0.2 pg/liter fibroblast growth factor. Cultures were stained as described above. The primary antibody used was B30 or B33 culture supernatant. For double-labeling analysis, B30 was used with 15 pg/ml of concentrated R24 (a mouse IgG3 that also identifies the GD3 ganglioside; the generous gift of Dr. K. Lloyd). Secondary antibodies were used at 1:200 dilution for 20 min. Gang&o&de preparation and analysis. Freshly dissected postnatal mouse tissue or cultured PC12GlO cells were homogenized in chloroform/methanol (2:1, vol/ vol) and extracted with 0.12 1MKC1 (Suzuki, 1965). The extracts were dried by rotary evaporation, taken up in water, desalted through Bond-Elut columns (Analytithem International), and lyophilized for storage at -20°C. TLC analysis of gangliosides was done on aluminiumbacked TLC plates (silica gel 60, 0.2-mm-thick, E. Merck). The plates were developed in chloroform/methanol/0.25% aqueous CaCl, (60:40:10). Ganglioside standards were obtained from Calbiochem and were detected by resorcinol stain. After drying, the plates were dipped in 0.05% poly(isobutylmethacrylate) (Polysciences) in hexane for 90 sec. The dried chromatograms were sprayed with PBS before soaking in a blocking solution (PBS, 3% BSA, and 5% NGS) for 30 min (all the incubations were done at room temperature). The chromatograms were overlaid with the first antibody (diluted hybridoma supernatant) for 3 hr. After repeated washes in PBS, the chromatograms were overlaid with alkaline phosphatase-conjugated second antibody (diluted 1:lOOO)for 1 hr. After repeated washes in PBS, BCIP and NBT (BRL) were used at a 1:l molar ratio in 0.1 M Tris-HCl (pH 9.5), 0.1 M NaCl, 50 mM MgCl, for the color reaction. Upon completion, the reaction was stopped in water. Enzyme treatments and chromatography on DEAESephadex A-25 (Pharmacia) were done as described by Schwarting et al. (1987). Adhesion assays. PClZGlO cells, from a 98% B30 positive PC12 subclone isolated by sequential dilution cloning, were assayed for adhesion on different substrates: collagen (Vitrogen loo), fibronectin (10 pg/ml), and laminin (20 pg/ml). (PC12 cells are rat pheochromocytoma cells). First, PClZGlO cells, labeled with B30 in suspension, were assayed for reattachment by visual inspection. Second, adherent PClBGlO cells were exposed to growth medium containing mAb B30 or mAb lC6 and
STAINIER, BILDER, AND GILBERT
examined for morphological changes. (Another control antibody, mAb 4C1, that binds to the surface of all PClZGlO cells was also tested in these assays and had no effect on PClZGlO cellular morphology or reattachment.) Cell viability was measured by trypan blue exclusion and was not affected by either treatment. PClZGlO cells are refractory to NGF induction. RESULTS
B30 is a Marker for Neural
Crest-Derived
Neurons
During the development of the mouse CNS, there is a specific and transient expression of the B30 antigen on the surface of MesV neurons, starting shortly after their differentiation until about 2 weeks after birth (Stainier and Gilbert, 1989). MesV neurons are the only primary sensory neurons in the CNS and they are thought to derive from the neural crest (Narayanan and Narayanan, 1978). In the periphery, primary sensory neurons (neurons of the cranial nerves and of the dorsal root ganglia (DRG)) exhibit B30 immunoreactivity on both their central and their peripheral projections. Figure 1 shows El0 mouse embryos stained as wholemounts with mAb B30 and viewed at the level of the thoracic spinal cord. Individual neural crest cells differentiate into B30 immunoreactive DRG neurons (Figs. 1A and lB), which segregate into cohesive and segmentally repeated ganglia (Figs. 1C and 1D). Other neural crestderived neurons such as sympathetic and parasympathetic neurons are also B30 immunoreactive (data not shown). In the trigeminal ganglion, neurons derive from both the neural crest and the epidermal placodes, and all are B30 immunoreactive (Stainier and Gilbert, 1989). Neurons of the VIIth and VIIIth ganglia which are largely placode derived (Verwoerd and van Oostrom, 1979; Altman and Bayer, 1982) are also B30 positive (see the B30-labeled vestibulocerebellar fibers in Fig. 5, Stainier and Gilbert, 1989). This distribution indicates that, in the mouse, mAb B30 stains the neural crest-derived neurons as well as the placode-derived neurons of the sensory ganglia; other placode-derived structures, such as the olfactory epithelium, do not display B30 immunoreactivity. The characteristics and time course of the B30 immunoreactivity in the developing PNS follow the pattern described for the trigeminal system (Stainier and Gilbert, 1989, 1990): immunoreactivity appears about 12 hr after neuronal differentiation and disappears around 2 weeks after birth. Neural Crest Diferentiation
In vitro
Neural tube explant culture. To study the specificity of mAb B30 for neural crest-derived neurons, we set up a serum-free culture system in which to follow the neuro-
179
Mouse Neural Crest Surface Markers
nal differentiation of neural crest cells. We dissected lower trunk fragments (two to three somites long) of E9 mouse embryos (13-20 somites), freed the neural tube from surrounding tissues after a mild dispase treatment, and cultured it in a serum-free chemically defined medium containing hormones, growth factors, and transferrin (BBM) (Ziller et al, 1983). Fibronectin and laminin gave the best adhesion and migration of mouse neural crest cells, whereas collagen alone, as a film or a gel, allowed little or no migration. Figures 2A and 2B show such E9 neural tube explants cultured on fibronectin-coated dishes. Both cellular migration and axonal outgrowth from the explant are extensive. Undifferentiated neural crest cells, flat and polygonal, express the GD3 ganglioside (as revealed by B33 or Ii24 staining, data not shown). The long axons extending from the explant (Fig. 2B, arrows) are the El.9 positive, B30 negative projections of the primary motor neurons of the neural tube. Characterization of the neuronal phenotype. By 18 hr in culture, numerous cells have migrated away from the explant and have acquired a neuronal morphology with round, refringent perikaria and long, thin unbranched processes (Figs. 2C-2E). Staining such cultures with mAb El.9 confirmed the neuronal identity of these cells (Fig. 2F). As shown in Fig. 3, l-day-old explanted neural tube cultures contain B30 positive (arrows) as well as B30 negative (arrowheads) neurons. All the B30 positive cells are El.9 positive, confirming their neuronal identity and the B30 positive and negative neurons express the GD3 ganglioside, which, in these cultures, supports their neural crest origin (data not shown). Biochemical and Functional Ganglioside
Analysis
of the B30
Biochemical analysis. As shown in Stainier and Gilbert (1989; Fig. 12), the B30 immunostaining of a TLC plate of acidic glycolipids reveals two discrete gangliosides expressed in the postnatal midbrain and brainstem, dissected to include the trigeminal sensory components, and in the postnatal cerebellum. One of the bands migrates slightly above GM1 and the other migrates between GDl, and GD3 in a chloroform/methanol/ aqueous CaCl, solvent system. This migration pattern is similar to that of the D1.l antigen (Cheresh et al, 1984) and the Jones antigen (Schlosshauer et al, 1988). These two antibodies were originally described in the developing rat CNS as staining germinal cells (D1.l) (Cheresh et al., 1984) and as associated with neural cell and process migration (Jones) (Mendez-Otero et al., 1988). Now they have been shown to have the same specificity for Oacetylated-GD3 and other acetylated gangliosides: base treatment of brain gangliosides converts O-acetylated-
180
DEVELOPMENTALBIOLOGYV0~~~~144,1991
STAINIER, BILDER, AND GILBERT
GD3 into GD3 as revealed by mAb R24 staining, and this treatment causes the loss of the D1.l and Jones immunoreactivity (Cheresh et aZ.,1984; Blum and Barnstable, 1987). We base treated a TLC plate of P12 acidic glycolipids and analyzed it by B30 (Fig. 4, lanes a and c) and R24 (Fig. 4, lane b) immunostaining. Figure 4 shows that the B30 epitope was not affected by base treatment (compare lane a to the control lane c) and that the B30 ganglioside was not converted into GD3 (lane b). However, base treatment revealed a third B30 immunoreactive band between the two original bands (lane a). On the basis of this biochemical analysis of the B30 epitope and of the different distribution of the D1.l (and Jones) antigen with the B30 antigen, we conclude that mAb B30 has a unique specificity. Functional analysis. We searched for a B30 immunoreactive cell line that would enable us to investigate the function of the B30 ganglioside. We found that 1 in lo4 PC12 cells was B30 positive and that the B30 expressing cells were always clustered. We isolated several B30 positive subclones by sequential dilution cloning and isolated their acidic glycolipids for analysis. Figure 5 shows the B30 immunostaining of acidic glycolipids from one of these subclones, PClZGlO. Only the faster migrating B30 ganglioside is expressed and it migrates as a doublet (lane b). Recent reports have provided evidence that cell surface gangliosides may be involved in cellular interactions and adhesion (Cheresh et ak, 1986; Sariola et al., 1988). Specifically, gangliosides are thought to redistribute actively into discrete areas of the cell, and they may act synergistically with cell surface receptors by creating an appropriate electrostatic environment (Cheresh and Klier, 1986; Cheresh et ah, 1986; Burns et al, 1988). In light of these studies, we investigated the effect of B30 on the adhesion of PClZGlO cells. In the first of a series of qualitative experiments, we observed that the B30 immunostaining of PC12GlO cells in suspension inhibited their subsequent reattachment to fibronectin or laminin. Then, to determine whether cells preattached to these substrates could be displaced by B30, PClZGlO cells were allowed to settle and bind on the plate for a minimum of 18 hr, after which they were overlaid with growth medium containing 50 pg/ml of mAb B30 or 50 pg/ml of mAb lC6, which binds to an antigen present on the surface of all PC12 cells (Stainier, 1990). Figure 6 shows that the addition of mAb B30 to these flattened
MmLseNeural Crest Surface Markers
181
cells caused a rapid and significant cellular retraction and rounding up. The speed of this effect increased with both concentration of antibody and temperature (complete release from the substrate in 15 min at 37°C). Figure ‘7 shows as a parallel control that cells treated with mAb lC6 exhibit no difference in their morphology or their adherence. Figure 8 shows the immunostaining pattern of the cells shown in Fig. 7. The mAb B30 caused the PClZGlO cells that express the B30 antigen to round up (arrow) and also to detach from the plate, while mAb lC6 stained all the cells but had no visible effect on their morphology. This effect was observed on a variety of substrates, including laminin, fibronectin, and collagen. In primary cultures, mouse trigeminal neurons show an even distribution of the B30 ganglioside on the cell body and axonal projections, including the growth cones (Stainier and Gilbert, submitted for publication). Addition of growth medium containing 50 pg/ml of mAb B30 on trigeminal neurons, sparsely plated on a laminin substrate, caused the cell bodies to detach although the neurites remained adherent (data not shown). These experiments suggest a differential adhesion along the extended trigeminal neuron. Also, when we used the confocal microscope to examine developing sensory ganglia, we observed a higher concentration of the B30 antigen at interneuronal junctions (data not shown). Similarly, GD2 and GD3 gangliosides have been reported to be heavily concentrated at the interface of cultured melanoma cells and their substrate (Cheresh et ah, 1984). These observations are consistent with the presumed role of gangliosides in cellular interactions and adhesion. DISCUSSION
Neural crest cells represent a unique system in which to study how uncommitted embryonic cells acquire particular developmental fates. These progenitor cells give rise to a wide variety of cell types, ranging from sensory and autonomic ganglia neurons, and Schwann cells, to facial bone, cartilaginous structures, and melanocytes (reviewed by Le Douarin, 1982). Recent experimental evidence, in vivo and in vitro, indicates that some neural crest cells are multipotent at the time of their initial migration (Baroffio et ah, 1988; Bronner-Fraser and Fraser, 1988). These progenitors proliferate as they migrate in the periphery and make further differentiation
FIG. 1. (A, B) mAb B30 stains developing DRG neurons. El0 (32~somite-stage) mouse embrvo stained with mAb B30 and viewed at the level of the thoraeic spinal cord (parasagittal view). Two newly differentiated and B30 immunoreactive DRG neurons project centrally where their axons run rostrally and caudally for a few segments. Their peripheral projections are not yet visible. Scale bars, (A) 50 pm; (B) 10 pm. (C, D) E10.5 (35- and 3%somite-stage) mouse embryos stained with mAb B30 and viewed at the level of the thoracic spinal cord (parasagittal view). B30 immunoreactive DRG neurons segregate into cohesive and segmentally repeated structures or ganglia. The arrows point to somitic boundaries. Scale bars, 50 Km.
182
DEVELOPMENTALBIOLOGY
VOLlJME144,1%1
FIG. 2. Neuronal differentiation of neural crest cells in culture. (A) Explanted neural tube after 12 hr in culture. Flat, polygonal neural crest cells migrate away from the neural tube on a fibronectin-coated dish. fp, floor plate. Scale bar, 300 pm. (B) Explanted neural tube after 24 hr in culture. Motor neurons project axons (arrows) from the neural tube while the neural crest cells are migrating. These projections are El.9 positive but B30 negative (data not shown). NT, neural tube. Scale bar, 50 pm. (C) Neurons appear by 18 hr of culture. Away from the explanted neural tube, differentiated neurons (arrows) mix with undifferentiated crest cells (arrowhead). Scale bar, 50 hrn. (D, E) Up to 15% of the neural crest cells differentiate into neurons after 24 hr in culture. Away from the explant, numerous neurons are clearly visible and extend long and thin processes. Scale bars, 50 pm. (F) Neural crest-derived neurons are El.9 positive. El.9 labels cells, and their processes, which have migrated away from the explanted neural tube. Scale bar, 50 pm.
STAINIER,BILDER,ANDGILBERT
Mouse Neural Crest Surface Markers
183
FIG. 3. (A-F) B30 immunostaining of neural crest-derived neurons. Explanted neural tube cultures were stained with mAb B30 24 hr after dissection. The cultures contain, away from the neural tube, B30 positive (arrows) and B30 negative (arrowheads) neurons. Scale bars, 50 pm.
184
DEVELOPMENTALBIOLOGY
a
b
c
FIG. 4. Biochemical analysis of the B30 epitope: base treatment of mouse acidic glycolipids reveals a third B30 immunoreactive band. P12 cerebellar acidic glycolipids containing 0.5-l pg sialic acid were applied to the TLC plate and developed with chloroform/methanol/ 0.25% aqueous CaCl, (60:40:10). The plate was then exposed to ammonium hydroxide vapor for 3 hr (or water vapor as a control, lane c) and then immunostained with mAb B30 (lanes a and c) or mAb R24 (lane b). (a) base-treated acidic glycolipids stained with B30. (b) basetreated acidic glycolipids stained with R24. (c) control acidic glycolipids stained with B30. Ganglioside standards from Calbiochem were run in parallel and stained with resorcinol. The arrowheads point to the two original B30 immunoreactive gangliosides.
choices in response to factors in their local environment (reviewed by Anderson, 1989). The isolation and characterization of surface markers for mouse crest cells (such as mAb B33) and some of their derivatives (such as mAb B30) is critical for the further understanding of the cellular mechanisms involved in lineage segregation in this system. Indeed, the
vOLUME144,1991
use of cell surface antibodies to distinguish and manipulate specific cell types and their precursors in culture has led to some elegant work on glial cell determination in the rat optic nerve (reviewed by Raff, 1989). Raff and co-workers were able to show that both cell-cell interactions and intrinsic cellular programs contribute to the diversification of glial cells; they also identified the signaling molecules that mediate some of the cell-cell interactions. Approaches such as cell sorting, immunoablation, and quantitative clonal analysis are being applied to analyze the differentiation of chick neural crest cells (Maxwell et al, 1988; Vogel and Weston, 1988). But, as a variety of molecular genetic strategies are now available to manipulate the mouse genome in both somatic and germ cells (reviewed by Rossant, 1990), there is considerable advantage in establishing a culture system for mouse neural crest differentiation. Using the serum-free chemically defined medium first described by Ziller et al. (1983), we were able to show neuronal differentiation within 18 hr of culture. The observation that, by 24 hr, some neurons are B30 positive while others are B30 negative may reflect the stage of differentiation of these neurons or may indicate the phenotypic diversity of the neuronal derivatives in vitro. Careful immunoablation experiments should resolve this question. The B30 staining pattern in viva indicates that the B30 antigen appears within 12 hr after neuronal differentiation and that all neural crest-derived neurons are B30 immunoreactive. In the CNS, MesV neurons exhibit a characteristic B30 staining pattern: like the neural crest-derived neurons of the PNS, they become B30 positive shortly after differentiation until about 2 weeks after birth. This specific B30 staining pattern of MesV neurons supports the notion that they are neural crest derived. Indeed, celltracing studies in quail-chick chimaeras provided the first line of evidence for the neural crest origin of the centrally located MesV neurons (Narayanan and Narayanan, 1978), although other experiments questioned these results (Straznicky and Rush, 1985). Functional
a
b
FIG. 5. B30 immunostaining of PC12GlO acidic glycolipids. Acidic glycolipids containing 0.5-l pg sialic acid were applied to the TLC plate and developed with chloroform/methanol/0.25% aqueous CaCl, (60:40:10). The plate was then immunostained with mAb B30. (a) acidic glycolipids from P15 mouse cerebellum. (b) acidic glycolipids from PClZGlO cells. Ganglioside standards were run in parallel.
Sign&ance
of the B30 Ganglioside
The B30 epitope is resistant to base treatment and is thus different from the D1.l (or Jones) epitope. It is carried by two minor gangliosides expressed on mouse neural crest-derived neurons (and certain cerebellar neurons) (Stainier and Gilbert, 1989). This restricted distribution raises obvious questions about the potential role and function of these rare gangliosides. While gangliosides have been implicated in a wide range of phenomena from neurotrophism to regeneration (reviewed by Hannun and Bell, 1989), there is good evidence for the involvement of cell surface gangliosides in cellular interactions and adhesion. The adhesion of
STAINIER, BILDER, AND GILBERT
Mcruse Neural
Crest Surface Markers
185
FIG. 6. mAb B30 causes PClZGlO cells to round up and detach from a laminin substrate. PClZGlO cells were allowed to adhere on a laminin substrate for at least 18 hr, after which they were overlaid with growth medium containing 50 rg/ml of B30 and kept at room temperature. Addition of the antibody caused the flattened cells to retract (arrow), round up, and sometimes detach from the substrate. The time course of the experiment is shown in minutes in the upper right hand corner of each micrograph. The micrographs at the bottom of the panel show a Hoffman phase view of the PClZGlO cells followed in this experiment and the distribution of B30 immunoreactivity on these cells.
186
DEVJ&OPMENTAL BIOLOGY
VOLUME 144.1991
FIG. 7. mAb B30 causes PClZGlO cells to round up (arrow), whereas mAb lC6 has no effect on cellular morphology. PClZGlO cells growing on laminin were overlaid with growth medium containing 50 pg/ml of mAb B30 (left) or 50 pg/ml of mAb lC6 (right) and kept at room temperature. The time course is in minutes.
postnatal rat cerebellar cells to fibronectin, inhibited by Arg-Gly-Asp peptides that block the recognition of fibronectin to its receptor, is partially inhibited by antibodies to the fibronectin receptor and is slowed by antibodies to the D1.l ganglioside (Stallcup et al., 1989). These studies implicate both the fibronectin receptor and the D1.l ganglioside in the adhesion of cerebellar cells to fibronectin. The B30 gangliosides are expressed on specific sets of neurons during axon outgrowth and synapse formation. Using PClZGlO cells, a B30 positive subclone of the neural crest-derived PC12 cells, we showed that B30 labeling inhibited the subsequent reattachment of immunoreactive cells, and that the addition of mAb B30 to the culture medium caused flattened cells to retract, round up, and even detach from various substrates. These experiments suggest a role for the B30 gangliosides (at least the faster migrating one) in the cellular adhesion process; however, the experiments on trigeminal neurons point to additional adhesive interactions for the
neuronal processes. The B30 gangliosides may thus be involved in maintaining the cohesion of specific groups of neurons. In the cerebellum, it may be involved in maintaining the cohesion of the postmitotic granule cell layer before migration through the molecular layer; in the periphery, it may be involved in holding together crest-derived neuronal structures, such as the sensory ganglia. We thank Dr. K. Lloyd for mAb R24, Dr. Simonneau for interesting discussions, and C. Alvarez and M. Grether for critical reading of the manuscript.
REFERENCES ALTMAN, J., and BAYER, S. A. (1982). Development of the cranial nerve ganglia and related nuclei in the rat. Adv. Anat. Embryo.!. Cell. Biol 74, l-89. ANDERSON, D. J. (1989). The neural crest cell lineage problem: Neuropoiesis? Neuron 3, l-12. BARROFIO, A., DUPIN, E., and LE DOUARIN, N. N. (1988). Clone-forming
STAINIER, BILDER, AND GILBERT
Mouse Neural Crest Surface Markers
18’7
FIG. 8. Distribution of B30 and lC6 immunoreactivity on PC12GlO cells. The PClZGlO cells shown in Fig. 7 were incubated with fluoresceinated secondary antibody for 15 min. The left panels show the distribution of B30 immunoreactivity. Around the cells that have retracted, B30 labels substrate-attached adhesion plaques. By this stage many cells have detached from the plate. The arrow points to a group of cells that rounded up but did not detach (see Fig. 7). The right panels show the distribution of lC6 immunoreactivity. The arrow points to a flat cell that showed no visible morphological changes throughout the experiment.
ability and differentiation potential of migratory neural crest cells. Proc. Natl. Acad. Sci. USA 85,5325-5329.
BLUM, A. S., and BARNSTABLE,C. J. (1987). O-acetylation of a cell-surface carbohydrate creates discrete molecular patterns during neural development. Proc. Natl. Acad. Sci. USA 84,8716-8720. BOISSEAU,S., and SIMONNEAU,M. (1989). Mammalian neuronal differentiation: Early expression of a neuronal phenotype from mouse neural crest cells in a chemically defined culture medium. Development 106,665-674. BRONNER-FRASER, M., and FRASER,S. E. (1988). Cell lineage analysis reveals multipotency of some avian neural crest cells. Nature 335, 161-164. BURNS,G. F., LUCAS,C. M., KRISSANSEN,G. W., WERKMEISTER,J. A., SCANLON,D. B., SIMPSON,R. J., and VADAS, M. A. (1988). Synergism between membrane gangliosides and Arg-Gly-Asp-directed glycoprotein receptors in attachment to matrix proteins by melanoma cells. J. Cell Biol. 107, 1225-1230. CHERESH, D. A., HARPER, J. R., SCHULZ, G., and REISFELD, R. A. (1984). Localization of the gangliosides GD2 and GD3 in adhesion plaques and on the surface of human melanoma cells. Proc. Natl. Acad. Sci. USA 81,5767-5771.
CHERESH,D. A., and KLIER, F. G. (1986). Disialoganglioside GD2 distributes into substrate-associated microprocesses on human melanoma cells during their attachment to fibronectin. J. Cell Biol 102, 1887-1897. CHERESH,D. A., PIERSCHBACHER, M. D., HERZIG,M. A., and MUJOO,K. (1986). Disialogangliosides GD2 and GD3 are involved in the attachment of human melanoma and neuroblastoma cells to extracellular matrix proteins. J. Cell Biol. 102.688-696. CHERESH,D. A., REISFELD,R. A., and VARKI, A. P. (1984). O-acetylation of disialoganglioside GD3 by human melanoma cells creates a unique antigenic determinant. Science 225,844-846. HANNUN, Y. A., and BELL, R. M. (1989). Function of sphingolipids and sphingolipid breakdown products in cellular regulation. Science 243, 500-507.
LE DOUARIN,N. M. (1982). “The Neural Crest,” pp. 60-67. Cambridge Univ. Press. Cambridge. MAXWELL, G. D., FORBES,M. E., and CHRISTIE,D. S. (1988). Analysis of the development of cellular subsets present in the neural crest using cell sorting and cell cultures. Neuron 1,557-568. MENDEZ-OTERO,R., SCHLOSSHAUER,B., BARNSTABLE,C., and CONSTANTINE-PATON,M. (1988). A developmentally regulated antigen
188
DEVELOPMENTALBIOLOGY
associated with neural cell and process migration. J. Neurosc~ 8, 564-579. NARAYANAN, C. H., and NARAYANAN, Y. (1978). Determination of the embryonic origin of the mesencephalic nucleus of the trigeminal nerve in birds. J. EmbyoL Exp. MwphoL 43,85-105. PUKEL, C. S., LLOYD,K. O., TRAVASSOS,L. R., DIPPOLD,W. G., OETPJXNGkN, H. F., and OLD, L. J. (1982). GD3, a prominent ganglioside of human melanoma: Detection and characterization by a mouse monoclonal antibody. J. Ezp. Med. 155,1133-114’7. RAFF, M. C. (1989). Glial cell diversification in the rat optic nerve. Science 243,1450-1455. ROSSANT, J. (1990). Manipulating the mouse genome: Implications for neurobiology. Neuron 2,323-334. SARIOLA, H., AUFDERHEIDE, E., BERNHARD, H., HENKE-FAHLE, S., DIPPOLD,W., and EKBLOM,P. (1988). Antibodies to cell surface ganglioside GD3 perturb inductive epithelial-mesenchymal interactions. Cell 54, 235-245. SCHLOSSHAUER,B., BLUM, A. S., MENDEZ-OTERO,R., BARNSTABLE, C. J., and CONSTANTINE-PATON,M. (1988). Developmental regulation of ganglioside antigens recognized by the Jones antibody. J. Neurosci. 8.580-592. SCHWARTING,G. A., JIJNGALWALA,F. B., CHOU,D. K., BOYER,A. M., and YAMAMOTO,M. (1987). Sulfated glucuronic acid-containing glycoconjugates are temporally and spatially regulated antigens in the developing mammalian nervous system. Dev. BioL 120,65-76. SCOTT,S. A. (1987). The development of skin sensory innervation patterns. Trends Neurosei. 10,468-473. STAINIER, D. Y. R. (1990). Neuronal differentiation and maturation in the mouse trigeminal sensory system. PhD thesis, Harvard University, Cambridge, MA.
VOLUME 144,1991
STAINIER,D. Y., and GILBERT,W. (1989). The monoclonal antibody B30 recognizes a specific neuronal cell surface antigen in the developing mesencephalic trigeminal nucleus of the mouse. J. Neuroaci 9,24682485. STAINIER, D. Y. R., and GILBERT, W. (1990). Pioneer neurons in the mouse trigeminal sensory system. Proc. NatL Ad. Sci. USA 87, 923-927. STALLCUP,W. B. (1988). Involvement of gangliosides and glycoprotein fibronectin receptors in cellular adhesion to fibronectin. Exp. Cell Res. 177,90-102. STALLCUP,W. B., PYTELA, R., and RUOSLAHTI,E. (1989). A neuroectoderm-associated ganglioside participates in fibronectin receptormediated adhesion of germinal cells to fibronectin. Dev. BioL 132, 212-229. STRAZNICKY,C., and RUSH, R. A. (1985). The effect of nerve growth factor on the developing primary sensory neurons of the trigeminal nerve in chick embryos. Anat. EmbyoL 171,91-95. SUZUKI, K. (1965). The pattern of mammalian brain gangliosides. III. Regional and developmental differences. J. Neurochem. 12,969-979. THEILER, K. (1989). “The House Mouse,” 2nd ed., Springer-Verlag, Berlin. VERWOERD,C. D. A., and VAN OOSTROM,C. G. (1979). Cephalic neural crest and placodes. EmbrgoL Cell BioL 58, l-75. VOGEL, K. S., and WESTON,J. A. (1988). A subpopulation of cultured avian neural crest cells has transient neurogenic potential. Neuron 1,569-577. ZILLER, C., DUPIN, E., BRAZEAU, P., PAIJLIN, D., and LE DOUARIN,N. (1983). Early segregation of a neuronal precursor cell line in the neural crest as revealed by culture in a chemically defined medium. Cell 32, 627-638.
