To identify and analyse precursor cells of neuronal and glial cell lineages during the early development of the chick peripheral nervous system, monoclonal antibodies were raised against a population of undifferentiated cells of E6 dorsal root ganglia (DRG).

Non-neuronal cells of E6 DRG express surface antigens that are recognized by four monoclonal antibodies, G1, G2, GLI 1 and GLI 2. The proportion of non-neuronal cells in DRG that express the GLI 1 antigen is very high during ganglion formation (80% at E4) and decreases during later development (15% at E14). GLI 2 antigen is expressed only on a minority of the cells at E6 and increases with development. The G1 and G2 antigens are expressed on about 60-80% of the cells between E6 and E14. All cells that express the established glia marker O4 are also positive for the new antigens. In addition, it was demonstrated that GLI 1positive cells from early DRG, which are devoid of O4 antigen, could be induced in vitro to express the O4 antigen. Thus, the antigen-positive cells are considered as glial cells or glial precursor cells.

Surprisingly, the antigen expression by satellite cells of peripheral ganglia is dependent on the type of ganglion: antigens G1, G2 and GLI 1 were not detectable on glial cells of lumbosacral sympathetic ganglia and GLI 2 was expressed only by a small subpopulation. These results demonstrate an early immunological difference between satellite cells of sensory DRG and sympathetic ganglia. As the antigens could however be induced in vitro also in sympathetic ganglion cells, it is suggested that the specific antigen expression is due to specific environmental cues acting on precursor cells in different types of ganglia rather than to intrinsic differences between sensory and sympathetic glial precursor cells.

The glial cells of the peripheral nervous system (PNS), including the Schwann cells of peripheral nerves, the satellite cells of sensory and autonomic ganglia and the enteric glial cells are derived from the neural crest. The development of peripheral glial cells has been studied in detail for the Schwann cells of sciatic nerve and for satellite cells of peripheral ganglia. It is characterized by the sequential appearance of the Ca2+-binding S100 protein (Holton and Weston, 1982 ), the cell surface glycolipids sulfatide and galactocerebroside, recognized by specific monoclonal antibodies O4 and O1 (Schachner et al., 1981; Rohrer and Sommer, 1983; Rohrer, 1985; Mirsky et al., 1990), by the expression of NGF receptors (Rohrer and Sommer, 1983; Zimmermann and Sutter, 1983; Jessen et al., 1990) and myelin specific proteins (Jessen and Mirsky, 1991 and references therein). Virtually all of these differentiation steps are dependent on signals provided by neurons and disruption of neuron-glia contact results in a rapid loss of most differentiated properties (Aguayo et al., 1976; Weinberg and Spencer, 1976; Mirsky et al., 1980; Poduslo, 1984). Whereas the nature and number of the signal molecules involved in the control of glia differentiation are unknown, cAMP mimics in vitro the action of such factors (Sobue and Pleasure, 1984; Sobue et al., 1986) and thus has been implicated in the signal transduction pathway of Schwann cell differentiation factors.

In contrast to these later stages of glial cell development the early steps of the glial cell lineage are largely unknown. In vivo cell lineage analysis (Bronner-Fraser and Fraser, 1989) and clonal neural crest cultures (Sieber-Blum and Cohen, 1980; Baroffio et al., 1991) provide evidence for different types of precursor cells in migrating neural crest. Precursor cells can give rise either to many different cell types, to both neurons and glial cells, to melanocytes and glial cells, or only to glial cells. It is unclear, however, up to which stage such cells exist during development and when the cells are finally committed to glial cell differentiation. In addition, it is unknown if neuron-derived signals are required to elicit or to stabilize glial commitment. To understand the factors involved during glial cell differentiation, it is important to identify and to select cells during early stages of development. To this end monoclonal antibodies were raised against an undifferentiated cell population in chick dorsal root ganglia (DRG) and selected for cell-type-specific antibodies expressed early during development.

In the present manuscript, we describe a set of newly developed monoclonal antibodies that define a sequence of glial cell differentiation stages in chick DRG. These antibodies also reveal an unexpected heterogeneity between glial cells of different ganglia. Whereas satellite cells from DRG are antigen-positive, sympathetic ganglion satellite cells are devoid of these antigens. In vitro, however, sympathetic glial cells acquire these antigens, suggesting that the lack of antigen expression in sympathetic ganglia is not an intrinsic property of the cells, but rather controlled by local cues.

Immunization

Balb/c mice were immunized with E6 DRG precursor cells prepared as described previously (Rohrer et al., 1985). For the primary immunization 3×106 cells were injected i.p. with complete Freund’s adjuvant. For boosting, mice received 3×106 cells, mixed with incomplete Freund’s adjuvant every 6 weeks over a period of 1 year. The last 3 days before fusion cells were injected daily (i.p.) without any further supplement. The splenocytes were fused with X63-Ag8-653 myeloma cells using PEG 6000 (50% w/v in PBS, Boehringer Mannheim), following the standard protocol of Köhler and Milstein (1975). Hybridoma supernatants were screened immunocytochemically, using cultures of E6 chick DRG ganglion cells 3 hours and 24 hours after plating of the cells. Hybridomas of interest were cloned by limiting dilution.

Cell preparation and culture

Cells from lumbosacral DRG were obtained and cultured as described in detail previously (Emsberger and Rohrer, 1988). Ganglion cells from trigeminal, nodose, sympathetic and ciliary ganglia were prepared as described for DRG, however, trypsinization was carried out for 20 minutes (trigeminal and nodose), 25 minutes (E7 sympathetic) and 30 minutes (ciliary). Undifferentiated cells were obtained from E6 DRG cell suspensions as described previously (Rohrer et al., 1985). Trunk neural crest cells were obtained according to the procedures of Cohen (1977). Neural tubes prepared from 20-24 somite quail embryos were plated on collagen-coated substrata which allows the neural crest cells to migrate onto the culture dish (DMEM, 15% horse serum, 15% chick embryo extract was used as culture medium). Conditioned medium was obtained from Ell DRG neuron cultures: E11 DRG ganglion dissociates were treated with 04 antibody and complement to eliminate O4-positive glial cells (Rohrer et al., 1985) and the remaining neuronal cell population was kept in culture for up to 10 days in the presence of NGF (20 ng/ml) and standard culture medium. Half of the medium volume was withdrawn every other day and replaced with fresh medium. The conditioned medium aliquots were combined and added to the cultures in the indicated experiments by diluting 1:1 with fresh medium.

Immunocytochemistry

To analyse the cell specificity of the monoclonal antibodies, cultures were stained with hybridoma supernatant after 24 hours in vitro using a modification of previously described (Emsberger and Rohrer, 1988) methods: bound antibodies were visualized by adding subsequently biotinylated anti-mouse Ig (Amersham, diluted 1:100) and FITC-labeled Streptavidin (Amersham, 1:100) for 30 and 15 minutes, respectively. For double-labeling, cells were first incubated with either Gl, G2, GLI1 or GLI2, followed by either Q211 or 04 antibodies. Rhodamine- or FITC-labeled secondary antibodies, which are specific for antibody isotypes (IgG or IgM) were used to distinguish between the different antibodies.

To detect antigens on tissue sections, E6 chick embryos were fixed for 24 hours (4% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4) at 4°C, washed several times and kept overnight in 30% sucrose (w/v in water). The tissue was frozen and 7 μm sections were air-dried onto gelatine-coated coverslips. All subsequent steps were then performed using Tris-phoshate buffer (40 mM Tris, 120 mM NaCl, 9 mM Na2HPO4/KH2PO4-buffer, pH 7.8) supplemented with 0.1% gelatine unless indicated otherwise. Dried sections were rehydrated, incubated with undiluted hybridoma supernatant overnight, followed by biotinylateded anti-mouse Ig and FITC-labeled streptavidin.

Antigen characterization

Cultured DRG cells were treated with proteases, neuraminidase and with organic solvents and the effect of the various treatments on antibody binding was then analysed. Lipid extraction with chloroform/methanol (1:1/ v:v) was done for 5 minutes, protease treatment (trypsin, 0.25 mg/ml or pronase E, 0.1 mg/ml in PBS) and neuraminidase incubation (0.4 U/ml in 50 mM Tris, 150 mM NaCl, pH 7.6 (TBS)) was done for 30 minutes at 37°C. The enzyme treatments required prior fixation of cells by 0.25% glutaraldehyde to prevent the detachment of the cells from the culture dish. Epitopes on membrane proteins were identified by western blotting. DRG were collected in sample buffer, solubilized by boiling and the equivalent of 10-20 ganglia was applied onto a SDS-polyacrylamide gel. After the run, proteins were transferred onto nitrocellulose sheets which were then incubated for 2 hours with antibodies of interest. Immuno reactive bands were disclosed by incubation with HRP-conjugated goat antimouse Ig (Bioscience, 1:1000 in TBS plus 5% milkpowder for 1 hour) and subsequent reaction with 4-chloronaphtol and H2O2 in PBS.

Antigens detected by monoclonal antibodies (mAb) G1, G2, GLI 1 and GLI 2 are markers for early glial cells in chick DRG

At embryonic day 7 (E7) and later, all cells of lumbosacral chick DRG cells can be identified as either neurons, glial cells or fibroblasts using the Q211 antigen and 04 antigen as neuronal and glial marker and fibronectin expression as characteristics for fibroblasts (Rohrer et al., 1985). During earlier development, however, the ganglia contain undifferentiated cells that are not recognized by these antibodies. To obtain markers that allow identification and selection of these cells, monoclonal antibodies were raised against the undifferentiated cell population from E6 chick DRG. To select for cell surface antigens, the antibodies were screened on cultured DRG cells using immunocytochemical methods. Four antibodies, mABs G1, G2, GLI 1 and GLI 2, which during early development specifically recognize glial cells in DRG, were obtained. In E6 DRG cultures only cells with non-neuronal morphology were stained by the antibodies, whereas neurons remained unlabeled (Fig. 1). Double-label experiments with the established glia-specific 04 antibody (Fig. 2) demonstrated that all O4-positive cells were also positive for the antigens Gl, G2 and GLI 1 and that about 80% of the O4-positive cell population expressed the GLI 2 antigen. The mAb Gl, G2 and GLI 1 bind however to a large number of undifferentiated, O4-negative non-neuronal cells (Fig. 2). As these cells are fibronectin-negative (data not shown), it is suggested that these O4-negative, G-antigen-positive cells are glia precursor cells which become O4-positive during normal development.

Fig. 1.

Immunostaining of E6 DRG cultures with the antibodies G1, G2, GLI 1 and GLI 2. E6 DRG cells were cultured for 1 day on a polyomithine-laminin sustrate and then incubated with the antibodies G1 (A,E), G2 (B,F), GLI 1 (C,G) and GLI 2 (D,H), followed by biotinylated anti-mouse antibodies and FITC-labeled streptavidin. Please note that neuronal cell bodies and neurites are not stained. (A-D) phase contrast, (E-H) fluorescence optics. Bar 10 μm.

Fig. 1.

Immunostaining of E6 DRG cultures with the antibodies G1, G2, GLI 1 and GLI 2. E6 DRG cells were cultured for 1 day on a polyomithine-laminin sustrate and then incubated with the antibodies G1 (A,E), G2 (B,F), GLI 1 (C,G) and GLI 2 (D,H), followed by biotinylated anti-mouse antibodies and FITC-labeled streptavidin. Please note that neuronal cell bodies and neurites are not stained. (A-D) phase contrast, (E-H) fluorescence optics. Bar 10 μm.

Fig. 2.

Double-immunolabeling of E6 DRG cells with the antibodies O4 and G2/GLI 1. E6 DRG cells were cultured for 1 day and then stained for both GLI 1 antigen (A) and O4 antigen (B) or G2 antigen (C) and O4 antigen (D) using subclass-specific secondary antibodies labeled either with FITC or RITC. Note that all O4-positive cells are also G2 and GLI1 positive. In addition cells are present which express G2 or GLI 2 antigens but are O4 negative, Bar 10 μm.

Fig. 2.

Double-immunolabeling of E6 DRG cells with the antibodies O4 and G2/GLI 1. E6 DRG cells were cultured for 1 day and then stained for both GLI 1 antigen (A) and O4 antigen (B) or G2 antigen (C) and O4 antigen (D) using subclass-specific secondary antibodies labeled either with FITC or RITC. Note that all O4-positive cells are also G2 and GLI1 positive. In addition cells are present which express G2 or GLI 2 antigens but are O4 negative, Bar 10 μm.

Expression of G1, G2, GLI 1 and GLI 2 antigens in trigeminal, nodose, ciliary and sympathetic ganglia and sciatic nerve of chick and quail embryos

To determine whether the antigens described are also expressed in other peripheral ganglia, trigeminal, ciliary and nodose ganglia and sympathetic ganglia of the lumbosacral sympathetic chain were analysed (Table 1). The satellite cells of sympathetic ganglia differed from all other peripheral glial cells investigated by the complete absence of Gl, G2, and GLI 1 antigens between E7 and E14. GLI 2 antigen was also absent at E7, but was detectable on 8±2% of non-neuronal cells at E10 and increased to 20±3% at E14. Surprisingly, in sympathetic ganglia the neuronal cell population was labeled with the antibodies G1 and G2. However, the structures of the G1and G2 antigens expressed by sympathetic neurons seem to differ from the DRG satellite cell antigens (see below). Antigen expression is not restricted to the ganglion satellite cells, but is also observed on Schwann cells (E7 sciatic nerve Schwann cells are positive for GLI 1; G2 and GLI 2 are weakly expressed).

Table 1.

Expression of G1, G2, GLI 1 and GLI 2 antigens by non-neuronal cells of different ganglia of the chick PNS

Expression of G1, G2, GLI 1 and GLI 2 antigens by non-neuronal cells of different ganglia of the chick PNS
Expression of G1, G2, GLI 1 and GLI 2 antigens by non-neuronal cells of different ganglia of the chick PNS

Immunohistological studies on tissue sections confirmed the tissue-specific expression of the antigens G1l and G2. A strong signal was observed in sensory DRG, sympathetic ganglia and peripheral nerves (Fig. 3). GLI 1 and GLI 2 antigens could not be demonstrated on sections, due to the fixation sensitivity of the epitopes.

Fig. 3.

Immunohistological detection of Gl antigen on cross-sections of E5 chick embryos. Frozen sections of E5 chick embryos were stained with G1 antibody followed by biotinylated anti-mouse antibodies and FTTC-labeled streptavidin as described in Materials and methods, nt, neural tube; drg, dorsal root ganglion. Bar 10 μm.

Fig. 3.

Immunohistological detection of Gl antigen on cross-sections of E5 chick embryos. Frozen sections of E5 chick embryos were stained with G1 antibody followed by biotinylated anti-mouse antibodies and FTTC-labeled streptavidin as described in Materials and methods, nt, neural tube; drg, dorsal root ganglion. Bar 10 μm.

Antigen characterization

To obtain some information about the molecular nature of the epitopes recognized by the different mAb, we analysed the effect of various chemical and enzymatic treatments of cultured E6 DRG ganglion cells on antibody binding (Table 2). These data indicate that the G1 antigen expressed by DRG satellite cells is a sialic acid-containing glycolipid. This conclusion is supported by the finding that the poly-sialoganglioside-antigen Q211 (Rösner et al., 1988) demonstrated similar sensitivity to the treatments (Table 2). The antigen recognized by the G1 antibody on sympathetic neurons differs from the antigen expressed by DRG cells by its resistance to neuraminidase treatment. The proteaseresistance and extractability by organic solvents are indicative for the lipid nature of the G2 antigen on sensory DRG satellite cells. The G2 antigen on sympathetic neurons however is not extracted by organic solvents, suggesting that this epitope may be shared by different molecules. The GLI 1 antigen was identified as a 160×103Mr protein on western blots, whereas the molecular nature of the GLI 2 antigen remained unclear.

Table 2.

Characterization of G1, G2, GLI 1, GLI 2 and Q211 antigens on E6 DRG cells by lipid extraction, protease and neuraminidase treatment and western blot

Characterization of G1, G2, GLI 1, GLI 2 and Q211 antigens on E6 DRG cells by lipid extraction, protease and neuraminidase treatment and western blot
Characterization of G1, G2, GLI 1, GLI 2 and Q211 antigens on E6 DRG cells by lipid extraction, protease and neuraminidase treatment and western blot

Developmental expression of the antigens

We have previously demonstrated that the expression of the glia-specific O4 antigen is characteristic for different developmental stages of the ganglia of the chick PNS (Rohrer and Sommer, 1983; Rohrer, 1985; Rohrer et al., 1985; Rohrer and Thoenen, 1987). We now investigated the developmental expression of the newly defined antigens, using the 04 antigen as point of reference.

To gain information about the earliest stages of glia development, the expression of the antigens in neural crest was investigated. The antigens G1, G2 and GLI 1 (but not 04) were detectable on subpopulations of neural crest cells with a non-neuronal morphology. Also in vivo, the G1 antigen was detected on migrating neural crest cells, about 24 hours after the first cells expressing the neural crest marker HNK-1 (Vincent and Thiery, 1984) were found (data not shown).

The antigen expression during ganglion development was studied in detail for sensory DRG. The ganglion cell population from E6, E10 and E14 DRG was stained after short-term culture with the newly developed antibodies or the O4 antibody to determine the proportion of antigen-expressing cells. Three different expression patterns were found (Fig. 4): The GLI 1 antigen is present on the majority of ganglion nonneuronal cells at E6 and the proportion of antigen-positive cells decreases during embryonic development. The antigens G1 and G2 are expressed throughout development on up to 80% of the ganglion non-neuronal cells, whereas the GLI 2 antigen is expressed on a minority of cells at E6, but labels a large proportion of glial cells in DRG at later developmental stages.

Fig. 4.

Developmental changes in the proportion of DRG non-neuronal cells expressing the antigens G1, G2, O4, GLI 1 and GLI 2. Ganglia were dissected at the embryonic stages indicated, dissociated to single cells, kept for 1 day in culture and then analysed for the proportion of antigen positive non-neuronal cells. Points represent the mean S.D. of at least 3 independent experiments.

Fig. 4.

Developmental changes in the proportion of DRG non-neuronal cells expressing the antigens G1, G2, O4, GLI 1 and GLI 2. Ganglia were dissected at the embryonic stages indicated, dissociated to single cells, kept for 1 day in culture and then analysed for the proportion of antigen positive non-neuronal cells. Points represent the mean S.D. of at least 3 independent experiments.

Double-label experiments demonstrated that at E6 all O4-positive cells also expressed the antigens G1, G2 and GLI 1. In contrast, only 79% of the O4-positive cells were positive for the GLI 2 antigen, indicating that this antigen is expressed even later than O4 during glial cell differentiation. A large proportion of the E6 DRG cells expresses the antigens G1, G2 and GLI 1 but lack O4, whereas at E10 and E14 virtually all G1-, G2- and GLI 1-positive cells are also O4-positive as analysed by double-label experiments (data not shown). These results thus allow us to define discrete steps in the early glial differentiation pathway. The earliest stage is defined by the presence of GLI 1 and the absence of O4. Intermediate stages express both antigens, whereas the late stage is characterized by the lack of GLI 1 and the presence of O4 and GLI 2.

Modification of antigen expression in culture

Since there is considerable evidence that in Schwann cells the expression of many differentiated properties is dependent on the interactions with neurons (Jessen and Mirsky, 1991), we investigated to what extent the expression of the antigens G1, G2, GLI 1 and GLI 2 on ganglion satellite cells can be manipulated in vitro. When undifferentiated, O4- and Q211-negative cells from E6 DRG (Rohrer et al., 1985) were kept under standard culture conditions, the GLI 1 antigen, the marker for undifferentiated cells, remained detectable in a high proportion of cells and the late glia markers 04 and GLI 2 were not expressed (Fig. 5). However, in the presence of medium conditioned by DRG neurons from older (E11) embryos, both O4 and GLI 2 antigens were induced (Fig. 5) and double-label experiments demonstrated directly the transition from GLI 1positive, O4-negative cells to O4-positive cells (Fig. 6). The glial differentiation factor(s) are not only present in DRG-conditioned medium, but also in extracts from Ell whole chick embryos (not shown).

Fig. 5.

Effect of E11 DRG neuron-conditioned medium on the expression of the antigens O4 (A), GLI 1 (B) and GLI 2 (C) by cultured E6 DRG glial precursor cells. Glial precursor cells were prepared by complement-mediated lysis of differentiated O4-positve glia cells (and Q211-positive neurons). Cells were cultured in the presence (•) or absence (○) of conditioned medium and the proportion of antigen-positive non-neuronal cells was determined at the time points indicated. (The proportion of GLI 1 antigen-positive cells in 3 hour cultures may be higher as in these cultures non-neuronal cells are difficult to identify by morphological criteria; the values given are a conservative estimation).

Fig. 5.

Effect of E11 DRG neuron-conditioned medium on the expression of the antigens O4 (A), GLI 1 (B) and GLI 2 (C) by cultured E6 DRG glial precursor cells. Glial precursor cells were prepared by complement-mediated lysis of differentiated O4-positve glia cells (and Q211-positive neurons). Cells were cultured in the presence (•) or absence (○) of conditioned medium and the proportion of antigen-positive non-neuronal cells was determined at the time points indicated. (The proportion of GLI 1 antigen-positive cells in 3 hour cultures may be higher as in these cultures non-neuronal cells are difficult to identify by morphological criteria; the values given are a conservative estimation).

Fig. 6.

Induction of O4 antigen in GLI 1-positive precursor cells: Double-labeling for O4 (B) and GLI 1 (C) after a 3 day incubation of O4 negative glial precursor cells with E11 DRG conditioned medium. (A) Phase contrast.

Fig. 6.

Induction of O4 antigen in GLI 1-positive precursor cells: Double-labeling for O4 (B) and GLI 1 (C) after a 3 day incubation of O4 negative glial precursor cells with E11 DRG conditioned medium. (A) Phase contrast.

The expression of GLI 1 and GLI 2 antigens in E7 sympathetic glial cells is also subject to extrinsic control: we observed in vitro a rapid appearance of the antigens GLI 1 and GLI 2 on sympathetic glia under standard culture conditions which could be further stimulated by Ell DRG-conditioned medium (Fig. 7). Similar findings were obtained for the antigens G1 and G2 (data not shown). These findings indicate that the absence of G1, G2, GLI1 and GLI 2 in sympathetic gial cells may be due to both negative control and a lack of induction factors in intact sympathetic ganglia.

Fig. 7.

Effect of E11 DRG neuron-conditioned medium on the expression of the antigens GLI 1 (A) and GLI 2 (B) by cultured E7 sympathetic glial cells. E7 sympathetic ganglion cells were cultured in the presence (•) or absence (○) of conditioned medium and the proportion of antigen-positive non-neuronal cells was determined at the time points indicated.

Fig. 7.

Effect of E11 DRG neuron-conditioned medium on the expression of the antigens GLI 1 (A) and GLI 2 (B) by cultured E7 sympathetic glial cells. E7 sympathetic ganglion cells were cultured in the presence (•) or absence (○) of conditioned medium and the proportion of antigen-positive non-neuronal cells was determined at the time points indicated.

Monoclonal antibodies directed against cell surface antigens provide a powerful tool to identify and to isolate cell populations which can then be studied with respect to their developmental potential, growth and differentiation factor requirements. With the aim of obtaining antibodies which enable identification of glia and neuron precursor cells, antibodies were raised against the undifferentiated cell population of chick DRG. This approach resulted in a series of monoclonal antibodies directed against cell surface antigens expressed on glial cells and undifferentiated precursors.

In contrast to the antigens GLI 1 and GLI 2, which are glia-specific in all ganglia, G1 and G2 antigens are in addition expressed on sympathetic neurons, which are the only antigen-positive cells in these ganglia. It should be pointed out, however, that the G1 and G2 antigens expressed by DRG satellite cells and sympathetic neurons must be different as demonstrated by the different susceptibility to neuraminidase and lipid extraction. The preliminary analysis of the antigens expressed on DRG glial cells indicated that most antibodies are directed against lipid antigens, whereas the antibody GLI 1 recognizes a cell surface protein epitope. The antibodies were screened on cells shortly after trypsinization of the ganglia, which may explain the high proportion of lipid antigens. The advantage of such trypsin-resistant cell surface antigens is, however, that the antibodies can be used to isolate cells immediately after trypsinization, either by negative selection i.e. complement-mediated cell lysis (Rohrer et al., 1985; Rohrer and Thoenen, 1987), or by positive selection i.e using a fluorescence-activated cell sorter.

The developmental changes in the proportion of antigen-positive cells in DRG are explained by the downregulation of GLI 1 and the increase of O4 and GLI 2 antigens during differentiation. The finding that all O4-positive cells at E6 are also GLI 1 positive, and the induction of O4 antigen in GLI 1-positive, O4-negative cells in vitro strongly argues in favour of a glia differentiation pathway involving the sequential appearance of GLI 1, O4 and GLI 2 antigens. Thus the new set of antibodies allows us to identify and select glial precursors from the earliest stages of ganglion formation which previously has not been possible. In addition, the developmental changes in antigen expression pattern allows one to define discrete steps of early glial cell differentiation in chick DRG.

While the antigens described are stage-specific and glia-specific in chick DRG, not all glial cells in the PNS express these antigens. Cells from the trigeminal or nodose ganglion are not or only weakly labeled by the antibody GLI 2 and the glial cells from the sympathetic ganglia are completely devoid of G1, G2 and GLI 1 antigen. This difference does not reflect different stages of maturation as sympathetic glial cells were devoid of the antigens during the whole time period analysed (E7-E14). Thus these data provide evidence for a surprising and unexpected heterogeneity of peripheral glial cells that, to our knowledge, has not been observed so far. In the CNS, however, there is ample evidence for heterogeneity of type 1 astrocytes obtained from different brain regions (Wilkin et al., 1990).

The sequence of appearance of antigens in DRG and the ganglion-specific expression of cell surface antigens may either be explained by intrinsic differentiation programs of sensory and sympathetic glial cells or specific local interactions in different ganglia. There is considerable evidence that the differentiation of Schwann cells and of ganglion satellite cells depends on interactions with neurons (Aguayo et al., 1976; Weinberg and Spencer, 1976; Mirsky et al., 1980; Mirsky et al., 1990). We now show that O4-negative glia precursor cells from E6 DRG require the presence of differentiation factor(s) in E11 DRG-neuron-conditioned medium to express the O4 antigen. The ability to influence the expression of O4 antigen in vitro suggests that also in vivo the initial expression of the O4 antigen may be controlled by the local environment.

The most interesting finding of the present study is the molecular difference between satellite cells of different ganglia, which is reflected by the absence of the antigens GLI 1, G1 and G2 in sympathetic ganglion satellite cells. In culture, these antigens are however rapidly expressed by sympathetic ganglion cells. Also the GLI 2 antigen, which is absent during early development, is expressed in vitro and is further increased in the presence of E11 DRG-conditioned medium. These results suggest that the lack of antigen expression in sympathetic ganglia may be due to both the presence of inhibitory factors, which repress antigen expression, and the absence of inducing factors.

Taken together, these data suggest that the described antigenic pattern depends on environmental instructions. It should be pointed out, however, that the results do not exclude the existence of intrinsic differences between sensory and sympathetic glial cells. Indeed, the analysis of the developmental potential of quail sensory and sympathetic ganglion cells by backtransplantation into 2-day-old embryos provided evidence for distinct glial cell lineages for sensory and sympathetic satellite cells (Le Lievre et al., 1984; Le Douarin, 1984). The development of sensory and sympathetic satellite cells may thus be due to an early commitment to two different lineages in the neural crest, which is followed then by local instructions within the developing ganglia, resulting in the expression of ganglion-specific differentiated properties. Satellite cells of different peripheral ganglia thus may have distinct identities and characteristic properties, as observed for astrocytes in different brain regions.

We thank J. Schnitzer and J. Christie for helpful comments on the manuscript and H. Thoenen for continuous support. The technical assistance of C. Krieger and H. Wilde-Launert is gratefully acknowledged. This work was supported by the Deutsche Forschungsgemeinschaft (SFB 220).

Aguayo
,
A. J.
,
Charron
,
L.
and
Bray
,
G. M.
(
1976
).
Potential of Schwann cells from unmyelinated nerves to produce myelin: a quantitative untrastructural and autoradiographic study
.
J. Neurocytol
.
5
,
565
573
.
BarofBo
,
A.
,
Dupin
,
E.
and
Le Douarin
,
N. M.
(
1991
).
Common precursors for neural and mesectodermal derivatives in the cephalic neural crest
.
Development
112
,
301
305
.
Bronner-Fraser
,
M.
and
Fraser
,
S. E.
(
1989
).
Developmental potential of avian trunk neural crest cells in situ
.
Neuron
3
,
755
766
.
Cohen
,
A. M.
(
1977
).
Independent expression of the adrenergic phenotype by neural crest cells in vitro
.
Proc. Natl. Acad. Sci., U.S.A
.
74
,
755
766
.
Emsberger
,
U.
and
Rohrer
,
H.
(
1988
).
Neuronal precursor cells in chick dorsal root ganglia: differentiation and survival in vitro
.
Dev. Biol
.
126
,
420
432
.
Holton
,
B.
and
Weston
,
J. A.
(
1982
).
Analysis of glial cell differentiation in peripheral nervous tissue. II. Neurons promote S100 synthesis by purified glial precursor cell populations
.
Dev. Biol
.
89
,
72
81
.
Jessen
,
K. R.
and
Mirsky
,
R.
(
1991
).
Schwann cell precursors and their development
.
Glia
4
,
185
194
.
Jessen
,
K. R.
,
Morgan
,
L.
,
Stewart
,
H. J. S.
and
Mirsky
,
R.
(
1990
).
Three markers of adult non-myelin-forming Schwann cells, 217c(Ran-l), A5E3 and GFAP: Development and regulation by neuron-Schwann cell interactions
.
Development
109
,
91
103
.
Köhler
,
G.
and
Milstein
,
C.
(
1975
).
Continuous cultures of fused cells secreting antibody of predefined specificity
.
Nature
256
,
495
497
.
Le Douarin
,
N. M.
(
1984
).
A model for cell line divergence in the ontogeny of the peripheral nervous system
.
In Cellular and Molecular Biology of Neuronal Development
. (
I.
Black
, ed.)
Plenum Press
,
New York
, pp.
3
28
.
Le Lievre
,
C.
,
Schweizer
,
G. G.
,
Ziller
,
C.
and
Le Douarin
,
N. M.
(
1980
).
Restriction of devlopmental capacities in neural crest cell derivatives as tested by in vivo transplantation experiments
.
Dev. Biol
.
77
,
362
378
.
Mirsky
,
R.
,
Dubois
,
C.
,
Morgan
,
L.
and
Jessen
,
K. R.
(
1990
).
Prenatal Schwann cell development: Appearance of 04 differentiation antigen in rat embryo sciatic nerve and its regulation by axon-schwann cell signals
.
Development
109
,
105
116
.
Mirsky
,
R.
,
Winter
,
J.
,
Abney
,
E. R.
,
Press
,
R. M.
,
Gavrilovic
,
J.
and
Raff
,
M. C.
(
1980
).
Myelin-specific proteins and glycolipids in rat Schwann cells and oligodendrocytes in culture
.
J. Cell Biol
.
84
,
483
494
.
Podulso
,
J. F.
(
1984
).
Regulation of myelination: biosynthesis of the major glycoprotein by Schwann cells in the presence and absence of myelin assembly
.
J. Neurochem
.
42
,
493
503
.
Rösner
,
H.
,
Greis
,
C.
and
Henke-Fahle
,
S.
(
1988
).
Developmental expression in embyryomc rat and chicken brain of a polysialoganglioside-antigen reacting with the monoclonal antibody Q211
.
Dev. Brain Res
.
42
,
161
171
.
Rohrer
,
H.
(
1985
).
Non-neuronal cells from chick sympathetic and dorsal root sensory ganglia express catecholamine uptake and receptors for nerve growth factor during development
.
Dev. Biol
.
111
,
95
107
.
Rohrer
,
H.
,
Henke-Fahle
,
S.
,
El-Sharkawy
,
T.
,
Lux
,
H. D.
and
Thoenen
,
H.
(
1985
).
Progenitor cells from embryonic chick dorsal root ganglia differentiate in vitro to neurons: biochemical and electrophysiological evidence
.
EMBO J
.
4
,
1709
1714
.
Rohrer
,
H.
and
Sommer
,
I.
(
1983
).
Simultaneous expression of neuronal and glial properties by chick ciliary ganglion cells during development
.
J. Neurosci
.
3
,
1683
1693
.
Rohrer
,
H.
and
Thoenen
,
H.
(
1987
).
Relationship between differentiation and terminal mitosis: chick sensory and ciliary neurons differentiate after terminal mitosis of precursor cells whereas sympathetic neurons continue to divide after differentaition
.
J. Neurosci
.
7
,
3739
3748
.
Schachner
,
M. S.
,
Kim
,
S. K.
and
Zehnle
,
R.
(
1981
).
Developmental expression in central and peripheral nervous system of oligodendrocyte cell surface antigens (O antigens) recognized by monoclonal antibodies
.
Dev. Biol
.
83
,
328
338
.
Sieber-Blum
,
M.
and
Cohen
,
A. M.
(
1980
).
Clonal analysis of quail neural crest cells: they are pluripotent and differentiate in vitro in the absence of noncrest cells
.
Dev. Biol
.
80
,
96
106
.
Sobue
,
G.
and
Pleasure
,
D.
(
1984
).
Schwann cell galactocerebroside induced by derivatives of adenosine 3’, 5’-monophosphate
.
Science
224
,
72
74
.
Sobue
,
G.
,
Shuman
,
S.
and
Pleasure
,
D.
(
1986
).
Schwann cell responses to cyclic AMP: Proliferation, change in shape, and appearance of surface galactocerebroside
.
Brain Res.
362
,
23
32
.
Vincent
,
M.
and
Thlery
,
J. P.
(
1984
).
A cell surface marker for neural crest and placodal cells: further evolution in peripheral and central nervous system
.
Dev. Biol
.
103
,
468
481
.
Weinberg
,
H.
and
Spencer
,
P. S.
(
1976
).
Studies on the control of myelinogenesis. II. Evidence for neuronal regulation of myelination
.
Brain Res
.
113
,
363
378
.
Wilkin
,
G. P.
,
Marriot
,
D. R.
and
Cholewinski
,
A. J.
(
1990
).
Astrocyte heterogeneity
.
Trends Neurosci
.
13
,
43
46
.
Zimmermann
,
A.
and
Sutter
,
A.
(
1983
).
β-Nerve growth factor (βNGF) receptors on glial cells. Cell-cell interaction between neurons and Schwann cells in cultures of chick sensory ganglia
.
EMBO J
.
2
,
879
885
.