Cell Tissue Res (1990) 260:147-159 Cell and

�9 Springer-Verlag 1990

Differentiation of embryonic chick sympathetic neurons in vivo: ultrastructure, and quantitative determinations of catecholamines and somatostatin Stefan Ross, Anseim Fischer, and Klaus Unsicker

Department of Anatomy and Cell Biology, University of Marburg, Marburg, Federal Republic of Germany

Accepted December 17, 1989

Summary. The ultrastructural and transmitter develop- ment of lumbar sympathetic ganglia was studied in em- bryonic day-6 through -18 chick embryos. At embryonic day 6, ganglia are populated by two morphologically distinct types of neuronal cells and Schwann cell precur- sors. The neuronal populations basically comprise a granule-containing cell and a developing principal neu- ron. Granule-containing cells have an irregularly shaped or oval nucleus with small clumps of chromatin attached to the inner nuclear membrane and numerous large (up to 300nm) membrane-limited granules. Developing principal neurons display a more rounded vesicular nu- cleus with evenly distributed chromatin, prominent nu- cleoli, more developed areas of Golgi complexes, and rough endoplasmic reticulum and large dense-core vesi- cles up to 120 nm in diameter. There are granule-con- taining cells with fewer and smaller granules which still display the nucleus typical for granule-containing cells. These granule-containing cells may develop toward de- veloping principal neurons or the resting state of gran- ule-containing cells found in older ganglia. Both gran- ule-containing cells and developing principal neurons proliferate and can undergo degeneration. At embryonic day 9 there are far more developing principal neurons than granule-containing cells. Most granule-containing cells have very few granules. Mitotic figures and signs of cell degeneration are still apparent. Synapse-like ter- minals are found on both developing principal neurons and granule-containing cells. Ganglionic development from embryonic day 11 through 18 comprises extensive maturation of developing principal neurons and a nu- merical decline of granule-containing cells. Some gran-

Acknowledgements. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (Un 34/10-1). We thank Heidi Hlawaty and Heike Reichert-Preibsch for excellent technical assis- tance.

Send offprint requests to: Dr. K. Unsicker, Department of Anato- my and Cell Biology, University of Marburg, Robert-Koch-Str. 6, D-3550 Marburg, FRG

ule-containing cells with very few and small granules still persist at embryonic day 18. The mean catechol- amine content per neuron increases from 0.044 femtomol at embryonic day 7 to 0.22 femtomol at embryonic day 15. Concomitantly, there is a more than 6-fold in- crease in tyrosine hydroxylase activity. Adrenaline has a 14% share in total catecholamines at embryonic day 15. Somatostatin levels are relatively high at embry- onic day 7 (1.82 attomol per neuron) and are 10-fold reduced by embryonic day 15. Our results suggest the presence of two morphologically distinct sympathetic neuronal precursors at embryonic day 6: one with a bi- nary choice to become a principal neuron or to die, the other one, a granule-containing cell, which alterna- tively may develop into a principal neuron, acquire a resting state or die.

Key words: Sympathoadrenal cell lineage Sympathetic neurons - Small granule-containing cells - Chick

Development of the nervous system results from the coordinated actions of humoral, matrix and cell surface- bound molecules that provide epigenetic cues for the differentiation of neurons and glial cells. Elucidation of these cues has been facilitated by analyzing the migra- tion and differentiation pathways of cell lineages that constitute the peripheral nervous system (LeDouarin et al. 1975; LeDouarin and Smith 1983). Cell surface markers specific for neurons, glial and non-neuronal cells have become important tools in these studies for identifying the respective cell types and undifferentiated cells (Mirsky 1982; Rohrer and Thoenen 1987). From neuronal progenitor cells in a given peripheral ganglion arise diverse neuronal subtypes that may be identified based on their transmitter phenotypes and expression of other markers (Ciment and Weston 1981 ; Dodd et al. 1984; Teitelman et al. 1984; Sommer et al. 1985; Gib- bins et al. 1987; Heym and Kummer 1988). How this

148

diversity and the frequently persisting plasticity (Un- sicker et al. 1978; Otten and Lorez 1983; Ka tz e ta l . 1986) o f neurona l phenotypes are archieved, is largely unknown.

In the sympa thoadrena l lineage three ma jo r neurona l and endocr ine cell types exist: principal neurons, chro- maffin cells, and small granule-conta in ing ( synonymous term: small intensely f luorescent (SIF)) cells (Coupland 1965; Grillo 1966; Er/ink6 1976; Landis and Pat terson 1981; Unsicker et al. 1989). Each category, again, can be subdivided accord ing to its expression o f catechol- aminergic, cholinergic and peptidergic properties. For the m a m m a l i a n sympa thoadrena l lineage it has been shown that g lucocor t icoid hormones , nerve growth fac- tor (NGF) , ciliary neuro t rophic factor, and f ibroblast g rowth fac tor interfere with morpho log ic pheno type de- cisions in this lineage (Unsicker et al. 1978, 1985a, b, c; 1989; D o u p e e ta l . 1985a, b ; A n d e r s o n and Axel 1986; Shaw and Le tourneau 1986; Stemple et al. 1988; Seidl and Unsicker 1989a, b). The impact o f these fac- tors for the development and diversification o f the sym- pa thoadrena l cell lineage in the chick has no t been fully explored. This m a y in par t be explained by the more p ronounced morpholog ica l heterogenei ty o f avian as c o m p a r e d to m a m m a l i a n sympa thoadrena l ceils and the lack o f adequate biochemical markers for this wider vari- ety o f cell types (Papka 1972; Unsicker 1973a, b, c; Gabel la 1976; Luckenbi l l -Edds and v a n H o r n 1980). Wi thou t such markers , u l t ras t ructural features are ex- t remely impor tan t for the classification o f avian sym- pa thoadrena l cell types. As an in t roduct ion o f a series o f studies in which we have analyzed the implications o f t rophic factors and ho rmones for the in vitro and in vivo differentiat ion o f chick embryon ic paraver tebral sympathet ic ganglia we repor t here on the ul t ras t ructural and t ransmit ter (catecholamines and somatos ta t in) de- ve lopment o f these ganglia start ing at embryon ic day 6.

Materials and methods

Animals

Chick (White Leghorn) eggs obtained from a local hatchery were incubated at 37 ~ C in a humidified (75%) atmosphere. Chicks from embryonic days (E) 6, 7, 9, 11, 15 and 18 - corresponding to stages 29, 31, 37, 38, 41 and 44 (Hamburger and Hamilton 1951) - respectively, were used throughout this study. Lumbar paraverte- bral ganglia located caudal to the articulation of the 7th rib with the vertebra (Haider 1960) were rapidly dissected under a binocular microscope and processed as described below.

Electron microscopy

The rostral five lumbar ganglia of embryonic day 6, 9, 11, 15 and 18 embryos were either fixed by immersion or brief perfusion (5 min) through the left ventricle of the heart in situ and removed afterwards. 2.5% phosphate-buffered (0.1 M, pH 7,4) glutaralde- hyde (Merck, Darmstadt) served as fixative. After an additional fixation period of 1 h at 4 ~ C ganglia were postfixed in 2% phos- phate-buffered OsO4 for 1 h. Ganglia were then washed twice in phosphate buffer, block-stained in a saturated alcoholic solution

of uranyl acetate (Merck, Darmstadt) for 1 h (Silva et al. 1968) and subsequently dehydrated in a graded series of ethanols fol- lowed by propylene oxide (Serva, Heidelberg). Semithin sections (0.5-1.0 ~tm) of Araldite-embedded material were stained with azur-II-methylene-blue (Richardson et al. 1960) and used for light- microscopical analysis. Ultrathin sections were cut on an ultramic- rotome OMU2 (Reichert-Sitte) and stained with uranyl acetate and lead citrate (Reynolds 1963).

Quantitative determinations of catecholamines, tyrosine hydroxyIase activity and somatostatin

The lumbosacral paravertebral sympathetic ganglia were removed from embryonic day 7, 9, 11 and t 5 chick embryos and collected in ice cold Ca 2 +-, Mg 2+-free Hanks' balanced salt solution (CMF). Ganglia were then centrifuged and incubated for 15 rain at 37 ~ in 0.125% trypsin-CMF, washed three times in Dulbecco's modi- fied Eagle's medium (Gibco) and dissociated into single cells using flame-constricted siliconized Pasteur pipettes. Yields of neuronal cells per ganglion (see below) were 3160_+ 189 (embryonic day 7), 5845 _+ 380 (embryonic day 9), 11418 + 288 (embryonic day 11), and 7864_+1755 (embryonic day 15) (neuronal cell numbers • n = 9 to 12). Percentages of neuronal cells varied from 85 (embryon- ic day 7) to 94% (embryonic day 9). Neurons were characterized by their phase-contrast morphology and tetanus toxin binding 3 h after seeding on polyornithine-coated coverslips (cf. Unsicker and Wiegandt 1988) and scored if they had a refractile halo, short processes and a tetanus toxin-labeled cell membrane. Atiquots of cell suspensions were either seeded for culture experiments or pro- cessed for determinations of catecholamines, tyrosine hydroxylase, and somatostatin.

Catecholamines

Quantitative determinations were performed as previously de- scribed (Miiller and Unsicker 1981, 1986). In brief, 300 lal of a mixture containing 0.1 M sodium acetate buffer (pH 6.0, 4 ~ C), 1% (wt/vol) Triton X-100, 25 mM EDTA, and as internal stan- dard, 12.5 ng N-methyldopamine were added to a 100 gl aliquot. The samples were vigorously vortex-mixed for 5 min, deproteinized with 25 gl of HC10, (70%) and stored at --80 ~ C until analysis. Determination of catecholamines was performed by high perfor- mance liquid chromatography (HPLC) with amperometric detec- tion.

Tyrosine hydroxylase

The activity of tyrosine hydroxylase was determined as described (Miiller and Unsicker 1986; Seidl et al. 1987). After centrifugation at 10000 xg at 6 ~ C a 140 gl aliquot was gel-filtrated at 6 ~ C. Fifty gl of the protein eluate was incubated for 15 rain at 37 ~ C with 50 gl of a reaction mixture containing 2 mM 6,7-dimethyltetrahy- dropterine, 5 mM ascorbic acid, 0.2 mM L-tyrosine, 2 mg/ml cata- lase, 0.5 mM o-phenylhydroxylamine and 0.2 M sodium acetate buffer, pH 6.0. Control samples were boiled for 10 min. Reactions were terminated with 5 ~tl 70% HC104. Adrenaline was added as an internal standard. Enzymatically formed DOPA was extracted using the same procedure as for dopamine, noradrenaline, and adrenaline, and the DOPA levels were measured by HPLC with amperometric detection.

Somatostatin

Somatostatin-like immunoreactivity was determined by radioim- munoassay. 3-(125j)iodotyrosyl")-Tyr"-somatostatin-14 and an an-

149

Fig. 1 a-e. Micrographs of azur II/methylene blue-stained semithin (0,5 I gm) sections showing portions of lumbar sympathetic gan- glia (all L1) at embryonic day 6 (a), embryonic day 9 (b), embryonic day 11 (e), embryonic day 15 (d), and embryonic day 18 (e). Arrows

mark developing sympathetic principal neurons, arrowheads a somewhat smaller cell with oval or irregular nuclei. V blood vessel; M mitotic figure, a-c x 350; d, e x 560

151

tiserum against somatostatin (code N.1611), according to a proto- col provided by Amersham (Braunschweig, FRG), were employed.

Statis t ics

Biochemical data were subject to statistical analysis using the rank- sum test according to Mann and Whitney (1947). Values are ex- pressed as Standard Error of the Means (SEM) and were consid- ered to be significantly different at the 1% level.

Results

Nomenclature

We basically distinguish between two types of neuronal cells, a granule-containing cell with dense-core vesicles up to 300 nm in diameter, and a developing principal sympathetic neuron with dense-core vesicles not exceed- ing 120 nm in diameter. Both cell types are subject to substantial morphological changes between embryonic day 6 and embryonic day 18.

Embryonic day 6

Light-microscopic examination (Fig. 1 a) reveals cells, assembled in clusters of different size, that are dispersed within fibrous areas presumably representing bundles of developing nerve fibers. Blood vessels are not discernible at this developmental stage. A thin sheath of connective tissue surrounds the ganglion. Two types of cells can be distinguished within the ganglion : (i) cells with a large vesicular nucleus and one or two prominent nucleoli usu- ally showing a narrow cytoplasmic rim; (ii) smaller cells with darker, often irregular nuclei and one or two nucle- oli.

Electron-microscopically, the most conspicuous cell type at this age is a granule-containing cell (Fig. 2a). Granules of high electron densities surrounded by a limi- ting membrane are round, oval or irregular in shape and 80-300 nm in diameter. They are found both within

Fig. 2a-c. Electron micrographs from L1 ganglia (embryonic day 6) showing granule-containing cells (a) and granule-containing cell forms that may represent transitional stages (GCT) toward developing principal neurons or a resting form of granule-contain- ing cell which is found in older ganglia, a Perikaryal areas of two granule-containing cells and numerous processes (P) containing the typical membrane-limited granules with diameters up to 300 nm. ER short profiles of rough endoplasmic reticulum. Arrows mark a stretch of basal lamina-like material covering the surface of a process, x 14490. b, e GCT cells differing with regard to the number and average size of granules (arrows). These cells still dis- play the irregularly contoured nuclei and chromatin arrangement which is typical for granule-containing cells (cf. Figs. 2a and 3a). MV multivesicular body; ER sparse profiles of rough endoplasmic reticulum; a loosely arranged axons. An axon with numerous mi- crotubules emerging from a granule-containing cell is marked by a large arrow in c. b x 16000; c • 16800

a small perikaryal area and in the numerous short slender processes, singly or more frequently in tightly packed groups. Another characteristic feature of the granule-containing cell is an irregularly contoured nucle- us with clusters of chromatin often lining the inner nucle- ar membrane. The cells have polyribosomes, little rough endoplasmic reficulum, few Golgi profiles and crista- type mitochondria. This cell type closely resembles the small granule-containing SIF cells (type II; Lu et al. 1976) of mammalian sympathetic ganglia (Er/inko 1976). As early as embryonic day 6 one can find granule-con- taining cells closely associated with basal lamina-like material (Fig. 2a).

The prominent structural features of another neuro- nal cell type (Fig. 3a) are the large vesicular nucleus with sparse, evenly distributed chromatin and large nu- cleoli, a larger perikaryon with polyribosomes and few already well-developed parallel stacks of ribosome-stud- ded endoplasmic reticulum, multiple Golgi-complexes and numerous crista-type mitochondria. Dense-core ves- icles never exceed 120 nm in diameter and are often found close to Golgi stacks. These cells may form loosely arranged groups with intervening spaces free of cells, but containing matrix.

Neuronal cells with the morphological characteristics of both granule-containing cells and developing princi- pal neurons apparently exist (Fig. 2b, c). Their number is low at embryonic day 6 as compared to later ages (see below). Since they may represent granule-containing cells in transition to developing principal neurons, the term " G C T cell" will be used in this study.

The structure of the nuclei of GCT cells is essentially that of granule-containing cells. However, their granules are smaller and less numerous than those of granule- containing cells resembling type I-SIF cells (Lu et al. 1976). Cellular elements ahnost entirely devoid of gran- ules display more Golgi complexes, multivesicular bodies and rough endoplasmic reticulum. Their processes are thicker and contain numerous parallel microtubules (Fig. 2c).

Profiles of nerve fibers have a heterogeneous appear- ance at this stage (Figs. 2a, c, 3a and inset). Some con- tain microtubules, others, with irregular outlines, display actin filaments often more densely packed beneath the cell membrane and numerous electron-lucent large vesi- cles. The latter profiles are tentatively classified as growth cones. Typical synaptic profiles with pre- and postsynaptic membrane specializations were not de- tected at this age. Schwann cell progenitors (Fig. 2b) could tentatively be identified by their close association with large bundles of developing nerve fibers and/or their small chromatin-rich nuclei. This observation cor- roborates the finding that a Schwann cell-specific cell surface protein can be detected as early as embryonic day 6 in the chick embryo (Ziller et al. 1989).

Both granule-containing cells and developing princi- pal neurons undergo mitotic cell divisions at this age (Fig. 3 b, c). At the same time, degenerating and macro- phage-like cells (Fig. 3 d) are found in the ganglia.

Since sufficient amounts of ganglia for the quantita-

153

Table 1. Development of total catecholamine-, noradrenaline-, dopamine -content, tyrosine hydroxylase activity, and somatostatin concen- tration

Values per neuron E7 n E9 n E11 n El5 n

Total catecholamine 0.044 _+0.005 24 0.104 _+0.009 30 0.110+0.007 54 0.22 -+0.007 36 content (fmol)

Noradrenaline 0.04 _+0.005 24 0.09 _0.009 30 0.09 -+0.007 54 0.17 -I-0.007 36 content (fmol)

Adrenaline 0.002 _+0.0005 24 0.006 _+0.0005 30 0.007_+0.0009 54 0.03 _+0.002 36 content (fmol)

Dopamine 0.002 _+0.0005 24 0.008 _+0.002 30 0.004_+0.0005 54 0.02 +0.001 36 content (fmol)

Noradrenaline (%) 91 87 89 77

Adrenaline (%) 4.5 6.0 7.0 14

Dopamine (%) 4.5 8.0 4.0 9.0

Tyrosine hydroxylase 0.0183 -+0.0018 19 0.0249-+0.0024 33 0.024_+0.0018 28 0.1154_+0.0016 15 activity (fmol Dopa/ neuron min- 1)

Somatostatiu 1.82 _+0.11 61 0.59 _+0.02 132 0.12 _+0.007 88 0.20 _+0.009 142 content (amol)

(Values represent means _+ S.E.M.)

tive determinations of catecholamines and somatostat in could not be dissected at embryonic day 6, the following figures refer to embryonic day 7. Total catecholamines amounted to 0.044 femtomol (fmol) per neuron, and noradrenaline had a share of 91% (Table 1). The propor- tions of dopamine and adrenaline were 4.5% each. Tyro- sine hydroxylase activity amounted to 0.0183fmol D O P A / n e u r o n . m i n - 1 . The somatostat in content per neuron was 1.82 _+ 0.11 a t tomol (amol; n = 61).

Embryonic day 9

At the light-microscopical level (Fig. 1 b), ganglia display blood vessels, strands of nerve fibers and still two types of neuronal cells. One has a large vesicular nucleus and

Fig. 3 a-d. a Developing principal neuron (DPN) from an embryon- ic day 6 L1 ganglion displaying a large vesicular nucleus with evenly distributed chromatin and a large perikaryon containing numerous Golgi complexes (G) and profiles of rough endoplasmic reticulum (ER). Arrowheads mark growing axons with typical membrane pro- files in a vesicular form; s nucleus of a developing Schwann cell tentatively identified by its shape and close apposition to axons. Inset: Portion fiom an axon bundle containing large dense-core vesicles (arrowheads) of the type that can also be found in perikarya of developing principal neurons, x 9500; inset • 25000. b, e Mitot- ic figures of developing principal neurons (b) and a granule-con- taining cell (e). b x 11000; e • 16000. d Degenerative changes, marked by arrows consisting of coalescent highly electron-dense material and a large vacuole containing membrane material. The identity of the cell is not entirely clear, but according to its nuclear structure it may be a granule-containing cell. x 22000

prominent nucleoli, the other is less frequent and charac- terized by a smaller irregular nucleus and less cytoplasm.

Compared to embryonic day 6 the number of gran- ule-containing cells as identified by their granules has dramatically decreased, more so in the rostral than in the caudal ganglia. However, when taking the nuclear morphology (clumps of chromatin in granule-containing cells as compared to homogeneously distributed chroma- tin in developing principal neurons, cf. Fig. 4a) as a criterion for the identification of granule-containing cells, they still occur in substantial numbers. This implies that the number of granules per cytoplasmic area has greatly decreased (Fig. 4a, b). Frequently, granules ap- pear in thin processes (Fig. 4b). By their size, electron density and shape of the vesicles they often resemble the large dense-core vesicles of developing principal neu- rons suggesting the existence of transitional cell forms (GCT cells) f rom granule-containing cells to developing principal neurons. Developing principal neurons have not undergone substantial morphological differentiation as compared to embryonic day 6. Large dense-core vesi- cles (80-120 nm in diameter) have shifted f rom the peri- nuclear region into the processes. Mitotic figures and signs of cell degeneration (Fig. 4a) are still apparent at embryonic day 9. Synapses (Fig. 4b, inset) are rarely found on both developing principal neurons and gran- ule-containing cells, which is in line with results reported by Ross et al. (1978) and Hruschak et al. (1982).

Catecholamine content per neuron has more than doubled as compared to embryonic day 7 and the pro- port ion of adrenaline, but not dopamine has slightly, but significantly increased by 1.5% (Table 1). Tyrosine hydroxylase has increased by 30% as compared to em-

154

Fig. 4a, b. Electron micrographs from embryonic day 9 L1 ganglia showing granule-containing cells with few scattered granules (ar- rowheads). By nuclear structure and granule equipment these cells are identified as granule-containing cells. G Golgi complex; C cen-

triole; d degenerating cell and axon. Inset: Presynaptic terminals with numerous clear vesicles abutting a granule-containing cell (granules marked by arrowheads). Note postsynaptic membran spe- cialisation (arrow). a x 11200; b x 19200; inset x 59000

155

Fig. 5a, b. Developing principal neuron (a), granule-containing cell in transition toward a principal neuron (b) and Schwann cells (s) in an embryonic day 18 L1 ganglion. G Golgi complexes; ER rough endoplasmic reticulum; C collagen fibrils; a axon profiles. Granules are marked by arrowheads, a • 10500; b x 21000

156

bryonic day 7. The somatostatin content has decreased by almost two thirds and now amounts to 0.59 amol per neuron.

Embryonic day 11

Figure I c shows a portion of a L1 paravertebral sympa- thetic ganglion. Most neurons now display a large vesi- cular nucleus and a large cell body. Cell-free spaces are less frequent and the wider spaces contain bundles of nerve fibers and blood vessels.

Granule-containing cells identified by electron mi- croscopy by their principal criteria (large granules and nuclear morphology) have further decreased in number and almost disappeared from the rostral ganglia. How- ever, GCT cells (cf. Fig. 5 b) resembling the former gran- ule-containing cells in terms of chromatin arrangement persist; they display large, dense-core vesicles up to about 150 nm in diameter. Developing principal neurons have increased in size and differentiation particularly with regard to the extensive development of the rough endoplasmic reticulum. Their perikarya are now nearly entirely devoid of large dense-core vesicles.

Processes display numerous microtubules and neu- rofilaments. Synapses exhibit all ultrastructural charac- teristics typical of mature terminals and postsynaptic sites including small (about 40 60 nm in diameter) clear vesicles at the presynaptic membrane. Schwann cells (cf. Fig. 5a) are discernible by their close association with developing principal neurons and a well-developed gran- ular endoplasmic reticulum. Mitosis figures can no lon- ger be seen at this stage, but degenerative cells still occur.

Neither the total catecholamine content per neuron nor the relative proportions of noradrenaline, adrena- line, and dopamine have significantly changed as com- pared to embryonic day 9. Tyrosine hydroxylase activity has remained unchanged. Somatostatin, in contrast to catecholamines, has dropped to about 20% of its aver- age neuronal content at embryonic day 9 (Table 1).

Embryonic days 15 and 18

Light microscopy reveals only one type of neuron with a voluminous cell body and large pale rounded nucleus (Fig. I d, e). These neurons still increase in size between embryonic day 15 and embryonic day 18. Ganglia at both ages contain numerous blood vessels and areas con- taining neuritic bundles.

Electron micrographs (Fig. 5a, b) reveal developing principal neurons as the predominant type of neuron. Fine structural features such as, e.g., the extensive rough endoplasmic reticulum, multiple Golgi complexes, multi- vesicular and dense bodies closely resemble those of post-hatching principal sympathetic neurons (Lucken- bill-Edds and vanHorn 1980). However, we also clearly identified the former granule-containing cells, still recog- nizible by their typical nuclei and a few remaining vesi- cles with electron-dense cores (GCT cell form; Fig. 5 b).

Quantitative determinations of catecholamines at embryonic day 15 reveal a more than 100% increase in the amount of total catecholamines and a overpropor- tional increase of adrenaline, which now makes up 14% of the total catecholamine content. The increase in total catecholamine is reflected by a 4.8-fold increase in tyro- sine hydroxylase activity. Somatostatin amounts to 0.2 amol per neuron at this age.

Discussion

Neuronal cells of sympathetic ganglia are heterogeneous at various levels. In the early chick embryo, electron microscopy permits to distinguish granule-containing cells - characterized by large dense-core vesicles (over 200 and up to 300 nm in diameter) and an irregularly shaped nucleus with peripherally condensed chromatin - and developing principal neurons, which lack the very large dense-core vesicles. Both populations are of ap- proximately equal size at embryonic day 6, but numbers of granule-containing cells subsequently decrease (cf. Luckenbill-Edds and vanHorn 1980); we found fewer granule-containing cells in embryonic day 9, and very few in embryonic day 11, 15 and 18 lumbar sympathetic ganglia. Interestingly, they maintain their characteristic nucleus even after their large granules have disappeared (cf. Fig. 5b, embryonic day 18). They may be induced by glucocorticoid hormones to re-express the granule- containing phenotype when grown in dissociated cell cul- ture from embryonic day 15 embryos. However, when taken at earlier stages, they cannot be maintained by glucocorticoids (unpublished results). These cells there- fore morphologically resemble the SIF cells of mamma- lian sympathetic ganglia, which also contain characteris- tic granules. They seem to differ from SIF cells, however, in that they are refractory to glucocorticoids during their early development (Er/ink6 1976). Granule-containing cells in the chick are intensely fluorescent (Luckenbill- Edds and vanHorn 1980) and their disappearance (i.e., disappearance of granules) from the developing ganglia can be monitored by glyoxyic acid-induced histofluores- cence, which allows to detect intensely fluorescent cells; already at embryonic day 7, they have been reported to constitute only 2% of lumbar sympathetic chain neu- rons (Ernsberger et al. 1989). The vast majority of devel- oping principal neurons in embryonic day 6/7 ganglia appears to belong to a noradrenergic phenotype: nor- adrenaline has a 91% share in total catecholamines at this age (this study) and about 86% of the neurons are tyrosine hydroxylase-immunoreactive (Ernsberger et al. 1989).

Further differentiation of embryonic chick sympa- thetic neurons results in at least 3 chemically defined subpopulations: one population starts to express vasoac- tive intestinal polypeptide (VIP) (Hayashi et al. 1985; New and Mudge 1986), becomes detectable at embryonic day 10 and constitutes about 10% of the neuronal popu- lation at embryonic day 14 (after one day in culture; Ernsberger et al. 1989). Vasoactive intestinal polypeptide

157

neurons are cholinergic, which is also a property of mammalian sympathetic vasoactive polypeptide neurons (Lundberg et al. ~979, 1982). In the present study we were not able to distinguish them from adrenergic neu- rons based on ultrastructural criteria. A second popula- tion expresses the adrenaline synthesizing enzyme PNMT (Teitelman et al. 1985) and also apparently syn- thesizes adrenaline as judged from the quantitative de- termination of A (14% of total catecholamines at embry- onic day 14). The size of this population is not entirely clear, but is unlikely to exceed 20% of all neurons. The third and largest population comprises the noradrener- gic neurons.

In addition to the above mentioned subpopulations there is a somatostatin-containing neuronal subset (Maxwell et al. 1984; Garcia-Arrars et al. 1984), which is already small at embryonic day 6/7 (Ernsberger et al. 1989). Somatostatin-immunoreactive cells are no longer apparent in older ganglia, which is in line with the sharp decrease in somatostatin between embryonic day 7 and embryonic day 11 as determined by radioimmunoassay. Somatostatin is expressed by tyrosine hydroxylase-im- munoreactive cells, but it has not been established whether somatostatin is contained in granule-containing cells, developing principal neurons or both.

The lineage relationships of these neuronal pheno- types and factors controlling their differentiation are only beginning to be explored. A granule-containing-, SIF-like cell has been proposed to be the common pro- genitor for both sympathetic neurons, chromaffin cells and intermediate cell forms (Landis and Patterson 1981; Unsicker et al. 1989). This proposition is based on stu- dies using cultured SIF cells from mammalian sympa- thetic ganglia and immature chromaffin cells (Unsicker et al. 1978; Doupe et al. 1985a, b; Seidl and Unsicker 1989a, b). These cells acquire neuronal characteristics under stimulation with nerve growth factor or become differentiated endocrine chromaffin cells in the presence of glucocorticoids. More detailed analysis of the devel- opmental fates of isolated rat sympathoadrenal precur- sors (Anderson and Axel 1986; Anderson 1988; Ander- son and Michelson 1989) has shown that they may dif- ferentiate along a neuronal pathway and die, or survive when supported by nerve growth factor, or become chro- maffin-like in the presence of glucocorticoids. Moreover, heterogeneity among sympathoadrenal precursors seems to exist even before they have invaded the adrenal anla- gen and have been exposed to elevated levels of gluco- corticoids.

Our results suggest that there may be a dual (or even more heterogeneous) progenitorship for chick sympa- thetic neuronal cells. At embryonic day 6, i.e., shortly after the coalescence of neurons in a secondary sympa- thetic chain, granule-containing cells and developing principal neurons are clearly distinguishable by ultra- structural criteria, and both cell types apparently prolif- erate. There is also morphological evidence for cell death of both cell types subsequent to embryonic day 6. The depletion of the granule-containing cell type from the ganglia, accordingly, must not necessarily reflect their

exclusive transition into the developing principal neuron form as suggested in the hypothesis of a SIF cell as a common progenitor for principal neurons and chro- maffin cells (Landis and Patterson 1981). Their eventual fate may also include death or they may return to a quiescent, non-proliferating, agranular state as we have found in embryonic day 15 and embryonic day 18 gan- glia. Results from our culture studies (unpublished) im- ply that such a resting granule-containing cell form may persist and may be induced by glucocorticoids to become chromaffin-like. However, the fate of granule-containing cells does not seem to be exclusively controlled by gluco- corticoids, since glucocorticoids cannot prevent the nu- merical decrease of granule-containing cells in embryon- ic day 7, -9 and -11 cultures grown for 2 days (unpub- lished results). In the adult avian adrenal, i.e., in a gluco- corticoid-rich environment (cf. Chester-Jones etal. 1959), all morphological variations of sympathoadrenal cells persist including principal neurons, chromaffin cells, "developing principal neurons", granule-contain- ing cells and several transitional forms (Unsicker 1973 a). This "in vivo experiment" also implies that some gran- ule-containing cells may be refractory to glucocorticoids and/or that additional factors are required for the full expression of the chromaffin phenotype.

Why do granule-containing cells disappear from chick sympathetic ganglia between embryonic day 6 and embryonic day 18 and how does this relate to the devel- opment of plasma glucocorticoids ? During embryogene- sis in the chick plasma glucocorticoids (corticosterone, cortisol and cortisone) become detectable at embryonic day 9 (about 20 ng/ml; Kalliecharan and Hall 1974), but in contrast to rat embryogenesis (Seidl and Unsicker 1989b), only moderately increase by about 3-fold up to embryonic day 15. At embryonic day 17.3, and short- ly before birth, levels of corticosterone alone in rat plas- ma are already twice as high. However, we have also previously found that plasma glucocorticoid levels in the rat embryo at embryonic day 16.3 (20 nM) would al- ready be sufficient to trigger the marker enzyme of chro- maffin cells, PNMT, but fail to do so because the cells still lack functional glucocorticoid receptors at this age. Do chick granule-containing cells at embryonic day 9 express glucocorticoid receptors 9. The available data sug- gest that initial expression of glucocorticoid binding shows marked temporal variations in different chick tis- sues. Thus, in the retina hydrocortisone binding is high- est between embryonic day 7 and 10, whereas in the brain a peak is reached at embryonic day 13 to 14 (Koehler and Moscona 1975). There is no information available with regard to specific glucocorticoid binding sites in embryonic chick sympathetic ganglia. However, results from our culture studies (unpublished) clearly in- dicate that three-fold more embryonic day 7 neurons survive in the presence of hydrocortisone as compared to controls, although granule-containing cells cannot be preferentially maintained. Thus, reasons for the deple- tion of granule-containing cells from embryonic chick sympathetic ganglia may include programmed neuron death, transition into developing principal neurons, and/

158

or a lack of factors other than glucocorticoids that, e.g., may be provided by the early adrenal environment and are essential for their survival.

References

Anderson DJ (1988) Cell fate and gene expression in the developing neural crest. In: Gorio Ae t al. (eds) Neural development and regeneration. NATO ASI Series, Vol H22, 187 198. Springer, Berlin Heidelberg New York

Anderson DJ, Axel R (1986) A bipotential neuroendocrine precur- sor whose choice of cell fate is determined by NGF and gluco- corticoids. Cell 47:1079-1090

Anderson DJ, Michelson AS (1989) Role of glucocorticoids in the chromaffin-neuron developmental decision. Int J Dev Neurosci (in press)

Chester-Jones I, Phillips JG, Holmes WN (1959) Comparative physiology of the adrenal cortex. In : Gorbman A (ed) Compar- ative Endocrinology. John Wiley & Sons, New York, Chapman & Hall, London, pp 582-612

Ciment G, Weston JA (1981) Immunochemical studies of avian peripheral neurogenesis. In: McKay R, Raft MC, Reichardt LF (eds) Monoclonal antibodies to neural antigens. Cold Spring Harbor Laboratory, Vol 2

Coupland RE (1965) The natural history of the chromaffin cell. Longmans, London

Dodd J, Soltr D, Jessell TM (1984) Monoclonal antibodies against carbohydrate differentiation antigens identify subsets of prima- ry sensory neurones. Nature 311:469~470

Doupe A J, Landis SC, Patterson PH (1985 a) Environmental influ- ences in the development of neural crest derivatives: Glucocor- ticoids, growth factors, and chromaffin cell plasticity. J Neuro- sci 5:2119-2142

Doupe AJ, Patterson PH, Landis SC (1985b) Small intensely fluo- rescent cells in culture: Role of glucocorticoids and growth factors in their development and interconversions with other neural crest derivatives. J Neurosci 5:2143-2160

Er/ ink60 (1976) SIF cells. Structure and function of the small, intensely fluorescent sympathetic cells. Fogarty International Center Proceedings No 30, DHEW Publication No (NIH) 76- 942

Ernsberger U, Sendtner M, Rohrer H (1989) Proliferation and differentiation of embryonic chick sympathetic neurons: Effects of ciliary neurotrophic factor. Neuron 2:1275-1284

Fukada K (1985) Purification and partial characterization of a cholinergic neuronal differentiation factor. Proc Natl Acad Sci USA 82: 8795-8799

Gabella G (1976) Structure of the autonomic nervous system. Chapmann and Hall, London. John Wiley & Sons, New York

Garcia-Arras JE, Chanconie M, Fontaine-Perus JE (1984) In vivo and in vitro development of somatostatin-like immuuoreactivi- ty in the peripheral nervous system of quail embryos. J Neurosci 4:1549-1558

Gibbins IL, Furness JB, Costa M (1987) Pathway-specific patterns of the co-existence of substance P, calcitonin gene-related peptide, cholecystokinin and dynorphin in neurons of the dorsal root ganglia of the guinea-pig. Cell Tissue Res 248:417- 437

Grillo MA (1966) Electron microscopy of sympathetic tissues. Pharmacol Rev 18:387-399

Haider W (1960) Das sympathische Nervensystem des Haushuhns (Gallus domesticus). Diss med vet, Berlin (Humboldt-Univer- sit fit)

Hamburger V, Hamilton JL (1951) A series of normal stages in the development of the chick embryo. J Morphol 88:49- 92

Hayashi M, Edgar D, Thoenen H (1985) Nerve growth factor

changes the relative levels of neuropeptides in developing senso- ry and sympathetic ganglia of the chick embryo. Dev Biol 118:49-55

Heym C, Kummer W (1988) Regulatory peptides in paraganglia. Progr Histochem Cytochem 18, No 2

Hruschak KA, Friedrich VL, Giacobini E (1982) Synaptogenesis in chick paravertebral sympathetic ganglia: A morphometric analysis. Dev Brain Res 4:229-240

Kalliecharan R, Hall RK (1974) A developmental study of the levels of progesterone, corticosterone, cortisol, and cortisone circulating in plasma of chick embryos. Gen Comp Endocrinol 33 : 147-159

Katz DM, Adler JE, Black IB (1986) Expression and regulation of tyrosine hydroxylase in adult sensory neurons in culture: Effects of elevated potassium and nerve growth factor. Brain Res 385 : 68-73

Koehler DE, Moscona AA (1975) Corticosteroid receptors in the neural retina and other tissues of the chick embryo. Arch Bio- chem Biophys 170 : 102-113

Landis SC, Patterson PH (1981) Neural crest cell lineages. TINS 7:172-175

LeDouarin NM, Smith J (1983) Differentiation of avian autonomic ganglia. In: Elfvin L-G (ed) Autonomic ganglia. John Wiley & Sons, Chichester New York Brisbane Toronto Singapore, pp 427452

LeDouarin NM, Renaud D, Teillet MA, LeDouarin GH (1975) Cholinergic differentiation of presumptive adrenergic neuro- blasts in interspecific chimeras after heterotopic transplanta- tions. Proc Nat Acad Sci USA 72: 728-732

Lu KS, Lever JD, Santer RM, Presley R (1976) Small granulated cell types in rat superior cervical and coeliac-mesenteric ganglia. Cell Tissue Res 172: 331-343

Luckenbill-Edds L, vanHorn C (1980) Development of chick para- vertebral sympathetic ganglia. I. Fine structure and correlative histofluorescence of catecholaminergic cells. J Comp Neurol 191:65-76

Lundberg JM, H6kfelt T, Schultzberg M, Uvn~s-Wallenstein K, K6hler C, Said SI (1979) Occurrence of vasoactive intestinal polypeptide (VIP)-like immunoreactivity in certain cholinergic neurons of the cat: Evidence from combined immunohisto- chemistry and acetylcholinesterase staining. Neuroscience 4:1539-1559

Lundberg JM, H6kfelt T, Anggard A, Terenius L, Elde R, Markey K, Goldstein M, Kimmel J (1982) Organisational principles in the peripheral sympathetic nervous system: Subdivision by coexisting peptides (somatostatin-, avian pancreatic polypep- tide-, and vasoactive intestinal polypeptide-like immunoreactive materials). Proc Natl Acad Sci USA 79:1303-1307

Mann HB, Whitney DR (1947) On a test of whether one of two random variables is stochastically larger than the other. Ann Math Statist 18:50

Maxwell GD, Sietz PD, Chenard PH (1984) Development of soma- tostatin-like immunoreactivity in embryonic sympathetic gan- glia. J Neurosci 4:576-584

Mirksy R (1982) The use of antibodies to define and study major cell types in the central and peripheral nervous system. In: Brockes J (ed) Neuroimmunology. Plenum Press, New York London, pp 141-181

Mfiller TH, Unsicker K (1981) High-performance liquid chroma- tography with electrochemical detection as a highly efficient tool for studying catecholaminergic systems. I. Quantification of noradrenaline, adrenaline and dopamine in cultured adrenal medullary cells. J Neurosci Methods 4:39-51

Mfiller TH, Unsicker K (1986) Nerve growth factor and dexameth- asone modulate synthesis and storage of catecholamines in cul- tured rat adrenal medullary cells: Dependence on postnatal age. J Neurochem 46: 516-524

New HV, Mudge AW (1986) Distribution and ontogeny of SP, CGRP, SOM, and VIP in chick sensory and sympathetic gan- glia. Dev Biol 116:337 346

159

Otten U, Lorez HP (1983) Nerve growth factor increases substance P, cholecystokinin and vasoactive intestinal polypeptide im- munoreactivities in primary sensory neurones of newborn rats. Neurosci Lett 34:153-158

Papka RE (1972) Ultrastructural and fluorescence histochemical studies of developing sympathetic ganglia in the rabbit. Am J Anat 134:337 363

Reynolds ES (1963) The use of lead citrate at high pH as an elec- tron-opaque stain in electron microscopy. J Cell Biol 17:208- 212

Richardson KC, Jarret C, Finke EH (1960) Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol 35:313 323

Rohrer H, Thoenen H (1987) Relationship between differentiation and terminal mitosis: Chick sensory and ciliary neurons differ- entiate after terminal mitosis of precursor cells whereas sympa- thetic neurons continue to divide after differentiation. J Neuros- ci 7 : 3739-3748

Ross LL, Cosio L, Witczak L, Smolen A (I978) Synaptic develop- ment in the avian sympathetic ganglion. Anat Rec 190:525

Seidl K, Unsicker K (1989a) Survival and neuritic growth of sym- pathoadrenal (chromaffin) precursor cells in vitro. Int J Dev Neurosci, Special Issue "Development and plasticity of sym- pathoadrenal cells", Unsicker K (ed), pp 465-473

Seidl K, Unsicker K (1989b) The determination of the adrenal medullary ceil fate during embryogenesis. Dev Biol (in press)

Seidl K, Manthorpe M, Varon S, Unsicker K (1987) Differential effects of nerve growth factor and ciliary neuronotrophic factor on catecholamine storage and catecholamine synthesizing en- zymes of cultured rat chromaffin cells. J Neurochem 49:169- 174

Shaw TJ, Letourneau PC (1986) Chromaffin cell heterogeneity of process formation and neuropeptide content under control and nerve growth factor - altered condition in cultures of chick embryonic adrenal gland. J Neurosci Res 16:337-355

Silva MT, Carvalho GF, Magalhaes MM (1968) The fixative action of uranyl acetate in electron microscopy. Experientia 20:1074

Sommer EW, Kazimierczak J, Droz B (1985) Neuronal subpopula- tions in the dorsal root ganglion of the mouse as characterized by combination of ultrastructural and cytochemical features. Brain Res 346:310-326

Stempte DL, Mahanthappa NK, Anderson DJ (1988) Basic FGF

induces neuronal differentiation, cell division, and NGF depen- dence in chromaffin cells : A sequence of events in sympathetic development. Neuron 1:517-525

Teitelmann G, Skaper S, Baker H, Park DH, Joh TH, Adler R (1984) Expression of phenylethanolamine N-methyltransferase in sympathetic neurons and extraadrenal chromaffln tissue of chick embryos in vivo and in vitro. Dev Brain Res 13:283-291

Unsicker K (1973a) Fine structure and innervation of the avian adrenal gland. I. Fine structure of chromaffin cells and ganglion cells. Z Zellforsch 145:389-416

Unsicker K (1973b) Fine structure and innervation of the avian adrenal gland. II. Cholinergic innervation of chromaffin cells. Z Zellforsch 145:417-422

Unsicker K (1973c) Fine structure and innervation of the avian adrenal gland. III. Non-cholinergic nerve fibres. Z Zellforsch 145:557-575

Unsicker K, Krisch B, Otten U, Thoenen H (1978) Nerve growth factor-induced fiber outgrowth from isolated rat adrenal chro- maffin cells: Impairment by gtucocorticoids. Proc Natl Acad Sci USA 75:3498-3502

Unsicker K, Skaper SD, Varon S (1985 a) Neurotrophic and neurite promoting factors. Effects on early postnatal chromaffin cells from rat adrenal medulla. Dev Brain Res 17:117-129

Unsicker K, Skaper SD, Varon S (1985b) Developmental changes in the responses of rat chromaffin cells to neuronotrophic and neurite promoting factors. Dev Biol 111 : 425-433

Unsicker K, Skaper SD, Varon S (1985c) Phenotypical changes of embryonic chick adrenal medullary cells in vitro induced by nerve growth factor and ciliary neuronotrophic factor. Neu- rosci Lett 60 : 127 132

Unsicker K, Seidl K, Hofmann H-D (1989) The neuro-endocrine ambiguity of sympatho adrenal cells. Internat J Dev Neurosci Special issue "Development and Plasticity of Sympathoadrenat Cells", Unsicker K (ed), pp 413-417

Unsicker K, Wiegandt H (1988) Promotion of survival and neurite outgrowth of cultured peripheral neurons by exogenous lipids and detergents. Exp Cell Res 178:377 389

Ziller C, Dupin E, Baroffio A, Dulac C, Cameron P, LeDouarin N (1989) New insights on the development of the peripheral nervous system from the neural crest. In: Post FEBS '89. "Reg- ulation of gene expression in nervous system". Catania, July 8 11, 1989