Cell Tissue Res (1983) 229:299-308 Cell and Tissue Research �9 Springer-Verlag 1983

Differentiation and transdifferentiation of adrenal chromaffin cells of the guinea pig III. Transplants under the kidney capsule*

K. Unsicker, U. Zwarg, and O. Habura-Flfih Department of Anatomy and Cell Biology, Philipps University of Marburg, Marburg, Federal Republic of Germany

Summary. Histochemical, ultrastructural and biochemical studies (quan- titative determinations of catecholamines (CA) and phenylethanolamine N-methyltransferase (PNMT) activity) were carried out on autologous transPlants of adult guinea-pig adrenal medulla under the kidney cap- sule, in order to investigate the specific influences of a virtually nerve-free environment in comparison to those mediated by a densely innervated one such as in the iris (cf. Unsicker et al. 1981). Three weeks after trans- plantation chromaffin cells survived well, most cells maintaining their morphological identity in terms of adrenaline (A) storage, although bio- chemically measured A and PNMT had dramatically decreased. Chro- maffin cells in transplants extended neurite-like processes in an identical fashion as seen in transplants to the anterior chamber of the eye and in culture. Chromaffin cells were frequently connected by synaptoid con- tacts, but did not receive cholinergic synapses as observed in transplants to the iris. It may be concluded that the growth factor(s) eliciting neurite outgrowth from transplanted chromaffin cells are rather ubiquitously present, independent of whether the transplantation site is sparsely or richly innervated.

Key words: Adrenal chromaffin cells Plasticity - Transplants - Kidney capsule - Ultrastructure - Catecholamine biochemistry

Previously it was shown (Unsicker et al. 1981) that chromaffin cells from the adrenal medulla of adult guinea pigs develop morphological features of sympathetic neurones or neurone-like small granule-containing cells when they are grafted to the anterior chamber of the eye. This transdifferentiation phenomenon involves extension of neurites, a decrease in the size and number of chromaffin granules, the appearance of small catecholamine

Send offprint requests to: Dr. K. Unsicker, Department of Anatomy and Ceil Biology, Philipps University, Robert-Koch-Str. 6, D-3550 Marburg, Federal Republic of Germany

300 K. Unsicker et al.

(CA)-storing vesicles typical of adrenergic nerves, formation of nerve fibre networks in the denervated host iris, and losses of the adrenaline (A)-synthe- sizing enzyme phenylethanolamine N-methyl-transferase (PNMT) and A. In tissue-culture experiments explants of guinea-pig adrenal medulla dis- played only sparse neurite outgrowth, which was not significantly increased in the presence of nerve growth factor (NGF; Unsicker 1981), suggesting that a factor(s) different from N G F probably induced neurite outgrowth of chromaffin cells in eye-chamber transplants. The present study was de- signed to evaluate (1) whether neuronal transdifferentiation of chromaffin cells in vivo is crucially dependent on a specific environment, e.g. that pro- vided by the iris, which has a rich supply of adrenergic, cholinergic and sensory nerves and is known to produce neurite growth-stimulating factor(s) after denervation (Ebendal et al. 1980), or (2) whether neurite formation by chromaffin cells can occur in any environment outside the intact adrenal gland. The space between the kidney parenchyma and capsule as a trans- plantation site was chosen because of its sparse nerve and rich blood supply (Carruba et al. 1974).

Materials and methods

Adult female guinea pigs weighing 301~500 g were used throughout the study. Under ether anesthesia the left adrenal gland was exposed and removed under sterile conditions, placed in a Petri dish containing ice-cold Hanks' balanced salt solution supplemented with 10% fetal calf serum (Flow Laboratories) and carefully freed from cortical tissue using a binocular microscope. Small pieces (0,5 I mm 3) were transplanted between kidney capsule and parenchy- ma. The animals were kept under constant laboratory conditions with free access to food and tap water and were sacrificed three weeks after surgery.

Electron microscopy. Six animals were perfused under ether anesthesia via the descending aorta with approximately 200 ml of 3.5% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4; 4 - 6 C) for 15 min. Transplants were removed together with adjacent kidney cortex and cap- sule, postfixed in phosphate-buffered glutaraldehyde for 2 h, and rinsed for at least 2 h in phosphate buffer. Tissues were then postfixed in aqueous 2% OsO4 and uranyl acetate for 2 h each, dehydrated in a graded series of ethanol and embedded in Araldite. Thin sections were cut on a Reichert-Sitte OmU 2 ultramicrotome, stained with uranyl acetate and lead citrate for 5 min each and viewed with a Zeiss EM 9A and a Siemens 101 electron microscopes. The contralateral intact adrenal gland and adrenals from four non-operated animals were used as controls and processed for electron microscopy as described above.

5-Hydroxydopamine (5-OHDA) treatment. Two animals received three intraperitoneal injec- tions of 300 mg/kg 5-OHDA-hydrochloride (Labkemi, Sweden) dissolved in 0.2% ascorbic acid 30, 18 and 6 h prior to sacrifice.

Light microscopy. 0.5-1 lam-thick sections were cut from Araldite-embedded material and stained with 1% toluidine blue or a mixture of 1% toluidine blue, 1% methylene blue and 1% borax for light-microscopic orientation.

Histochemistry. The argentaffin reaction was carried out on semithin sections of Araldite- embedded, glutaraldehyde-fixed postosmicated material (Gorgas and B6ck 1976) using the ammoniac silver nitrate solution described by Singh (1964). Noradrenaline (NA) forms unsolu- ble precipitates with glutaraldehyde and, hence, can be identified in sections (Hopwood 1971). Adrenaline (A) and dopamine (DA) are lost during the fixation procedure and, accordingly, cannot be shown with this method.

Transplants of adrenal chromaffin cells: Differentiation and transdifferentiation 301

For formaldehyde-induced fluorescence of catecholamines (CA) adrenal glands and trans- plants were processed according to the method of Falck and Hillarp (cf. Bj6rklund et al. 1972). The paraformaldehyde used had previously been equilibrated in air at 70% relative humidity. After vacuum-embedding in paraffin, specimens were sectioned, mounted in liquid paraffin and analyzed in a Zeiss fluorescence microscope fitted with BG 3 or BG 12 (Schott) primary and Zeiss 47-50 secondary filters. In addition, fresh specimens of kidney capsule with adhering transplanted cells were treated as whole mounts and processed according to the glyoxylic acid method (De La Torre and Surgeon 1976).

Determ&ations of catecholamines (CA) and phenylethanolamine N-methyltransferase (PNMT). Adrenal medullary transplants (n=3) and intact adrenals from five animals were carefully dissected from the kidney surface, freed from adhering tissue and homogenized in 6 ml of 5 mM tris-HCL buffer (pH Z0), containing 0.2% Triton X-100 and 0.2% bovine serum albu- min (w/v both). The homogenate was divided into two parts for the estimation of PNMT activity and CA and further processed as described earlier (Unsicker et al. 1981).

Results

Controls

The ul t ras t ructural appearance o f the adrenal medulla o f guinea pigs was as described by Unsicker et al. (1978). There was a high predominance of A-storing cells, which conta ined granular vesicles with cores o f medium electron densities and an average diameter o f 180 nm. Most NA-s tor ing cells resembled small granule-containing (SGC) cells. They amoun ted to abou t 10% of the total chromaff in cell popula t ion and were character ized by a small cell body, high nuclear-cytoplasmic ratio and small granular vesicles (average core diameters 80 nm).

Adrenal medulla transplants under the kidney capsule

Light microscopy and histochemistry. Toluidine blue-stained semithin sec- tions (Fig. 1) revealed densely packed polygonal chromaff in cells in the centre and a more disorderly ar ray o f chromaff in cells at the per iphery o f the transplants, where loose strands o f connective tissue separated the chromaffin-cell cords. Cells appeared vital, wi thout any signs o f degenera- tion, and were sur rounded by a dense capillary network. Intact cort ical cells were not observed. Al though no counts o f chromaff in cells were per- formed, it was apparen t that they numbered less per cross-sectional area than in the adrenal medulla in situ. Very occasionally areas o f necrotic tissue containing abundan t macrophages , but no degenerat ing chromaff in cells bordered a transplant . When the argentaff in react ion was applied to semithin sections (Fig. 2) chromaff in cells did not display any silver precipi- tates, indicating that p r imary amines were either absent or present in low amounts only. Mast cells were positively stained.

Formaldehyde- induced fluorescence was intensely yellow in chromaff in cells. Yellow-green fluorescent processes, which originated f rom spindle- shaped chromaff in cells, were observed in the per iphery o f the transplants and in whole mounts o f kidney capsule with adherent t ransplanted cells (Fig. 3). Thei r final dest inat ion could not be elucidated. Extrinsic adrenergic nerve fibres were not seen to invade the transplants.

302 K. Unsicker et al.

Fig. I. Adrenal medullary cells in a transplant under the kidney capsule 3 weeks after transplan- tation. The chromaffin cells (ch) show no signs of degenerative changes and are loosely arranged in cords and clusters. The transplant is well vascularized (v). Toluidine blue-stained semithin section, x 125

Fig. 2. Ammoniacal silver nitrate staining of an Araldite-embedded transplant. No specific reaction product can be detected in chromaffin cells (arrows). Mast cells (m) are stained deep brown, x 125

Fig. 3. Whole mount of a kidney capsule with attached transplanted chromaffin cells (ch) processed for the visualisation of catecholamines. Processes (p) can be seen to originate from intensely fluorescent cells, x 420

Electron microscopy

T r a n s p l a n t e d chromafJin cells r e t a ined m o s t o f the u l t r a s t r u c t u r a l fea tures charac te r i s t i c for this cell type (Figs. 4, 5), b u t a d o p t e d m o r e e l o n g a t e d shapes. C A - s t o r i n g g r a n u l a r vesicles were s ign i f ican t ly smal ler t h a n in con - trols, core d i ame te r s v a r y i n g f r o m 50 to 170 nm. Cores were r o u n d or oval

Transplants of adrenal chromaffin cells : Differentiation and transdifferentiation 303

Fig. 4. Group of transplanted chromaffin cells (ch l-ch6) embedded into loose connective tissue (ct). All the cells contain the specific "ch romaf f in" storage vesicles, however, of different densities. It can be seen that two chromaffin cells (ch4, chs), which extend processes, have relatively few "chromaf f in" granules in their perikarya • 7800

304 K. Unsicker et al.

in shape, of low or medium electron density and surrounded by an evenly wide electron-lucent halo indicative of the storage of a secondary amine (Fig. 5). Granular vesicles containing cores of high electron density were found only rarely in Golgi areas and never in a cell type of their own. The number and size distribution (Fig. 5) of granular vesicles showed greater variations between chromaffin cells than in controls. In addition to the large storage vesicles, vesicles 50 nm in diameter with or without dense cores occurred in clusters (Fig. 7). Vesicles of this type are abundant in noradrenergic nerve varicosites and in SGC cells of the guinea-pig adrenal medulla (Unsicker et al. 1978), but they are extremely rare in A-storing cells of the guinea-pig adrenal medulla.

Chromaffin cells in transplants formed processes (Figs. 4, 6-8) that could be traced for more than 100 ~tm in thin sections. These processes contained the populations of vesicles described above within their varicosities and microtubules and neurofilaments in the intervaricose areas.

5-OHDA caused an increase in electron density of the chromaffin gran- ules in approximately 40% of the cells and their processes, and was also taken up into small (50 nm) vesicles.

Nerve cells with the ultrastructural characteristics of sympathetic neurones (Grillo 1966; Unsicker 1967) were occasionally found in transplants in clus- ters of two to three cells or as single cells. They had large oval nuclei (diameters approximately 10 ~tm) with one to two prominent nucleoli and scanty chromatin. Polyribosomes and Nissl bodies were abundant in the cytoplasm, and dense-core vesicles (80-120 nm) were regularly observed in the well-developed Golgi areas.

Nerve fibres were frequently encountered in transplants. Axons containing the large "chromaff in" and small (50 nm) granular and agranular vesicles (Figs. 6 8) within their varicose regions could be distinguished from profiles that displayed both 50 nm agranular, 50 nm granular, and few dense-core vesicles measuring 80-120 nm (Fig. 8), but not the large "chromaff in" storage granules. These two types of neurites were thought to represent processes of chromaffin and nerve cells, respectively. Nerve fibres were in- vested by Schwann cells. 5-OHDA was taken up by most of the small and large vesicles in nerve and chromaffin cell processes and produced the characteristic electron density to the cores (Figs. 7, 8).

Synapses and synaptoid contacts, which were regularly observed on grafted chromaffin and nerve cells, have been previously described in detail (Un- sicker et al. 1977).

Biochemical determinations of adrenaline ( A ), noradrenaline ( NA ) and phenylethanolamine N-methyltransferase (PNMT) activity in transplants

The results given in Table 1 show that transplantation drastically reduced the catecholamine (CA) content of the chromaffin cells. A decreased to

Figs. 5--8. Transplanted chromaffin cells and chromaffin cell processes. Fig. 5, Three chromaffin cells (chl-ch3) with storage vesicles of different diameters. Irrespective of the size differences, all vesicles have the typical appearance of adrenaline-storing granules, with cores of low to medium electron densities. • 18 000. Fig. 6. Process containing two populations of dense-core vesicles. Small granular vesicles (arrows) are smaller than those found in the cell bodies of transplanted chromaffin cells (cf. Fig. 5). Two larger granular vesicles (arrowheads), which are in the size-range of chromaffin granules, suggest an origin of this process from a chromaffin cell. • 18000. Fig. 7. Uptake of 5-OHDA into processes containing small (arrows) and large (arrowhead) vesicles. The diameters of the large granular vesicles may be indicative of their "chromaffin" nature. • 18000. Fig. 8. Axon profiles that contain large (arrowhead) and small (arrows) granular vesicles (5-OHDA treatment). The profile showing the large granules possibly represents a chromaffin cell process. Other axon profiles (a) are devoid of storage vesicles. • 36000

306 K. Unsicker et al.

Table 1. Quantitative determinations of adrenaline (A), noradrenaline (NA) and PNMT in controls and transplanted adrenal glands. Catecholamine levels are expressed in ng per gland or transplant, respectively. Values for PNMT activity are given in pmol.min-1 per gland or transplant, respectively. All values are the means of three to five independent determinations with their standard errors

n A NA A i n % o f A + N A PNMT

Controls 5 12,100+_1,600 840_+280 94 920+_100

Transplants 3 100 + 25 72 +_ 20 60 6 _+ 3

approximately 1%, and NA to approximately 10% of control levels. In relation to the total CA content, a dropped from 94 to 60%. The relative decrease of A was reflected by a sharp decrease in PNMT activity.

Discussion

The present investigation shows that the chromaffin cells in the adrenal medulla of adult guinea pigs develop in an identical manner in autologous transplants beneath the kidney capsule as in grafts to the anterior chamber of the eye (Unsicker et al. 1981): (i) phenotypically the bulk of the cells still belongs to the A-storing type as judged by the negative silver-nitrate staining and the low electron density of the vesicular cores, although PNMT activity decreases dramatically and A constitutes only 60% of the total CA as compared to 94% in controls; (ii) transplanted chromaffin cells re- semble SGC-cells and sympathetic neurons in that their storage granules become smaller than in "mature" chromaffin cells and they extend neurites.

The decrease of A and PNMT in adrenal medullary transplants may be explained by the lack of intact adrenocortical tissue in the immediate environment of the transplants, since high levels of glucocorticoids are re- quired for the induction of the A-synthesizing enzyme PNMT (Kirshner and Goodall 1957; Wurtman 1966). It is generally agreed that PNMT de- clines substantially in transplants (e.g. anterior chamber of the eye: Pohor- ecky et al. 1970; Unsicker et al. 1981), but it remains controversial whether the A-storing cells maintain their morphological identity or transform to NA-storing cells. Guinea-pig chromaffin cells in transplants to the anterior chamber of the eye (Unsicker et al. 1981) and beneath the kidney capsule (present study) do not stain with ammoniac silver nitrate and, at the ultra- structural level, still contain typical A-storing granules, although an increase in electron density of the cores is quite obvious in a few cells. In the adrenal medulla of hypophysectomized rats, Coupland (1982) observed long-term maintenance of typical A- and NA-storing cells in the same proportion and distribution as in controls. In contrast, Piezzi and Cavicchia (1973) described NA-storing cells to be the predominant cell type in transplants of rat adrenal medulla to the anterior chamber of the eye. Pohorecky et al. (1970) have reported that CA in transplants consist almost entirely of NA. These results might be interpreted in terms of a transformation or selective death of A-storing cells. In this context it should be remembered that prima-

Transplants of adrenal chromaffin cells: Differentiation and transdifferentiation 307

ry expression of PNMT activity, but not its subsequent increase, in the embryonic adrenal gland occurs independently from the influence of gluco- corticoid hormones (Bohn et al. 1981 ; Teitelman et al. 1982). This observa- tion might also include the possibility that A-storing cells maintain their identity even under reduced glucocorticoid influence.

It is now well established that chromaffin cells may occur in vivo and in vitro in two phenotypically different forms, namely as an endocrine and as a neuron-like cell. This is best exemplified in vivo by the existence of two types of SIF-cells that can be found in sympathetic ganglia, where they may act as interneurons or secretory cells producing DA or NA, in addition to various peptides (see Taxi 1979, for review). It is still unknown what signals govern this choice of a particular phenotype in vivo. In vitro, numerous factors have been identified, which favour the expression of the neuronal or the endocrine phenotype, respectively (for an extensive discus- sion of the literature, see Unsicker 1981, 1982; Unsicker et al. 1981 ; Un- sicker and Hofmann 1982).

The factors that induce neurite outgrowth from cultured chromaffin cells cover a large spectrum of well-identified and still enigmatic proteins, such as nerve growth factor (NGF), a factor contained in C6 rat glioma cell-conditioned medium (Unsicker et al. 1982; Vey et al. 1982), and factors present in media conditioned over adrenal fibroblast-like (Unsicker and Hofmann 1982) or SV-3T3 cells (to be published). In the context of the present study it is interesting to note that in vivo the capacity of a particular environment, such as the iris or the kidney capsule, to produce a neurite- outgrowth-inducing factor does not necessarily depend on whether this envi- ronment originally has a rich nerve supply (iris) or is virtually devoid (kidney capsule) of nerve fibres. Ebendal and co-workers (1980) have provided evi- dence that the denervated rat iris produces NGF, a possible candidate for eliciting neurite outgrowth from transplanted chromaffin cells. However, in a previous paper of this series (Unsicker 1981) we have shown that NGF does not induce neurite outgrowth in guinea-pig chromaffin cells in vitro. Thus, we conclude that a hitherto unidentified molecule, that may be rather ubiquitous and certainly different from NGF, is reponsible for neurite out- growth from transplanted guinea-pig chromaffin cells. Alternatively, an action of different growth factor(s) might be postulated.

Acknowledgements. This work was supported by grants from the Deutsche Forschungsgemein- schaft. The authors wish to thank Professor K. L6ffelholz, Dr. R. Lindmar and Mrs. U. Wolf, Department of Pharmacology, University of Mainz, for conducting the biochemical determinations of catecholamines and PNMT. Skilful technical and secretarial assistance was provided by R. Korff, Ch. Fiebiger, H. Schneider and W. B6rner.

References

Bj6rklund A, Falck B, Owman Ch (1972) Fluorescence microscopic and microspectrofluoro- metric techniques for the cellular localization and characterization of biogenic amines. In: Berson SA (ed) Methods of investigative and diagnostic endocrinology, Vol 1. The thyroid and biogenic amines (Rall JE, Kopin IJ (eds). North Holland Publ Co, Amsterdam, pp 318-368

Bohn MC, Goldstein M, Black IB (1981) Role of glucocorticoids in expression of the adrenergic phenotype in rat embryonic adrenal gland. Dev Biol 82:1-10

308 K. Unsicker et al.

Carruba M, Ceccarelli B, Clementi F, Mantegazza P (1974) Selectivity in the reinnervation of iris and adrenal medulla by superior cervical ganglion after transplantation under the kidney capsule. Brain Res 77 : 39-53

Coupland RE (1982) The identity of adrenaline- and noradrenaline-storing chromaffin cells. Symp Molecular Neurobiology of Peripheral Catecholaminergic Systems, Ibiza, Spain, 5 10th September 1982

De La Torre JC, Surgeon JW (1976) Histochemical fluorescence of tissue and brain mono- amines: results in 18 rain using the sucrose-phosphate-glyoxylic acid (SPG) method. Neu- roscience 1:451-453

Ebendal T, Olson L, Seiger ~, Hedlund KO (1980) Nerve growth factors in the rat iris. Nature 286 : 25-28

Gorgas K, B6ck P (1976) Identification of chromaffin and enterochromaffin cells in semithin sections by means of the argentaffin reaction. Mikroskopie 32 : 57-63

Grillo MA (1966) Electron microscopy of sympathetic tissues. Phanuacol Rev 18:387-399 Hopwood D (1971) The histochemistry and electron histochemistry of chromaffin tissue. Progr

Histochem Cytochem 3/1 Kirshner N, Goodall McC (1957) The formation of adrenaline from noradrenaline. Biochem

Biophys Acta 24:658-659 Piezzi RS, Cavicchia JC (1973) Explants of rat adrenal medulla. A light and electron microscop-

ic study. Anat Rec 175:77-86 Pohorecky LA, Piezzi RS, Wurtman RJ (1970) Steroid induction of phenylethanolamine-N-

methyl transferase in adrenomedullary explants: Independence of adrenal innervation, En- docrinology 86:1466-1468

Singh J (1964) A modification of the Masson-Hamperl method for staining of argentaffin cells. Anat Anz 115:81 82

Taxi J (1979) The chromaffin and chromaffin-like cells in the autonomic nervous system. lnt Rev Cytol 57:283-343

Teitelman G, Joh TH, Park D, Brodsky M, New M, Reis DJ (1982) Expression of the adrener- gic phenotype in cultered fetal adrenal medullary cells: role of intrinsic and extrinsic factors. Dev Biol 89:45(L459

Unsicker K (1967) I~lber die Ganglienzellen im Nebennierenmark des Goldhamsters (Mesocrice- tus auratus). Z Zellforsch 76:187-219

Unsicker K (1981) Differentiation and transdifferentiation of adrenal chromaffin cells of the guinea pig. II. Adrenal medullary explants grown in tissue culture. Cell Tissue Res 219:339 353

Unsicker K (1982) Cell and tissue culture studies on the sympatho-adrenal system. In: Elfvin L (ed) Autonomic ganglia. Wiley and sons, New York (in press)

Unsicker K, Hofmann HD (1982) Phenotypic plasticity of cultured bovine chromaffin cells. I. Morphological changes induced by nonchromaffin cells and organ extracts, but not by mouse and bovine nerve growth factor. Dev Brain Res (in press)

Unsicker K, Zwarg U, Habura O (1977) Electron microscopic evidence for the formation of synapses and synaptoid contacts in adrenal medullary grafts. Brain Res 120:533-539

Unsicker K, Habura-Flueh O, Zwarg U (1978) Different types of small granule-containing cells and neurons in the guinea pig adrenal medulla. Cell Tissue Res 189:109-130

Unsicker K, Tschechne B, Tschechne D (1981) Differentiation and transdifferentiation of adrenal chromaffin cells of the guinea pig. I. Transplants to the anterior chamber of the eye. Cell Tissue Res 215:341 367

Unsicker K, Hofmann HD, Millar TJ, Mfiller T, Schmidt R, Vey J, Ziegler W (1982) Neural crest lineages and chromaffin cell plasticity. Symp Molecular Neurobiology of Peripheral Catecholaminergic Systems, lbiza, Spain, 5-10th September 1982

Vey J, Hofmann HD, Unsicker K (1982) C 6-Gliomzellen produzieren einen Faktor, der Neuritenwachstum und ~berleben chromaffiner Zellen in vitro stimuliert. 3. Arbeitstagung der Anatomischen Gesellschaft, Wiirzburg, 6.-8. October 1982

Wurtman RJ (1966) Control of epinephrine synthesis in the adrenal medulla by the adrenal cortex: Hormonal specificity and doseresponse characteristics. Endocrinology 79:608-614

Accepted October 27, 1982