Histochemistry (1989) 91 : 265-272 Histochemis �9 Springer-Verlag 1 9 ~

Light- and electron-microscopical localization of calcitonin, calcitonin gene-related peptide, somatostatin and C-terminal gastrin]cholecystokinin immunoreactivities in rat thyroid J. Arias 1., L. Seopsi a, J.A. Fischer 2, and L.-I. Larsson 1,3. .

Unit of Histochemistry, University Institute of Pathology, Copenhagen, Denmark 2 Research Laboratory for Calcium Metabolism, Departments of Orthopedic Surgery and Medicine, Forchstrasse 340, CH-8008 Ziirich, Switzerland 3 Department of Molecular Cell Biology, State Serum Institute, Building 81, Amager Boulevard 80, DK-2300 Copenhagen, Denmark

Accepted October 14, 1988

Summary. Parafollicular C cells of the rat thyroid contain several immunoreactive peptides including calcitonin (CT), calcitonin gene-related peptide (CGRP), somatostatin and a C-terminal gastrin/CCK immunoreactive epitope as shown at the light- and electron-microscopical levels. Adult thyroid C cells are strongly immunoreactive to CT and most of the ceils also react strongly with CGRP antisera and weakly with a gastrin/CCK antiserum. The latter antiserum may cross-react with CGRP. This cross-reactivity probably only occurs at very high concentrations of CGRP observed in adult thyroid C cells, but not in intrathyroidal CGRP- containing nerves, nor in early neonatal C cells. In neonatal rats, somatostatin immunoreactive C cells are numerous and most of these cells are also CT and CGRP immunoreac- tive. In contrast, only few C cells display somatostatin im- rnunoreactivity in adult rat thyroids. Sequential staining experiments revealed that some thyroidal C cells simulta- neously express all four types of immunoreactivity. At the electron microscopical level, all of these immunoreactivities were observed in secretory granules of C cells. Double- and triple-staining experiments, moreover, documented that some peptides are co-localized in the same granules.

Introduction

Calcitonin (CT) producing C cells of the thyroid gland con- tain several other immunoreactive regulatory peptides such as calcitonin gene-related peptide (CGRP), somatostatin and an epitope reacting with antisera to the C-terminal tetrapeptide amide of gastrin and cholecystokinin (CCK) (Buffa et al. 1979; Bussolati and Pearse 1967; H6kfelt et al. 1975; Kameda et al. 1982; Larsson 1985a and b; Sabate et al. 1985; Van Noorden et al. 1977). Messenger-RNA spe- cies encoding the precursors of CT, CGRP, somatostatin and CCK have also been recognized in C cells of normal thyroid glands and in medullary thyroid carcinoma (Amara et al. 1985; Deschenes et al. 1984; Goodman et al. 1982; Hrfler et al. 1987; Rosenfeld et al. 1983; Sabate et al. 1985).

Two forms of CGRP, differing in one of the 37 amino acid residues are present in the rat thyroid (Amara et al.

* Dr. Arias is now at Universidad de Oviedo, Depto. de Psicobiolo- gia, Facultad de Psicologia, E-3307 Oviedo, Spain and Dr. Scopsi is at Anatomia Patologica Istituto Nazionale Tumori, Via G. Vene- zian 1, 1-20133 Milano, Italy ** To whom offprint requests should be sent

1985). c~CGRP and CT are formed as a result of alternative splicing of the initial messenger transcript of the CT/ eCGRP gene. The second form, flCGRP, is the product of a different gene. mRNA encoding for the precursor of CT is the most abundant, followed by those encoding ~CGRP and flCGRP.

Somatostatin is also synthesized by thyroidal C cells (Hrfler et al. 1987), but its expression is species-variable. Thus, in rabbit and guinea pig thyroids, all C cells display both CT and somatostatin immunoreactivity, whereas, in the rat, only a minority of CT immunoreactive cells stain for somatostatin (Kameda et al. 1984; Larsson 1985a and b). In neonatal rats, however, nearly all CT immunoreactive cells contain somatostatin (Larsson 1985b). Thus, the ex- pression of somatostatin by rat C cells is developmentally regulated, implying a functional role for this peptide during early thyroid development.

CCK m R N A has been detected in a rat C cell tumour (Deschenes et al. 1984), but immunocytochemical analyses using region-specific antisera have not revealed the presence of true CCK in normal rat thyroid glands (Larsson 1985 b). However, an antiserum recognizing the C-terminal tetra- peptide amide, common to gastrin and CCK, was found to react with rat thyroid C cells (Larsson 1985b). Most likely, a cross-reacting peptide was detected. A possible can- didate is CGRP. Thus, as a result of some similarities be- tween the carboxyl terminus of CCK-4 and CGRP, absorp- tions of the antiserum with large quantities of eCGRP abol- ished staining of nerve terminals, previously believed to contain CCK (Ju et al. 1986).

In the present report, the ultrastructural localization, ontogeny and the coexistence of CT, CGRP, somatostatin- 14 and -28, and the CCK-4-1ike epitope have been examined in normal rat thyroid glands.

Materials and methods

Tissue processing

Adult Wistar rats (n = 19) and rat pups, collected from timepaired Wistar rats at l, 4, 8 and 21 days post partum, were lightly anaes- thesized with diethyl ether and intracardially perfused with Ringer solution. For light microscopy, rats were perfusion fixed with either 4% paraformaldehyde (Merck, Darmstadt, FRG) in 0.1 M sodium phosphate buffer pH 7.4 (PB), or with similarly buffered commer- cial formalin (4% formaldehyde). For electron microscopy, rats were perfused with a mixture of 2% paraformaldehyde and 3% glutaraldehyde in PB at 4 ~ C. For light microscopy thyroid glands

266

were post-fixed overnight at 4 ~ C in either 4% paraformaldehyde or formalin, and further processed for cryostat sectioning (Larsson 1983). Specimens for electron microscopy were post-fixed for 15-30 min in the paraformaldehyde-glutaraldehyde mixture at 4 ~ C, rinsed overnight in 0.1 M sodium cacodylate buffer pH 7.3, and embedded in Lowicryl K4M (Carlemalm et al. 1982).

Antisera and reagents

Primary antisera included a goat anti-rat CGRP antiserum (Ab 216; Tschopp et al. 1985), and rabbit antisera to synthetic human CT (Immuno Nuclear Corporation, Stillwater, MI, USA), to soma- tostatin-14 (Ab R37/3; Larsson et al. 1979), to somatostatin-2811- 14] (Peninsula Laboratories, Belmont, CA, USA), and to synthetic human gastrin I (Ab 4562). The latter antiserum recognizes the C-terminal tetrapeptide amide common to gastrin and cholecystok- inin (cf. Larsson 1985b).

Secondary antibodies included unlabelled sheep anti-rabbit IgG (SAR, Statens Bakteriologiska Laboratorium, Stockholm, Sweden) and rabbit anti-goat IgG (RAG, Dakopatts, Copenhagen, Denmark). These two IgG preparations were also labelled with colloidal gold particles of 5 nm (SARG5 and RAGG5, respectively) and 10 nm (SARG10) (Wang et al. 1985). Colloidal gold particles were fabricated according to Slot and Geuze (1985). Peroxidase- antiperoxidase (PAP) complex was purchased from Dakopatts (Co- penhagen, Denmark).

Primary antisera, secondary unlabelled antibodies and PAP complexes were diluted in 0.05 M Tris buffer pH 7.4, containing 0.15 M NaC1 (0.05 M TBS) and 0.25% BSA. Gold-labelled second- ary antibodies were diluted to an optical density of 0.3 at 520 nm in 0.02 M Tris buffer pH 8.2 containing 0.15 M NaC1, 1% BSA and 0.02% Na-azide.

Immunocytochemical staining

Light microscopy. Cryostat sections were soaked in 0.05 M TBS containing 1% triton X-I00 (TBS-triton) for 15 rain at room tem- perature, pretreated for 1 h with 1% human serum albumin (HSA) in 0.05 M TBS and incubated with the primary antisera at dilutions ranging from 1:600 to 1:12800 for 20 h at 4~ followed by 1 h at room temperature. After 3 x 10 rain rinses with TBS-triton, a two-step method with SAR (1:30) and PAP complex (1:75) was used with the primary rabbit antisera, and a three-step method with RAG (1:240), SAR (1:30) and PAP complex (1:75) was used with the primary goat antiserum. All these immunoreagents were applied for 30 rain at room temperature. Peroxidase activity was demonstrated with the 3-amino-9-ethylcarbazole (3-A-9-EC) pro- cedure (Graham et al. 1965).

For the demonstration of two different antigens occurring in the same cell, both the elution method of Tramu et al. (1978), and the mirror section procedure were used. In the latter procedure, two consecutive sections were mounted on chromealum-gelatine- coated glass slides with the cut surfaces facing upwards. Thus, the complementary sides of the two sections were exposed to the immunoreagents.

Electron microscopy'. Ultrathin sections were collected on naked Ni-grids and exposed to 1% HSA-0.05 M TBS for 10 min at room temperature. Incubation with primary antisera was then carried out as described for light microscopy. The site of the immunologi- cal reaction was localized using the immunogold method of Wang and Larsson (1985). In the case of CT and somatostatin antisera, SARG5 and SARG10 were used. To enhance labelling efficiency with antiserum 4562, the sections were sequentially reacted with SARG5 and RAGG5. With the goat anti-CGRP antiserum, a two- step procedure was used with RAG followed by SARG5 or SARG10. Double and triple immunostaining was done according to the paraformaldehyde vapour method (Wang and Larsson 1985). In the double-labelling experiments, the sequence of the primary antibodies and secondary gold-labelled probes was varied, and the best combination with respect to labelling efficiency of

both antigens was determined (e.g.: CT-5 nm colloidal gold+ CGRP-10 nm colloidal gold; CT-10 nm colloidal gold+CGRP- 5 nm colloidal gold; CGRP-5 nm colloidal gold + CT-10 nm colloi- dal gold; CGRP- 10 nm colloidal gold + CT-5 nm colloidal gold). For triple staining experiments, silver enlargement of 5 nm colloi- dal gold particles to 15 nm was carried out (Scopsi et al. 1986).

Controls. Controls included conventional staining controls (Stern- berger 1979; Larsson 1983) as well as specificity controls using primary antisera preabsorbed with synthetic human CT (Ciba- Geigy, Basel, Switzerland), somatostatin-14, somatostatin-28[l-14], somatostatin-2811-12], tetragastrin and synthetic rat ~CGRP (Pen- insula, Belmont, CA and Cambridge Research Biochemicals, Cam- bridge, UK) at 15-50 gg/ml diluted antiserum.

Quantitation

For the demonstration of two different peptides (CT/CGRP, CGRP/somatostatin, CGRP/gastrin-CCK) in the same cell, 700 cells in four adult thyroid glands were examined and immunoreac- tire cells counted in 9 x 13 cm prints (final magnification 192 x ) taken from sequentially stained sections (using the Tramu's elution method). In neonatal rats, the number of immunoreactive C cells were counted in adjacent sections.

In electron micrographs (63000 x, final magnification), the density of colloidal gold labelling over secretory granules was deter- mined by counting and by stereological procedures (Weibel 1969). Quantitation of labelling densities was performed on double- stained specimens using different combinations of primary antisera.

Results

Adul t rat thyroid glands contained numerous parafol l icular C cells located mainly in the central por t ion of each lobe. These cells stained strongly with antibodies to CT and to CGRP. Examinat ion of adjacent sections and sequential staining, using Tramu 's elution method, revealed the co- existence of CT, C G R P , somatostat in-14 and -28 and CCK- 4-like immunoreact ivi ty (Figs. 1-3). In over 90% of the C cells, both immunoreact ive CT and C G R P were demon- strated. Of these, only 0.4% and 0.7%, respectively, con- tained either only CT or CGRP. C G R P immunoreact ive nerve terminals were observed a round thyroid follicles.

The somatosta t in- immunoreact ive cells occurred in small clusters. 3.1% of the C G R P immunoreact ive cells also contained somatostat in. Of the somatostat in-~4 and -28 containing cells only 4.6% and 1.3%, respectively, did not contain CGRP. All cells reactive to the C-terminal gas- t r in /CCK antiserum contained C G R P immunoreact ivi ty.

The electron-microscopic observations extended the l ight-microscopic results. C cells contained numerous secre- tory granules, displaying both CT (Fig. 4a) and C G R P (Fig. 4b) immunoreactivit ies. Using double staining proce- dures, the presence of both C G R P and CT was revealed in the same granules (Fig. 4c). The staining intensity was the same in dark and clear granules. The number of gold particles was evaluated in 150 granules from different cells. The highest labelling intensity of individual granules was always found with the first ant iserum used. Labelling den- sity obtained with ant iserum 4562 was invariably low in relation to CT and CGRP. Moreover, immunoreact ive granules were not present in all C cells examined. However, when present, 4562 immunoreact ivi ty co-localized with CT and C G R P (see insert, Fig. 4c). Somatostat in-14 immuno- reactivity also co-localized with both CT/CGRP, and with somatostatin-28. A small number of C cell granules did not stain with any of the antisera tested.

267

Fig. 1. a and b Sections of adult rat thyroid gland, stained consecu- tively for C G R P (a) and for CT (b) using the Tramu elution proce- dure. e and d Mirror sections, stained for CGRP (e) and CT (d).

Note that CT and CGRP co-exist in nearly all C cells. Scale bar 20 gm

268

269

Fig. 3 a and b. Mirror sections of adult rat thyroid gland, stained with the gastrin/CCK antiserum (a) or for CGRP (b). Note the corresponding immunoreactive cells. Scale bar 20 btm

In neonatal rats, the number of CT and C G R P immuno- reactive cells increased in parallel. In nearly all CT-contain- ing cells the presence of C G R P was also revealed at all stages studied. Somatostatin immunoreactive cells were nu- merous in thyroids of 1-, 4- and 8-day-old rats. At this early neonatal stage, many parafollicular cells stained for calcitonin, C G R P and somatostatin-14. At around day 8, the number of somatostatin immunoreactive cells started to decrease progressively to attain the low frequency ob- served in adult rats. The number of CCK-4 immunoreactive cells, on the other hand, increased markedly after the neon- atal period.

Discussion

The present results confirm and extend previous findings on the co-existence of CT, CGRP, somatostatin and the gastrin/CCK-4-1ike antigen in thyroidal C cells. In the adult

4

Fig. 2. a and b Sections of adult rat thyroid gland, stained consecu- tively for CGRP (a) and for somatostatin (b) using the Tramu elution procedure, e and d Mirror sections, stained for CGRP (c) and somatostatin (d). Note co-existence of CGRP and somatostatin in some C cells. Note also in a the presence of a varicose CGRP fiber not immunoreactive for somatostatin in b. Scale bars 20 p.m

rat, the vast majority of C cells stain strongly with CT antisera. Most of these cells also react strongly with antisera to C G R P and react weakly with an antiserum recognizing the C-terminal gastrin/CCK tetrapeptide amide. In con- trast, a small number of cells react with antisera to somatos- tatin-14 and somatostatin-28. Numerous C G R P immunore- active nerve terminals were detected in the thyroid, confirm- ing recent findings of others (Grunditz et al. 1986; Zabel et al. 1987a). The nerve fibres did not display CT, somatos- tatin or gastrin/CCK-4-1ike immunoreactivities. They prob- ably correspond to previously described sensory C G R P nerve terminals (Gibbins et al. 1985; Lee et al. 1985).

As revealed here for the first time at the electron micro- scopical level, all immunoreactive peptides studied were de- tected in secretory granules. The intensity of staining over the granules varied with the different peptides. We therefore undertook double- and triple-staining procedures to deter- mine the degree of co-localization of the peptides. As a result, CT was found to co-exist with CGRP, somatostatin and gastrin/CCK-4-1ike immunoreactivity within individual granules.

A direct quantitative comparison between the granular contents of the different peptides was not feasible since the sequence of application of primary antibodies was im- portant for the ultimate labelling efficiency. This result was expected as the peptides are much smaller than the IgG molecules used for their detection, causing steric hindrance. Moreover, the antisera employed have variable titers and affinities. Thereby, their sensitivity and detection efficiency varies.

271

Our data demonstrate that several secretory peptides, encoded by distinct genes are co-localized in the same gran- ules of rat thyroidal C cells. The number of CT and C G R P immunoreactive cells increased in parallel during ontogeny (Nitta et al. 1986). The number of somatostatin immunore- active cells, on the other hand, gradually decreased during the postnatal period (Alumets et al. 1980; Larsson 1985b; Zabel et al. 1987b).

The number of CCK-4-1ike immunoreactive cells in- creased in parallel with the decrease in somatostatin cells. The identity of the CCK-4-1ike antigen is not known, but most likely does not represent authentic cholecystokinin, as antisera recognizing other regions of CCK do not stain C cells (Larsson 1985b). Its pattern of development differs from that of CGRP. Ju et al. (1986) recently reported that absorption of C-terminal gastr in/CCK antisera with C G R P abolished staining of nerve terminals in the spinal cord. It is therefore possible that the antisera cross-reacted with C G R P in thyroid C cells. Thyroidal nerve fibers, however, stained with only CGRP antisera and not with the C-termi- nal gastr in/CCK antiserum. The absent gastr in/CCK stain- ing in nerve fibers could be due to lower C G R P concentra- tions in the nerve fibers. Similarly, the discrepant develop- mental patterns of C G R P and the gastr in/CCK immunore- activity in the C cells could have a similar explanation.

Acknowledgements. The Unit of Histochemistry and Department of Molecular Cell Biology are sponsored by the Danish Medical Research Council, the Lundbeck Foundation and Cancer Society. Dr. J. Arias was on a fellowship granted by the Spanish Govern- ment and Dr. L. Scopsi was on a fellowship granted by the Danish Medical Research Council. The Research Laboratory for Calcium Metabolism is supported by the Swiss National Science Founda- tion grant 3.957-0.84 and the Kanton of Zurich. We thank Mrs. A.-M. Ohlsen for technical assistance.

References

Alumets J, H~tkanson R, Lundquist G, Sundler F, Thorell J (1980) Ontogeny and ultrastructure of somatostatin and calcitonin cells in the thyroid gland of the rat. Cell Tissue Res 206:193-201

Amara SG, Ariza JL, LeffSE, Swanson LW, Evans RM, Rosenfeld MG (1985) Expression in brain of a messenger RNA encoding a novel neuropeptide homologous to calcitonin gene-related peptide. Science 229 : 1094-1097

Buffa R, Chayvialle JA, Fontana P, Usellini L, Capella C, Solcia E (1979) Parafollicular cells of rabbit thyroid store both calcito- nin and somatostatin and resemble gut D cells ultrastructurally. Histochemistry 62:281-288

Bussolati G, Pearse AGE (1967) Immunofluorence localization of

4

Fig. 4a-e. Electron micrographs of C cells of adult rat thyroid, stained for CT using 5.7 nm colloidal gold particles as markers (a), for CGRP using 5.7 nm colloidal gold particles as markers (b and insert), double-stained for calcitonin (5 nm particles) and CGRP (10 nm particles) (c) and triple-stained for the gastrin/CCK- like antigen (silver-enhanced 5 nm particles corresponding to the largest particles), CGRP (10 nm particles) and CT (5 nm particles) (insert in e). Note that most granules contain both CT and CGRP (a, b and e) and that some granules contain both CGRP, CT and gastrin/CCK-like immunoreactivities. Scale bars 400 nm in a, b and e and 100 nm in the inserts in b and e

calcitonin in the "C" cells of pig and dog thyroid. J Endocrinol 37 : 205-209

Carlemalm E, Garavito RM, Villiger W (1982) Resin development for electron microscopy and an analysis of embedding at low temperature. J Microsc (Oxford) 126:123 143

Deschenes R J, Lorenz LJ, Haun RS, Roos BA, Collier K J, Dixon JE (1984) Cloning and sequence analysis of a cDNA encoding rat preprocholecystokinin. Proc Natl Acad Sci USA 81 : 726-730

Gibbins IL, Furness JB, Costa M, MacIntyre I, Hillyard C J, Girgis S (1985) Co-localization of calcitonin gene-related peptide-like immunoreactivity with substance P in cutaneous, vascular and visceral sensory neurons of guinea pigs. Neurosci Lett 57:125-130

Goodman RH, Jacobs JW, Dee PC, Habener JF (1982) Somatosta- tin-28 encoded in a cloned cDNA obtained from a rat medul- lary thyroid carcinoma. J Biol Chem 257:1156-1159

Graham RC, Lundhoim U, Karnowsky MJ (1965) Cytochemical demonstration of peroxidase activity with 3-amino-9-ethylcar- bazole. J Histochem Cytochem 13:15~152

Grunditz T, Ekman K, Hfikanson R, Rerup C, Sundler F, Uddman R (1986) Calcitonin gene-related peptide in thyroid nerve fibers and C cells: effects on thyroid hormone secretion and response to hypercalcemia. Endocrinology 119 : 2313-2324

H6fler H, Childers H, Montminy MR, Goodman RH, Lechan RM, DeLellis RA, Tischler AS, Wolfe HJ (1987) Localization of somatostatin mRNA in the gut, pancreas and thyroid gland of the rat using antisense RNA probes for in situ hybridization. Acta Histochem (Suppl) 34:$101-SI05

H6kfelt T, Efendib S, Hellerstr6m C, Johansson O, Luft R, Ari- mura A (1975) Cellular localizatin of somatostatin in endocrine- like cells and neurons of the rat with special reference to the Al-cells of the pancreatic islets and to the hypothalamus. Acta Endocrinol 80 (Suppl 200):1-41

Ju G, H6kfelt T, Fischer JA, Frey P, Rehfeld JF, Dockray GJ (1986) Does cholecystokinin-like immunoreactivity in rat pri- mary sensory neurons represent calcitonin-gene-related pep- tide? Neurosci Lett 68:305-310

Kameda Y, Oyama H, Endoh M, Horino M (1982) Somatostatin immunoreactive C cells in thyroid glands from various mamma- lian species. Anat Rec 204:161-170

Kameda Y, Oyama H, Horino M (1984) Ontogeny of immunoreac- rive somatostatin in thyroid C cells from dogs and guinea pigs. Anat Rec 208:89-101

Larsson L-I (1983) Methods for immunocytochemistry of neu- rohormonal peptides. In: Bj6rklund A, H6kfelt T (eds) Hand- book of chemical neuroanatomy, vol 1 : Methods in chemical neuroanatomy. Elsevier, Amsterdam, pp 147-209

Larsson L-I (1985 a) Distribution and morphology of somatostatin cells. In: Patel YC, Tannenbaum GS (eds) Advances in experi- mental medicine and biology, vol 188: Somatostatin. Plenum Press, New York, pp 383-402

Larsson L-I (1985 b) Differential changes in calcitonin, somatosta- tin and gastrin/cholecystokinin-like immunoreactivities in rat thyroid parafollicular cells during ontogeny. Histochemistry 82:121-130

Larsson L-I, Golterman N, De Magistris L, Rehfeld F J, Schwartz TW (1979) Somatostatin cell processes as pathways for parac- fine secretion. Science 205:1393-1395

Lee Y, Takami K, Kawai Y, Girgis S, Hillyard C J, Maclntyre I, Emson PC, Tohyama M (1985) Distribution of calcitonin gene-related peptide in the rat peripheral nervous system with reference to its coexistence with substance P. Neuroscience 15:1227-1237

Nitta K, Kito S, Kubota Y, Girgis SY, Hiltyard C J, MacIntyre l, Inagaki S (1986) Ontogeny of calcitonin gene-related peptide and calcitonin in the rat thyroid. Histochemistry 84:139 143

Rosenfeld MG, Mermod J-J, Amara SG, Swanson LW, Saw- chenko PE, Rivier J, Vale WW, Evans RM (1983) Production of a novel neuropeptide encoded by the calcitonin gene via tissue-specific RNA processing. Nature 304:129-135

272

Sabate MI, Stolarsky LS, Polak JM, Bloom SR, Varndell IM, Gathei MA, Evans RM, Rosenfeld MG (1985) Regulation of neuroendocrine gene expression by alternative RNA process- ing. Colocalization of calcitonin and calcitonin gene-related peptide in thyroid C-cells. J Biol Chem 260:2589-2592

Scopsi L, Larsson L-I, Bastholm L, Nielsen MH (1986) Silver- enhanced colloidal gold probes as markers for scanning electron microscopy. Histochemistry 86:35-41

Slot JW, Geuze HJ (1985) A new method for preparing gold probes for multiple-labeling cytochemistry. Eur J Cell Biol 38 : 87-93

Sternberger LA (1979) Immunocytochemistry, 2nd edn. John Wiley and Sons, New York

Tramu G, Pillez A, Leonardelli J (1978) An efficient method of antibody elution for the successive or simultaneous localization of two antigens by immunocytochemistry. J Histochem Cyto- chem 26 : 322-324

Tschopp FA, Henke H, Petermann JB, ToNer PH, Tauzer R, H6k- felt T, Lundberg JM, CueUo CA, Fischer TA (1985) Calcitonin gene-related peptide and its binding site in the human central

nervous system and pituitary. Proc Natl Acad Sci USA 82: 248-252

Van Noorden S, Polak JM, Pearse AGE (1977) Single cellular origin of somatostatin and calcitonin in the rat thyroid gland. Histochemistry 53:243-247

Wang B-L, Larsson L-I (1985) Simultaneous demonstration of multiple antigens by indirect immunofluorescence or immuno- gold staining. Histochemistry 83 : 47-56

Wang B-L, Scopsi L, Nielsen MH, Larsson L-I (1985) Simplified purification and testing of colloidal gold probes. Histochem- istry 83 : 109-115

Weibel ER (1969) Stereological principles for morphometry in elec- tron microscopy cytology. Int Rev Cytol 26:235-302

Zabel M, Biela-Jacek I, Surdyk J, Dietel M (1987a) Studies on localization of calcitonin gene-related peptide (CGRP) in the thyroid-parathyroid complex. Virchows Arch A 411 : 569-573

Zabel M, Surdyk J, Biela-Jacek I (1987b) Immunocytochemical studies on thyroid parafollicular cells in postnatal development of the rat. Acta Anat 130:251-256