Proc. Nati. Acad. Sci. USA Vol. 83, pp. 5316-5320, July 1986 Neurobiology Immunohistochemical demonstration of choline acetyltransferase- immunoreactive preganglionic fibers in guinea pig autonomic ganglia (acetylcholine/neuropeptide Y/enkephalin/) BJORN LINDH*t, WILLIAM STAINES*, TOMAS HOKFELT*, LARS TERENIUSt, AND PAUL M. SALVATERRA§ Departments of *Histology and tAnatomy, Karolinska Institutet, S-104 01 Stockholm, Sweden; tDepartment of Pharmacology, Uppsala University, Biomedicum, P.O. Box 573, S-751 24 Uppsala, Sweden; and §Division of Neurosciences, Beckman Research Institute of the City of Hope, Duarte, CA 91010 Contributed by Tomas Hokfelt, March 11, 1986

ABSTRACT The distribution of choline acetyltransferase Ketalar (Parke Davis, Gwent, Great Britain) and Rompun (ChoAcTase)-like immunoreactivity in the superior cervical Vet (Bayer, Leverkusen, F.R.G.) anesthesia. In two animals, and the stellate ganglion was analyzed with im- the cervical part ofthe was cut on one side munohistochemistry. A dense network of ChoAcTase-immu- immediately caudal to the superior cervical ganglion, and the noreactive nerve fibers was observed in both ganglia studied. animals were killed 5 days later. All guinea pigs were The ChoAcTase-positive fibers were found in all parts of the perfused through the ascending aorta with 100 ml ofTyrode's ganglia but had a differential distribution, giving rise to a solution, and the superior cervical ganglion and the stellate patchy staining pattern. After preganglionic denervation of the ganglion were rapidly dissected and immersed in 10% superior , all ChoAcTase-positive fibers disap- formalin with 0.4% picric acid in 0.16 M sodium phosphate peared. For comparison, the distributions of enkephalin- and buffer (12) for 90 min. The specimens were transferred to a neuropeptide Y-like immunoreactivity were analyzed. In the phosphate-buffered 15% sucrose solution and stored at 4°C for stellate ganglion, the enkephalin-positive fibers had a marked 48 hr. They were cut on a cryostat (Dittes, Heidelberg) at 14 Am, regional distribution. Areas with the highest enkephalin-flber and the sections were immediately removed from the cryostat density seemed to contain low numbers of ChoAcTase-positive and soaked for 30 min in phosphate-buffered saline (PBS, pH fibers and vice versa. The findings give evidence for occurrence 7.4) containing 0.3% Triton X-100 (13). The sections were then of acetylcholine synthesis in preganglionic fibers in paraver- processed for indirect immunohistochemistry according to tebral sympathetic ganglia. Coons and collaborators (see ref. 14). In brief, the sections were incubated in a humid atmosphere at 4°C for 24 hr with a More than 30 years ago, Koelle and Friedenwald (1) devel- monoclonal ChoAcTase antibody (5) (1E6; ascites fluid diluted oped a histochemical technique for visualization of acetyl- 1:10) or with antiserum raised in rabbits against NPY (15) cholinesterase (AcChoEase; acetylcholine acetylhydrolase, (dilution 1:400) or Enk (16) (dilution 1:100). After rinsing, the EC 3.1.1.7), the enzyme responsible for breakdown of the sections were incubated at 37°C for 30 min with fluorescein neurotransmitter acetylcholine (AcCho). By use of this isothiocyanate (FITC)-conjugated sheep anti-mouse antibodies method, presumptive cholinergic neurons have been studied (Amersham International; dilution 1:10), for ChoAcTase anti- both in the periphery and in the central nervous system of serum, or FITC-conjugated swine anti-rabbit (Dakopatts, numerous species (see refs. 2 and 3). More recently, several Copenhagen; dilution 1:10) or donkey anti-rabbit antibodies groups have succeeded in purifying the AcCho-synthesizing (Amersham International; dilution 1:20), for Enk and NPY enzyme choline acetyltransferase (ChoAcTase; acetyl- antisera. AU sera used in this study, except the ChoAcTase CoA:choline O-acetyltransferase, EC 2.3.1.6) (4-7). Almost antiserum, contained 0.3% Triton X-100 (13). As a control, all studies carried out using antibodies to ChoAcTase have, sections were incubated with normal mouse serum (for however, focused on the analysis of cholinergic neurons in ChoAcTase) or with Enk and NPY antisera preabsorbed with, the central nervous system (8). In 1983, Furness et al. (9) respectively, [methionine]Enk or NPY (10 ,.g ofpeptide per ml reported using ChoAcTase antiserum to study cholinergic of diluted antiserum). cell bodies in the gastrointestinal tract of the guinea pig (9). Restaining experiments were also performed. Thus, after In the present work, we have analyzed some autonomic microscopic examination and photography, the slides were ganglia in the guinea pig with a monoclonal antibody to immersed in PBS, the coverslips were removed, and, after ChoAcTase and have carried out a comparison with rabbit rinsing, the sections were incubated with NPY antiserum and antisera raised to two neuronal peptides, methionine-enkeph- appropriate secondary antibodies as described above. alin (Enk) (10) and "neuropeptide tyrosine" (NPY) (11). We All sections were examined in a Zeiss fluorescence micro- observed dense, ChoAcTase-positive fiber networks with a scope equipped with a scanning dark-field oil-immersion patchy distribution both in the superior cervical and the condenser (17). To visualize the fluorescence, a Schott KP stellate ganglia, probably representing preganglionic cholin- 500 excitation filter and an LP 520 stop filter were used. For ergic fibers. photography, Kodak Tri-X black-and-white film was used. MATERIALS AND METHODS RESULTS Dense networks of ChoAcTase-immunoreactive nerve fibers All experiments were performed on male guinea pigs were observed around the principal ganglion cells in the (200-300 g) under Stesolid (Dumex, Copenhagen) and superior cervical ganglion (Fig. 1A). The ChoAcTase-pos- Hypnorm Vet (Leo, Helsingborg, Sweden) or, alternatively, itive fibers were fine and had a varicose appearance (Fig. 1D).

The publication costs of this article were defrayed in part by page charge Abbreviations: AcCho, acetylcholine; ChoAcTase, choline acetyl- payment. This article must therefore be hereby marked "advertisement" transferase; AcChoEase, acetylcholinesterase; NPY, neuropeptide in accordance with 18 U.S.C. §1734 solely to indicate this fact. Y ("neuropeptide tyrosine"); Enk, enkephalin.

5316 Downloaded by guest on September 26, 2021 Neurobiology: Lindh et al. Proc. Natl. Acad. Sci. USA 83 (1986) 5317 Although the ChoAcTase-immunoreactive nerve fibers were fibers (compare Fig. 2 A and B). The majority of principal found in all parts of the ganglion, they often seemed to have ganglion cells contained NPY-like-immunoreactivity (Fig. a somewhat regional distribution, forming more intensely 20. In some regions almost all cells contained this peptide, fluorescent networks around certain groups ofcells (Fig. 1A). whereas in other areas NPY cells formed small groups, After cutting of the cervical sympathetic trunk caudal to the intermingled with NPY-negative neurons (Fig. 2C). The superior cervical ganglion, no ChoAcTase-positive fibers highest numbers of NPY-containing cell bodies were seen in were seen in the ganglion (Fig. 1C). The majority ofprincipal areas where the Enk-positive fibers were most numerous ganglion cells were NPY-immunoreactive (Fig. 1B). Many (compare Fig. 2 B and C). NPY-positive cells were surrounded by ChoAcTase-immu- No ChoAcTase-immunoreactive cell bodies were ob- noreactive fibers. However, in some instances, some of the served within the ganglia studied. None of the immunoreac- most dense ChoAcTase-positive networks were found tive structures described above were seen after incubation around NPY-negative cells (compare Fig. 1 A and B). Only a with any of the control sera. few Enk-positive fibers were seen. In the stellate ganglion, a dense network of varicose ChoAcTase-immunoreactive nerve fibers was observed, DISCUSSION here also with a regional distribution (Fig. lE and Fig. 2A). It is generally accepted that AcCho is transmitter in auto- A sparse network of varicose, Enk-positive fibers could be nomic ganglia both in mammalian and nonmammalian spe- observed in certain parts of the ganglion (Fig. 2B). Compar- cies, being responsible both for fast and slow excitatory ison of consecutive sections revealed that the two fiber postsynaptic potentials (18-20). This implies that the networks had a somewhat complementary distribution. preganglionic neurons in the sympathetic lateral column Thus, although ChoAcTase-positive fibers were seen in all projecting to the para- and prevertebral ganglia are cholin- parts of the ganglion, the densest ChoAcTase networks were ergic. In fact, these neurons contain intense AcChoEase seen in areas poor in Enk-immunoreactive fibers. In contrast, activity (21) and ChoAcTase (22, 23). However, no unequiv- Enk-rich regions had low numbers of ChoAcTase-positive ocal immunohistochemical results on the cholinergic nature

I

FIG. 1. Immunofluorescence micrographs of the superior cervical ganglion (A-D) and the stellate ganglion (E) of the guinea pig after incubation with antiserum to ChoAcTase (ChAT; A, C-E) and NPY (B). A and B show the same section. (A) A very dense network of ChoAcTase-immunoreactive nerve fibers with a patchy staining pattern can be seen. (B) The majority of the ganglion cells are NPY- immunoreactive. Arrows point to two NPY-negative cells which are surrounded by an extremely dense ChoAcTase-positive fiber network. (C) No ChoAcTase-positive fibers can be seen after transection of the cervical sympathetic trunk. (D and E) The ChoAcTase-positive fibers seen in the superior cervical ganglion (D) and the stellate ganglion (E) are fine and have a varicose appearance. (Bars = 50 Aum.) Downloaded by guest on September 26, 2021 5318 Neurobiology: Lindh et al. Proc. Natl. Acad. Sci. USA 83 (1986)

FIG. 2. Immunofluorescence micrographs (montages) of three consecutive sections of the stellate ganglion of the guinea pig after incubation with antiserum to ChoAcTase (ChAT; A), Enk (B), or NPY (C). The ChoAcTase- (A) and the Enk-immunoreactive (B) fiber networks both seem to have a regional distribution within the ganglion, with a somewhat complementary distribution. Most of the ganglion cells are NPY-immunoreactive (C). In some regions the NPY-positive cells are more prevalent and almost all ganglion cells contain this peptide (to the right). Analysis of consecutive sections reveals that these regions correspond to areas where the Enk fiber network dominates (compare B and C). In areas where the ChoAcTase-positive fiber network dominates, the NPY-positive cells are intermingled with NPY-negative cells (to the left) (compare A and C). (Bars = 50 ,um.) Downloaded by guest on September 26, 2021 Neurobiology: Lindh et al. Proc. Natl. Acad. Sci. USA 83 (1986) 5319 of the preganglionic fibers in the autonomic ganglia have been cholinergic presynaptic control is exerted on noradrenergic presented (see 2). Early studies using the AcChoEase tech- ganglion cells lacking this peptide. The functional signifi- nique originally developed by Koelle and Friedenwald (1) cance of this is unclear, but it is interesting that NPY- focused on staining present in the postganglionic neurons in containing norepinephrine neurons may be specifically in- autonomic ganglia (24-28). However, dense networks of volved in control of blood flow (31). AcChoEase-positive fibers also have been described in Preliminary studies on two prevertebral ganglia, the celiac paravertebral ganglia (25, 28). This type of AcChoEase superior mesenteric ganglion and the inferior mesenteric activity disappears after preganglionic denervation, suggest- ganglion, have so far shown only few and weakly ChoAc- ing localization in preganglionic fibers (25, 28). Tase-immunoreactive fibers (unpublished results). This may The recent development in several laboratories of antibod- be related to the low intensity ofAcChoEase staining in fibers ies raised against the AcCho-synthesizing enzyme ChoAc- in these ganglia (28) and may also indicate that our technique Tase has opened up new possibilities to localize cholinergic is not sensitive enough to demonstrate ChoAcTase in all systems with immunohistochemistry. However, with the cholinergic systems. exception of neurons in the gastrointestinal wall ofthe guinea In conclusion, our results provide direct histochemical pig (9), studies so far have dealt with the localization of evidence that one of the classical cholinergic systems, the ChoAcTase in the central nervous system (8). In the present preganglionic fibers in autonomic ganglia, contains ChoAc- paper, we have described ChoAcTase-positive fiber net- Tase and thus has the capacity to synthesize AcCho. works with varying density in two paravertebral sympathetic ganglia, the superior cervical ganglion and the stellate gan- The skillful technical assistance of Ms. W. Hiort and Ms. S. glion of the guinea pig. The distribution of these ChoAcTase- Nilsson and the excellent secretarial help of Ms. Elisabet Bjorklund positive fibers is patchy, but principally the entire ganglion are gratefully acknowledged. We thank Professor L.-G. Elfvin for this . After support and valuable discussion. This study was supported by grants receives type of transection of the sympa- from the Karolinska Institute and the Swedish Medical Research thetic trunk caudal to the superior cervical ganglion, all these Council (04X-2887, 12X-5189), Ruth and Richard Julins Stiftelse, fibers disappeared, strongly underlining their preganglionic Alice and Knut Wallenbergs Stiftelse, and Magnus Bergvalls nature and in agreement with the early AcChoEase studies Stiftelse. (25, 28). The low numbers of ChoAcTase-positive fibers in some parts of the two ganglia suggest that not all pregan- 1. Koelle, G. B. & Friedenwald, J. S. (1949) Proc. Soc. Exp. glionic fibers have been "labeled." This may indicate that Biol. Med. 70, 617-622. some preganglionic neurons contain another transmitter or, 2. Silver, A. (1974) The Biology of Cholinesterases (American perhaps more likely, that these fibers have ChoAcTase levels Elsevier, New York). too low to be detected with the present technique. Although 3. Butcher, L. L. & Wolf, N. J. (1984) in Handbook ofChemical Neuroanatomy, eds. Bjorklund, A., Hokfelt, T. & Kuhar, strongly AcChoEase-positive ganglion cell bodies were ob- N. J. (Elsevier, Amsterdam), Vol. 3, pp. 1-45. served in early studies (24-28), no ChoAcTase-positive cell 4. Kimura, H., McGeer, P. L., Peng, J. H. & McGeer, E. D. bodies were seen under the present experimental conditons (1981) J. Comp. Neurol. 200, 151-200. (no colchicine treatment), suggesting low levels ofenzyme in 5. Crawford, G. D., Correa, L. & Salvaterra, P. M. (1982) Proc. the somata. Natl. Acad. Sci. USA 79, 7031-7035. Autonomic ganglia contain varying amounts of Enk-immu- 6. Eckenstein, F. & Thoenen, H. (1982) EMBO J. 1, 363-368. noreactive fibers (16), probably representing preganglionic 7. Levey, A. I., Armstrong, D. M., Atwich, S. F., Terry, R. D. fibers (29). Therefore the distribution of the ChoAcTase- & Wainer, B. H. (1983) J. Neurosci. 3, 1-9. positive fibers was compared with the Enk-immunoreactive 8. Wainer, B. H., Levey, A. I., Mufson, E. J. & Mesulam, ones. M. M. (1984) Neurochem. Int. 6, 163-182. In the superior cervical ganglion, very few Enk-positive 9. Furness, J. B., Costa, M. & Eckenstein, F. (1983) Neurosci. fibers were observed, in agreement with earlier studies (16). Lett. 40, 105-109. In the stellate ganglion, the Enk-positive fibers formed a 10. Hughes, J., Smith, T. W., Kosterlitz, H. W., Fothergill, moderately dense network with a marked regional distribu- L. H., Morgan, B. A. & Morris, H. R. (1975) Nature (London) tion. Areas with the highest Enk fiber density seemed to 258, 577-579. contain low numbers of ChoAcTase-positive fibers, suggest- 11. Tatemoto, K., Carlquist, M. & Mutt, V. (1982) Nature (Lon- ing a complementary distribution of Enk- and ChoAcTase- don) 296, 659-660. positive fibers. Although the comparatively few Enk-positive 12. Zamboni, L. & de Martino, C. (1967) J. Cell Biol. 35, 148A. fibers also may contain ChoAcTase-like immunoreactivity, it 13. Hartman, B. K., Zide, D. & Udenfriend, S. (1972) Proc. Natl. Acad. Sci. USA 69, 2722-2726. is apparent that the vast majority of ChoAcTase-positive 14. Coons, A. H. (1958) in General Cytochemical Methods, ed. fibers do not contain Enk-like immunoreactivity as defined Danielli, J. F. (Academic, New York), pp. 399-422. by our conditions of analysis. 15. Lundberg, J. M., Terenius, L., Hbkfelt, T. & Tatemoto, K. These findings are ofinterest in view ofearlier speculations (1984) J. Neurosci. 4, 2376-2386. that AcCho and Enk-like peptides may be localized in the 16. Schultzberg, M., Hckfelt, T., Terenius, L., Elfvin, L.-G., same preganglionic neurons (16, 30) and the actual demon- Lundberg, J. M., Brandt, J., Elde, R. P. & Goldstein, M. stration of colocalization of ChoAcTase- and Enk-like (1979) Neuroscience 4, 249-270. immunoreactivity in neurons in the rat sympathetic lateral 17. MArtensson, R. & Bjorklund, A. (1984) in Handbook ofChem- column (23). Further analysis at the spinal cord level will ical Neuroanatomy, eds. Bjorklund, A. & Hokfelt, T. therefore be to decide whether the (Elsevier, Amsterdam), Vol. 2, pp. 380-386. necessary finally 18. Skok, V. I. (1983) in Autonomic Ganglia, ed. Elfvin, L.-G. preganglionic sympathetic neurons innervating the superior (Wiley, Chichester, U.K.), pp. 265-279. cervical ganglion and the stellate ganglion in the guinea pig 19. Kobayashi, H. & Tosaka, T. (1983) in Autonomic Ganglia, ed. contain both compounds or whether AcCho and the Enk-like Elfvin, L.-G. (Wiley, Chichester, U.K.), pp. 281-307. peptide are present in separate neurons. 20. Weight, F. F. (1983) in Autonomic Ganglia, ed. Elfvin, L.-G. A large proportion of the noradrenergic ganglion cells in (Wiley, Chichester, U.K.), pp. 309-344. the sympathetic ganglia contain NPY (31, 32), a 36 amino acid 21. Navaratnam, V. & Lewis, P. R. (1970) Brain Res. 18, 411-425. peptide (11). These cell bodies also exhibit a somewhat 22. Houser, C. R., Crawford, G. D., Barber, R. P., Salvaterra, patchy distribution within the ganglion as shown in this P. M. & Vaughn, J. E. (1983) Brain Res. 266, 97-119. study. Restaining experiments indicate that the highest den- 23. Kondo, H., Kuramoto, H., Wainer, B. H. & Yanaihara, N. sity of ChoAcTase-positive fibers often is found around (1985) Brain Res. 335, 309-314. NPY-negative cell bodies, suggesting that the strongest 24. Koelle, G. B. (1951) J. Pharmacol. Exp. Ther. 103, 153-171. Downloaded by guest on September 26, 2021 5320 Neurobiology: Lindh et al. Proc. Nati. Acad. Sci. USA 83 (1986)

25. Koelle, G. B. (1955) J. Pharmacol. Exp. Ther. 114, 167-184. 30. Schultzberg, M., Lundberg, J. M., Hokfelt, T., Terenius, L., 26. Holmstedt, B. & Sjoqvist, F. (1957) Acta Physiol. Scand., Brandt, J., Elde, R. P. & Goldstein, M. (1978) Neuroscience 3, Suppl. 145 42, 72-73. 1169-1186. 27. Holmstedt, B. & Sjoqvist, F. (1959) Acta Physiol. Scand. 47, 31. Lundberg, J. M., Terenius, L., Hokfelt, T., Martling, C. R., 284-2%. Tatemoto, K., Mutt, V., Polak, J., Bloom, S. & Goldstein, M. 28. Fredricsson, B. & Sjoqvist, F. (1962) Acta Morphol. Neerl.- (1982) Acta Physiol. Scand. 116, 477-480. Scand. 5, 140-166. 32. Lindh, B., Hokfelt, T., Elfvin, L.-G., Terenius, L., 29. Dalsgaard, C.-J., Hokfelt, T., Elfvin, L.-G. & Terenius, L. Fahrenkrug, J., Elde, R. & Goldstein, M. (1986) J. Neurosci., (1982) Neuroscience 7, 2039-2050. in press. Downloaded by guest on September 26, 2021