Proc. Natl. Acad. Sci. USAVol. 92, pp. 11819-11823, December 1995Neurobiology

Localization of choline acetyltransferase in rat peripheralsympathetic neurons and its coexistence with nitric oxidesynthase and neuropeptidesMIGUEL A. MORALES*t, KRISTINA HOLMBERG*, ZHI-QING XU*, COSTATINO COZZARIt, BOYD K. HARTMANt,PIERS EMSON§, MENEK GOLDSTEIN1, LARS-GOSTA ELFVIN*, AND ToMAs HOKFELT*II*Department of Neuroscience, Karolinska Institute, S-171 77 Stockholm, Sweden; tlnstituto de Investigaciones Biomedicas, Universidad Nacional Autonoma deMexico, 04510, Mexico D.F.; tDepartment of Psychiatry, University of Minnesota, Minneapolis, MN 55455; §Department of Neurobiology, The BabrahamInstitute, Cambridge, CB2 4AT, United Kingdom; and "Department of Psychiatry, New York University Medical Center, New York, NY 10016

Contributed by Tomas Hokfelt, August 31, 1995

ABSTRACT Indirect immunofluorescence methods using amouse monoclonal antibody raised to rat choline acetyltrans-ferase (ChAT) revealed dense networks of ChAT-immunoreac-tive fibers in the superior cervical ganglion, the stellate ganglion,and the celiac superior mesenteric ganglion ofthe rat. Numerousand single ChAT-immunoreactive cell bodies were observed inthe stellate and superior cervical ganglia, respectively. Themajority of ChAT-immunoreactive fibers in the stellate andsuperior cervical ganglia were nitric oxide synthase (NOS)positive. Some ChAT-immunoreactive fibers contained en-kephalin-like immunoreactivity. Virtually all ChAT-positivecell bodies in the stellate ganglion were vasoactive intestinalpolypeptide (VIP)-positive, and some were calcitonin gene-related peptide (CGRP)-positive. After transection of thecervical sympathetic trunk almost all ChAT- and NOS-positive' fibers and most enkephalin- and CGRP-positivefibers disappeared in the superior cervical ganglion. Theresults suggest that most preganglionic fibers are cholin-ergic and that the majority of these in addition can releasenitric oxide, some enkephalin, and a few CGRP. Acetylcho-line, VIP, and CGRP are coexisting messenger molecules insome postganglionic sympathetic neurons.

Acetylcholine (ACh) and norepinephrine are the classic trans-mitters in parasympathetic and preganglionic sympathetic neu-rons and in postganglionic sympathetic neurons, respectively. Thenoradrenergic neurons could early on be demonstrated with aspecific histochemical fluorescence method developed by Falck etal. (1). Koelle and Friedenwald (2) described a histochemicaltechnique for detecting ACh esterase, the enzyme responsible forbreakdown ofACh, but this method is not specific for cholinergicneurons (3). Therefore, the purification of the ACh-synthesizingenzyme, choline acetyltransferase (ChAT), by several groupsoffered considerable promise for tracing cholinergic systems inbrain and periphery with immunohistochemistry, and severalmapping studies on central cholinergic systems have been pub-lished (see ref. 4). In the peripheral nervous system Costa andcoworkers (5, 6) visualized cholinergic cell bodies and fibers in theguinea pig gastrointestinal tract, Lindh et al. (7) found denseChAT-positive networks in the superior cervical ganglion (SCG)and stellate ganglion (SG) of the guinea pig Kasa et al. (34) andMatthews et al. (35) found ChAT-positive fibers in mouse and ratSCG, respectively, and Suzuki et al. (8) demonstrated ChAT inpostganglionic rat parasympathetic neurons. However, the situ-ation with regard to identification of cholinergic neurons in theperiphery is less satisfactory than that for the central cholinergicsystems.Autonomic ganglia contain in both the pre- and the post-

ganglionic neurons a number of peptides showing distinct

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

distribution patterns and often coexisting with classical trans-mitters (9). Moreover, nitric oxide (NO) has during recentyears emerged as a messenger candidate in both the centraland peripheral nervous systems (10, 36-40), and its synthe-sizing enzyme, NO synthase (NOS), is present in preganglionicfibers in both prevertebral and paravertebral ganglia (11-13,41, 42).We have used a mouse monoclonal antibody raised against

purified rat brain ChAT (14) to study sympathetic ganglia of therat. The results demonstrate that both cholinergic cell bodies andfibers can be distinctly visualized in pre- and some postganglionicsympathetic neurons. Moreover, several coexistence situations ofChAT with NO and peptides are defined.

MATERIALS AND METHODSMale Sprague-Dawley rats (200-250 g) were used. In threerats the SCG was denervated by cutting the cervical sym-pathetic trunk 15 mm caudal to the ganglion under sodiumpentobarbital anesthesia (Mebumal; 40 mg/kg, i.p.). Five to7 days later the animals that had undergone this surgery aswell as the untreated rats were anesthetized (as above) andperfused with 2% paraformaldehyde in 0.1 M phosphatebuffer (pH 7.0). The SCG, the SG, and the celiac superiormesenteric ganglion (CSMG) were dissected out, postfixed,and cut at 14-,um thickness in a cryostat (Microm, Heidel-berg). The sections were processed for the indirect immu-nohistochemical method-that is, incubated in a humidatmosphere at 4°C overnight with a mouse monoclonalantibody raised against rat brain ChAT (diluted 1:1,000) (14)or with one of several rabbit antisera raised against enkepha-lin (1:400) (15), vasoactive intestinal polypeptide (VIP)(1:400; Cambridge Research Biochemicals), calcitonin gene-related peptide (CGRP) (1:400; Peninsula Laboratories),the catecholamine-synthesizing enzyme tyrosine hydroxylase(TH) (1:1,000) (16), or goat antiserum to NOS (1:2,000; P.E.,unpublished work). The ChAT antibody is characterized bya high affinity for the enzyme (3 x 1011 M-1; C.C. and K.H.,unpublished work). All antisera were diluted in phosphate-buffered saline containing 0.3% Triton X-100 (17). Afterrinsing, the sections were incubated with fluorescein iso-thiocyanate (FITC)-conjugated sheep (1:20; Amersham) ordonkey (1:40; Jackson ImmunoResearch) anti-mouse anti-bodies for the ChAT antibody and with FITC-conjugatedgoat anti-rabbit antibodies (1:80; Boehringer-Mannheim) forthe other antisera. The goat NOS antiserum was followed by

Abbreviations: ACh, acetylcholine; ChAT, choline acetyltransferase;CGRP, calcitonin gene-related peptide; CSMG, celiac superior mes-enteric ganglion; NOS, nitric oxide synthase; SCG, superior cervicalganglion; SG, stellate ganglion; TH, tyrosine hydroxylase; VIP, vaso-active intestinal polypeptide.I'To whom reprint requests should be addressed.

11819

Dow

nloa

ded

by g

uest

on

June

3, 2

021

11820 Neurobiology: Morales et al.

Lissamine rhodamine sulfonyl chloride (LRSC)-conjugateddonkey anti-goat antibodies (1:40; Jackson ImmunoResearch)or FITC-conjugated donkey anti-goat antibodies (1:50; NordicLausanne, Switzerland). For colocalization studies, sectionsprocessed for visualization of ChAT or NOS were subse-quently incubated with one of the above-mentioned antisera

followed by LRSC-conjugated goat anti-rabbit antibodies(1:40) or LRSC-conjugated donkey anti-rabbit antibodies(1:40), or aminomethylcoumarin (AMCA)-conjugated donkeyanti-goat antibodies (all from Jackson ImmunoResearch).For control purposes all peptide antisera were preabsorbedovernight at 4°C with the respective peptide at 1 ,uM

FIG. 1. Immunofluorescence micrographs of the SG (a and g-m), CSMG (b), and SCG (c-f) after incubation with antiserum to ChAT (a-c,e, g, j, and 1), NOS (d andf), TH (h), VIP (i and k), or CGRP (m). Thus, cld, e/f g/h, j/k, and l/m are double-stained sections. ChAT-immunoreactivecell bodies (a, g, j, and 1) are found in SG, and dense fiber networks are seen in SG, SCG, and CSMG (a-c and e). In SG and SCG there is mostlya close overlap between ChAT- and NOS-positive fiber networks (c and d), with an almost complete coexistence (e and f), although in restrictedareas NOS-positive fibers are absent or express low NOS-like immunoreactivity (asterisk). Note a single ChAT-immunoreactive cell body in SCG(c). SG contains numerous VIP neurons (i), and ChAT (g) coexists with VIP (k) and CGRP (m) but not TH (h). (Bars = 50 ,um for the groupsa/b, c/d, elf, and g-m.)

_=.

Proc. Natl. Acad. Sci. USA 92 (1995)

Dow

nloa

ded

by g

uest

on

June

3, 2

021

Proc. Natl. Acad. Sci. USA 92 (1995) 11821

(Peninsula Laboratories). Normal mouse, rabbit, and goatserum were used as control for ChAT, TH, and NOSstaining, respectively. The sections were examined in aNikon Microphot FX epifluorescence microscope equippedwith filters for FITC, LRSC, and AMCA fluorescence.

RESULTS

In control SG (Fig. la), SCG (Fig. 1 c and e; Fig. 2 a, c, e, andi), and CSMG (Fig. lb), dense networks of ChAT-immunoreactive fibers were observed often surrounding the

FIG. 2. Immunofluorescence micrographs of SCG (a-h and k-n) and SG (i and j) after incubation with antiserum to ChAT (a, c, e, i, k, and1), enkephalin (b, d, f h, and n), NOS (g and m), or CGRP (1). Thus, a/b, c/d, e/f, g/h, ilj, and l/m are double-stained sections. ChAT and enkephalinoverlap partially (a and b), but only in a few cases can coexistence in fibers/varicosities be convincingly demonstrated (arrowheads in c-f). NOS(thin arrows ing) and enkephalin (thick arrow in h) do not overlap (two neurons are labeled 1 and 2 for orientation). ChAT (i) and CGRP (/) coexistin fibers around a cell body. After axotomy, only single ChAT (k and 1) and NOS (m) fibers remain, and they contain both ChAT- and NOS-likeimmunoreacti"ities (arrowheads in k-m). Note a NOS-positive and ChAT-negative cell body (arrow in m). Sympathetic-trunk transection causesdisappearance of enkephalin fibers ipsilaterally in SCG (asterisk in n). Control ganglion is seen to the left (n). (Bar = 50 ,um for the groups a/b,c-J, k/n, and l/m.)

Neurobiology: Morales et al.

Dow

nloa

ded

by g

uest

on

June

3, 2

021

11822 Neurobiology: Morales et al.

ganglion cell bodies in a basket-like manner as well as formingtracts with multiple parallel fibers. In addition there were inSG numerous ChAT-immunoreactive cell bodies (Fig. 1 a, g,j, and 1), whereas only single cells showed this immunoreac-tivity in SCG (Fig. lc). The ChAT-immunoreactive cell bodieswere TH-negative (Fig. 1 g and h). There were many VIP-immunoreactive cell bodies in the SG (Fig. ii), and virtually all ofthem were ChAT-immunoreactive (Fig. 1 j and k). Smallernumbers of ChAT-immunoreactive cells contained CGRP-likeimmunoreactivity (Fig. 1 1 and m). Varying numbers of enkepha-lin-immunoreactive cell bodies were observed in SG and espe-cially SCG, and they were ChAT-negative and TH-positive.A very dense network of NOS-immunoreactive fibers was

observed in all three ganglia (Fig. 1 d and f; Fig. 2g) with adistribution overlapping with the ChAT-immunoreactive fi-bers (Fig. 1 c and e), and these fibers were mostly identical (Fig.1 e and f). However, in restricted areas of the SCG, ChAT-immunoreactive fibers were NOS-negative or at most veryweakly NOS-immunoreactive (Fig. 1 c and d). There wereenkephalin-positive (Fig. 2 b, d, fJ h, and n) and CGRP-positive(Fig. 2j) fibers in SCG (Fig. 2 a-h), SG (Fig. 2j), and CSMG.Enkephalin-immunoreactive fibers were more numerous inSCG than in SG. Numerous enkephalin-positive fibers formedbaskets in the SCG (Fig. 2 b, d, and f), and at least some ofthese fibers were ChAT-immunoreactive (Fig. 2 c, d, e, andf).In the SG, CGRP-immunoreactive fibers formed baskets (Fig.2g), and they were frequently ChAT-immunoreactive (Fig. 2i),whereas other CGRP-positive fibers (single or in bundles)lacked ChAT-like immunoreactivity. No coexistence betweenNOS (Fig. 2g) and enkephalin (Fig. 2h) immunoreactivitiescould be established. After sectioning of the sympathetic trunkcaudal to the SCG, virtually all ChAT (Fig. 2 k and 1), NOS(Fig. 2m), 'and most enkephalin (Fig. 2n) fibers disappeared inthe SCG. The remaining ChAT-positive fibers were NOS-positive but enkephalin- and CGRP-negative. None of theimmunoreactivities described above were observed after in-cubation with control serum.

DISCUSSIONUsing an antibody raised against purified rat brain ChAT (14),we here demonstrate the presence of this enzyme in thesympathetic autonomic nervous system, in both paravertebraland prevertebral ganglia of the rat. This antibody has previ-ously been used to demonstrate cortical cholinergic systems inrat brain with an excellent sensitivity and resolution (18). Thepresent findings show dense networks of fibers containingChAT-like immunoreactivity in the SCG, SG, and CSMG.Surgical transection of the sympathetic trunk caused an almostcomplete disappearance of these fibers in the SCG, confirmingtheir preganglionic nature. The source(s) of the dense fibernetworks in the rat SG and CSMG remains to be analyzed.Moreover a population of principal ganglion neurons in theSG, and single neurons in SCG, expressed ChAT but not TH,strongly supporting their cholinergic nature. The large numberof ChAT-positive fibers in all parts of the SCG and SG suggeststhat perhaps all preganglionic fibers in fact are cholinergic, inagreement with numerous physiological studies showing thatACh is the principal transmitter in sympathetic ganglia (19).The present findings confirm the existence of numerous

NOS-positive fibers in paravertebral ganglia (11) and theirdisappearance after transection of the sympathetic trunk (13),supporting the view of a preganglionic origin in cell bodies inthe sympathetic lateral column (13, 20-22). Here we show thatvirtually all of these NOS fibers in SG and SCG are cholinergicand thus presumably release both ACh and NO. However,there are patches of ChAT-positive fibers which seem to lack,or at least contain only very low levels of, NOS-like immuno-reactivity. It is thus possible that a population of the pregan-glionic fibers release ACh but not NO.

A small population of sympathetic neurons express AChesterase (23, 24). In the cat these neurons contain VIP andproject to salivary glands, where VIP and ACh interact in thecontrol of secretion and blood flow (25, 26). The presentfindings clearly demonstrate that virtually all VIP-immunore-active neurons in the rat SG are ChAT-immunoreactive andvice versa, giving final proof for the cholinergic nature of thesesympathetic neurons. A few VIP-positive neurons were alsoobserved in SCG (27, 28), and they were also ChAT-positive.Small numbers of CGRP-positive neurons in rat paravertebralganglia contain VIP and have been suggested to be cholinergic(29). In agreement, the present results show that CGRP- andChAT-like immunoreactivities coexist in these neurons.Nerve fibers containing various peptides have been identi-

fied in pre- and paravertebral ganglia (see ref. 9). They areparticularly abundant in prevertebral ganglia and originate inspinal cord, sensory ganglia, and the gastrointestinal tract (seerefs. 9 and 30). Here we focus on enkephalin and CGRP fibersin paravertebral ganglia. At least part of the CGRP fibersseems to be preganglionic (31). Since enkephalin coexists withChAT in cell bodies in the sympathetic lateral column (32),some enkephalin fibers in the SCG may represent pregangli-onic cholinergic fibers. In agreement with these views, manyCGRP and enkephalin fibers disappear after immunologicallyinduced sympathectomy (33). The enkephalin fibers in theSCG exhibit a patchy distribution (15) overlapping with, butbeing much more sparse than, the ChAT-immunoreactiveones. Under high-power objectives it was found that in normalganglia only some enkephalin-immunoreactive terminals in theSCG were clearly ChAT-immunoreactive but that CGRP-immunoreactive baskets in SG often were ChAT-positive.However, to monitor coexistence of ChAT and NOS versusenkephalin and CGRP in single nerves/varicosities may bedifficult, partly due to their different intracellular localization:ChAT and NOS are cytoplasmic enzymes and "fill out" theentire axons and nerve endings, whereas the peptides arestored in vesicles and are mainly present in the nerve endings.We may therefore underestimate the degree of coexistence of,for example, ChAT and enkephalin. Finally, enkephalin-likeimmunoreactivity was not observed in NOS-immunoreactivefibers. Thus, it is possible that ChAT/NOS fibers and ChAT/enkephalin fibers represent two subpopulations of pregangli-onic fibers. Whether or not there is a population of enkephalin-positive, ChAT-negative, and NOS-negative fibers with anunknown principal transmitter remains to be investigated.Clearly some of the enkephalin fibers in the SCG could arisefrom enkephalin-containing noradrenergic neurons in thisganglion (15).

Note Added in Proof. After submission of this paper, we have becomeaware of a publication by Sann et al. (43) which provides a detailedstudy on ChAT-immunoreactive neurons and peptide coexistence inthe inferior mesenteric ganglion of several species.

This work was supported by the Swedish Medical Research Council(2887,5189) and DGAPA-IN204093; PEC del IBM, UNAM (Mexico)to M.A.M.

1. Falck, B., Hillarp, N.-A., Thieme, G. & Torp, A. (1962) J.Histochem. Cytochem. 10, 348-354.

2. Koelle, G. B. & Friedenwald, J. S. (1949) Proc. Exp. Biol. Med.70,617-622.

3. Silver, A. (1974) The Biology of Cholinesterases (Elsevier, NewYork).

4. Wainer, B. H., Levey, A. I., Mufson, E. J. & Mesulam, M. M.(1984) Neurochem. Int. 6, 163-182.

5. Furness, J. B., Costa, M., Emson, P. C., Hakanson, R.,Moghimzadeh, E., Sundler, F., Taylor, I. L. & Chance, R. E.(1983) Cell Tissue Res. 234, 71-92.

6. Steele, P. A., Brookes, S. J. H. & Costa, M. (1991) Neuroscience45, 227-239.

Proc. Natl. Acad. Sci. USA 92 (1995)

Dow

nloa

ded

by g

uest

on

June

3, 2

021

Proc. Natl. Acad. Sci. USA 92 (1995) 11823

7. Lindh, B., Staines, W., Hokfelt, T., Terenius, L. & Salvaterra,P. M. (1986) Proc. Natl. Acad. Sci. USA 83, 5316-5320.

8. Suzuki, N., Hardbo, J. E. & Owman, C. (1990) J. Cereb. BloodFlow Metab. 10, 399-408.

9. Elfvin, L.-G., Lindh, B. & Hokfelt, T. (1993)Annu. Rev. Neurosci.16, 471-507.

10. Garthwaite, J. & Boulton, C. L. (1995) Annu. Rev. Physiol. 57,683-706.

11. Anderson, C. R., Edwards, S. L., Furness, J. B., Bredt, D. S. &Snyder, S. H. (1993) Brain Res. 614, 78-85.

12. Anderson, C. R., Furness, J. B., Woodman, H. L., Edwards, S. L.,Crack, P. J. & Smith, A. I. (1995) J. Autonom. Nerv. Syst. 52,107-116.

13. Ceccatelli, S., Lundberg, J. M., Zhang, X., Aman, K. & Hokfelt,T. (1994) Brain Res. 656, 381-395.

14. Cozzari, C., Howard, J. & Hartman, B. (1990) Soc. Neurosci.Abstr. 16, 200.

15. Schultzberg, M., H6kfelt, T., Terenius, L., Elfvin, L.-G., Lund-berg, J. M., Brandt, J., Elde, R. P. & Goldstein, M. (1979)Neuroscience 4, 249-270.

16. Markey, K. A., Kondo, S., Shenkman, L. & Goldstein, M. (1980)Mol. Pharmacol. 17, 79-85.

17. Hartman, B. K., Zide, D. & Udenfriend, S. (1972) Proc. Natl.Acad. Sci. USA 69, 2722-2726.

18. Umbriaco, D., Watkins, K. C., Descarries, L., Cozzari, C. &Hartman, B. K. (1994) J. Comp. Neurol. 348, 351-373.

19. Skok, V. I. (1993) inAutonomic Ganglia, ed. Elfvin, L.-G. (Wiley,Chichester), 265-279.

20. Anderson, C. R. (1992) Neurosci. Lett. 139, 280-284.21. Blottner, D. & Baumgarten, H.-G. (1992) J. Comp. Neurol. 316,

45-55.22. Valtschanoff, J. G., Weinberg, R. J. & Rustioni, A. (1992) J.

Comp. Neurol. 321, 209-222.23. Koelle, G. B. (1995) J. Pharmacol. Exp. Ther. 103, 153-171.

24.

25.

26.

27.

28.

29.30.

31.

32.

33.

34.35.

36.37.38.39.40.

41.

42.

43.

Holmstedt, B. & Sjoqvist, F. (1959) Acta Physiol. Scand. 47,284-296.Lundberg, J. M., Hokfelt, T., Schultzberg, M., Uvnas-Wallensten,K., Kohler, C. & Said, S. I. (1979) Neuroscience 4, 1539-1559.Lundberg, J. M., Anggard, A., Fahrenkrug, J., Hokfelt, T. &Mutt, V. (1980) Proc. Natl. Acad. Sci. USA 77, 1651-1655.Hokfelt, T., Elfvin, L.-G., Schultzberg, M., Fuxe, K., Goldstein,M. & Park, D. (1977) Neuroscience 2, 885-896.Sasek, C. A. & Zigmond, R. E. (1989) J. Comp. Neurol. 280,522-532.Landis, S. C. & Fredieu, J. R. (1986) Brain Res. 377, 177-181.Costa, M. & Furness, J. B. (1987) The Enteric Nervous System(Churchill Livingstone, Edinburgh).Yamamoto, K., Senba, E., Matsunaga, T. & Tohyama, M. (1989)Brain Res. 487, 158-164.Kondo, H., Kuramoto, H., Wainer, B. H. & Yanaihara, N. (1985)Brain Res. 335, 309-314.Dagerlind, A., Zhang, X., Brimijoin, S., Lindh, B. & Hokfelt, T.(1994) NeuroReport 5, 909-921.Kasa, P., Dobo, E. & Wolff, J. R. (1991) Tissue Res. 265, 151-158.Matthews, M. R., Ribeiro-da-Silva, A. & Cuello, C. (1991) Neu-rocytochemical Methods, eds. Calas, A. & Eugene, D. (Springer,Heidelberg), 279-282.Furchgott, R. F. (1990) Acta Physiol. Scand. 139, 257-270.Garthwaite, J. (1991) Trends Neurosci. 14, 60-67.Bredt, D. S. & Snyder, H. (1992) Neuron 8, 3-11.Moncada, S. (1992) Acta Physiol. Scand. 145, 201-227.Vincent, S. R. & Hope, B. T. (1992) Trends Neurosci. 15, 108-113.Dun, N. J., Dun, S. L., Fostermann, U. & Tseng, L. F. (1992)Neurosci. Lett. 147, 217-220.Dun, N. J., Dun, S. L., Wu, S. Y. & Fostermann, U. (1993)Neurosci. Lett. 158, 51-54.Sann, H., McCarthy, P. W., Schemann, M., Jurzak, M., Poethke,R. & Pierau, F.-K. (1995) J. Autonom. Nerv. Syst. 54, 195-205.

Neurobiology: Morales et al.

Dow

nloa

ded

by g

uest

on

June

3, 2

021