Immunohistochemical localization of choline acetyltransferase using a monoclonal antibody: A radioautographic method

h’euroscience Vol. IO, No. 3, pp. 907-922, I983 Printed in Great Britain

0306-4522/83 $3.00 + 0.00 Pergamon Press Ltd IBRO

IMMUNOHISTOCHEMICAL LOCALIZATION OF CHOLINE ACETYLTRANSFERASE USING A MONOCLONAL ANTIBODY: A RADIOAUTOGRAPHIC METHOD M. E. Ross, D. H. PARK, G. TEITELMAN, and T. H. JOH*

V.

M.

PICKEL, D.

J. REIS

Laboratory of Neurobiology, Cornell University Medical College, 1300 York Avenue, New York, NY 10021, U.S.A. Abstrati-Monoclonal antibodies to rat striatal choline acetyltransferase were produced by fusion of sensitized mouse lymphocytes with murine plasrnacytoma (NSl) cells. Two stable anti-choline acetyltransferase lines were established by limiting dilution cloning. Specificity of antibody was established by the following criteria: (1) on an enzyme linked immunosorbant assay, antibodies reacted against choline acetyltransferase which was highly purified; (2) by immunoprecipitation, monoclonal antibody bound to its antigen and precipitated choline acetyltransferase activity from solution, when used in conjunction with

rabbit antimouse IgG; and (3) monoclonal antibody was shown to specificallylocalize cholinergic neurons. The monoclonai antibody to choline acetyltransferase was radiolabeled in culture by incubating hybridomas in medium containing 3H-labeled amino acids. This 3H-labeled antibody was used for radioautography on cryostat sections of rat peripheral and central nervous systems. In a sampling of areas, highly specific labeling of cholinergic structures was afforded at both light and electron microscopic levels. Double labeling of tyrosine hydroxylase, a catecholaminergic marker, and choline acetyltransferase was carried out by reacting sections first with the ‘H-labeled antibody to choline acetyltransferase and then with rabbit antibody to tyrosine hydroxylase. The choline acetyltransferase label was radioautographically processed and tyrosine hydroxylase was visualized by the peroxidase-antiperoxidase method. The combined techniques of peroxidase and radioautographic histochemistry provide permanent electron dense labels which can be examined simultaneously within a single histologic section.

Choline acetyltransferase (ChAT) (acetyl-CoA: choline-0-acetyltransferase, EC 2.3.1.6) is the enzyme catalyzing the biosynthesis of acetylcholine. As such, it is a specific marker protein of cholinergic neurons. Over the past several years numerous laboratories have endeavored to purify ChAT for the production of specific antisera for immunofew have ob_ histochemistry. 2.4.11.27.28.34,35.36,39 However, tained homogeneous enzyme and considerable argument has arisen as to the specific activity that must be attained to assure antigenic purity,3,33.38specificity of the conventional antisera, and hence, specificity of staining. The development of cell fusion methods for the production of monoclonal antibodieszO provides a technique allowing for production of monospecific antibodies to defined antigens regardless of the purity of the antigens used for immunization.24*25 Thus, it is possible to select for specificity of antibodies rather than antigens, a methodological advantage in production of antisera to proteins which are difficult to isolate in a highly purified state.32 An additional attraction of the hybridoma technique is the ability to generate de novo in culture isotopically-labeled antibodies of remarkably high specific activity which can *To whom requests for reprints should be sent. ChAT. choline acetvitransferase: DAB. 3-3’-diaminobenzidine; ELISA, e&me linked immu: nosorbant assay; PAP, peroxidase anti-horseradish peroxidase; PAGE, polyacrylamide gel electrophoresis; TH, tyrosine hydroxylase.

Abbreviations:

907

be used for histologic radioautography at both light and electron microscopic levels. Thus it is possible to combine the radiolabeled antibody with a second immunohistochemical method, such as the peroxidase antiperoxidase (PAP) method, to localize two antigens in a single tissue section. In the present study, we have sought to develop in this manner a monoclonal antibody to ChAT, to examine its utility for immunocytochemistry, both by light and electron microscopy and finally to establish whether the radiolabeled antibody can be used in conjunction with a conventionally produced antibody to a catecholaminergic neuronal marker, tyrosine hydroxylase (TH),16 to label two neuro-transmitter specific macromolecules in the same section. EXPERIMENTAL PROCEDURES PuriJication of choline acetyltransferase

Choline acetyltransferase activity was assayed for both purification and immunoprecipitation experiments by our modification of the method of Schrier and Schuster.“’ To 100 ~1 of enzyme was added 100 ~1 of a reaction mixture containing 0.05% bovine serum albumin 0.4OmM acetyl co-enzyme A including 0.04 PCi l-‘4C-acetyl co-enzyme A, 15 mM choline bromide, 4OOmM NaCI, 0.2mM eserine sulfate, 0.2mM neutralized EDTA and 400mM Kphosphate, pH 7.5. Samples were incubated at 37” for 10 min and the reaction was stopped by immersing reaction tubes in an ice water bath. Samples were then passed over Dowex l- x 8 columns (0.5 x 3.0 cm) and each tube was washed with 1.2ml of water which was added to the appropriate column. Column effluents were collected and scintillation counted.

908

M. Ross et al Table Purificatmn step Homogenate Suoernatant

Sepharose 4B Hydroxylapatite

1 Purification

of cholme

acetyltransferase

from rat strlata*

Total protem

Total actwlty

Specdic actlvlty

Yield

,*,:ogo 3130 1040 115

ulutst 38,200 27.500 24,800 17,800

unm,‘mg 2 1: 8.8 23 8 155.0

106 72 65 47

42 II 3 73 8

11.900

643.0

31

30h.O

17 10

1160.0 14.000.0

Purification -fold

(1) Hydroxylapatite (11)

Sephadex G 150 CM-Sepbadex

18.5

2.59 0.128

6310 3780

2440.0 29,500.O

‘The purification was carried out by using 1OOg of rat stciata. tA unit of enzyme activity is expressed as 1 urn01 of acetylcholine formed per min at 37°C. fNote that the specific activity of striatal homogenate is approximately 18 times that of pig brain homogenate.

Choline acetyltransferase was purified from rat striatum by methods described in detail elsewhere*’ and summarized in Table 1. Rats were stunned, decapitated and the striata removed by dissection. Striatia from 500 rat brains were homogenized in 10 volumes of 50 mM potassium phosphate buffer, pH 7.5, with a Waring blender. The tissue homogenate was centrifuged at 100,OOOgfor 60min and the supematant concentrated by lyophilization. Approximately 7aO/,of the total ChAT activity assayed as described above was recovered in the supernatant. The concentrated supematant was then dialyzed in 10 mM potassium phosphate buffer, pH 7.0, and applied to a spharose 4B column (5 x 88 cm) which had been previously equilibrated in the same buffer. Protein was eluted in the same buffer and 10ml fractions were collected. Enzyme activity was recovered in a single peak with the highest specific activity accounting for 65% of the total activity. A 2.7-fold purification was achieved in this step. Active fractions were concentrated and dialyzed against 10 mM potassium phosphate buffer pH 6.8, and layered on a hydroxylapatite (fast flow) column (2.6 x 33 cm) which had been equilibrated in the same buffer. Protein was eluted with a linear 10400 mM buffer gradient (880 ml each) and 10 ml fractions were collected. The ChAT activity was again eluted in a single peak representing 470/, of the total activity and a 6.5-fold purification. Fractions possessing the highest specitic activity were concentrated and dialyzed in IOmM potassium phosphate buffer, pH 6.8. The enzyme was then put on a hydroxylapatite (high resolution) column (1.6 x 22 cm), eluted with a linear concentration gradient from 10 to 400mM potassium phosohate buffer. OH 6.8 (220 ml each1 and 4.4 ml fractions were &lected. choline a&tyltransfer&e activity was eluted in a single protein peak at approximately 170 mM phosphate. Fractions containing the highest specific activity, representing 31% of the total activity and a 4. l-fold purification, were concentrated. The enzyme was then applied to a Sephadex G-150 column (2.6 x 92cm), which had been equilibrated with 10 mM potassium phosphate buffer, pH 7.0, containing 10% glycerol. Proteins were eluted with the same buffer, and 5 ml fractions were collected. Choline acetyltranaferase activity was eluted in two peaks. The second peak, containing the highest specific activity, was concentrated, dialyzed in 10 mM citrate-sodium phosphate bu!Ter, pH 7.2, which contained 0.1 mM dithiothreitol, 0.1 mM EDTA and 10% glycerol. This second peak rqresented 1TA of the total activity and a further puri&c&ion of 3.8 fold. Theeenzyme was fur&r purified on a CM-Saphadex C-50 column (0.9 x 25 cm) with a 1 cm layer of sephadex G25 on top of it. Both mol&ular sieve and cation &&onger were equilibrated in the above citrate-sc%dium phoqhate buffer before building the column. The protein was &ted in a linear gradient from 0-25OmM NaCl in the same buffer (75ml each). Choline acetyltransferase activity was re-

covered as a single peak, representing 9.9% of the total activity and a 12 fold purification. The overall puri&cation of ChAT by the above procedure was 14,000 fold, with a 10% recovery of the total activity. The specific activity of the purified enzyme was 29,500 nmol ACh/min/mg protein. This preparation of partially purified ChAT po%essed no non-specific myltransferase activity as the specific ChAT inhibitor 4-( I-Naphthyl vinyl) pyridine (Cd B&hem-&l&g Corp., LaJolla. CA) abolished all measurable activitv in aliauots of this prebaraiion. Choline acetyltransferase was further purified for subsequent identification of the specificity of monoclonai antibodies. Enzyme, partially purified as described above, was subjected to polyacrylamide gel electrophoresis (PAGE) in 0.4 M Tris-glycine buffer, pH 8.2. One gel was cut into 2 mm slices, and protein was eluted and assayed for ChAT activity. The area corresponding to the enzymaticalty active protein band was cut from 34 gels and ChAT was ~i+trophoretically eluted using an ISCO model 1750 apparatus. The ChAT enzyme protein eluted from PAGE gels displayed a single protein band when subjected once agaiq to electrophoresis and is referred to hereafter as highly p&&d ChAT. Production transferase

of monoclonal

antibodies to choline acetyl-

Ten CB6 mice (Jackson Laboratory, Bar Harbor, ME) were immunized with 40 fig of the partially purified ChAT, described above, which was emulsified in an equal volume of complete Freund’s adjuvant. Six series of injections were given intradermally in the backs of the animals at 2 week intervals. Sera were tested 7 days after each series of injections for the presence of ChAT antibody by immunodouble diffusion against the antigen used for injection. Those animals, in which immuno-double diffusion was positive, received 72 h before death 40 pg iv. of the same ChAT antigen de-scribed above, suspended in normal satine. At the same time, plasmacytoma cells of the 8-aza-gqnine resistant nonimmunoglobuiin secreting iine, NSI, were seeded in 60 mm petri dishes (Falcon) to a density of 2 x 10’ cells per dish and feeder laters of normal spleen cells in selective medium were priqad in 20 culture p&eag&ar 24 cluster) so that each of the 480 w&s contained 2 x IO’ cells. Culture medium consisted of KPMl It440 sdemented with lo”/, fetal calf 6erum, 1% of lOOmM~s&ittm pymvate and n&essential zimiua a&& (Giancf Island Nolo&al Co.. Grand Island. NYl and %QIJ of uwi&#in&ptomy&. Selective m&i& for killing pa&tat myeloma cells after fagion contained the a&ve basic. supp%mented with 2% of a soiution coia&nin*.idkM hypoxae, 0.043 mM aminopterin -and i.6 m)irl thymidine (HAT).23 Fusions were carried out in a manner sin&r to Hammerling’s13 modification of the method of Pontecorvo.”

Immunohistochemical

localization of choline acetyltransferase

Seventy-two h after administering the final i.v. antigen dose, sensitized mice were killed by cervical dislocation, their spleen cells were harvested and mixed in a 5:l spleen to tumor ratio with the NSl cells seeded 72 h earlier. The cells were pelleted and supematant was thoroughly aspirated. Under constant agitation at 37”C, 2ml was added of a 41.7% (w/w) solution of polyethyleneglycol (M.W. 4000, Baker, Fisher Scientific) in Delbucco’s phosphate buffered saline containing 15% dimethylsulfoxide. After 2min the fusion reaction was stooned bv dronwise addition of RPM1 1640, slowly at first and ihen acceleiating so that 30 ml were added over 3 min. The fusion mixture was pelleted and resuspended in a small volume of HAT medium and cells were distributed onto feeder layers at a density of 2 x 10s cells per well. After 8 days in a 5% CO, humidified incubator kept at 37°C medium was half replaced with the basic feed containine no aminouterin (HT). Colonies of hybrid cells appeared over 14-30 days in cuiture. Screening of hybridoma cultures for the secretion of ChAT antibodies was performed using an enzyme linked immunosorbant assay (ELISA) adapted from the method of Engvall and Perlmannn and detailed elsewhere by Ross, Reis and Joh.32 In brief, an appropriate amount of ChAT (50 &well), using either partially purified enzyme for initial screening of cultures (100 ng ChAT/well) or highly purified enzyme (approximately 10-20 ng/well) for specificity testing, was incubated in the wells for 1 h at 37°C. Areas of the wells to which antigen molecules were not adsorbed were then blocked with a phosphate buffered saline solution containing 5% bovine- serum albumin, 0.5% Tween 20 (Sigma Chemical Co.. St. Louis. MO) and 0.025X Na Azide. Supematant (50 ~1 each)’ from’ hybridoma’ cultures was incubated in the wells for 1 h at 37°C. Antimouse immunoglobulin, conjugated with alkaline phosphatase, which was produced in our laboratory by standard techniques of glutaraldehyde conjugation or purchased from Miles Laboratories, was then incubated in the wells for 30min at 37°C. Finally, 50 ~1 of a 2.3 mM p-nitrophenylphosphate solution in 9.6% diethanolamine buffer, pH 9.8, was added as a substrate for alkaline phosphatase to obtain a yellow reaction product. Colonies-secreting anti ChAT antibody were cloned bv limiting dilution as described by Reid,” again using normal mouse spleen cells as feeder layers at a density of 5 x lo5 cells per ml. Clones secreting ChAT antibody were injected i.p. into CB6 mice which were pre-treated i.p. with 1 ml of pristane (2,6,10,14-tetramethylpentadecane, Aldrich Chemical Co., Milwaukee, WI) and high titers of monoclonal antibody in resultant ascitic fluid were harvested for use in immunochemical precipitation. Clones were also expanded for direct harvesting of antibody from culture medium and for internal isotopic labeling of antibody for immunocytochemical radioautography. Immunochemical precipitation

To 20 ~1 of partially purified rat striatal ChAT was added 100~1 of IgG from either control mouse ascites or ascites of a mouse inoculated in. with a hvbridoma secretina ChAT antibody. The mixture was incubated for 1 h at room temperature. Either 20~1 of 0.9% NaCl or 20 ~1 of rabbit anti-mouse antibody was then added and this mixture was incubated for 1h at room temperature. Samples were centrifuged at 27,OOOgfor 15 min and 100 ~1 of each supematant was assayed for ChAT activity. Production of radiolabeled antibody

Monoclonal ChAT antibody was labeled with tritium in order to ensure a high specific activity of endogeneously-labeled immunoglobulin, hybridoma clones in the log phase of growth were harvested, and washed in RPM1 1640 which did not contain the five amino acids, leucine, lysine, phenylalanine, proline and tyrosine. Cells (5 x 106)were collected into 1 ml of the amino acid deficient situ.6 In

909

RPM1 which had bean supplemented with 20% fetal calf serum and 1.5-2.0 mCi of a mixture of radiolabeled amino acids, [3H]leucine, [3H]lysine, [3H]phenylalanine, [3H]proline and [)H]tyrosine (TRK 550, Amersham, Arlington Heights, IL, sp. act. 70-100 Ci/mmol). In order to eliminate ethanol in the isotopic solution, the radiochemical was lyophilized to dryness and resuspended in 1 ml of the deficient medium before addition to cells. After 24 h in culture, the condition of cells was checked and a second bolus of 5 x IO6 cells, similarly washed and pelleted, was added to the 1 ml culture. At the end of another 18-24 h, the radioactive culture was checked for viability of cells and supematant was collected. If less than 50% of the cells added to culture were viable when the supematant was harvested, significant amounts of nonspecific radiolabeled proteins were likely to be present in the supematant. The labeled ChAT antibody from-such cultures was purified on a Sephadex G 200 column. Supematant collected from healthy radioactive cultures was dialyzed extensively against 0.9% saline suffered in 10 mM potassium phosphate, pH 7.0, over a 72 h period. Ten microliters of the radiolabeled antibody was then digested in 1 ml of Protosol (New England Nuclear) at 37°C for 3 h and radioactivity measured after addition of scintillation counting fluid. Like aliquots from preparations were used for radioimmunocytochemical labeling of tissue sections. Immunocytochemical radioautography

Endogenously-labeled monoclonal antibodies have been used by this laboratory and others6.7,8 for immunocytochemical localization of specific antigens at both light and electron microscopic levels. For immunocytochemical localization of ChAT, 260 g male Sprague Dawley rats were anesthetized i.p. with 6Omg/kg of nembutal and perfused through the heart with 4% (w/v) parafotmaldehyde in 0.1 M sodium phosphate buffer, pH 7.3. Blocks of tissue (l-2 mm thick) were cut and post-fixed in the paraformaldehyde solution for 1 h at room temperature and subsequently immersed in 30% sucrose overnight at 4°C for cutting into 8 pm sections at -20°C on a cryostat (Bright Microtome cryostat, Hacker Instruments, Faitfield, NJ) and mounted on slides coated with gelatin. Sections were allowed to thaw and were processed for radioautography. Briefly, mounted tissue sections were transferred to a Tris-saline solution (0.9% NaCl in 0.1 M Tris buffer, pH 7.6) for 15 min and then transferred seauentiallv to: (1) 0.25X Triton x 100 in T&saline for 15 mm; (2) a 1:30 dilution-of normal rabbit serum in Tris-saline for 30 min; (3) a 1:20 to 1:50 dilution of equivalent fractions of radiolabeled monoclonal ChAT antibody which had been harvested in culture supematants and dialyzed or passed over a Sephadex G200 column and concentrated was prepared in Tris-saline containing 1% fetal calf serum and incubated for 24-48 h at 4°C. All steps were spaced by two 15 min washes in T&-saline. Slides were washed twice in distilled water and Tris-saline. In order to remove some lipids and reduce background labeling, they were then dehydrated, immersed in xylene, and rehydrated in serial alcohols (30, 50, 70, 95, 100x, xylene). Slides were taken from 30% alcohol to a final waterwash before coating with radiographic emulsion. Ilford L4 emulsion (Polysciences, Warrington, PA) was prepared in the dark under an OC orange filter safe light (Sodium lamp, Thomas) using a 1:4 (w/v) dilution in double distilled water kept at 37-4o”C in a stationary water bath. Once the gel was melted and thoroughly mixed, immunoreacted slides were taken individually from the final water wash, dipped in the emulsion and allowed to dry upright, on end, on a plexiglass slide rack to ensure even coating. Slides were then placed in a desiccated dark box, tightly sealed and stored upright at 4°C for l&14 days. Silver grains were then developed with D19 developer (Eastman Kodak, Rochester, NY) for 5 min at 2o”C, washed twice in distilled water, and fixed for 30 s in Rapid Fix (Eastman Kodak). Tissues were

M. Ross et al.

YIO Table 2 lmmunopreqxtatmn

of chohne acetyltransferase monoclonal antibodv ChAT actwlty

Fluid ChAT + Control ascnes + RAM ChAT + An&ChAT awtes + 0 9:” N&l ChAT + AntKhAT asutes + RAM

(ChAT) actwty

pmol ACh:30 mm

0 0 decrea,e

7050

0

6558

7

689

90.1

by

To 20 ~1 of ChAT, partially purified as described in the text, was added 100 ~1 of IgG from either control mouse ascites or ascites of a mouse inoculated I p. wth a hybndoma secreting anti-ChAT antibody. Samples were mcubated for 1 h at room temperature and 20~1 of either rabbit antimouse antibody (RAM) or 0.9% NaCl was added. After incubating another 30 min at room temperature, precipitates were sedimented at 27,000-g for 15 min and 100~1 of each supernatant was assayed for ChAT activity. As a further control, a sample of the ChAT used for immunoprecipitation was assayed m the presence of the specific CbAT inhIbitor 4(1-Naphthyl vinyl) pyridine to demonstrate that only ChAT and not a non-specific transferase activity was measured in this system

serially in ethyl alcohol (30, 50, 70, 95 and 100% w/v) and washed twice in xylene. With a coverslip in place, sections were examined for radiographic localization of ChAT. For electron micrographic radioautography, animals were perfused in the same paraformaldehyde solution containing 0.2% glutemk-lehyde. Following the 5 min perfusion and a subsequent 1 h immersion in the same fixative, coronal sections through the dorsal motor nucleus of the vagus were cut with a vibrating microtome (Vibratome).14 Vibratome sections were then labeled with antibody as described above, except that a 1:lOO dilution of ‘H-monoclonal ChAT antibody was used in an overnight incubation followed by two 15 min washes in Tris-saline, 1 h incubation in 2% osmium tetroxide, dehydration and flat embedding in Epon 812.29Ultra-thin sections were collected from the surface of the Vibratome sections to process for radioautography using L4 emulsion. After an exposure period of l-3 months the autoradiograms were developed with paraphenylene diamine or Microdal-X and examined with a Philips 201 ekcctron microscope. The criteria used to establish specific labeling in profiles examined under the electron microscope were rigorous. A structure was considered labeled if 3-5 silver grains in contrast to O-1 for a corresponding region of the neuropil was found over that structure. Moreover, the structures presented as specifically labeled could be followed with their label through three or more serial sections. For double label studies, the procedure was essentially as above except that following incubation of the tissue sections in radiolabeled antibody to ChAT, they were then sequentially incubated, (a) overnight in a 1: 1000 dilution of TH rabbit antiserum, (b) for 30 min in a 1: 50 dilution of anti-rabbit antibody, (c) for 30 min in peroxidase antihorseradish peroxidase (PAP) complex. The PAP label was visualized by reacting bound peroxidase for 6min in a solution conkining 44mg of 3-j’-diaminobenzidine (DAB) and 0.04ml of 30”/, hydrogen ueroxide m 0.1 M Tris HCl, pH 7.6 producing’ a brown reaction product?9,‘2 Finally, sections were coated with radiographic emulsion, as described earlier, for development of the radioautographic label. For controls used in ail studies, hybridomas were selected which secreted antibodies which reacted on ELISA testing with proteins, elukd from PAGE gels, containing no ChAT activity and which did not react-with the highly purified ChAT antigen. These hybridomas were cultured in isotonic medium to produce ikemally labeled control a&b&y which in every step was treated in a fashion parallel to the labeled ChAT antibody. In one set of experiments as well. mtemally labeled TH antibody produced from our monothen dehydrated

clonal lines3* was used as control for labeling of ChATcontaining cells. RESULTS

Two out of ten mice inoculated produced antibody. Following cell fusion, less than 30% of the cultures plated grew hybridoma colonies. Three percent of these secreted antibody to ChAT. Of those, two stable clonal lines secreting ChAT antibody were established. The ChAT antibodies of both linm are high molecular weight immunoglobins, probably of the IgM class as they elute in the void volume of a Sephadex G200 column during the purification of the antibodies. Specificity of the antibodies was determined by the following criteria: first, on ELISA, they reacted against highly purified ChAT, second, also on ELISA, they did not react to contaminant proteins electrophoretically separated from ChAT after the final column in the purification procedure; third, ChAT activity was eliminated from partially purified rat striatal ChAT by immunoprecipitation (Table 2). Light microscopy

The distribution of labeled specific antibody was examined by light microscopy in areas of brain, spinal cord and peripheral tissues weil established as containing the cell bodies and/or terminals of cholinergic neurons. ‘J~J’J~*~~*~’ Sampled were cholinergic motor neurons of the spinal cord and brainstem, neurons in the myenteric plexus, neuromuscular junctions in diaphragm and brainstem areas of-heavy cholinergic innervation. In the central nervous system, cholinergic motor neurons of the anterior horn of the cervical spinai cord were heavily labeled with gra& cuneentr~ted over their perikaryal cytdplasm (Fig. 1). ?%e ace& newopil, contai&g proximal &r&&es, was &so densely labeled. In con@&, ad-t &eas of the spinal white matter co&ah& few grains (Figs IA,B). In control sections taken through the same area of

Fig. 1. Immunoradioautographic labeling of spinal motor neurons and diaphragmatic neuromuscular junction in rat using monoclonal antibody to choline acetyltransferase. Monoclonal ChAT antibody labeled with tritium in situ and harvested from cloned hybridoma culture supematants was used for radioautography as detailed in Experimental Procedures. (A) Motor neurons of the cervical anterior horn. Bar = 50 pm. (B) Low power view of the ventral horn of cervical spinal cord. Large cholinergic perikarya are easily delineated. Note the absence of labeling in white matter. Bar = 1OO~m. (C) Control section through the anterior horn of the spinal cord. Tissue was incubated in tritium labeled control hybridoma antibody which does not react to highly purified ChAT on ELISA and processed in the same manner as sections in A and B. Bar = 50 pm. (D) Labeling of a choline+ neuromuscular junction of the diaphragm. Bar = 50 pm.

911

Fig. 2. Immunoradioautograpbic labeling of motor nuclei of the rat brainstem using monoclonal antibody of to ChAT in Fig. 2. (A) Low newer view of the h -nt;lcfeus. call BRbbesW

stru6tures m the rat mid& using mono&oat Fig. 3. Immunoradioautographic in the red nuckus. Bar= 1OOflm. (I&) Higher antibody to choline acetyhransferase. majpifkation of the same nucleus. Bar = 60 jtrn. (C) Labeling of the cbohnergic terminal A&d in the interpeduncuiar nucleus. Camrgi A10 cells, dorsal to the interpedtmeutar mu&us, am staked for tyrosine hydroxylase using the peroxidase-antiprkdase (PAP) method. magnitkation of area cimumsert ‘kdinC.NotetbedMnetioubetweentheh of catecholamine cell bodies and coarse silver grams correspondmg to choiinergte Bar= 5Opm. Fig. 4. Electron microscopic radioautograpbic locahxatioa using mono&ma1 antibody to Cl&l” in the dorsal motor nuckns of the vagus. (A) Sliver grains am locaWed in the cytoplasm of a neuronat e&l body (nc) and are completely absent from the cytopkm of an Bar = 0.5 pm, (B) Transverse (dt) and lo@tudinal (dl) section tkougb two le&ive toc&ixation of silver grams. Surrounding neuropii is devoid of sil Spm. Fig. 5. Electron microscopic radioautograpbic locakttion using monoclonal antibody to C&AT in tbe dorsal motor nucleus of the vagus. (A) Low magnikaion micrograph showing sifver grains predominantly over identitiable neu the neuropil as in&cat&by the dendrite enoWed by a dashed line. Bar = 0.5 pm. ion electron micrograph showing reduced silver grains over a myelinated axon. Bar = 05 pm.

912

Fig. 2. 913

Fig. 3 914

Fig. 6. Combined peroxidase-antiperoxidase localization of tyrosine hydroxylase and radioautographic localization of choline acetyltransferase in a single section through the locus ceruleus. (A) Double label in the area of the locus ceruleus. Slides were first incubated with 3H-labeled antiChAT monoclonal antibody, washed and reacted with rabbit anti-tyrosine hydroxylase antibody followed by layering of the PAP complex and reaction with diaminobenxidine as described in Experimental Procedures. Slides were then coated with radiographic emulsion, incubated for 10 to 14 days and developed for visualization of reduced silver grains corresponding to ChAT. Note coarse silver grains abutting on TH-labeled perikarya. (Choline@ perikarya labeled only with silver grains were also seen in other regions of the same tissue section.) Bar = 12 pm. (B) Single label using ChAT antibody in the area of the locus ceruleus. The tissue section was reacted only with tritiated monoclonal ChAT antibody and processed as described in Experimental Procedures. Note the absence of labeling of catecholaminergic perikarya and the rather coarse, heavy labeling of surrounding neuropil. Bar = 30 pm.

917

Immunohistochemical

localization of choline acetyltransferase

the cord and incubated with a ‘H-labeled hybridoma antibody which did not react with ChAT, neither neurons nor neuropil were labeled (Fig. 1C). Heavy concentrations of silver grains were distributed over nuclear groups of the brain stem containing motor neurons of cranial nerves including the hypoglossal nucleus (Figs 2A,B), the dorsal motor nucleus of the vagus (Fig. 2A), the nucleus ambiguous, motor nucleus of the trigeminal nerve (Fig. 2C), facial nucleus and abducens and oculomotor nuclei (Fig. 2D). In the periphery there was dense labeling of neurons in the myenteric ganglia and plexus of the ileum. There was also dense labeling of the neuromuscular junction in the diaphragm (Fig. 1D). Additional areas containing labeled neurons or neuropil were seen in selected sections through the brain stem. In the medulla and pons, labeled cells were observed in the area of the nucleus reticularis tegmenti pontis and the nucleus ventralis tregmenti. In the nucleus locus ceruleus there was a dense labeling of the neuropil but not of perikarya (Figs 6A,B). In the midbrain, perikarya were labeled in the caudal levels of the red nucleus (Figs 3A,B), although the density of labeling was less than that observed in cranial nerve nuclei. In the interpeduncular nucleus (Fig. 3C), there was dense labeling restricted to the neuropil, presumably representing the terminations of the cholinergic habenula-interpeduncular pathway described by Kataoka et al.” Dense labeling of neuropil was also seen surrounding the perikarya in the substantia nigra pars compacta and in the region dorsal to the interpeduncular nucleus corresponding to the A10 dopaminergic neurons. In these regions cell bodies were unlabeled. Dense label was also seen throughout periaqueductal grey matter. Electron microscopy To prove that radioautographic labels were contained within neuronal elements, the area of the dorsal motor nucleus of the vagus was examined in tissues processed for electron microscopic radioautography. In electron microscopic sections through the dorsal motor nucleus of the vagus (Figs 4, 5) ChAT was localized predominantly in the dendrites, myelinated and unmyelinated axons. Radioactivity was also detected in perikarya where it was diffusely distributed throughout the cytoplasm (Fig. 4). In serial sections the grain count was substantially greater within neurons and processes than in adjacent areas of the neuropil. Striking was the absence of labeling in glia (Fig. 4) indicating that the grains observed in the neuropil in light microscopic sections represent staining of ChAT in neuronal processes. Combined staining with antibody to tyrosine hydroxylase To examine whether a purported

interaction

be-

919

tween cholinergic and catecholaminergic neurons could be demonstrated in the same section, double label studies were undertaken through locus ceruleus, substantia nigra and ventral tegmentum containing the A10 group. Tissues prepared for radioautographic localization of ChAT were counterstained for TH labeled by the PAP method (see Experimental Procedures for details). In both the substantia nigra (A10 area) and locus ceruleus, TH-stained perikarya corresponded to intrinsic catecholaminergic neurons.9*15*22,26*43 In all areas there was a rich deposit of silver grains around TH-stained elements (Figs 3C,D). In the substantia nigra there was substantial ChAT labeling along the processes as well as perikarya of the dopaminergic neurons possibly indicating cholinergic projections onto soma and dendrites of the nigral neurons. A similar localization of silver grains to soma and dendrites was found among noradrenergic neurons of the locus ceruleus (Fig. 6). To determine whether it was necessary to prepare sections for radioautography before staining with the PAP technique, sections through the substantia nigra were first incubated with rabbit antibody to TH and taken through the peroxidase reaction with DAB prior to incubation with 3H-monoclonal antibodies. When this sequence was used, high concentrations of silver grains were found within all perikarya which contained peroxidase label. When the order of antibody labeling was reversed, i.e. sections were first incubated in 3H-monoclonal antibody, then rabbit antibody to TH, developing the PAP label and finally processing slides for radioautography, cells of the Al0 and locus ceruleus contained only background levels of silver grains. Hence, artifactual distribution of ChAT within catecholaminergic perikarya was avoided by using the latter sequence in developing the antibody labels. DISCUSSION

Use of antibodies against ChAT to localize cholinergic neurons by immunocytochemistry has a long and often controversial history.37 A number of difficulties have been encountered by numerous investigators. First, the production of specific antisera to ChAT has proven difficult by conventional immunochemical methods because of the need for relatively large amounts of a pure antigen which is not readily obtainable. Second, purified ChAT appears to be poorly antigenic, as affirmed in the present study by the relatively long period of time required to induce an antibody in sensitized mice and also the low number of hybridomas producing a ChAT antibody. The low antigenicity of ChAT contrasts with our experience in producing monoclonal antibodies to catecholamine synthesizing enzymes.32 Finally, the purity of the enzyme has been challenged by several investigators,3~33~38raising the possibility that antibodies may be produced against closely allied but

920

M. Rossetal

non-specific macromolecules which may lead to spurious staining. Using hybridoma technology, a single antibodysecreting cell is selected and cloned so that a monoclonal antibody is, by definition, monospecific in the sense that it reacts against a single antigenic determinant.20.2’.25Providing the selection of such an antibody is performed carefully, using several criteria, one can be confident of the specificity of the reagent. Unfortunately, the presence in a tissue of proteins other than ChAT sharing a similar antigenic determinant is difficult to rule out absolutely, and a number of monoclonal antibodies should be generated and screened to this transferase to establish the specificity of localization using several different antibodies. The unique ability of monoclonal antibodies to be isotopically labeled in situ has produced powerful reagents for histological radioautography at both the light and electron microscopic levels6*‘,’ Regarding this antibody to ChAT, the specificity of labeling for choline& neurons has been demonstrated by the heavy concentration of silver grains visualized by light microscopy over perikarya and processes of neurons well established to be cholinergic, including motor neurons of spinal cord and brain stem, neurons of the myenteric plexus and in the neuromuscular junction. That the label is neuronal within the CNS was demonstrated by electron microscopic radioautography in which silver grains were highly concentrated over the processes and perikarya of neurons and were not concentrated within glia. In addition, ChAT was localized to other neuronal groups of the brain stem. Of particular interest in this respect was labeling of neurons of the red nucleus. The localization of ChAT to this structure confirms the observations of Kimura et al.” in mapping the cat brain with a rabbit antiserum raised against human ChAT. Moreover, Kobayashi et aLI9 reported moderate levels of ChAT activity by biochemical assay in micropunches through the nucleus. The localization of ChAT to neurons of this nucleus stands in contrast to studies in which the presence of cholinergic systems in the brain stem have been inferred by the presence of staining for acetylcholinesterase,26,4’ for in these cytochemical investigations the red nucleus was not included in the survey as a cholinergic structure. These observations raise the prospect that there may be groups of choline@ neurons in the brain which either do not express cholinesterase, or contain the esterase in concentrations too low to be detected by the standard histochemical procedures. The discrepancy is important in two respects. First, it indicates that the results of mapping cholinergic neurons by a direct assessment of the distribution of ChAT and indirectly by mapping cholinesterase may differ. While others have pointed this out, the argument has been that non-choline@ neurons will also contain cholinesterase, for example, dopamine neurons of the substantia nigra. 26Our study indicates the converse,

that a population of cholinergic neurons may not express cholinesterase. The prospect is therefore raised that there may be classes of cholinerglc neurons which are relatively atypical in that they do not appear to use acetylcholinesterase to inactivate their transmitter. In electron microscopic sections of this series, a striking contrast between grain count over labeled structures and surrounding neuropil was observed which was consistent through serial sections. This undoubtedly reflects the lower levels of background labeling afforded by monoclonal antibodies.5J2 This characteristic of localization with monoclonal antibodies obviates the need for statistical analysis of grain count in electron micrographic profiles to identify specifically labeled structures. The unique ability of monoclonal antibodies to be isotopically labeled in situ for use in histologic radioautography at both light and electron microscopic levels is demonstrated here. In combination with a second immunohistochemical method such as PAP, the possibilities for double label studies examining anatomical interrelationships between choline+ and catecholaminergic systems is currently being explored. Results presented here at the light microscopic level are encouraging, particularly in the locus ceruleus where the coarse aggregates of silver grains abutting on cell bodies appear rather typical of those found radioautographically at the light level in synaptic contacts positively identified at the electron microscopic level. The stability and electron-dense nature of this radioactive immunohistochemical label. as well as the obvious distinction between reduced silver grains corresponding to ChAT and the smaller, more uniform grains of the PAP product, localizing TH. makes this double label technique ideal for ultrastructural studies. While the possibility of a synaptic interaction can only be suggested at the light microscopic level, double label studies employing this technique can unequivocally identify synaptic relationships at the electron microscopic level. Indeed, such a capability has been demonstrated by Cuello et al. in a recent publication.* An important pitfall in this procedure was found in our current studies. When the catecholaminergic localization was carried out first, and taken through the peroxidase reaction with DAB prior to incubation with 3H-monoclonal antibody, the isotopic immunoglobulin could be ‘trapped’ by the PAP complex and/or DAB reaction product, giving the artifactual ‘localization’ of ChAT within catecholaminergic perikarya. When the order of immunoreaction was reversed, i.e. with ‘H-ChAT antibody incubated on tissue sections before localization of TH, no such labeling artefact was generated. An alternative explanation for the appearance of silver grains over DABlabeled cells is that the peroxidase reaction product initiates a chemical reduction of the radiographic emulsion. If this was the case. no difference would have been observed between the two series as in both

Immunohistochemical

localization of choline acetyltransferase

the PAP label was developed before coating the slides with emulsion. Furthermore, when control slides which contained PAP-stained tissue and no isotopic antibody labeling were coated with emulsion, incubated in parallel with radioautographic slides and developed, no increase in silver grain numbers was found over the cells containing DAB, indicating that the brown reaction product does not cause chemical reduction of the emulsion.

instances

921

Using the sequence of antibody application presented, it was evident that the rabbit PAP complex was not binding to the tritiated mouse monoclonal ChAT antibody because cells labeled only for ChAT were seen in other areas of the same sections containing double label. Experience gained from the work reported here should prove useful in studies of cholinergic interactions with other neuronal systems.

REFERENCES

1. Briggs C. A. and Cooper J. R. (1982) Choline& modulation of the release of [3H] acetylchohne from synaptosomes of the myenteric plexus. J. Neurochem. 38, 501-508. 2. Chao L.(1980) Choline acetyltransferase: purification and characterization. J. Neurosci. Res. 5, 85-l 15. 3. Chao L. P.. Wolfaram F. J. and Ena L. F. (1977) Immunohistochemicai localization of choline acetvltransferase: real or artifact?’ A resy. Neurochem. R& 2, 323-325. 4. Cozzari C. and Hartman B. K. (1980) Preparation of antibodies specific to choline acetyltransferase from bovine caudate nucleus and immunohistochemical localization of the enzyme. Proc. natn. Acud. Sci. U.S.A. 77, 7453-7457. 5. Cue110A. C., Galfre G. and Milstein C. (1979) Detection of substance P in the central nervous system by a monoclonal antibody. Proc. natn. Acad. Sci. U.S.A. 76(7), 3532-3536. 6. Cuello A. C. and Milstein C. (1981) Use of internally labeled monoclonal antibodies. In Physiological Peptides and New Trends in Radioimmunolonv (ed. C. A. Bizollon) DD. 293-305. Elsevier-North Holland. Amsterdam. 7. Cuello A. C., Milstein C. %dXPriestley J. V. (198bj Use of monoclonal antibodies in immunocytochemistry with special reference to the central nervous system. Br. Rex Bull. 5, 575-587. 8. Cue110 C. A., Priestley J. V. and Milstein C. (1982) Immunocytochemistry with internally laveled monoclonal antibodies. Proc. natn. Acad. Sci. U.S.A. 79, 6655669. 9. Dahlstrom A. and Fuxe K. (1964) Evidence for the existence of monoamine containing neurons in the central nervous system--I. Demonstration of monoamines in the cell bodies of brainstem neurons. Actu physiol. stand. 62 suppl. 232, l-55. 10. Dzieniszewski P. and Kilbinger H. (1978) Muscarinic modulation of acetylcholine release evoked by dimethylphenyl piperazinium and high potassium from guinea-pig myenteric plexus. Eur. J. Pharmac. 50, 385-391. 11. Eckenstein F., Barde Y. A. and Thoenen H. (1981) Production of specific antibodies to choline acetyltransferase purified from pig brain. Neuroscience 6, 993-1000. Quantitation by enzyme-labeled 12. Engvall E. and Perlman P. (1972) Enzyme linked immunosorbant assay (ELISAFIII. antiimmunoglobulin in antigen-coated tubes. J. Immunol. 109, 129-135. 13. HammerlingG. J. (1977) T-lymphocyte tissue culture lines produced by cell hybridization. Eur. .I. Immunof. 7,743-746. 14. Hijkfelt T. and Ljungdahl A. (1972) Modification of the Falck-Hillarp fluorescence method using the vibratome. Simple rapid and sensitive localization of catecholamines in sections of unfixed or formahn fixed brain tissue. Hisrochemie 29, 315339.

15. Hokfelt T., Fuxe K., Goldstein M. and Johansson 0. (1974) Immunohistochemical evidence for the existence of adrenaline neurons in the rat brain. Brain Res. 66, 235-251. 16. Joh T. H. and Ross M. E. (in press) Preparation of catecholamine synthesizing enzymes as immunogens for immunohistochemistry. In Immunocytochemistry IBRO Handbook Series: Methods in Neurosciences. (ed. Cue110A. C.) John Wiley, New York. 17. Kataoka K., Nakamura Y. and Hassler R. (1973) Habenulo-interpeduncular tract: a possible cholinergic neuron in rat brain. Brain Res. 62, 264267. 18. Kimura H., McGeer P. L., Peng J. H. and McGeer E. G. (1981) The central cholinergic system studied by choline acetyltransferase immunohistochemistry in the cat. J. camp. Neural. 200, 151-201. 19. Kobayashi R. M., Brownstein M., Saavedra J. M. and Palkovits M. (1975) Choline acetyltransferase content in discrete regions of the rat brain stem. J. Neurochem. 24, 637-640. 20. Kohler G. and Milstein C. (1975) Continuous cultures of fused cells secreting antibody of predefined specificity. Nature, Land. 256, 495-497.

21. Kohler G. and Milstein C. (1976) Derivation of specific antibody-producing

tissue culture and tumor lines by cell fusion.

Eur. J. Immunol. 6, 511-519. 22. Lindvall 0. and Bjorklund A. (1974) The organization of the ascending catecholamine neuronal systems in the rat brain as revealed by the glyoxylic acid fluorescence method. Acra physioi. scund. suppl. 412, l-48. 23. Littlefield J. W. (1964) Selection of hybrids from matings of fibroblasts in vitro and their presumed recombinants.

Science, N.Y. 145, 709. 24. Milstein C., Adetugbo K., Cowan J., Kohler G., Secher D. S. and Wilde C. D. (1977) Somatic cell genetics of antibody secreting cells: studies of clonal diversification and analysis by cell fusion. Cold Spring Harbor Symp. Quanr. Eioi. 41, 793-803.

25. Milstein C., Galfre G., Scher D. S. and Springer T. (1979) Monoclonal antibodies and cell surface antigens. Cell Eiol. Int. Reports 3, I-16.

26. Palkovits M. and Jacobowitz D. M. (1974) Topographic atlas of catecholamine and acetylchohnesterase-containing neurons in the rat brain-II. Hindbrain (mesencephalon, rhombencephalon). J. camp. Neural. 157, 29-42. 27. Park D. H., Ross M. E., Pickel V. M., Reis D. J. and Joh T. H. (1982) Antibodies to rat choline acetyltransferase for immunochemistry and immunocytochemistry. Neurosci. Let?. 34, 129-135. 28. Peng J. H., McGeer P. L., Kimura H., Sung S. C. and McGeer E. G. (1980) Purification and immunochemical properties of choline acctyltransferase from human brain. Neurochem. Res. 5, 943-962.

922

M. Ross et al.

29. Pickel V. M. (1981) Immunocytochemical methods. In Neuroanatomical Tract-Tracing Methods (eds. Heimer L. and Robards M. J.) pp. 483-509. Plenum, New York. 30. Pontecorvo G. (1975) Production of mammalian somatic cell hybrids by means of polyethylene glycol treatment Somat. Cell Genet. I, 397-400. 31. Reid L. C. M. (1979) Cloning. In Methods zn Enzymology (eds. Jakoby W. B. and Pastan J. H.) Vol. 58, pp. 151-164. 32. Ross M. E., Reis D. J. and Joh T. H. (1981) Monoclonal antibodies to tyrosine hydroxylase: production and characterization. Brain Res. 208, 493498. 33. Rossier J. (1975) Immunohistochemistry localization of choline acetyltransferase: real or artifact? Brazn Res. 98, 619-622. 34. Rossier J. (1976) Purification of rat brain choline acetyltransferase. J. Neurochem. 26, 543-548. 35. Rossier J. (1976) Biophysical properties of rat brain choline acetyltransferase. J. Neurochem. 26, 555-559. 36. Rossier J. (1976) Immunological properties of rat brain choline acetyltransferase. J. Neurochem. 26, 549-553. 37. Rossier J. (1977) Choline acetyltransferase: a review with special reference to its cellular and subcellular localization.

ht. Rec. Neurobiol. 20, 283-337. 38. Rossier J. (1981) Serum monospecificity:

a prerequisite for reliable immunohistochemical locahzation of neuronal markers including choline acetyltransferase. Neuroscience 6, 989-991. 39. Ryan R. J. and McClure W. 0. (1979) Purification of choline acetyltransferase from rat and cow brain. Biochemistry 18, 5357-5365. 40. Schrier B. K. and Shuster L. (1967) A simplified radiochemical assay for vholine acetyltransferase. J. Neurochem. 19, 977-985. 41. Shute C. C. D. and Lewis P. R. (1967) The ascending choline@ reticular system: neocortical, olfactory and subcortical projections. Brain 90, 497-520. 42. Stemberger L. (1979) Zmmunocytochemisrry 2nd edn. Wiley, New York. 43. Ungerstedt U. (1971) Stereotaxic mapping of the monoamine pathways in the rat brain. Acfa physiol. stand. suppl.

367. l-78. (Accepted 25 April 1983)