Distribution of neurons expressing calcitonin gene-related peptide mRNAS in the brain stem, spinal cord and dorsal root ganglia of rat and guinea-pig

Neuroscience Vol. 29, No. 1, pp. 225-239, Printed in Great Britain

1989

0306-4522/89 $3.00 + 0.00 Pergamon Press plc 0 1989 IBRO

DISTRIBUTION OF NEURONS EXPRESSING CALCITONIN GENE-RELATED PEPTIDE mRNAS IN THE BRAIN STEM, SPINAL CORD AND DORSAL ROOT GANGLIA OF RAT AND GUINEA-PIG M. RI?THELYI,* C. B. METZ and P. K. LUND Department of Physiology, University of North Carolina at Chapel Hill, Chapel Hill, NC 27514, U.S.A. Abstract-1n situ hybridization histochemistry was used to localize calcitonin gene-related peptide mRNAs in spinal cord, brain stem and dorsal root ganglion neurons of the rat and guinea-pig. A 32P-labeled 23-base-long (23mer) oligodeoxyribonucleotide (oligomer) complementary to calcitonin generelated peptide mRNA sequences encoding residues 23-30 of calcitonin gene-related peptide was used primarily as a probe (CGRP I probe). A 32mer complementary to mRNA sequences for residues l&20 of calcitonin gene-related peptide (CGRP II probe) was also used as a positive control for specificity of the 23mer for calcitonin gene-related peptide mRNA. In both the guinea-pig and rat calcitonin gene-related peptide mRNA was localized specifically to neurons of the dorsal root ganglion, to spinal motoneurons and to motoneurons of the hypoglossal, facial and accessory facial motor nuclei. Differences in the distribution of calcitonin gene-related peptide mRNA between the rat and guinea-pig included a higher proportion of rat dorsal root ganglion neurons containing calcitonin gene-related peptide mRNA and the localization of calcitonin gene-related peptide mRNA to motoneurons of the ambiguus motor nucleus, parabrachial and peripeduncular nucleus of the rat but not the guinea-pig. In the guinea-pig, in contrast, calcitonin gene-related peptide mRNA was localized also to motoneurons of the abducens, trigeminal, trochlear and oculomotor nerves. The neuronal groups in the intact rat found here to contain calcitonin gene-related mRNA have also been shown previously to contain calcitonin gene-related peptide immunoreactivity in colchicine-treated rats. Colchicine-treated rats, however, have been found to contain additional groups of calcitonin gene-related peptide immunoreactive neurons which, in the intact rats used in the present study, showed no detectable hybridization with the calcitonin gene-related peptide probe.

Calcitonin gene-related peptide (CGRP) is a 37-amino acid neuropeptide encoded by a messenger RNA (mRNA) that results from alternate processing of the calcitonin gene primary transcript.‘.28.32 CGRP has been isolated and characterized in rat and human.20,22 Radioimmunoassay revealed the presence of CGRP in the central nervous system of the rat3’ as well as in the human thyroid, pituitary and brain.40 CGRP-like immunoreactivity has been localized to various neuronal groups and fiber tracts of the rat brain,‘2.28.30 as well as to neurons of the spinal cord and dorsal root ganglion (DRG) of several species including man.* A surprising observation in all of these studies was the occurrence of CGRP-like immunoreactivity in two developmentally unrelated and functionally different neuronal populations: the primary sensory neurons (PSN) and the spinal and brain stem motoneurons. A common feature shared *Visiting professor, to whom correspondence should be addressed at the Second Department of Anatomy, Semmelweis University Medical School, Tuzoho utca 58, 1094 Budapest, Hungary. Abbreviations: CGRP, calcitonin gene-related peptide; DRG, dorsal root ganglion; EDTA, ethylenediaminetetra-acetate; PSN, primary sensory neurons; SDS, sodium dodecyl sulphate; SSC, standard saline citrate. 225

by these two neuron populations is the arborization of their axons to the periphery which could theoretically allow the uptake, retrograde transport and perikaryal accumulation of blood-born substances. Several studies have demonstrated CGRP immunoreactive innervation of various viscera (eye;38 tongue, epiglottis, and pharynx;” iris and cerebral arteries;‘0*‘6 pancreas;23 cardiovascular system;2’,42 respiratory tract;* esophagus;26 palate;*’ thyroid;’ gastrointestinal tract;3 urinary tract34). Since some of these studies also demonstrated substantial reduction or elimination of CGRP immunoreactivity by capsaicin treatment,2.8.34%3sit is generally agreed that the CGRP innervation originates mostly from sensory fibers, i.e. distal branches of spinal and brain stem primary sensory neurons. CGRP immunoreactivity has also been localized to motor end plates and to nerve fibers innervating striated muscles.37.39 Information about the cellular sites of neuropeptide biosynthesis is of value in the design of experiments to address the functional significance of neuropeptides. In the present study therefore, to complement existing information about the localization of CGRP immunoreactivity, we have mapped the distribution of CGRP mRNA in neurons of the rat spinal cord, dorsal root ganglion and brain stem. We have also studied the distribution of CGRP

M. R~THELYI et al.

226 IO his-arg-leu-ala-gLy

AA

-leu-leu-ser-arg-ser-gly-gly-g~y-val-lys-asp-asn-phe-va~-pro-thr

25

mRNA CAU-CGG-CUG-GCA-GGC-UUG-CUG-AGC-P1GG-UCG-GGA-GGU-GUG-GUG-AAG-GPL:-AAC-UUU-GUG-CCC-AC 3’

GTA-GCC-GAC-CGT-CCG-AAC-GAC-TCG-TCC-AX-CC CGRP

II

PROBE

5’

3’

CAC-TTC-CTG-TTG-AAA-CAC-GGG-TG

(32mer)

CGRP

I

PROBE

5’

(23mer)_

Fig. I. Schematic representation of the CGRP oligomer probes. Top line: amino acid (AA) sequence of CGRP molecule. Middle line: base sequence of the mRNA encoding the precursor peptide. Bottom line: base sequence of the two CGRP probes (32- and 23-base-long); 5’ and 3’ indicate the polarity of the molecules.

mRNA in neurons of the guinea-pig, a species commonly used in functional studies but for which there is little available information about the distribution of CGRP immunoreactivity.8,9.34

chrome alum subbed slides from water/gelatin, dried overnight at 37°C and kept at 4°C for IO days-2 weeks before hybridization.2s Prehybridization and hybridization

A synthetic 23-base-long probe complementary to CGRP mRNA sequences encoding amino acid residues 23-30 of CGRP was synthesized on an Applied Biosystems DNA synthesizer (CGRP I probe). In addition to CGRP I which was routinely used as a probe for CGRP mRNA in these studies, a 32mer that is complementary to a nonoverlapping region of CGRP mRNA (encoding residues I@-20 of CGRP; CGRP II probe) was also used in selected experiments as a positive control for specificity. Probe sequences are shown in Fig. I. The probes were labeled at the 5’ end with 32P using T4 polynucleotide kinase and j2P ATP (7000 Ci/mmol ICN) followed by gamma purification on 15% polyacrylamide gels containing 7M urea.” Probe specific activity was 0.5-1.0 x 10” counts per minute (cpm)/pmol. Routinely, aliquots of 3-5 x IO6 cpm of probe were lyophilized and resuspended in 2X hybridization buffer for application to each slide. This 2X hybridization buffer contained: 6X standard saline citrate (SSC), 2X Denhardts, 10% Dextran sulphate, 50mM Tris-HCl pH 7.0, 5 mM EDTA pH 7.0, 0.1% sodium dodecyl sulphate (SDS), lOOpg/ml denatured salmon testis DNA (Sigma) and lOO~g/ml denatured yeast transfer RNA (Bethesda Research Laboratories). [20X SSC = 3 M NaCl, 0.3 M Na-citrate pH 7.0; 20X Denhardts = 0.4% each bovine serum albumin, ficoll and polyvinylpyrrolidine (all purchased from Sigma Biochemicals).] Probes were labeled 3-5 days before use for in situ hybridization with isotope less than one half-life old.

In situ hybridization was carried out by a modification of procedures described previously.24 Cryostat as well as Vibratome sections were washed in PBS containing Triton-X 100 (0.2%) for 60min. Slides were removed from the PBS-Triton wash, left to stand a few minutes and then carefully wiped on the back and top edges with filter paper. For prehybridization, a 50-~1 aliquot of 2X hybridization buffer was boiled for 2 min, quenched on ice and mixed with 50 ~1 of 60% deionized formamide (Fluka) then spread onto each slide. The slides were covered with Parafilm,29 and incubated at 37°C for 1.5-2h in humid chambers. At the end of prehybridization, sections were dehydrated in ascending series of ethanol containing 0.33 M ammonium acetate and air-dried for 10 min. For hybridization, aliquots of 2X hybridization buffer containing labeled probe were boiled, quenched on ice and mixed with an equal volume of 60% deionized formamide before application to the slides. Hybridization was for 48-72 h at 37°C and during this time the slides were coverslipped and kept in a moist chamber. Following hybridization, the coverslips were removed by submerging the glass slides in 4 x SSC. Sections were washed in decreasing concentrations of salt solution (4X, 2X, and 0.1X SSC) at room temperature for 2-3 h, dehydrated through ascending ethanol series containing 0.33 M ammonium acetate and air-dried for 5-10 min. In some experiments, modified 2X prehybridization and hybridization buffers were used to test their effects on signal intensity and background. In these experiments the routinely used 2X hybridization buffer (buffer A) was modified by elimination of tRNA (buffer B) or by elimination of tRNA and addition of lOOpmol/ml of an unlabeled oligomer that was unrelated in sequence to CGRP mRNA (buffer C).

Tissue preparation

Autoradiography

Three rats (I 50-300 g b.w., strain: Sprague-Dawley) and four guinea-pigs (150-300 g b.w., strain: Hartley, Charles River Breeding Laboratories) were anesthetized by intraperitoneal (i.p.) injection of pentobarbital(60 mg/kg for rats and 50mg/kg for guinea-pigs) and then perfused with a solution containing 4% paraformaldehyde and 0.2% glutaraldehyde in 0.1 M phosphate buffer (pH = 7.4). The aldehyde perfusion was preceded with a brief vascular rinse with PBS. After perfusion, the brain stem, selected portions of the spinal cord (cervical, thoracic, lumbar and sacral regions) and DRG of cervical, thoracic and lumbar segments were dissected and immersed for 24 h in the same fixative as used for perfusion. Tissue blocks were cryoprotected in 0.1 M phosphate buffer containing 30% sucrose. Serial lo-pm-thick frozen cryostat sections were prepared from all tissues. Sections were mounted onto gelatin chrome alum subbed slides and stored at 4°C for 2-3 weeks. The brain stem and spinal cord of one rat and two guinea-pigs were cut with a Vibratome and 30-pm-thick corona1 sections were collected in PBS, mounted onto

Slides were covered with Kodak XAR-5 X-ray film and exposed for 24 h at room temperature. Slides were dipped into Kodak NTB-3 emulsion diluted 2:1 (v:v) in 0.33 M ammonium acetate and then dried for 1.5-2 h at 28°C and 80% relative humidity. After complete drying, slides were exposed at 4°C in light-tight plastic boxes containing a drying agent (Drierite, anhydrous calcium sulphate, W. A. Hammond Drierite Co.). The exposure time was based on intensity of hybridization signals observed on X-ray film exposures, but never exceeded 3 weeks. Sections were developed in Kodak D-19 developer diluted I: 1 with distilled water, fixed in Kodafix and coverslipped without counterstaining.

EXPERIMENTAL PROCEDURES

Preparation and radiolabeling of oligodeoxyribonucleotide (oligomer) probes

Microscopical examination A low-power dark-field microscope (Leitz Dialux 22) was used to scan the sections. Areas with silver grains in clusters (apparent hybridization signal) were photographed. Coverslips were then removed, the sections were stained with 0.5% Thionin. recoverslipped and rephotographed. The

CGRP mRNAs in rat and guinea-pig topo~raphi~l relationship between high densities of silver grains and perikarya was examined after staining. Data anai_vses

Exposed and developed X-ray films gave a low-resolution, low-sensitivity estimate of hybridization signal intensity over the sections and served as a guide to optimal duration of exposure of sections to photoemulsion. Low-power dark-field viewing of the unstained sections after in situ hybrjdization hist~hemistry and photoemulsion development revealed three patterns of silver gram distribution as follows. 1. ~w-density randomly dist~but~ silver grains (IO-15 grains/2500 rrn3; this was interpreted as background hybridization and was used as a reference for specific hybridization. Great care was taken to ensure uniform autoradiography procedures across experiments, and background hybridization was similar across experiments. 2. Clustered high densities of silver grains that lay over neuronal perikarva (300-1500 grains/2500 Um2). This nattern was interpreied is positivehybridizatidn with a pa&cular probe. Neurons with the clustered high densities of silver grains over the perikaryon will be termed labeled neurons subsequently in the text. 3. High densities of randomly distributed silver grains (SO-100 grains/2500 ym2) that occurred occasionally over selected brain areas. No definitive interpretation of this pattern of silver grains was made and the possible explanations for this observation are discussed. Silver grains were counted manually at 800 x magni~~tion under bright-field illumination. Cenrrof experimenfs (i) In each experimental series one slide was used as an autoradio~phic control. This control slide was treated as were experimental slides except that no labeled probe was added to the hybridi~tion buffer. (ii) Adjacent Vibratome sections from rat spinal cord were treated with two doses of ribonuclease (5Opg and lOO~g/ml) for 30 min at 37°C immediately before prehybridization. Elimination of positive hyb~dizatio~ by ribonuclease pretreatment is indicative that hybridization is due to RNA. (iii) Varied probe concentrations (5, 2.5, 1.25 and 0.6 x lo6 cpm) were tested on adjacent brain stem Vibratome sections to establish whether intensity of hybridization signals was concentration-de~nd~nt. (iv) A number of synthetic oligomer probes unrelated in sequence to CGRP mRNA (a somatostatin 4lmer, a cholecystokinin 26mer, and a substance K 23mer) were used on serially adjacent sections as negative controls. These probes were similar or longer than the CGRP probe and contained similar or higher G/C content. Lack of hyb~di~tion of these control probes with the same neurons as the CGRP probes, therefore, controlled for the possibility of nonspecific hyb~dization of the CGRP probe to neuronal substances other than CGRP mRNA. (v) While the 23-base-long CGRP oligomer was used as a probe in most experiments, a second 32-base-long ohgomer that is complementary in sequence to a non-overlapping portion of the CGRP mRNA was used as a positive control for hybridjzation specificity in selected experiments.

Immunohistochemistry The distributions of CGRP immunoreactive neurons and those that hybridized with CGRP oligomer were compared. Thirty-micrometer-thick Vibratome sections adjacent to those used for in situ hybri~tion and containing various brain stem motoneuronal groups were processed for immunohistochemistry reaction. For better diffusion of the antisera, sections were briefly treated with ethanol.” Immunohist~hem~strv was performed with a polyclonal CGRP antiserum that was raised in rabbits agaiust7synthetic l-37 human CGRP.’ Free-floating sections were incubated in a

227

f :80@)dilution of the antiserum at 4°C for 24-36 h and were immunostained using the peroxidase-antiperoxidase method.“) Nickel ammonium sulfate was used to intensify the di~ino~nzidine reaction product.‘* RESULTS

Specificity

oflocalization of CGRP mRNA

Using the 23-base-long CGRP oligomer numerous motoneurons in the ambiguous nucleus of rat brain were labeled by in situ hybridization hist~hemistry (Fig. 2A). Neurons in the ambiguous nucleus in an adjacent section showed positive staining with the CGRP antiserum (Fig. 2B). Ribonuclease treatment decreased, but did not entirely eliminate the labeling of both the cervical and lumbar motoneurons in the rat. Whereas moto~eurons in untreated sections were covered with dense groups of silver grains (Figs 3A and B), the grain density over motoneurons decreased proportionally with the concentration of ribonuclease. At both low and high concentrations of the enzyme, the background labeling increased with respect to the control sections (Figs 3A, 4A and SA). The degradation of RNA by ribonuclease pretreatment is also apparent in that the Nissl substance decreased within the motoneurons as the ribonuclease concentration was increased (Figs 3B, 4B and 5B). Figure 6A-D shows the intensity of hybridization signal vs background in four adjacent Vibratome sections of the trigeminal motor nucleus of the guinea-pig after hybridizations with probe concentrations ranging from 5 x IO6to 0.6 x t06cpm. Signal intensity over motoneurons and general background decreased in parallel with decreasing probe concentration. Based on this, and similar test experiments in the rat, 3 x lo6 cpm/slide of the labeled CGRP probe was used routinely. Synthetic oligomer probes complementary to mRNAs encoding CGRP, somatostatin, and substance K labeled different ~pulations of neurons as in examples shown in Fig. 7A-C where adjacent sections of rat DRG were incubated with three different probes. Two different oligomers complementary to nonoverlapping portions of CGRP mRNA recognized neurons of the same size and distribution in the DRG of rats. Localization of neurons expressing CGRP mRNA activity in the rat Dorsal root ganglion. A large number of small (20 pm in diameter) and medium size neurons (up to 45 pm in diameter) at all spinal cord levels (cervical, thoracic, lumbar and sacral) showed intense hybridization with the CGRP oligomer {Fig. SA and B). In each ganglion, labeled neurons occurred intermingled with unlabeled ones. Spinal cord. Almost all large neurons in the ventral horn of the cervical and lumbar regions of the spinal

M. UTHELYI

228

cord were labeled with the CGRP probe (Fig. 9A and B). In thoracic and sacral segments only some of the large neurons were labeled, while adjacent large neurons showed silver grain densities that could not be distinguished from the background. No labeled neurons were seen either in the thoracic or in the sacral preganglionic neuronal groups. Brain stem. Two kinds of neurons were labeled in the brain stem: (i) groups of large neurons corresponding to motoneurons within the well-known brain stem motor nuclei,i5 and (ii) small neurons that occurred interspersed with unlabeled cells. Motoneurons. All along the caudal and rostra1 portions of the medulla, the motoneurons of the XIIth nerve were labeled to various degrees (Fig. 10A and B). A compact group of neurons in the ventrolateral portion of the medulla corresponding to the ambiguus nucleus contained intensely labeled nerve cells (Fig. 11A-C). The majority of neurons in the large motor nucleus of the VIIth nerve located superficially in the ventrolateral medulla were labeled. More rostrally, at the medullary-pontine junction, small groups of labeled neurons were found immediately medial to the exiting facial nerve. This localization of cells corresponds to the accessory facial nucleus.” No labeled cells were found in the motor nuclei of the more rostra1 cranial nerves (VIth, Vth, IVth and IIIrd). Small neurons. Small labeled neurons were found in two locations: in the dorsolateral pons (parabrachial nucleus) and in the cranial portion of the mesencephalon immediately ventral and dorsal to the emerging medial geniculate body (peripeduncular nucleus). In the parabrachial nucleus the densities of the silver grains over perikarya barely exceeded the background level. The clusters of silver grains could, however, be correlated to groups of neurons rather than to individual cells. In contrast, small but dense clusters of silver grains in the peripeduncular nucleus (Fig. l2A) clearly overlaid individual small neurons (Fig. 12B) that were scattered among non-labeled ones. Detection of the signal over the small neurons in both the parabrachial and peripeduncular nuclei

Fig. 2. (A) Motoneurons Motoneurons from immunoreactivity.

ef al

was largely dependent on the use of hybridization buffer B (hybridization buffer without added tRNA; Fig. 13A and B). No labeled cells were seen in the mesencephalic nucleus of the trigeminal nerve. Contrary to our expectations based on previous immunohistoof silver grains chemical results,” no accumulations above background were found in the nucleus of the solitary tract or in the superior olivary nucleus. Localization

oj’ neurons

expressing

CGRP

Using an identical probe (CGRP I probe), probe concentration and other in situ hybridization histochemistry procedures, sections from guinea-pig routinely showed a higher background than sections from the rat. The higher background did not however, impede the unequivocal localization of hybridization signal over neuronal perikarya. Dorsal root ganglion. CGRP mRNA labeling was predominant in small-sized neurons. The proportion of labeled neurons was lower than that in the rat. The use of hybridization buffer C (no tRNA and inclusion of an unlabeled oligomer unrelated in sequence to CGRP probe) resulted in an improved signal to background. Spinal cord. Large ventral horn neurons were moderately but distinctly labeled at the levels of cervical and lumbar enlargements. Brain stem. In contrast to our findings in the rat, only large neurons in well identified motoneuronal groups were labeled in the guinea-pig brain stem. Most of the neurons in the XIIth nerve nucleus of the caudal medulla were intensely labeled (Fig. 14A and B), while only scattered neurons were labeled in the rostra1 portion of the nucleus. In distinct contrast to the rat, no labeling could be detected over the neurons of the ambiguous nucleus. Neurons in both the VIIth nerve motor nucleus and the small neuron group of the accessory facial nucleus (Fig. 15A and B) were labeled. The labeled cells in this latter group seemed to be continuous with a large group of equally well labeled neurons in the Vth

from the ambiguus nucleus

the ambiguus Arrows point

nucleus to CGRP

mRNA

activity in the guinea-pig

of a rat showing hybridization signal. (B) [closely adjacent section to (A)] showing CGRPimmunoreactive nerve fibres. Scale bars = IOOpm.

Figs 3-5. Photomicrogaphs showing the effect of ribonuclease digestion on the hybridization signal of rat motoneurons (lumbar region, horizontal sections). Figure 3. (A) Dark-field picture, intact control, arrows point to accumulation of silver grains overlying motoneurons. (B) High-power bright-field view of the motoneurons marked by arrows on (A). Figure 4. (A) Dark-field picture after digestion with 50 pg/ml ribonuclease, arrow points to accumulation of silver grains overlying a motoneuron. (B) High-power bright-field view of the motoneuron marked by arrow on (A). Figure 5. (A) Dark-field picture after digestion with lOO~g/ml ribonuclease, arrow points to accumulation of silver grains overlying a motoneuron. (B) High-power bright-field view of the motoneuron marked by arrow on (A). Ribonuclease digestion in increasing concentration gradually eliminates the staining of the Nissl substance (ribosomes) in the motoneurons (compare 3B, 4B and 5B). Scale bars: 3A, 4A and 5A, 100 pm; 3B, 4B and 5B, 10 pm. Fig. 6. Dark-field photographs showing the effect of decreasing probe concentration on the hybridization signal in trigeminal motoneurons of an intact guinea-pig on immediately adjacent Vibratome sections. (A) 5 x 10’cpm; (B) 2.5 x 106cpm; (C) 1.25 x 106cpm; (D) 0.6 x 106cpm. Clusters of silver grains (some of them marked by arrows in A-C) overlay neuronal perikarya. Scale bars = 100pm.

229

230

M.

R~THELYI et al.

Fig. 7. Dark-field photographs showing the distribution of labeled neurons in rat DRG using synthetic oiygodeoxyribonucleotide probes complementary to mRNAs encoding CGRP (A), substance K (B) and somatostatin (C) on closely adjacent sections. Scale bars = 1OO~m.

Fig. 8. Hybridization signal in rat dorsal root ganglion. (A) Dark-field photograph showing labeled small and medium sized neuronal perikarya. (B) Higher-power detail from (A). Arrows point to identical labeled neurons in both figures. Asterisks in (B) indicate unlabeled neurons. Scale bars: (A) 100 pm; (B) 10 pm. Fig. 9. Hybridization signal in rat spinal cord (lumbar segment, ventral horn). (A) Dark-field picture showing the distribution of groups of silver grains (arrows). (B) Bright-field view of the same area as in (A). Arrows point to neurons showing hybridization signal. Curved arrow in (B) points to an apparently unlabeled large neuron. Scale bars = 100pm. 231

M. R~THELYI et al.

232

nerve motor nucleus (Fig. 6B and C). A group of neurons ventral and lateral to the facial knee at the pontomedullary junction, i.e. motoneurons of the VIth nerve, were labeled and showed variable labeling intensity (Fig. 16A and B). More rostrally in the mesencephalon, numerous motoneurons in the IVth and IIIrd nerves (Fig. 17A and B) were labeled.

DISCUSSiON

CGRP probe specificity and methodological considerations

The 23-base-long synthetic CGRP oligomer was designed based on a conserved nucleotide sequence between rat and human mRNAs and was used for in situ hybridizations with sections of rat and guinea-pig brain, spinal cord and DRG. The specificity of hybridization of the probe with neurons containing CGRP mRNA was supported by the following. (a) Reduced hybridization signal intensity over labeled neurons after ribonuclease treatment of sections prior to hybridization. Wybridization signal intensity decreased in parallel with a dose of ribonuclease but ribonuclease treatment did not completely eliminate hybridi~tion signal. This may be due to residual CGRP mRNA fragments large enough in size to hybridize with the oligomer probe. (b) Increased hybridization signal intensity in both rat and guineapig neurons with increased probe con~ntration. (c) Findings that probes unrelated to CGRP (somatostatin, cholecystokinin, substance K oligomers) labeled different neuron populations in both rat and guinea-pig brain than did the CGRP probes. (d) The demonstration in the rat of CGRP immunostaining of neurons of the same size and distribution as those that hybridized with the CGRP probe when Vibratome sections containing the motoneurons of the ambiguous nucleus of the rat were processed for immunohistochemistry and in situ hybridization. The locations of hybridizing neurons found in the rat in the present study were also in good agreement with previously reported locations of CGRP immu(e) Findings that a second noreactive neurons.“~‘2.28~‘0 probe complementary in sequence to a different portion of rat CGRP mRNA than the routinely used 23mer labeled similar neurons to the 23mer when used on adjacent sections of rat DRG. Taken together, these findings suggest that a high density of silver grains over neuronal perikarya corresponds to specific hyb~dization of the CGRP probe

with neuronal CGRP mRNAs in both rat and guinea-pig. The CGRP I oligomer sequence was based on rat and human CGRP mRNA sequences but cross-hybridized with guinea-pig CGRP mRNA, suggesting conservation of CGRP mRNA sequences across human, rat and guinea-pig, at least in the region of the mRNA to which the CGRP I probe corresponds. In some brain regions we observed diffuse increases in silver grain density that were not clustered over individual neurons. We favor the possibility that such signals result from non-specific binding of the CGRP probe to some cellular component unrelated to CGRP mRNA as we have observed similar phenomena with a control oligomer complementary to cholecystokinin mRNA.25 As some control probes such as somatostatin and substance K oligomers did not show the diffuse labeling pattern, we cannot entirely exclude the possibility that the diffuse labeling represents specific hybridization to high densities of cells with low CGRP mRNA. As available data do not provide a definitive explanation for the diffuse labeling, we have chosen to limit our discussion of the neuronal localization of CGRP mRNA to regions where hybridization signals were significantly higher than background and clearly overlay neuronal perikarya. Transfer RNA is commonly used as a carrier nucleic acidI with the premise that it might help inhibit non-specific probe binding to high abundance cellular RNAs. In certain nuclei, however, the addition of tRNA to hybridization buffer resulted in reduced hybridization signal or eliminated hybridization signal entirely. In other brain regions inclusion of tRNA in hybridization buffer gave no improvement of background hybridization with the CGRP oligomer. Using probes for other mRNAs such as oxytocin or glucagon we have found no benefits in the addition of tRNA to prehybridization and hybridization buffer (unpublished observations). We certainly have no conclusive information as to the explanation for the apparent reduction in hybridization signal intensity in specific brain regions when tRNA is present in the hybridization buffer. One possibility is that tRNA may associate with the probe by base pairing and impede probe penetration into tissue. The apparent selectivity of such a problem for neurons with specific brain nuclei is puzzling but might

be

due to the fact that

in areas

with

low

abundance of CGRP mRNA any reduction in probe availability may be s~~cient to severely reduce or

Fig. 10. Hybridization signal in the hypoglossal nucleus of the rat. (A) Dark-held picture showing accumulation of silver grains (arrows). (B) Bright-field view of the same area as in (A). Arrows point to some of the neurons showing hybridization signal. Scale bars = 100 pm. cc, central canal. Fig. 11. Strong hybridization signal in the ambiguus nucleus of the rat. (A) Dark-field picture showing the high density of silver grains in a circumscribed area and some weaker hybridization in two adjacent spots ventrally and laterally (arrows). (B) Bright-field view of the same area as in (A). Arrows point to neurons showing weak hybridization signal. (C) High-power picture showing the motoneurons with strong hybridization signal in the main portion of the nucleus. Scale bars: (A, B) IiIO~m: (C) 10vm.

234

M.

R~THELYI

et

al

Fig. 12. Hybridization signal in the peripeduncular nucleus of the rat. (A) Dark-field picture showing groups of silver grains (arrows). (B) Bright-field view of the same area as in (A). Arrows point to neurons showing strong hybridization signal. Insert (top): high-power picture of two neurons with hybridization signal labeled with ringed arrows. Scale bars: (A, B) IOOpm; insert, IOpm. Fig. 13. The effect of special buffer B on the intensity of hybridization in the peripeduncular nucleus of the rat. (A) Dark-field picture from a section treated during both prehybridization and hybridization with hybridization buffer not containing tRNA. Arrows point to groups of silver grains indicating labeled neurons. (B) Dark-field picture of the immediately adjacent section (mirror image position) treated with regular hybridization buffer. No clear hybridization signal is detectable. Asterisks on both figures indicate the same large vessel. Scale bars = 100 pm.

eliminate the hybridization signal. Whatever the explanation we conclude that the benefits of the use of tRNA as carrier nucleic acid for in situ hybridization with oligomer probes are questionable. In contrast,

we did find evidence for benefits of the use of oligomers as carrier nucleic acid based on reduced background and improved signal-to-noise ratio in the guinea-pig dorsal root ganglion when an unlabeled

Fig. 14. Hybridization signal in the hypoglossal nucleus of the guinea-pig. (A) Dark-field picture showing groups of silver grains, some of them are marked by arrows. (B) Bright-field view of the same area as in (A). Arrows point to neurons showing strong hybridization signal. Scale bars = 100pm. cc, central canal. Fig. 15. Hybridization signal in the accessory facial nucleus of the guinea-pig. (A) Dark-field picture showing clusters of silver grains (arrows). (B) Bright-field view of the same area as in (A). Arrows point to neurons showing strong hybridization signal. Asterisk in the upper right hand corner indicates the fibers of the facial nerve. Scale bars = 100pm. Fig. 16. Hybridization signal in the abducens nucleus of the guinea-pig. (A) Dark-field picture showing clusters of silver grains (arrows). (B) Bright-field view of the same area as in (A). Arrows point to neurons showing strong hybridization signal. Scale bars = 1OO~m. VII, knee of the facial nerve. 235

236

M. R~THELYI et al.

Fig. 17. Hybridization signal in the oculomotor nucleus of the guinea-pig. (A) Dark-field picture showing clusters of silver grains (arrows). (B) Bright-field view of the same areas as in (A). Arrows point to neurons showing hybridization signal. Scale bars = 100 pm.

oligomer

unrelated

was used as carrier

to the

labeled

CGRP

oligomer

DNA.

Distribution of neuronal CGRP mRNA in the rat compared with neuronal CGRP immunoreactivity The present in situ hybridization results in intact rats localized CGRP mRNA to spinal motoneurons, motoneurons in the VIIth and XIIth cranial nerve nuclei, motoneurons in the ambiguus nucleus, DRG neurons of various size and to small neurons in the parabrachial and peripeduncular nuclei. These findings are in close agreement with the distribution of cellular CGRP immunoreactivity in the rat re-

ported by Rosenfeld et aL2’ These neurons, therefore, appear to synthesize and translate CGRP mRNA. There were discrepancies, however, in the distribution of neuron populations labeled here by in situ hybridization and some reports of the distribution of CGRP immunoreactive neurons in rat CNS. CGRP immunoreactive neurons have been reported in the spinal parasympathetic nucleus,’ inferior colliculus, superior olive, motor nuclei of the IIIrd, IVth, Vth nucleus, and and VIth nerves, 3o Edinger-Westphal cerebellar Purkinje cells I2 in the rat. In our study these neurons did not show detectable labeling with the CGRP probe. Further studies will be necessary to

CGRP mRNAs in rat and guinea-pig

237

Table 1. Distribution of CGRP mRNA in cranial nerve motor nuclei Cranial nerve nuclei XII Nucleus ambiguus VII VII act. VI V IV III

Rat

Guinea-pig

Almost all cells are labeled Almost all cells labeled No labeled cells Large proportion of the cells are labeled Cells are labeled No labeled cells Many labeled cells No labeled cells Many intensively labeled cells No labeled cells Many labeled cells No labeled cells Many labeled cells

XII, hypoglossal nerve; VII, facial nerve; VII act., accessory facial nerve nucleus; VI, abducens nerve; V, trigeminal nerve; IV, trochlear nerve; III, oculomotor nerve.

address the reasons for these discrepancies. One possibility is that the abundance of CGRP in some or all of these regions is below the detection limit of the in situ hybridization procedure. Immunohistochemical localizations are performed mainly on colchicine-treated rat brain and the present in situ hybridization studies were performed on untreated rats. Another possible explanation for the discrepancies, therefore, is that colchicine in some way induces CGRP synthesis or higher levels of synthesis of neurons that were not labeled in the present study. This latter possibility is currently under investigation. Comparison of the distribution of neurons with CGRP mRNA between rat and guinea-pig

In the guinea-pig previous immunohistochemical data refer only to the detection of the peptide in the spinal cord,” DRG7.8,34 and sensory fibers of the heart,6 small intestine,4 cardiovascular system42 and urinary tract.34 The present data on localization of CGRP mRNA therefore provide new information about the distribution of CGRP synthesis in the guinea-pig nervous system and about similarities and differences in CGRP synthesis between the rat and guinea-pig nervous systems. The rat and guinea-pig showed similarities in the distribution of CGRP mRNA in that CGRP mRNA was localized to DRG neurons and to motoneurons of the spinal cord and brain stem. Both quantitative and qualitative differences were observed, however, in the distribution of CGRP mRNA in the two species. Quantitative differences included the finding that more DRG neurons were labeled with the CGRP probe in the rat compared with the guinea-pig. In addition, hybridization signals in neurons of guineapig DRG and spinal ventral horn were weaker than found in neurons of the same regions in the rat. The present in situ hybridization procedures are not strictly quantitative in that signal intensity can be related to absolute cellular mRNA abundance. It seems likely, however, that these lower signal intensities in guinea-pig DRG and ventral horn neurons reflect a lower cellular abundance of CGRP mRNA as the difference was reproducible across different animals and neurons in other brain regions

such as the hypoglossal nucleus showed similar hy-

bridization signal intensity in the rat and guinea-pig. More notable differences in the distribution of CGRP mRNA in the cranial nerve motor nuclei of the rat and guinea-pig suggest true qualitative species differences in sites of expression of CGRP mRNA or at least major differences in abundance of CGRP. As summarized in Table 1 the CGRP probe labeled neurons of particular cranial nerve motor nuclei of both the rat and guinea-pig, yet neurons of other nuclei were labeled only in the rat or in the guineapig. Another apparent species difference was the absence of detectable CGRP mRNA in guinea-pig parabrachial and peripeduncular nuclei while these regions were labeled in the rat. The nervous systems of the rat and guinea-pig also show differences in distribution of other peptides. Cholecystokinin occurs in several groups of neurons in the rat mesencephalon,” whereas cholecystokinincontaining neurons were not detected in guinea-pig mesencephalon. I3 There are many differences in the distribution of somatostatin immunoreactive neurons between the rat and guinea-pig.36 The motor nuclei innervating the extraocular muscles (IIIrd, IVth and VIth perves) are especially noteworthy. By not receiving the arborization of somatostatin-positive fibers they vary not only from homologous structures in the rat brain, but also from other cranial nerve nuclei in the guinea-pig brain stem. Based on the present data CGRP may be added to the list of peptides with differential distribution within the nervous systems of different species. Significance

of neuronal CGRP

While some peptides in the CNS are involved in control of the anterior pituitary function or have been implicated in interneuronal communication or modulation, the precise funtional significance of most neuropeptides, including CGRP, is yet to be defined. In the present and other studies, CGRP immunoreactivity and/or CGRP mRNA were localized to neurons of widely different function in the rat, such as primary sensory neurons and motoneurons. An area of concordance between CGRP mRNA localization in the present study and peptide localization in other studies suggests de ~OOO synthesis of cGRP in these neuronal groups. Some discrepancies be-

238

M. R~THELYI et al.

tween the distribution of CGRP mRNA found in intact rats in the present study and the distribution of CGRP immunoreactivity found in previous studies of colchicine-treated rats, suggest that definitive conclusions about some postulated sites of CGRP synthesis in the rat will require comparisons of CGRP mRNA and peptide distribution in both intact and colchicine-treated rats. Interspecies comparisons of peptide mRNA distribution do not directly address functional significance but the species differences in neuronal sites of CGRP expression found here between the rat and guinea-pig do have implications for the functional significance of CGRP. For example, if CGRP has a general role in neurons that innervate striated muscle, it is difficult to understand why CGRP-positive neurons innervate tongue, laryngeal, pharyngeal and facial musculature in the rat, whereas extraocular muscles do not receive CGRP-positive neurons. This is especially intriguing since identical extraocular muscles in the guinea-pig are innervated primarily by motoneurons that contain CGRP mRNA.

The above reasonings imply that there is not a generalized function of CGRP in motoneurons, but rather a specific role or roles in individual groups of neurons that will need to be sought by carefully designed experiments. Although the present study has not furnished any direct evidence for the function of CGRP in the CNS, it adds to our information about the similarities and differences in CGRP expression in diverse neuronal populations within and across species. Such information for CGRP and other neuropeptides may provide insight into the potential roles of neuropeptides and facilitate experimental design to elucidate their precise functional

significance.

Acknowledgements-The authors are thankful to Dr E. R. Per1 for support; to Mr D. McGehee and Mr H. Wooller for the labeling of the oligomer probes; to Drs P. Petrusz and 1. Merchenthaler for their help in the immunohistochemical procedure; and to Mrs R. Fervagner for technical assistance. This work was supported by NIH grant No. NS23804 (P.K.L.). M.R.‘s stay was sponsored by a collaborative arrangement between NIH, NINCDS (grant No. NSl0321) and the Ministry of Health. Budapest, Hungary.

REFERENCES 1. Amara S. G., Jonas V., Rosenfeld M. G., Ong E. S. and Evans R. M. (1982) Alternative RNA processing in calcitonin gene expression generates mRNAs encoding different polypeptide products. Nature 298, 24&244. 2. Cadieux A., Springall D. R., Mulderry P. K., Rodrigo J., Ghatei M. A., Terenghi G., Bloom S. R. and Polak J. M. (1986) Occurrence, distribution and ontogeny of CGRP immunoreactivity in the rat lower respiratory tract: effect of capsaicin treatment and surgical denervations. Neuroscience 19, 605627. 3. Clague J. R., Sternini C. and Brecha N. C. (1985) Localization of calcitonin gene-related peptide-like immunoreactivity in neurons of the rat gastrointestinal tract. Neurosci. Lest. 56, 63-69. 4. Feher E., Burnstock G., Varndell I. M. and Polak J. M. (1986) Calcitonin gene-related peptide immunoreactive tibres in the small intestine of the guinea pig: electron microscopic immunohistochemistry. Cell Tiss. Res. 245, 3533358. 5. Fujimura M., Hancock M. B., Cooper C. W., Reumont K., Greeley G. H. Jr and Thompson J. C. (1985) Immunocytochemical localization of calcitonin gene-related peptide in pancreatic islet cells of the rat. Gasfroenterolog~ 88, 1390. 6. Gerstheimer F. P. and Metz J. (1986) Distribution of calcitonin gene-related peptide like immunoreactivity in the guinea pig heart. Anal. Embryol. 175, 255-260. I. Gibbins 1. L., Furness J. B., Costa M., MacIntyre I., Hillyard C. J. and Girgis S. (1985) Co-localisation of calcitonin gene related peptide-like immunoreactivity with substance P in cutaneous, vascular and visceral sensory neurons of guinea pigs. Neurosci. Left. 57, 125-130. 8. Gibson S. J., Polak J. M., Bloom S. R., Sabate I. M., Mulderry P. M., Ghatei M. A., McGregor G. P., Morrison J. F. B., Kelley J. S., Evans R. M. and Rosenfeld M. G. (1984) Calcitonin gene-related peptide immunoreactive in the spinal cord of man and of eight other species. J. Neurosci. 4, 3101-31 Il. R., Rerup C., Sundler F. and Uddman R. (1986) Calcitonin gene-related peptide 9. Grunditz T., Ekman R., Hakanson in thyroid nerve fibers and C cells: effects on thyroid hormone secretion and response to hypercalcemia. Endocrinology 119, 2313-2314. 10. Hanko J., Hardebo J. E., Kahrstrom J., Owman C. and Sundler F. (1985) Calcitonin gene-related peptide is present in mammalian cerebrovascular nerve fibres and dilates pial and peripheral arteries. Neurosci. Let!. 57, 91-95. II. Honda C. N., Rethelyi M. and Petrusz P. (1983) Preferential immunohistochemical localization of vasoactive intestinal polypeptide (VIP) in the sacral spinal cord of the cat: light and electron microscopic observations. J. Neurosci. 3, 2183-2196. 1. and Tohyama M. (1985) 12. Kawai Y. Takami K., Shiosaka S., Emson P. C., Hillyard C. J., Girgis S., MacIntyre Topographic localization of calcitonin gene-related peptide in the rat brain: an immunohistochemical analysis. Neuroscience 15, 741-163. 13. Larsson L. I. and Rehfeld J. F. (1979) Localization and molecular heterogeneity of cholecystokinin in the central and peripheral neurons systems. Brain Res. 165, 201-218. J. (1982) Molecular Cloning-A Laboratory Manual. Cold Spring Harbor, 14. Maniatis T., Frisch E. F. and Sambrook New York. 15. Matesz C. and Szekely G. (1983) The motor nuclei of the glossopharyngeal, vagal and the accessorius nerves in the rat. Acta biol. hung. 34, 215-230. 16. Matsuyama T., Wanaka A., Yoneda S., Kimura K., Kamada T., Girgis S., MacIntyre I., Emson P. C. and Tohyama M. (1986) Two distinct calcitonin gene-related peptide-containing peripheral neurons systems: distribution and quantitative differences between the iris and cerebral artery with special refrence to substance P. Brain Res. 373, 205-212. 17. Maxam A. and Gilbert W. (1977) A new method for sequencing DNA. Proc. naln. Acad. Sri. U.S.A. 74, 560-564.

CGRP mRNAs in rat and guinea-pig

239

18. Merchenthaler I., Csernus V., Petrusz P. and Mess B. (1988) New data on the immunocytochemical localization of thyrotropin releasing hormone (TRH) in the rat central nervous system. Am. J. Anat. 181, 359-376. 19. Morishima Y., Takagi H., Akai F., Tohyama M., Emson P. C., Hillyard C. J., Girgis S. I. and MacIntyre I. (1985) Light and electron microscopic studies of calcitonin gene-related peptide-like immunoreactive neurons and axon terminals of the nucleus of the tractus solitarius of the rat. Brain Res. 344, 191-195. 20. Morris G. R., Panic0 M., Etienne T., Tippins J., Girgis S. I. and MacIntyre I. (1984) Isolation and characterization of human calcitonin gene-related peptide. Nature 308, 746748. 21. Mulderry P. K., Ghatei M. A., Rodrigo J., Allen J. M., Rosenfeld M. G., Polak J. M. and Bloom S. R. (1985) Calcitonin gene-related peptide in cardiovascular tissues of the rat. Neuroscience 14, 947-954. 22. Nelkin B. D., Rosenfeld K. I., De Bustros A., Leong S. S., Roos B. A. and Baylin S. B. (1984) Structure and expression of a gene encoding human calcitonin and calcitonin gene related peptide. Biochem. biophys. Res. Commun. 123,648-655. 23. Petterson M., Ahren B., Bottcher G. and Sundler F. (1986) Calcitonin gene-related peptide: occurrence in pancreatic islets in the mouse and the rat and inhibition of insulin secretion in the mouse. Endocrinology 119, 865-869. 24. Priestley J. V., Hynes M. A., Han V. K. M., Rtthelyi M., Per1 E. R. and Lund P. K. (1988) In siru hybridization using 32P labelled oligodeoxyribonucleotides for the cellular localisation of mRNA in neuronal and endocrine tissue. Histochemistry 89, 467479.

25. Rethelyi M., Metz C. B., Gijrcs T. J. and Lund P. K. (1989) In siru hybridization of peptide mRNA-s on vibratome sections. Acta morph. hung. (in press). 26. Rodrigo J., Polak J. M., Femandez L., Ghatei M. A., Mulderry P. and Bloom S. R. (1985) Calcitonin gene-related peptide (CGRP) immunoreactive sensory and motor nerve of the rat, cat and monkey oesophagus. Gasrroenrerology -_ _. 27. Rodrigo J., Polak J. M., Terenghi G., Cervantes C., Ghatei M. A., Mulderry P. and Bloom S. R. (1985) Calcitonin gene-related peptide (CGRP) immunoreactive sensory and motor nerves of the mammalian palate. Hisrochemisrry 82, 61-74.

28. Rosenfeld M. G., Mermod J.-J., Amara S. G., Swanson L. W., Sawchenko P. E., Rivier J., Vale W. W. and Evans R. M. (1983) Production of a novel neuropeptide encoded by the calcitonin gene via tissue-specific RNA processing. Narure 304, 129-135.

29. Siegel R. E. and Young W. S. III (1985) Detection of preprocholecystokinin and preproenkephalin A mRNAs in rat brain by hybridization histochemistry using complementary RNA probes. Neuropeptides 6, 573-580. 30. Skofitsch G. and Jacobowitz D. M. (1985) Calcitonin gene-related peptide: detailed immunohistochemical distribution in the rat central nervous system. Peprides 6, 721-745. 31. Skofitsch G. and Jacobowitz D. M. (1985) Quantitative distribution of calcitonin gene-related peptide in the rat central nervous system. Peprides 6, 1069-1073. 32. Steenbergh P. H., Hoppener J. W. M., Zandberg J., Van de Ven W. J. M., Jansz H. S. and Lips C. J. M. (1984) Calcitonin gene-related peptide coding sequence is conserved in the human genome and is expressed in medullary thyroid carcinoma. J. clin. Endocrinol. Merab. 59, 358-360. 33. Sternberger L. A., Hardy P. H. Jr, Cuculis J. J. and Meyer H. G. (1970) The unlabeled antibody enzyme method of immunohistochemistry. Preparation and properties of soluble antigen-antibody complex (horseradish peroxidaseantiperoxidase) and its use in identification of spirochetes. J. Hisrochem. Cyrochem. 18, 315-333. 34. Su H. C., Wharton J., Polak J. M., Mulderry P. K., Ghatei M. A., Gibson S. J., Terenghi G., Morrison J. F. B., Ballesta J. and Bloom S. R. (1986) Calcitonin gene-related peptide imunoreactivity in afferent neurons supplying the urinary tract: combined retrograde tracing and immunohistochemistry. Neuroscience 18, 727-747. 35. Szekely G. and Matesz C. (1982) The accessory motor nuclei of the trigeminal, facial and abducens nerves in the rat. J. camp. Neural. 210, 258-264.

36. Taber-Pierce E., Lichtenstein E. and Feldman S. C. (1985) The somatostatin systems of the guinea pig brainstem. Neuroscience 15, 215-235. 37. Takami K., Kawai Y., Uchida S., Tohyama M., Shiotani Y., Yoshida H., Emson P. C., Hillyard C. J. and Maclntyre I. (1985) Effect of calcitonin gene-related peptide on contraction of striated muscle in the mouse. Neurosci. Lerr. 60, 227-230.

38. Terenghi G., Polak J. M., Ghatei M. A., Mulderry P. K., Butler J. M. Unger W. G. and Bloom S. R. (1985) Distribution and origin of calcitonin gene-related peptide (CGRP) immunoreactivity in the sensory innervation of the mammalian eye. J. camp. Newrol. 233, 506516. 39. Terenghi G., Polak J. M., Rodrigo J. M., Mulderry P. K. and Bloom S. R. (1986) Calcitonin gene-related peptide immunoreactive nerves in the tongue, epiglottis and pharynx of the rat: occurrence, distribution and origin. Brain Res. 365, I-14.

40. Tschopp F. A., Tobler P. H. and Fischer J. A. (1984) Calcitonin gene-related peptide in the human thyroid, pituitary and brain. Molec. cell. Endocrinol. 36, 53-51. 41. Vanderhaeghen J.J. (1985) Neuronal cholecystokinin. In Handbook of Chemical Neuroanatomy, Vol. 4: GABA and Neuropeptides in rhe CNS, Part I (eds A. Bjijrklund and T. Hiikfelt). Elsevier, Amsterdam. 42. Wharton J., Gulbenkian S., Mulderry P. K., Ghatei M. A., McGregor G. P., Bloom S. R. and Polak J. M. (1986) Capsaicin induces a depletion of calcitonin gene-related peptide (CGRP) immunoreactive nerves in the cardiovascular system of the guinea pig and rat. J. Autonomic Nero. Sysr. 16, 289-309. (Accepted 20 July 1988)