- Email: [email protected]
Immunoelectron microscopy of β-endorphinergic synaptic innervation of nitric oxide synthase immunoreactive neurons in the dorsal raphe nucleus
BRAIN RESEARCH Brain Research684 (1995) 185-193
ELSEVIER
Research report
Immunoelectron microscopy of ¢l-endorphinergic synaptic innervation of nitric oxide synthase immunoreactive neurons in the dorsal raphe nucleus O.-P. Wang, Y. Nakai * Department of Anatomy, Showa University School of Medicine, Hatanodai 1-5-8, Shinagawa-ku, Tokyo 142, Japan Accepted 14 March 1995
Abstract On the basis of the comparing of the distribution of fl-endorphin-like immunoreactive neuronal fibers and nitric oxide synthase-like immunoreactive neurons in the dorsal raphe nucleus, the synapses between the two immunocytochemically identified neurons were studied with a modified DAB-silver-gold intensification double immunostaining technique at the electron microscopic level. Although both of them can be found in the mediodorsal and medioventral parts of the dorsal raphe nucleus, the synapses between them could only be found in the mediodorsal part. The majority of the /3-endorphin-like immunoreactive neuronal fibers contained many dense-cored vesicles. The synapses made by fl-endorphin-like immunoreactive neuronal axon terminals on nitric oxide synthase-like immunoreactive neurons were both symmetrical and asymmetrical with the former predominant, especially in the axo-dendritic ones. /3-Endorphin-like immunoreactive perikarya could only be found in the ventrobasal hypothalamus. These findings suggest the possibility that the /3-endorphin- producing neurons in the ventrobasal hypothalamus could influence nitric oxide synthase-containing neurons in the dorsal raphe nucleus by synaptic relations.
Keywords: Dorsal raphe nucleus; /3-Endorphin; Nitric oxide synthase; Immunocytochemistry; Synapse; Electron microscopy 1. Introduction Dorsal raphe nucleus (DRN) is one of the most important nuclei in pain modtdation [37]. The serotonergic, GABAergic and enkephalinergic neurons in the DRN have been known as involved in this modulation and the synaptic relations between these neurons have also been studied [36,38,39] and reviewed [2,7]. Besides these, /3-endorphin (EP) in the DRN is also involved in the processes of pain modulation [23,26,27]. Recently, nitric oxide synthase (NOS)-containing neurons in the spinal cord have been reported to be involved in nociceptive processing in spinal
Abbreviations: ABC, avidin-biotin-peroxidase complex; ACTH, adrenocorticotropic hormone; DAB, 3,3'-diaminobenzidine tetrahydrochloride; DRN, dorsal raphe nucleus; ENK, enkephalin; EP, fl-endorphin; EP-LI, /3-endorphin-like immunoreactive;GABA, ~/-aminobutyricacid; 5-HT, serotonin; L-NAME, N'°-nitro-L-argininemethyl ester; NO, nitric oxide; NOS, nitric oxide syntha~'se;NOS-LI, nitric oxide synthase-like immunoreactive; NTS, nucleus tractus solitarii; PAP, peroxidase-antiperoxidase; SGI, silver-goldintensification; VBH, ventrobasalhypothalamus
* Corresponding author. Fax: (81) (3) 3784-6815. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 5 ) 0 0 4 1 8 - 1
cord [21,22,28,40] and as spinalthalamic projection neurons [16]. However, in the higher central nervous system, NOS may activate the antinociception induced by /3-endorphin [31], although NOS itself has been reported to cause hyperalgesia in the central nervous system [12]. Among the three most antinociceptive central nervous structures, the ventrolateral periaqueductal gray, DRN, and nucleus raphe magnus with its adjacent structures, only the DRN has a large number of NOS-containing neurons whereas the periaqueductal gray has its NOS-containing neurons only at its dorsal part [7,24,34]. On the other hand, the DRN also contained /3-endorphin-like immunoreactive (EP-LI) fibers [8,11,14,25]. Thus, the direct synaptic relation between the two kinds of neurons, if any, may occur within the nucleus. For this reason, a study was made into such a synaptic relation by preembedding double immunocytochemistry at the electron microscopic level.
2. Materials and methods Experiments were performed on eight adult, Wistar albino rats of 180-220 g in weight. Two were used for
186
Q.-P. Wang, Y. Nakai / Brain Research 684 (1995) 185-193
light microscopic study of EP and NOS immunostaining by the avidin-biotin-peroxidase complex (ABC) method [10]. Six were used for study of their synaptic relations at ultrastructural level by using the combined A B C - D A B silver-gold intensification and double PAP method.
2.1. Antisera Rabbit anti-NOS antibody was obtained from Euro-Diagnostica, Batch no. R 922301, made from the C-terminal of the cloned rat cerebella NOS, coupled to bovine serum albumin. Specificity was confirmed by the absorption test. The staining was abolished by 10 to 100 mg immunogen per ml diluted antiserum. Rabbit anti-EP antibody was Peninsula Lab. Lot no. 025659-1, made from rat EP, having 100% cross-reactivity to porcine, horse, and camel EP, but no cross-reaction with ACqTq, Met-ENK, 3'-EP, c~-MSH and PACAP 38. The ABC Kit purchased from Vector Labs contained biotinylated anti-rabbit IgG, and ABC reagents A and B.
2.2. Immunostaining procedures for light microscopy For the light microscopic study, the ABC method was used. The animals were deeply anesthetized with pentobarbital (50 mg/kg, i.p.), and perfused through the ascending aorta with 100 ml cold 0.9% sodium chloride, followed by 300 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were removed immediately and postfixed in a 4% paraformaldehyde fixative at 4°C overnight. The blocks of brain containing the DRN, the ventrobasal hypothalamus (VBH) and nucleus tractus solitarii (NTS) in the whole oral-caudal length were cut into serial 20 /zm thick coronal sections with a cryostat. The sections were rinsed several times with 0.1 M phosphate buffer containing 0.3% Triton X-100 (pH 7.4), and immersed in the same solution at 4° C overnight. The sections of midbrain were divided into two groups, each having all levels of the DRN. After preincubating in 5% normal goat serum for 10 min at room temperature, one section group was incubated with 1:2000 diluted rabbit anti-NOS, and the other was incubated with 1:1000 diluted rabbit anti-EP antibody whereas the sections containing VBH and NTS were only incubated with 1:1000 diluted rabbit anti-EP antibody. Both of the antibody solutions contained 0.3% Triton X-100. The incubation lasted for 2 h at room temperature and overnight at 4° C. The sections were then thoroughly rinsed with phosphate buffer containing 0.3% Triton X-100 (pH 7.4), and incubated with biotinylated anti-rabbit IgG for 1.5 h, and then with ABC complex for 2 h at room temperature. The sections were then treated with 3,3'-diaminobenzidine (DAB) in 0.05 M Tris-HC1 buffer including 0.005% hydrogen peroxide for 5 to 10 min in a microscope until the immunoreactivity was visible. Between the above steps, each section was carefully
rinsed three times with 0.1 M phosphate buffer containing 0.3% Triton X-100 (pH 7.4), except between the normal goat serum and the first antiserum. After the DAB reaction, all sections were carefully mounted on gelatin-coated microscope slides, dried at room temperature, and sealed. All the sections were photographed for comparing the distribution of the NOS-LI neurons with that of the EP-LI fibers in the DRN.
2.3. Immunostaining procedures for electron microscopy The experimental procedure was modified from the article reported previously [33]. Under deep Nembutal anesthesia (40 m g / k g b.wt., intraperitoneally), the animals were perfused through the ascending aorta with 100 ml of 0.9% saline followed by 300 ml of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were removed immediately and postfixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) at 4° C for 2 - 4 h. The blocks of brain containing the DRN were cut into serial 40 gm-thick coronal sections with an Oxford vibratome. The sections were rinsed several times with 0.1 M phosphate buffer (pH 7.4), and immersed in the same solution at 4 ° C overnight. Sections were treated with 0.1% Triton X-100 containing phosphate buffer for 20 min and incubated in 2% normal goat serum for 20 min at room temperature. For studying fl-endorphin, the first antigen being detected, ABC method [10] was applied. On the other hand, for studying NOS, the second antigen being detected, PAP double bridge technique [18,32] was used. In brief, the sections were incubated in the 1:1000 diluted anti-/3-endorphin antiserum, for 2 h at room temperature and overnight at 4° C. The next day, the sections were incubated with biotinylated anti-rabbit IgG for 45 min, and then with ABC complex for 1 h at room temperature. The sections were then treated for 5 - 7 min in the dark with DAB in 0.05 M Tris-HC1 (pH 7.6) buffer including 0.005% hydrogen peroxide. Between the above steps, each section was carefully rinsed three times with 0.1 M phosphate buffer, except between the normal goat serum and the first antibody. The sections that had good PAP-DAB reaction were treated with the silver-gold intensification [17,35,36] to darken the brown-colored DAB reaction material to black in a light microscope. After that, the sections were transferred for detecting NOS. First, the sections were incubated in 1:2000 diluted NOS antiserum, for 2 h at room temperature and 2 days at 4° C. After that, the sections were immunostained with goat-anti-rabbit IgG for 1.5 h, peroxidase-anti-peroxidase complex (PAP) for 2 h and again with goat-anti-rabbit IgG for 1.5 h, PAP for 2 h at room temperature and then for DAB reaction similar to the procedure previously mentioned with detecting EP, but no silver-gold-intensifyin procedures were made. For examination in the electron microscope, the sections were posffixed with 1% OsO 4 in 0.1 M phosphate buffer (pH 7.4) for 1 h at 4° C, dehydrated in an ethanol series, and
Q.-P. Wang, Y. Nakai/Brain Research 684 (1995) 185-193
187
Fig. 1. and Fig. 2. 1: Light micrographs taken from the neighboring sections showing the distribution of EP-LI fibers (A) and NOS-LI neurons (B). Note that, although the majority of the EP-LI fibers are in the ventrolateral part of the PAG, some varicosity-like fibers are also in the DRN. The inset in the A showing the EP-LI neuronal peril~u3,a in the arcuate nucleus. Bars = 0.1 ram. 2: Electron micrographs (no electron staining) showing EP-LI axons (EP) in the medioventral DRN. A, B: the very thin, long fibers; C, D: the axon terminals containing large dense-cored vesicles. Bars = 0.2 pro. Fig. 2. Electron micrographs (no electron staining) showing EP-LI axon-terminals (EP) synapse (arrows) on non-immunoreactive dendrites. Note the synapse in the A is an asymmetrical one. Bars = 0.2 /~m.
188
Q.-P. Wang, Y. Nakai /Brain Research 684 (1995) 185-193
embedded in a mixture of Epon-Araldite. Ultrathin sections were cut and examined under a JEM-1200EXII electron microscope. Neither lead citrate and uranyl acetate staining nor carbon coating was made on these ultrathin sections.
3. R e s u l t s
3.1. Distribution
Under light microscopy, although clustered immunoreactive fibers were seen mainly in the ventrolateral part of the periaqueductal gray (Fig. 1A), the EP-LI fibers were also seen in the DRN especially in the mediodorsal part of the DRN (Fig. 1A). Only the most anterior part of the DRN had the EP-LI fibers. Most of the fibers had varicosity appearance. On the other hand, many immunoreactive NOS neurons were also seen in both the mediodorsal and medioventral cell groups in the anterior part of the DRN in the adjacent section (Fig. 1B), although the cell bodies were also seen in the posterior part. Many immunoreactive EP-LI perikarya were found in the V B H (inset in Fig. 1A) but not in NTS.
3.2. Ultrastructure and synaptic relations
Under electron microscopy, the NOS-LI elements showed moderate immunostaining by the appearance of DAB reacted products whereas the EP-LI axon terminals showed many silver-gold particles loaded on the DAB reacted products. The loading of the silver-gold particles on the dense-cored vesicles was much stronger than that on the small, clear vesicles. Sometimes, small, clear vesicles that far from dense-cored vesicles also showed some silver-gold particles on them. Occasionally, some EP-LI axon terminals were found to make synapses (Fig. 3). No immunoreactive EP-LI perikarya and dendrites could be found. Although there were some EP-LI axon terminals near blood vessels, none were found near the central aqueduct. In this study, the mediodorsal part (Fig. 3 Fig. 4 Fig. 5) and medioventral part (Fig. 2) of the DRN were observed separately. Most of the EP-LI fibers (73%, 41 out of 56) in the medioventral part of the DRN seemed to be very thin, long axons (Fig. 2A, B) in the dorsoventral direction, although some varicosities containing large dense-cored vesicles (25%, 1 4 / 5 6 ) were also found (Fig. 2C,D), whereas the terminal-like EP-LI fibers, which had mainly
Fig. 3. Electron micrographs (no electron staining) showing EP-LI axon-terminals (EP) synapse (arrows) on NOS-LI neurons (NOS). The synapses
between EP-LI axon-terminals and NOS-LI dendrites are asymmetrical (A) or symmetrical (B). C shows two EP-LI axon-terminals contacting a NOS-LI dendrite, but only one can be identified as a synapse of symmetrical type. D shows an EP-LI axon terminal synapse on NOS-LI neuron perikaryon. Note all the five axon terminals in the figure containing dense-cored vesicles. Bar = 0.2 /.tin.
Q.-P. Wang, Y. Nakai/Brain Research 684 (1995) 185-193 small, clear vesicles, were rare (2%, 1 / 5 6 ) . About onetenth ( 6 / 5 6 ) of those fibers were found to contact with NOS-LI elements by the appearance of cell bodies ( 2 / 5 6 ) or dendrites ( 3 / 5 6 ) and even axon terminals ( 1 / 5 6 ) , but
189
none of them showed synaptic appearance as no active zones of the membranes were found in the present observation. In the mediodorsal part of the DRN, the quantity of the
Fig. 4. Electron micrograph (no electron staining) showing an EP-LI axon terminal receiving a synapse(curved arrow) from a non-immunoreactiveaxon terminal and in turn, making a synapse(arrow) to a non-immunoreactivedendrite. Note the axon terminalcontainingsome dense-coredvesicles. Bar = 0.2 /,I,m.
190
Q.-P. Wang, Y. Nakai / Brain Research 684 (1995) 185-193
ing synapse from a non-immunoreactive axon terminal, and in turn, making a synapse with a non-immunoreactive dendrite (Fig. 5).
4. Discussion 4.1. Methodology
Fig. 5. Electronmicrograph(no electronstaining)showingan EP-LIaxon terminal receiving a synapse (curved arrow)from a non-immunoreactive axon terminal and in turn, making a synapse (arrow) to a non-immunoreactive dendrite. Note the axon terminal containing some dense-cored vesicles. Bar = 0.2 /zm. EP-LI fibers was much more than that in the medioventral part. In total, 497 EP-LI axon profiles were photographed. Although the long, thin fibers were also seen in this part, the frequency was lower (7%, 37/497). Most of the EP-LI axon terminals (87%, 398/460) contained a mixture of large dense-cored vesicles and small clear vesicles or predominant dense-cored vesicles, although the terminals containing only small, clear vesicles were also found with lower frequency (13%, 62/460). The occurrence of identifiable synapse made by EP-LI terminals was not very much (31%, 141/460). Greater than the ratio in the medioventral part of the DRN, about one-third (35%, 160/460) of the EP-LI axon terminals were found to be contacted by NOS-LI profiles. Some EP-LI terminals were found making synapses (12%, 56/460) with NOS-LI dendrites (Fig. 4A,B,C) by the appearance of synaptic active zone in the contact region. Most of the terminals contained dense-cored vesicles whereas the vesicles immediately near the presynaptic membranes were usually of the small, clear type. The synapses were both asymmetrical (Fig. 4A; 4%, 20/460) or symmetrical (Fig. 4B,C) with the later predominant (8%, 36/460). The synapses between EP-LI terminal and NOS-LI perikarya were rare (Fig. 4D; 1%, 4/360). In this case, the synapses were also both asymmetrical (2 cases) and symmetrical (2 cases). In only one case, an EP-LI terminal was found receiv-
The DAB-SGI method has been proved as a useful tool to study the synaptic relations between two immunocytochemically identified neurons [33,36]. Its advantages and disadvantages have been discussed [35]. One of the disadvantages of the method is that the severe pH condition in the procedure often made the detecting of second antigen difficult, especially big molecules. In this study, we also encountered this problem because NOS, the enzyme that catalyzes L-arginine into L-citrulline and nitric oxide (NO) [15], is a large molecule. The immunoreactivity of the NOS, the second antigen detected in the present study, is not so satisfactory after DAB-SGI treatment. To overcome this problem, we used the double bridge PAP method [18,32], and the result showed that the modification is effective. Also, we omitted the usual electron staining of the ultrathin sections with uranium and lead. The omissions reduced the clearance of the cell membranes to some extent and induced the difficulty of focusing while taking photographs, but the contrast between the immunopositive and immunonegative neuronal elements was enhanced. The fact that loading of the silver-gold particles on the dense-cored vesicles, which contained /3-endorphin molecules, was much more than that on the small, clear vesicles suggests the reaction specificity of the DAB-SGI method in another aspect. 4.2. Morphology
EP-LI fiber has been described as existing in the DRN of rat [8], monkey [11] and human brain [25]. In the present study, we confirmed this finding in rat. Although we studied the whole rostral-caudal length of the DRN, only the rostral-most part of the DRN was found having EP-LI fibers. In the present study, we found 50% or more of the EP-LI axon terminals in the DRN were rich in dense cored-vesicles. This finding is identical to the report of ACTH-immunoreactive varicosities, which were thought to coexist with fl-endorphin (EP) in periaqueductal gray [4], although the ultrastructure of the EP-LI varicosities has been described as containing many small, clear vesicles and occasionally dense-cored vesicles in the hypothalamus [9]. However, the ultrastructural morphology of a vast majority ACTH-immunoreactive varicosities in the rat DRN was described as containing many small clear vesicles and only a very small number of varicosities principally filled with dense core vesicles [19]. We do not know
Q.-P. Wang, Y. Nakai/Brain Research 684 (1995) 185-193
why the morphological difference exists between the two peptides in the same nucleus. The difference could perhaps be because although they are coexistent in the same neurons in the VBH, they are not coexistent in the same varicosities in the DRN of the rat. The difference of the morphology of EP-LI axon varicosities between the mediodorsal and medioventral parts of the DRN is noticeable. This discrepancy maybe implies different functions of EP-LI axons in the different regions. We found some EP-LI axon terminal-making synapses with NOS-LI neurons in the mediodorsal part of the DRN and with the negative result within the medioventral part, the main part of the DRN, although we cannot describe this result as having no such synapses because the EP-LI axon terminals were rare. With regard to the synapses made by EP-or ACTH-immunoreactive axon terminals, there is also a discrepancy. In the hypothalamus, they were described as mainly symmetrical by EP immunostaining [9], whereas in the DRN they were described as usually asymmetric [19]. In the present study, we found most of the synapses made by EP-LI axon terminals were symmetrical although a few seemed asymmetrical. This difference between EP and ACTH immunoreactivity maybe also imply that the two neurotransmitters were not coexistent in the same axon terminals. We tend to agree that most of the synapse-made EP-LI axon terminals on the NOS-LI neurons are symmetric because DAB reacted products often adherent on the postsynaptic membranes to cause the pseudo asymmetriclike synapse. However, because asymmetrical synapses were also found between the EP-LI axon terminals and immunonegative neurons in the present study, and NOS-LI neurons often received asymmetrical synaptic afferent [34], asymmetrical synapses on the NOS-LI neurons by EP-LI axon terminals are also possible. /3-EP-producing neuronal perikarya were first reported within the VBH [2] and were confirmed recently [8,11]. Also recently, a small neuron group was reported in NTS as /3-EP-LI [29] and was confirmed by in situ hybridization study [3]. However, in the present study, we did not find any /3-EP-LI neurons; in the NTS, thus, it would appear that the /3-EP-LI axons in the DRN may originate from the VBH.
191
receptor than to other receptors [30]. Recently, it has been discovered that, in the higher central nervous system, NOS-containing neurons may be important in modulating the antinociception induced by fl-endorphin because Larginine potentiates antinociception induced by fl-endorphin, and the potentiating effect of L-arginine is attenuated by N'%nitro-L-arginine methyl ester (L-NAME), a selective inhibitor of NOS [31]. In the present study, we found some direct synapses between EP-LI axon terminal and NOS-LI neurons. This finding may provide morphological evidence of the physiological suggestion by Tseng et al. [31]. Because arcuate nucleus stimulation could inhibit the reflex of parafascicular neurons caused by noxious stimulation [3], and treatment of DRN lesion, microinjection of naloxone, or anti-EP antiserum could reverse the effect of arcuate stimulation [4], the DRN has been suggested to be involved in arcuate nucleus stimulation-caused pain inhibition [6]. In the present study, we found EP-LI neuron perikarya only in the arcuate nucleus, so the finding of the synapses made by EP-LI axon terminals in the DRN may provide morphological evidence of the physiological studies by Chen et al. [5,6]. In the DRN, the enkephalinergic neurons predominantly influence the serotonergic neurons indirectly through an intemeuron, the GABAergic neuron, by synapses on the GABAergic neurons [36] and in turn, by the synapses between GABAergic axon terminals and serotonergic neurons [39]. In the present study, EP-LI axon terminals made fewer synapses on NOS-LI neurons in the DRN than on non-immunoreactive neurons. Thus, it is not impossible that, similar to the enkephalinergic synaptic influence on serotonergic neurons, the EP-LI axon terminals could also influence NOS-LI neurons indirectly, although what is the interneuron between the EP-LI axon terminals and NOS-LI neurons needs to be studied further. It is noticeable that some authors [13] suggested that NO, the gas made by NOS, may cause the hyperalgesia itself, and the intracerebroventricular administration of L-arginine produced antinociception through another route, although they did not show if the effect of L-arginine was attenuated by L-NAME or not [13] as showed by Tseng et al. [31]. The mechanism of the involvement of NOS in the central nervous system in pain perception and pain modulation should be studied further.
4.3. Possible function
EP is one of the opioids that involved in the pain modulation [1,37]. Physiologically, EP has been reported to be involved in ventral periaqueductal gray (including the DRN) stimulation-induced analgesia [23]. Neurochemically, noxious stimulation elevated EP-LI level in the ventral periaqueductal gray including the DRN [26,27]. This concept has been further confirmed by the fact that the DRN has a large quantity of /z receptor [20], because /~ receptor is more effective in pain modulation than K and 8 receptors and EP has a higher bonding to /~
Acknowledgements This study was supported by research grants from the Ministry of Education, Science and Culture, Japan.
References [1] Basbaum, A.I. and Fields, H.L., Endogenouspain control system: brainstem spinal pathways and endorphin circuitry, Annu. Rev. Neurosci., 7 (1984) 309-338.
192
Q.-P. Wang, Y. Nakai / Brain Research 684 (1995) 185-193
[2] Bloom, F., Battenberg, E., Rossier, J., Ling, N. and Ouillemin, R., Neurons containing fl-endorphin in rat brain exist separatery from those containing enkephalin: Immunocytochemical studies, Proc. Natl. Acad. Sci. USA, 75 (1978) 1591-1595. [3] Bronstein, D.M., Schafer, M.K., Watson, S.J. and Akil, H., Evidence that beta-endorphin is synthesized in cells in the nucleus tractus solitarius: detection of POMC mRNA, Brain Res., 587 (1992) 269-275. [4] Buma, P., Veening, J. and Nieuwenhuys, R., Ultrastructural characterization of adrenocorticotrope hormone (ACTH) immunoreactive fibers in the mesencephalic central grey substance in the rat, Eur. J. Neurosci., 1 (1989) 659-672. [5] Chen, X.-Y., Duanmu, Z.-X., Jiang, X.-H., Yin, W.-P. and Yin, Q.-Z., Effect of hypothalamic arcuate stimulation on the pain-evoked unit discharges of thalamic parafascicular nucleus: A preliminary analysis, Acta Physiol. Cin., 38 (1986) 26-32. [6] Chen, X.Y., Yin, W.P. and Yin, Q.Z., The dorsal raphe nucleus is involved in the inhibitory effect of hypothalamic arcuate stimulation on pain-evoked unit discharges of the thalamic parafascicular nucleus, Acta Physiol. Cin., 39 (1987) 46-53. [7] Dun, N.J., Dun, S.L. and Frrstermann, U., Nitric oxide synthase immunoreactivity in rat pontine medullary neurons, Neuroscience, 59 (1994) 429-445. [8] Finley, J.C., Lindstrom, P. and Petrusz, P., Immunocytochemical localization of beta-endorphin-containing neurons in the rat brain, Neuroendocrinology, 33 (1981) 28-42. [9] Horvath, T.L., Naftolin, F. and Leranth, C., Beta-endorphin innervation of dopamine neurons in the rat hypothalamus: a light and electron microscopic double immunostaining study, Endocrinology, 131 (1992) 1547-1555. [10] Hsu, S.-M., Raine, L. and Fanger, H., Use of avidin-biotin- peroxidase complex (ABC) in immunoperoxidase techniques. A comparison between ABC and unlabeled antibody (PAP) procedures, J. Histochem. Cytochem., 29 (1981) 577-580. [11] Ibuki, T., Okamura, H., Miyazaki, M., Yanaihara, N., Zimmerman, E.A. and Ibata, Y., Comparative distribution of three opioid systems in the lower brainstem of the monkey (Macaca fuscata), J. Comp. Neurol., 279 (1989) 445-456. [12] Kawabata, A. and Takagi, H., The dual role of L-arginine in nociceptive processing in the brain: involvement of nitric oxide and kyotophin. In H. Takagi, N. Toda and R.D. Hawkins (Eds.), Nitric Oxide: Roles in Neuronal Communication and Neurotoxicity, Japan Scientific Societies Press, Tokyo, 1994, pp. 115-125. [13] Kawabata, A., Umeda, N. and Takagi, H., L-Arginine exerts a dual role in nociceptive processing in the brain: involvement of the kyotophin-Met-enkephalin pathway and NO-cyclic GMP pathway, Br. J. Pharmacol., 109 (1993) 73-79. [14] Khachaturian, H., Lewis, M.E., Tsou, K. and Watson, S.J., ~Endorphin, o~-MSH, ACTH, and related peptides. In A. BjSrklund and T. H5kfelt (Eds.), GABA and Neuropeptides in the CNS, Handbook of Chemical Neuroanatomy, Vol. 4, Elsevier, Amsterdam, 1985, pp. 216-272. [15] Knowles, R.G., Palacios, M., Palmer, R.M.J. and Moncada, S., Formation of nitric oxide from L-arginine in the central nervous system: A transduction mechanism for stimulation of the soluble guanylate cyclase, Proc. Natl. Acad. Sci. USA, 86 (1989) 5159-5162. [16] Lee, J.-H., Price, R.H., Williams, F.G., Mayer, B. and Beitz, A.J., Nitric oxide synthase is found in some spinothalamic neurons and in neuronal processes that appose spinal neurons that express Fos induced by noxious stimulation, Brain Res., 608 (1993) 324-333. [17] Liposits, Zs., Sherman, D., Phelix, C. and Paull, W.K., A combined light and electron microscopic immunocytochemical method for the simultaneous localization of multiple tissue antigens. Tyrosine hydroxylase immunoreactive innervation of corticotropin releasing factor synthesizing neurons in the paraventricular nucleus of the rat, Histochemistry, 85 (1986) 95-106. [18] Liu, Y.C., Tanaka, S., Inoue, K. and Kukosumi, K., Localization of
[19]
[20]
[21] [22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
fibronectin in the folliculo-stellate cells of the rat anterior pituitary by the double bridge peroxidase-antiperoxidasc method, Histochemistry, 92 (1989) 43-46. Lrger, L., Bonnet, C., Cespugiio, R. and Jouvet, M., Immunocytochemical study pf the CLIP/ACTH-immunoreactive nerve fibers in the dorsal raphe nucleus of the rat, Neurosci. Left., 174 (1994) 137-140. Mansour, A., Khachaturian, H., Lewis, M.E., Akil, H. and Watson, S.J., Autoradiographic differentiation of mu, delta, and kappa opioid receptors in the rat forcbrain and midbrain, J. Neurosci., 7 (1987) 2445-2564. Meller, S.T. and Gebhart, O.F., Nitric oxide (NO) and uociceptive processing in the spinal cord, Pain, 52 (1993) 127-136. Meller, S.T., Pechman, P.S., Gebhart, G.F. and Maves, T.J., Nitric oxide mediates the thermal hyperalgesia produced in a model of neuropathic pain in the rat, Neuroseience, 50 (1992) 7-10. Millan, M.H., Millan, M.J. and Herz, A., Depletion of central ~-endorphin blocks midbrain stimulation produced analgesia in the freely-moving rat, Neuroscience, 18 (1986) 641-649. Onstott, D., Mayer, B. and Beitz, A.J., Nitric oxide synthase immunoreactive neurons anatomically define a longitudinal dorsolateral column within the midbrain periaqueductal gray of the rat: Analysis using laser confocal microscopy, Brain Res., 610 (1993) 317-324. Pilcher, W.H., Joseph, S.A. and McDonald, J.V., Immunocytochemical localization of pro-opiomelanocortin neurons in human brain areas subserving stimulation analgesia, J. Neurosurg., 68 (1988) 621-629. Porto, C.A., Facchinetti, F., Pozzo, P., Benassi, C., Biral, G.P. and Genazzani, A.R., Tonic pain time-dependently affects beta-endorphin-like immunoreactivity in the ventral periaqueductal gray matter of the rat brain, Neurosci. Lett., 86 (1988) 89-93. Porto, C.A., Tassinari, G., Facchinetti, F., Panerai, A.E. and Carli, G., Central beta-endorphin system involvement in the reaction to acute tonic pain, Exp. Brain Res., 83 (1991) 549-554. Radhakrishnan, V. and Henry, J.L., L-NAME blocks responses to NMDA, substance P and noxious cutaneous stimuli in cat dorsal horn, Neuroreport, 4 (1993) 323-326. Schwartzberg, D.G. and Nakane, P.K., ACTH-related peptide containing neurons within the medulla oblongata of the rat, Brain Res., 276 (1983) 351-356. Terenius, L., Opioid peptides, Pain and stress. In J. Joosse, R.M. Buijs and F.J.H. Tilders (Eds.), The Peptidergic Neuron, Progress in Brain Research, Vol. 92, Elsevier, Amsterdam, 1992, pp. 375-383. Tseng, L.F., Xu, J.Y. and Pieper, G.M., Increase of nitric oxide production by L-arginine potentiates i.c.v, administered fl-endorphin-induced antinociception in the mouse., Eur. J. Pharmacol., 212 (1992) 301-303. Vacca, L.L., Abrahams, S.J. and Naftchi, N.E., A modified peroxidase-antiperoxidase procedure for improved localization of tissue antigens: localization of substance P in rat spinal cord, J. Histochem. Cytochem., 28 (1980) 297-307. Wang, Q.-P., Guan, J.-L. and Nakai, Y., Immunoelectron microscopy of enkephalinergic innervation of GABAergic neurons in the periaqueductal gray, Brain Res., 665 (1994) 39-46. Wang, Q.-P., Guan, J.-L. and Nakai, Y., Distribution and synaptic relations of NOS neurons in the dorsal raphe nucleus: a comparison to 5-HT neurons, Brain Res. Bull., 37 (1995) 177-187. Wang, Q.-P. and Nakai, Y., Double preembedding immunostaining for the study by electron microscopy of synaptic relationships between immunocytochemically identified neurons, Neurosci. Prot., in press. Wang, Q.-P. and Nakai, Y., Enkephalinergic innervation of GABAergic neurons in the dorsal raphe nucleus of the rat, Brain Res. Bull., 32 (1993) 315-320. Wang, Q.-P. and Nakai, Y., Dorsal raphe: an important nucleus in the pain modulation, Brain Res. Bull., 34 (1994) 575-585.
Q.-P. Wang, Y. Nakai /Brain Research 684 (1995) 185-193 [38] Wang, Q.-P., Nakai, Y. and Shioda, S., Enkephalinergic synaptic axon terminals on serotoninergic neurons in the dorsal raphe nucleus of the rat: An electron microscopic study by the double immunostaining method, Neurosci. Lett., 124 (1991) 129-132. [39] Wang, Q.-P., Ochiai, H. and Nakai, Y., GABAergic innervation of serotonergic neurons in the dorsal raphe nucleus of the rat studied by
193
electron microscopy double immunostaining, Brain Res. Bull., 29 (1992) 943-948. [40] Zbuo, M., Meller, S.T. and Gebbart, G.F., Endogenous nitric oxide is required for tonic cholinergic inhibition of spinal mechanical transmission, Pain, 54 (1993) 71-78.
ELSEVIER
Research report
Immunoelectron microscopy of ¢l-endorphinergic synaptic innervation of nitric oxide synthase immunoreactive neurons in the dorsal raphe nucleus O.-P. Wang, Y. Nakai * Department of Anatomy, Showa University School of Medicine, Hatanodai 1-5-8, Shinagawa-ku, Tokyo 142, Japan Accepted 14 March 1995
Abstract On the basis of the comparing of the distribution of fl-endorphin-like immunoreactive neuronal fibers and nitric oxide synthase-like immunoreactive neurons in the dorsal raphe nucleus, the synapses between the two immunocytochemically identified neurons were studied with a modified DAB-silver-gold intensification double immunostaining technique at the electron microscopic level. Although both of them can be found in the mediodorsal and medioventral parts of the dorsal raphe nucleus, the synapses between them could only be found in the mediodorsal part. The majority of the /3-endorphin-like immunoreactive neuronal fibers contained many dense-cored vesicles. The synapses made by fl-endorphin-like immunoreactive neuronal axon terminals on nitric oxide synthase-like immunoreactive neurons were both symmetrical and asymmetrical with the former predominant, especially in the axo-dendritic ones. /3-Endorphin-like immunoreactive perikarya could only be found in the ventrobasal hypothalamus. These findings suggest the possibility that the /3-endorphin- producing neurons in the ventrobasal hypothalamus could influence nitric oxide synthase-containing neurons in the dorsal raphe nucleus by synaptic relations.
Keywords: Dorsal raphe nucleus; /3-Endorphin; Nitric oxide synthase; Immunocytochemistry; Synapse; Electron microscopy 1. Introduction Dorsal raphe nucleus (DRN) is one of the most important nuclei in pain modtdation [37]. The serotonergic, GABAergic and enkephalinergic neurons in the DRN have been known as involved in this modulation and the synaptic relations between these neurons have also been studied [36,38,39] and reviewed [2,7]. Besides these, /3-endorphin (EP) in the DRN is also involved in the processes of pain modulation [23,26,27]. Recently, nitric oxide synthase (NOS)-containing neurons in the spinal cord have been reported to be involved in nociceptive processing in spinal
Abbreviations: ABC, avidin-biotin-peroxidase complex; ACTH, adrenocorticotropic hormone; DAB, 3,3'-diaminobenzidine tetrahydrochloride; DRN, dorsal raphe nucleus; ENK, enkephalin; EP, fl-endorphin; EP-LI, /3-endorphin-like immunoreactive;GABA, ~/-aminobutyricacid; 5-HT, serotonin; L-NAME, N'°-nitro-L-argininemethyl ester; NO, nitric oxide; NOS, nitric oxide syntha~'se;NOS-LI, nitric oxide synthase-like immunoreactive; NTS, nucleus tractus solitarii; PAP, peroxidase-antiperoxidase; SGI, silver-goldintensification; VBH, ventrobasalhypothalamus
* Corresponding author. Fax: (81) (3) 3784-6815. 0006-8993/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 0 6 - 8 9 9 3 ( 9 5 ) 0 0 4 1 8 - 1
cord [21,22,28,40] and as spinalthalamic projection neurons [16]. However, in the higher central nervous system, NOS may activate the antinociception induced by /3-endorphin [31], although NOS itself has been reported to cause hyperalgesia in the central nervous system [12]. Among the three most antinociceptive central nervous structures, the ventrolateral periaqueductal gray, DRN, and nucleus raphe magnus with its adjacent structures, only the DRN has a large number of NOS-containing neurons whereas the periaqueductal gray has its NOS-containing neurons only at its dorsal part [7,24,34]. On the other hand, the DRN also contained /3-endorphin-like immunoreactive (EP-LI) fibers [8,11,14,25]. Thus, the direct synaptic relation between the two kinds of neurons, if any, may occur within the nucleus. For this reason, a study was made into such a synaptic relation by preembedding double immunocytochemistry at the electron microscopic level.
2. Materials and methods Experiments were performed on eight adult, Wistar albino rats of 180-220 g in weight. Two were used for
186
Q.-P. Wang, Y. Nakai / Brain Research 684 (1995) 185-193
light microscopic study of EP and NOS immunostaining by the avidin-biotin-peroxidase complex (ABC) method [10]. Six were used for study of their synaptic relations at ultrastructural level by using the combined A B C - D A B silver-gold intensification and double PAP method.
2.1. Antisera Rabbit anti-NOS antibody was obtained from Euro-Diagnostica, Batch no. R 922301, made from the C-terminal of the cloned rat cerebella NOS, coupled to bovine serum albumin. Specificity was confirmed by the absorption test. The staining was abolished by 10 to 100 mg immunogen per ml diluted antiserum. Rabbit anti-EP antibody was Peninsula Lab. Lot no. 025659-1, made from rat EP, having 100% cross-reactivity to porcine, horse, and camel EP, but no cross-reaction with ACqTq, Met-ENK, 3'-EP, c~-MSH and PACAP 38. The ABC Kit purchased from Vector Labs contained biotinylated anti-rabbit IgG, and ABC reagents A and B.
2.2. Immunostaining procedures for light microscopy For the light microscopic study, the ABC method was used. The animals were deeply anesthetized with pentobarbital (50 mg/kg, i.p.), and perfused through the ascending aorta with 100 ml cold 0.9% sodium chloride, followed by 300 ml of 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were removed immediately and postfixed in a 4% paraformaldehyde fixative at 4°C overnight. The blocks of brain containing the DRN, the ventrobasal hypothalamus (VBH) and nucleus tractus solitarii (NTS) in the whole oral-caudal length were cut into serial 20 /zm thick coronal sections with a cryostat. The sections were rinsed several times with 0.1 M phosphate buffer containing 0.3% Triton X-100 (pH 7.4), and immersed in the same solution at 4° C overnight. The sections of midbrain were divided into two groups, each having all levels of the DRN. After preincubating in 5% normal goat serum for 10 min at room temperature, one section group was incubated with 1:2000 diluted rabbit anti-NOS, and the other was incubated with 1:1000 diluted rabbit anti-EP antibody whereas the sections containing VBH and NTS were only incubated with 1:1000 diluted rabbit anti-EP antibody. Both of the antibody solutions contained 0.3% Triton X-100. The incubation lasted for 2 h at room temperature and overnight at 4° C. The sections were then thoroughly rinsed with phosphate buffer containing 0.3% Triton X-100 (pH 7.4), and incubated with biotinylated anti-rabbit IgG for 1.5 h, and then with ABC complex for 2 h at room temperature. The sections were then treated with 3,3'-diaminobenzidine (DAB) in 0.05 M Tris-HC1 buffer including 0.005% hydrogen peroxide for 5 to 10 min in a microscope until the immunoreactivity was visible. Between the above steps, each section was carefully
rinsed three times with 0.1 M phosphate buffer containing 0.3% Triton X-100 (pH 7.4), except between the normal goat serum and the first antiserum. After the DAB reaction, all sections were carefully mounted on gelatin-coated microscope slides, dried at room temperature, and sealed. All the sections were photographed for comparing the distribution of the NOS-LI neurons with that of the EP-LI fibers in the DRN.
2.3. Immunostaining procedures for electron microscopy The experimental procedure was modified from the article reported previously [33]. Under deep Nembutal anesthesia (40 m g / k g b.wt., intraperitoneally), the animals were perfused through the ascending aorta with 100 ml of 0.9% saline followed by 300 ml of 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M phosphate buffer (pH 7.4). The brains were removed immediately and postfixed in 4% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4) at 4° C for 2 - 4 h. The blocks of brain containing the DRN were cut into serial 40 gm-thick coronal sections with an Oxford vibratome. The sections were rinsed several times with 0.1 M phosphate buffer (pH 7.4), and immersed in the same solution at 4 ° C overnight. Sections were treated with 0.1% Triton X-100 containing phosphate buffer for 20 min and incubated in 2% normal goat serum for 20 min at room temperature. For studying fl-endorphin, the first antigen being detected, ABC method [10] was applied. On the other hand, for studying NOS, the second antigen being detected, PAP double bridge technique [18,32] was used. In brief, the sections were incubated in the 1:1000 diluted anti-/3-endorphin antiserum, for 2 h at room temperature and overnight at 4° C. The next day, the sections were incubated with biotinylated anti-rabbit IgG for 45 min, and then with ABC complex for 1 h at room temperature. The sections were then treated for 5 - 7 min in the dark with DAB in 0.05 M Tris-HC1 (pH 7.6) buffer including 0.005% hydrogen peroxide. Between the above steps, each section was carefully rinsed three times with 0.1 M phosphate buffer, except between the normal goat serum and the first antibody. The sections that had good PAP-DAB reaction were treated with the silver-gold intensification [17,35,36] to darken the brown-colored DAB reaction material to black in a light microscope. After that, the sections were transferred for detecting NOS. First, the sections were incubated in 1:2000 diluted NOS antiserum, for 2 h at room temperature and 2 days at 4° C. After that, the sections were immunostained with goat-anti-rabbit IgG for 1.5 h, peroxidase-anti-peroxidase complex (PAP) for 2 h and again with goat-anti-rabbit IgG for 1.5 h, PAP for 2 h at room temperature and then for DAB reaction similar to the procedure previously mentioned with detecting EP, but no silver-gold-intensifyin procedures were made. For examination in the electron microscope, the sections were posffixed with 1% OsO 4 in 0.1 M phosphate buffer (pH 7.4) for 1 h at 4° C, dehydrated in an ethanol series, and
Q.-P. Wang, Y. Nakai/Brain Research 684 (1995) 185-193
187
Fig. 1. and Fig. 2. 1: Light micrographs taken from the neighboring sections showing the distribution of EP-LI fibers (A) and NOS-LI neurons (B). Note that, although the majority of the EP-LI fibers are in the ventrolateral part of the PAG, some varicosity-like fibers are also in the DRN. The inset in the A showing the EP-LI neuronal peril~u3,a in the arcuate nucleus. Bars = 0.1 ram. 2: Electron micrographs (no electron staining) showing EP-LI axons (EP) in the medioventral DRN. A, B: the very thin, long fibers; C, D: the axon terminals containing large dense-cored vesicles. Bars = 0.2 pro. Fig. 2. Electron micrographs (no electron staining) showing EP-LI axon-terminals (EP) synapse (arrows) on non-immunoreactive dendrites. Note the synapse in the A is an asymmetrical one. Bars = 0.2 /~m.
188
Q.-P. Wang, Y. Nakai /Brain Research 684 (1995) 185-193
embedded in a mixture of Epon-Araldite. Ultrathin sections were cut and examined under a JEM-1200EXII electron microscope. Neither lead citrate and uranyl acetate staining nor carbon coating was made on these ultrathin sections.
3. R e s u l t s
3.1. Distribution
Under light microscopy, although clustered immunoreactive fibers were seen mainly in the ventrolateral part of the periaqueductal gray (Fig. 1A), the EP-LI fibers were also seen in the DRN especially in the mediodorsal part of the DRN (Fig. 1A). Only the most anterior part of the DRN had the EP-LI fibers. Most of the fibers had varicosity appearance. On the other hand, many immunoreactive NOS neurons were also seen in both the mediodorsal and medioventral cell groups in the anterior part of the DRN in the adjacent section (Fig. 1B), although the cell bodies were also seen in the posterior part. Many immunoreactive EP-LI perikarya were found in the V B H (inset in Fig. 1A) but not in NTS.
3.2. Ultrastructure and synaptic relations
Under electron microscopy, the NOS-LI elements showed moderate immunostaining by the appearance of DAB reacted products whereas the EP-LI axon terminals showed many silver-gold particles loaded on the DAB reacted products. The loading of the silver-gold particles on the dense-cored vesicles was much stronger than that on the small, clear vesicles. Sometimes, small, clear vesicles that far from dense-cored vesicles also showed some silver-gold particles on them. Occasionally, some EP-LI axon terminals were found to make synapses (Fig. 3). No immunoreactive EP-LI perikarya and dendrites could be found. Although there were some EP-LI axon terminals near blood vessels, none were found near the central aqueduct. In this study, the mediodorsal part (Fig. 3 Fig. 4 Fig. 5) and medioventral part (Fig. 2) of the DRN were observed separately. Most of the EP-LI fibers (73%, 41 out of 56) in the medioventral part of the DRN seemed to be very thin, long axons (Fig. 2A, B) in the dorsoventral direction, although some varicosities containing large dense-cored vesicles (25%, 1 4 / 5 6 ) were also found (Fig. 2C,D), whereas the terminal-like EP-LI fibers, which had mainly
Fig. 3. Electron micrographs (no electron staining) showing EP-LI axon-terminals (EP) synapse (arrows) on NOS-LI neurons (NOS). The synapses
between EP-LI axon-terminals and NOS-LI dendrites are asymmetrical (A) or symmetrical (B). C shows two EP-LI axon-terminals contacting a NOS-LI dendrite, but only one can be identified as a synapse of symmetrical type. D shows an EP-LI axon terminal synapse on NOS-LI neuron perikaryon. Note all the five axon terminals in the figure containing dense-cored vesicles. Bar = 0.2 /.tin.
Q.-P. Wang, Y. Nakai/Brain Research 684 (1995) 185-193 small, clear vesicles, were rare (2%, 1 / 5 6 ) . About onetenth ( 6 / 5 6 ) of those fibers were found to contact with NOS-LI elements by the appearance of cell bodies ( 2 / 5 6 ) or dendrites ( 3 / 5 6 ) and even axon terminals ( 1 / 5 6 ) , but
189
none of them showed synaptic appearance as no active zones of the membranes were found in the present observation. In the mediodorsal part of the DRN, the quantity of the
Fig. 4. Electron micrograph (no electron staining) showing an EP-LI axon terminal receiving a synapse(curved arrow) from a non-immunoreactiveaxon terminal and in turn, making a synapse(arrow) to a non-immunoreactivedendrite. Note the axon terminalcontainingsome dense-coredvesicles. Bar = 0.2 /,I,m.
190
Q.-P. Wang, Y. Nakai / Brain Research 684 (1995) 185-193
ing synapse from a non-immunoreactive axon terminal, and in turn, making a synapse with a non-immunoreactive dendrite (Fig. 5).
4. Discussion 4.1. Methodology
Fig. 5. Electronmicrograph(no electronstaining)showingan EP-LIaxon terminal receiving a synapse (curved arrow)from a non-immunoreactive axon terminal and in turn, making a synapse (arrow) to a non-immunoreactive dendrite. Note the axon terminal containing some dense-cored vesicles. Bar = 0.2 /zm. EP-LI fibers was much more than that in the medioventral part. In total, 497 EP-LI axon profiles were photographed. Although the long, thin fibers were also seen in this part, the frequency was lower (7%, 37/497). Most of the EP-LI axon terminals (87%, 398/460) contained a mixture of large dense-cored vesicles and small clear vesicles or predominant dense-cored vesicles, although the terminals containing only small, clear vesicles were also found with lower frequency (13%, 62/460). The occurrence of identifiable synapse made by EP-LI terminals was not very much (31%, 141/460). Greater than the ratio in the medioventral part of the DRN, about one-third (35%, 160/460) of the EP-LI axon terminals were found to be contacted by NOS-LI profiles. Some EP-LI terminals were found making synapses (12%, 56/460) with NOS-LI dendrites (Fig. 4A,B,C) by the appearance of synaptic active zone in the contact region. Most of the terminals contained dense-cored vesicles whereas the vesicles immediately near the presynaptic membranes were usually of the small, clear type. The synapses were both asymmetrical (Fig. 4A; 4%, 20/460) or symmetrical (Fig. 4B,C) with the later predominant (8%, 36/460). The synapses between EP-LI terminal and NOS-LI perikarya were rare (Fig. 4D; 1%, 4/360). In this case, the synapses were also both asymmetrical (2 cases) and symmetrical (2 cases). In only one case, an EP-LI terminal was found receiv-
The DAB-SGI method has been proved as a useful tool to study the synaptic relations between two immunocytochemically identified neurons [33,36]. Its advantages and disadvantages have been discussed [35]. One of the disadvantages of the method is that the severe pH condition in the procedure often made the detecting of second antigen difficult, especially big molecules. In this study, we also encountered this problem because NOS, the enzyme that catalyzes L-arginine into L-citrulline and nitric oxide (NO) [15], is a large molecule. The immunoreactivity of the NOS, the second antigen detected in the present study, is not so satisfactory after DAB-SGI treatment. To overcome this problem, we used the double bridge PAP method [18,32], and the result showed that the modification is effective. Also, we omitted the usual electron staining of the ultrathin sections with uranium and lead. The omissions reduced the clearance of the cell membranes to some extent and induced the difficulty of focusing while taking photographs, but the contrast between the immunopositive and immunonegative neuronal elements was enhanced. The fact that loading of the silver-gold particles on the dense-cored vesicles, which contained /3-endorphin molecules, was much more than that on the small, clear vesicles suggests the reaction specificity of the DAB-SGI method in another aspect. 4.2. Morphology
EP-LI fiber has been described as existing in the DRN of rat [8], monkey [11] and human brain [25]. In the present study, we confirmed this finding in rat. Although we studied the whole rostral-caudal length of the DRN, only the rostral-most part of the DRN was found having EP-LI fibers. In the present study, we found 50% or more of the EP-LI axon terminals in the DRN were rich in dense cored-vesicles. This finding is identical to the report of ACTH-immunoreactive varicosities, which were thought to coexist with fl-endorphin (EP) in periaqueductal gray [4], although the ultrastructure of the EP-LI varicosities has been described as containing many small, clear vesicles and occasionally dense-cored vesicles in the hypothalamus [9]. However, the ultrastructural morphology of a vast majority ACTH-immunoreactive varicosities in the rat DRN was described as containing many small clear vesicles and only a very small number of varicosities principally filled with dense core vesicles [19]. We do not know
Q.-P. Wang, Y. Nakai/Brain Research 684 (1995) 185-193
why the morphological difference exists between the two peptides in the same nucleus. The difference could perhaps be because although they are coexistent in the same neurons in the VBH, they are not coexistent in the same varicosities in the DRN of the rat. The difference of the morphology of EP-LI axon varicosities between the mediodorsal and medioventral parts of the DRN is noticeable. This discrepancy maybe implies different functions of EP-LI axons in the different regions. We found some EP-LI axon terminal-making synapses with NOS-LI neurons in the mediodorsal part of the DRN and with the negative result within the medioventral part, the main part of the DRN, although we cannot describe this result as having no such synapses because the EP-LI axon terminals were rare. With regard to the synapses made by EP-or ACTH-immunoreactive axon terminals, there is also a discrepancy. In the hypothalamus, they were described as mainly symmetrical by EP immunostaining [9], whereas in the DRN they were described as usually asymmetric [19]. In the present study, we found most of the synapses made by EP-LI axon terminals were symmetrical although a few seemed asymmetrical. This difference between EP and ACTH immunoreactivity maybe also imply that the two neurotransmitters were not coexistent in the same axon terminals. We tend to agree that most of the synapse-made EP-LI axon terminals on the NOS-LI neurons are symmetric because DAB reacted products often adherent on the postsynaptic membranes to cause the pseudo asymmetriclike synapse. However, because asymmetrical synapses were also found between the EP-LI axon terminals and immunonegative neurons in the present study, and NOS-LI neurons often received asymmetrical synaptic afferent [34], asymmetrical synapses on the NOS-LI neurons by EP-LI axon terminals are also possible. /3-EP-producing neuronal perikarya were first reported within the VBH [2] and were confirmed recently [8,11]. Also recently, a small neuron group was reported in NTS as /3-EP-LI [29] and was confirmed by in situ hybridization study [3]. However, in the present study, we did not find any /3-EP-LI neurons; in the NTS, thus, it would appear that the /3-EP-LI axons in the DRN may originate from the VBH.
191
receptor than to other receptors [30]. Recently, it has been discovered that, in the higher central nervous system, NOS-containing neurons may be important in modulating the antinociception induced by fl-endorphin because Larginine potentiates antinociception induced by fl-endorphin, and the potentiating effect of L-arginine is attenuated by N'%nitro-L-arginine methyl ester (L-NAME), a selective inhibitor of NOS [31]. In the present study, we found some direct synapses between EP-LI axon terminal and NOS-LI neurons. This finding may provide morphological evidence of the physiological suggestion by Tseng et al. [31]. Because arcuate nucleus stimulation could inhibit the reflex of parafascicular neurons caused by noxious stimulation [3], and treatment of DRN lesion, microinjection of naloxone, or anti-EP antiserum could reverse the effect of arcuate stimulation [4], the DRN has been suggested to be involved in arcuate nucleus stimulation-caused pain inhibition [6]. In the present study, we found EP-LI neuron perikarya only in the arcuate nucleus, so the finding of the synapses made by EP-LI axon terminals in the DRN may provide morphological evidence of the physiological studies by Chen et al. [5,6]. In the DRN, the enkephalinergic neurons predominantly influence the serotonergic neurons indirectly through an intemeuron, the GABAergic neuron, by synapses on the GABAergic neurons [36] and in turn, by the synapses between GABAergic axon terminals and serotonergic neurons [39]. In the present study, EP-LI axon terminals made fewer synapses on NOS-LI neurons in the DRN than on non-immunoreactive neurons. Thus, it is not impossible that, similar to the enkephalinergic synaptic influence on serotonergic neurons, the EP-LI axon terminals could also influence NOS-LI neurons indirectly, although what is the interneuron between the EP-LI axon terminals and NOS-LI neurons needs to be studied further. It is noticeable that some authors [13] suggested that NO, the gas made by NOS, may cause the hyperalgesia itself, and the intracerebroventricular administration of L-arginine produced antinociception through another route, although they did not show if the effect of L-arginine was attenuated by L-NAME or not [13] as showed by Tseng et al. [31]. The mechanism of the involvement of NOS in the central nervous system in pain perception and pain modulation should be studied further.
4.3. Possible function
EP is one of the opioids that involved in the pain modulation [1,37]. Physiologically, EP has been reported to be involved in ventral periaqueductal gray (including the DRN) stimulation-induced analgesia [23]. Neurochemically, noxious stimulation elevated EP-LI level in the ventral periaqueductal gray including the DRN [26,27]. This concept has been further confirmed by the fact that the DRN has a large quantity of /z receptor [20], because /~ receptor is more effective in pain modulation than K and 8 receptors and EP has a higher bonding to /~
Acknowledgements This study was supported by research grants from the Ministry of Education, Science and Culture, Japan.
References [1] Basbaum, A.I. and Fields, H.L., Endogenouspain control system: brainstem spinal pathways and endorphin circuitry, Annu. Rev. Neurosci., 7 (1984) 309-338.
192
Q.-P. Wang, Y. Nakai / Brain Research 684 (1995) 185-193
[2] Bloom, F., Battenberg, E., Rossier, J., Ling, N. and Ouillemin, R., Neurons containing fl-endorphin in rat brain exist separatery from those containing enkephalin: Immunocytochemical studies, Proc. Natl. Acad. Sci. USA, 75 (1978) 1591-1595. [3] Bronstein, D.M., Schafer, M.K., Watson, S.J. and Akil, H., Evidence that beta-endorphin is synthesized in cells in the nucleus tractus solitarius: detection of POMC mRNA, Brain Res., 587 (1992) 269-275. [4] Buma, P., Veening, J. and Nieuwenhuys, R., Ultrastructural characterization of adrenocorticotrope hormone (ACTH) immunoreactive fibers in the mesencephalic central grey substance in the rat, Eur. J. Neurosci., 1 (1989) 659-672. [5] Chen, X.-Y., Duanmu, Z.-X., Jiang, X.-H., Yin, W.-P. and Yin, Q.-Z., Effect of hypothalamic arcuate stimulation on the pain-evoked unit discharges of thalamic parafascicular nucleus: A preliminary analysis, Acta Physiol. Cin., 38 (1986) 26-32. [6] Chen, X.Y., Yin, W.P. and Yin, Q.Z., The dorsal raphe nucleus is involved in the inhibitory effect of hypothalamic arcuate stimulation on pain-evoked unit discharges of the thalamic parafascicular nucleus, Acta Physiol. Cin., 39 (1987) 46-53. [7] Dun, N.J., Dun, S.L. and Frrstermann, U., Nitric oxide synthase immunoreactivity in rat pontine medullary neurons, Neuroscience, 59 (1994) 429-445. [8] Finley, J.C., Lindstrom, P. and Petrusz, P., Immunocytochemical localization of beta-endorphin-containing neurons in the rat brain, Neuroendocrinology, 33 (1981) 28-42. [9] Horvath, T.L., Naftolin, F. and Leranth, C., Beta-endorphin innervation of dopamine neurons in the rat hypothalamus: a light and electron microscopic double immunostaining study, Endocrinology, 131 (1992) 1547-1555. [10] Hsu, S.-M., Raine, L. and Fanger, H., Use of avidin-biotin- peroxidase complex (ABC) in immunoperoxidase techniques. A comparison between ABC and unlabeled antibody (PAP) procedures, J. Histochem. Cytochem., 29 (1981) 577-580. [11] Ibuki, T., Okamura, H., Miyazaki, M., Yanaihara, N., Zimmerman, E.A. and Ibata, Y., Comparative distribution of three opioid systems in the lower brainstem of the monkey (Macaca fuscata), J. Comp. Neurol., 279 (1989) 445-456. [12] Kawabata, A. and Takagi, H., The dual role of L-arginine in nociceptive processing in the brain: involvement of nitric oxide and kyotophin. In H. Takagi, N. Toda and R.D. Hawkins (Eds.), Nitric Oxide: Roles in Neuronal Communication and Neurotoxicity, Japan Scientific Societies Press, Tokyo, 1994, pp. 115-125. [13] Kawabata, A., Umeda, N. and Takagi, H., L-Arginine exerts a dual role in nociceptive processing in the brain: involvement of the kyotophin-Met-enkephalin pathway and NO-cyclic GMP pathway, Br. J. Pharmacol., 109 (1993) 73-79. [14] Khachaturian, H., Lewis, M.E., Tsou, K. and Watson, S.J., ~Endorphin, o~-MSH, ACTH, and related peptides. In A. BjSrklund and T. H5kfelt (Eds.), GABA and Neuropeptides in the CNS, Handbook of Chemical Neuroanatomy, Vol. 4, Elsevier, Amsterdam, 1985, pp. 216-272. [15] Knowles, R.G., Palacios, M., Palmer, R.M.J. and Moncada, S., Formation of nitric oxide from L-arginine in the central nervous system: A transduction mechanism for stimulation of the soluble guanylate cyclase, Proc. Natl. Acad. Sci. USA, 86 (1989) 5159-5162. [16] Lee, J.-H., Price, R.H., Williams, F.G., Mayer, B. and Beitz, A.J., Nitric oxide synthase is found in some spinothalamic neurons and in neuronal processes that appose spinal neurons that express Fos induced by noxious stimulation, Brain Res., 608 (1993) 324-333. [17] Liposits, Zs., Sherman, D., Phelix, C. and Paull, W.K., A combined light and electron microscopic immunocytochemical method for the simultaneous localization of multiple tissue antigens. Tyrosine hydroxylase immunoreactive innervation of corticotropin releasing factor synthesizing neurons in the paraventricular nucleus of the rat, Histochemistry, 85 (1986) 95-106. [18] Liu, Y.C., Tanaka, S., Inoue, K. and Kukosumi, K., Localization of
[19]
[20]
[21] [22]
[23]
[24]
[25]
[26]
[27]
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
fibronectin in the folliculo-stellate cells of the rat anterior pituitary by the double bridge peroxidase-antiperoxidasc method, Histochemistry, 92 (1989) 43-46. Lrger, L., Bonnet, C., Cespugiio, R. and Jouvet, M., Immunocytochemical study pf the CLIP/ACTH-immunoreactive nerve fibers in the dorsal raphe nucleus of the rat, Neurosci. Left., 174 (1994) 137-140. Mansour, A., Khachaturian, H., Lewis, M.E., Akil, H. and Watson, S.J., Autoradiographic differentiation of mu, delta, and kappa opioid receptors in the rat forcbrain and midbrain, J. Neurosci., 7 (1987) 2445-2564. Meller, S.T. and Gebhart, O.F., Nitric oxide (NO) and uociceptive processing in the spinal cord, Pain, 52 (1993) 127-136. Meller, S.T., Pechman, P.S., Gebhart, G.F. and Maves, T.J., Nitric oxide mediates the thermal hyperalgesia produced in a model of neuropathic pain in the rat, Neuroseience, 50 (1992) 7-10. Millan, M.H., Millan, M.J. and Herz, A., Depletion of central ~-endorphin blocks midbrain stimulation produced analgesia in the freely-moving rat, Neuroscience, 18 (1986) 641-649. Onstott, D., Mayer, B. and Beitz, A.J., Nitric oxide synthase immunoreactive neurons anatomically define a longitudinal dorsolateral column within the midbrain periaqueductal gray of the rat: Analysis using laser confocal microscopy, Brain Res., 610 (1993) 317-324. Pilcher, W.H., Joseph, S.A. and McDonald, J.V., Immunocytochemical localization of pro-opiomelanocortin neurons in human brain areas subserving stimulation analgesia, J. Neurosurg., 68 (1988) 621-629. Porto, C.A., Facchinetti, F., Pozzo, P., Benassi, C., Biral, G.P. and Genazzani, A.R., Tonic pain time-dependently affects beta-endorphin-like immunoreactivity in the ventral periaqueductal gray matter of the rat brain, Neurosci. Lett., 86 (1988) 89-93. Porto, C.A., Tassinari, G., Facchinetti, F., Panerai, A.E. and Carli, G., Central beta-endorphin system involvement in the reaction to acute tonic pain, Exp. Brain Res., 83 (1991) 549-554. Radhakrishnan, V. and Henry, J.L., L-NAME blocks responses to NMDA, substance P and noxious cutaneous stimuli in cat dorsal horn, Neuroreport, 4 (1993) 323-326. Schwartzberg, D.G. and Nakane, P.K., ACTH-related peptide containing neurons within the medulla oblongata of the rat, Brain Res., 276 (1983) 351-356. Terenius, L., Opioid peptides, Pain and stress. In J. Joosse, R.M. Buijs and F.J.H. Tilders (Eds.), The Peptidergic Neuron, Progress in Brain Research, Vol. 92, Elsevier, Amsterdam, 1992, pp. 375-383. Tseng, L.F., Xu, J.Y. and Pieper, G.M., Increase of nitric oxide production by L-arginine potentiates i.c.v, administered fl-endorphin-induced antinociception in the mouse., Eur. J. Pharmacol., 212 (1992) 301-303. Vacca, L.L., Abrahams, S.J. and Naftchi, N.E., A modified peroxidase-antiperoxidase procedure for improved localization of tissue antigens: localization of substance P in rat spinal cord, J. Histochem. Cytochem., 28 (1980) 297-307. Wang, Q.-P., Guan, J.-L. and Nakai, Y., Immunoelectron microscopy of enkephalinergic innervation of GABAergic neurons in the periaqueductal gray, Brain Res., 665 (1994) 39-46. Wang, Q.-P., Guan, J.-L. and Nakai, Y., Distribution and synaptic relations of NOS neurons in the dorsal raphe nucleus: a comparison to 5-HT neurons, Brain Res. Bull., 37 (1995) 177-187. Wang, Q.-P. and Nakai, Y., Double preembedding immunostaining for the study by electron microscopy of synaptic relationships between immunocytochemically identified neurons, Neurosci. Prot., in press. Wang, Q.-P. and Nakai, Y., Enkephalinergic innervation of GABAergic neurons in the dorsal raphe nucleus of the rat, Brain Res. Bull., 32 (1993) 315-320. Wang, Q.-P. and Nakai, Y., Dorsal raphe: an important nucleus in the pain modulation, Brain Res. Bull., 34 (1994) 575-585.
Q.-P. Wang, Y. Nakai /Brain Research 684 (1995) 185-193 [38] Wang, Q.-P., Nakai, Y. and Shioda, S., Enkephalinergic synaptic axon terminals on serotoninergic neurons in the dorsal raphe nucleus of the rat: An electron microscopic study by the double immunostaining method, Neurosci. Lett., 124 (1991) 129-132. [39] Wang, Q.-P., Ochiai, H. and Nakai, Y., GABAergic innervation of serotonergic neurons in the dorsal raphe nucleus of the rat studied by
193
electron microscopy double immunostaining, Brain Res. Bull., 29 (1992) 943-948. [40] Zbuo, M., Meller, S.T. and Gebbart, G.F., Endogenous nitric oxide is required for tonic cholinergic inhibition of spinal mechanical transmission, Pain, 54 (1993) 71-78.