Brain Research, 559 (1991) 145-148 © 1991 Elsevier Science Publishers B.V. All fights reserved. 0006-8993/91/$03.50 ADONIS 000689939124839E
145
BRES 24839
Light and electron microscopical Iocalisation of 5-HT-immunoreactive boutons in the rat trigeminal motor nucleus Sikha Saha 1, Kwabena Appenteng I and Trevor F.C. Batten 2 Departments of JPhysiology and 2Cardiovascular Studies, University of Leeds, Leeds (U. K.) (Accepted 18 June 1991)
Key words: Trigeminal; Serotonin; Electron microscopy; Synapse; Motoneuron
We have used pre-embedding EM immunohistoehemical methods to obtain quantitative data on the frequency and post-synaptic targets of 5-hydroxtryptamine-immunoreactive (5-HT-IR) boutons within the rat V motor nucleus. Thirteen percent (69/531) of all synaptic contacts in the motor nucleus involved 5-HT-IR boutons. Seventy-four percent of 5-HT-IR boutons made axo-dendritic contacts, 20% axo-somatie contaets, and 6% axo-axonic contacts. We conclude that a significant fraction of boutons in the motor nucleus are 5-HT-IR and most contribute to postsynaptie rather than presynaptie effects on trigeminal motoneurones.
LM studies have indicated that in common with other somatic motoneurone pools, the trigeminal motor nucleus receives a dense innervation from serotonin (5HT)-IR fibres 4"6'1°'12'13. The input arises from the raphe nuclei and the ventrolateral medullary reticular formation 5'6, among other sites, and has been suggested to provide a means of altering the gain or sensitivity of trigeminal motoneurones 6. However, there is little quantitative data on either the postsynaptic targets or the proportion of synaptic contacts involving 5-HT-IR boutons and so it has not been possiblie to obtain any assessment of the overall importance of 5-HT-IR fibres in the control of trigeminal motoneurone behaviour. We therefore set out to obtain quantitative data on the distribution, frequency and postsynaptic targets of 5-HT-IR boutons within the trigeminal motor nucleus. A preliminary abstract of this work has been communicated to the Physiological Society 9. The experiments were performed on 6 rats which were anaesthetised with pentobarbitone (dose = 60 mg/kg, i.v.) and perfused with 5% glutaraldehyde in 0.1 M phosphate buffer (pH = 7.4). The brains were removed and 1-2 mm blocks cut and post-fixed in the same fixative for 1-2 h at 4 °C. LM immunohistochemistry. Serial sections were cut on a vibratome (50 gm) and collected in phosphate-buffered saline (PBS, pH = 7.4). They were pre-incubated in normal (goat) serum for 30 min and then incubated overnight in a rabbit antiserum to glutaraldehyde-conjugated 5-HT 3, diluted 1/2000 in 0.01 M Tris-buffered sa-
line (TBS), containing 0.2% bovine serum albumin. To improve penetration of antibodies into the vibratome section 0.1% saponin was added to the dilution buffer 11. Bound antibody was visualised by incubating for 24 h at 4 °C with Fab-fragments of anti-rabbit IgG labelled with horseradish peroxidase (Sera Lab), diluted 1/300 in TBSsaponin, and the reaction visualised using DAB as the chromogen. EM immunohistochemistry. The brains of 4 rats were used for this phase of the study. 5-HT-immunostained vibratome sections were postflxed in 1% osmium tetroxide containing 1.5% potassium ferricyanide, dehydrated through 50% ethanol, 70% ethanol (containing 5% uranyl acetate), 90%, 100% (x2) ethanols, and flat-embedded on glass slides under plactic coverslips in Spurr's resin. The areas containing the trigeminal motor nucleus were cut out and re-embedded onto blocks. Only ultrathin sections taken from the surface 10 gm of each block were utilised for the ultrastructural analysis. EM data analysis. Somata were identified by the presence of dense stacks of rough endoplasmic reticulum and a nucleus. Axons were differentiated from dendrites by the absence of endoplasmic reticulum and ribosomes, and boutons identified by the presence of densely packed vesicles. The criteria used for the identification of synapses were the presence of an electron dense thickening associated with presynaptic vesicles. The proportion of 5-HT-IR boutons in the trigeminal motor nucleus was determined from a survey of 10 grids containing immunostained sections from 4 different animals. The grids
Correspondence: K. Appenteng, Department of Physiology, University of Leeds, Leeds LS2 9NQ, U.K. Fax: (44) (532) 334381.
146
e
D
f
D
Fig. 1. LM views of 5-HT-IR cells and fibres. A and B show lowand high-power views of labelling obtained in the nucleus raphe magnus. C and D show labelling obtained in the trigeminal motor nucleus. Note the apparent termination of 5-HT-IR fibres around neurones (marked by stars). E and F show locations of 5-HT-IR structures at the level of the trigeminal motor nucleus (E) and at the level of the genu of the VII nerve (F). The reconstructions in E and F based on serial sections taken from two animals. Reconstructions in E derived from sections spanning the entire length of the trigeminal motor nucleus (approximately 800/~m), while reconstructions in F derived from sections starting at the caudal border of the trigeminal motor nucleus and extending approximately 500 /~m caudal of this. Circles in E and F identify locations of dusters of neurones and crosses mark location of clusters of fibres. Bars in A and B = 100 #m, in C and D = 50/zm.
were r a n d o m l y selected prior to initial examination of the material under E M . E a c h grid was then placed into the E M and the b e a m r a n d o m l y focused o n t o two squares in the grid containing sections from the trigeminal motor nucleus. A l l b o u t o n s forming synaptic contacts within these squares were then surveyed and the numbers of 5 - H T - I R and non-immunoreactive boutons counted. In c o m m o n with others (e.g. ref. 4) we found, at the L M level, a profusion of 5 - H T - I R fibres within the trigeminal m o t o r nucleus (Fig. 1 C - E ) . Some of these a p p e a r e d to form b o u t o n s en passant and terminal boutons a r o u n d neurones in the m o t o r nucleus (Fig. 1C,D). 5 - H T - I R fibres were also o b s e r v e d in the nearby trigeminal main sensory nucleus and trigeminal nucleus oralis but the innervation of these a r e a s was very sparse in comparison to the m o t o r nucleus. N o 5 - H T - I R neurones were o b s e r v e d in any of these areas but clearly labelled neurones were observed in the raphe complex (raphe magnus and pontis) and the v e n t r o - m e d i a l portion of the
Fig. 2. Synaptic contacts formed by 5-HT-IR boutons within the trigeminal motor nucleus. A and B show two examples of axo-somatic contacts (S, soma), and C and D show two examples of axodendritic contacts (D, dendrite). Open arrowheads in A-D mark points of synaptic contact. Note the presence of dense-cored vesicles in all the 5-HT-IR boutons. The gold particles over the soma (S) in B were deposited in the course of an attempt to combine 5-HT immunostaining with post-embedding immunohistochemistry. All bars = 0.5/zm.
adjacent medullary reticular formation (Fig. 1 A - B , E F). All these observations are entirely in accord with previous reports from a n u m b e r of authors. A t E M level, 5 - H T - I R boutons within the trigeminal m o t o r nucleus were observed to m a k e axo-somatic (Fig. 2 A , B ) , axo-dendritic (Fig. 2 C , D ) , and axo-axonic (Fig. 3) synaptic contacts. We could not specifically identify the postsynaptic targets involved in axo-somatic and axodendritic synapses, as we did not a t t e m p t to combine neuroanatomical tracing of neurones in the m o t o r nucleus with 5-HT immunohistochemistry. H o w e v e r , o u r assumption is that the majority of such neurones are likely to have b e e n m o t o n e u r o n e s , as few interneurones have b e e n r e p o r t e d in the m o t o r nucleus. A consistent ultrastructural feature of 5 - H T - I R boutons was the presence of dense-cored vesicles (see Figs. 2 and 3). Such vesicles were also o b s e r v e d in 33% of a sample of seri-
147
Fig. 3. An example of an axo-axonic contact. A filled arrowhead marks the point of axo-axonic contact between a 5-HT-IR bouton and an urdabelled bouton while an open arrowhead marks the point of contact between unlabelled bouton and dendrite. Ax, axon; D, dendrite. Bar = 0.5/~m.
ally reconstructed ~-aminobutyric acid ( G A B A ) - I R boutons forming axo-somatic or axo-dendritic contacts within the motor nucleus s. They were also observed in G A B A - I R boutons forming axo-axonic contacts but no serial reconstructions of these were made so the true incidence of dense-cored vesicles could not be ascertained. The proportion of boutons immunolabelled for 5-HT and which all formed synaptic contacts was assessed from a random sample of 531 boutons. Overall 69 (i.e. 13% of total) boutons were 5-HT-IR and of these 74% made axo-dendritic contacts, 20% axo-somatic contacts, and 6% axo-axonic contacts. The incidence of 5-HT-IR boutons in the different animals was 9.1% (16/175), 12.3% (18/146), 14.9% (69/462), 23.5% (21/89), and 26.9% (14/52) of the total number of boutons. The variation in these figures between animals suggests that the overall figure given for the incidence of 5-HT-IR boutons should be treated as a minimum estimate because of the possibility that not all fibres are consistently labelled by pre-embedding immunohistochemistry, though we believe that the error may have been partly mitigated by examining only sections from the surface 10/~m of each block (see methods). Nevertheless it is clear that a significant fraction of synaptic boutons in the trigeminal motor nucleus may contain 5-HT. For comparison we have estimated, using the post-embedding immunogold labelling method, that 28% (i.e. 106/375) of boutons in the trigeminal motor nucleus contain G A B A and that 14% of G A B A - I R boutons formed axo-axonic synapses s. Two conclusions may be drawn from this; first 5-HT synaptic contacts in the trigeminal motor nucleus may be less prevalent than GABA-mediated synaptic contacts, and second any presynaptic 5-HT interactions will be much rarer than presynaptic G A B A actions.
In an earlier EM radioautographic study of 5-HT-IR varicosities in the rat trigeminal motor nucleus, Schaffar et al 1° reported that, overall, only 13-14% of 5-HT-IR varicosities formed synaptic contacts within the motor nucleus and that these contacts were exclusively axodendritic. Their failure to observe any axo-axonic axosomatic contacts represents an important difference from our data, but perhaps a more significant difference concerns their report of a lower incidence of synapses formed by 5-HT labelled terminals. We attempted to assess the possible incidence of non-synaptic release of 5-HT in a separate survey of 46 5-HT-IR profiles from our sample of four animals. The 5-HT-IR profiles all had the ultrastructural characteristics of terminals, namely the presence of densely packed vesicles coupled with the absence of endoplasmic reticulum. Thirty-four formed synaptic contacts but 12 did not. This may indicate the potential for non-synaptic release of transmitter but the absence of synapses can only be confirmed by serial sections. However this data does differ from that reported by Schaffar et a1.1° in that it suggests any non-synaptic release of 5-HT would only involve at best 26% of boutons as opposed to the figure of 86-87% suggested by Schaffar et al.l°. Their estimate of the incidence of 5-HT-labelled terminals was not based on a survey of boutons forming synapses, but simply on a survey of axonal profiles. Their observations were that 4.9% (91182) of 5-HT-labelled profiles formed clear synapses whereas 36.2% (1271351) of unlabelled profiles formed synapses. They used an empirical, and as yet an unproven formula originally developed by Beaudet and Sotelo 1 to estimate the incidence of 5-HT-labelled varicosities which would be expected to exhibit an active zone in random thin sections, assuming that all such varicosities formed a single synaptic contact. The probability (P) of observing active zones was given by d
2
P=--x--+-D ~
co D
where d is the mean active zone length, D the mean diameter of the varicosities or axonal profiles, and ~o the section thickness. The mean diameter of labelled varicosities in the motor nucleus was determined as 0.9 #m (range = 0.3-2), the mean active zone length was assumed to be 0.4 # m though no information was given as to how this was arrived at, and the section thickness taken as 0.08 #m. This gave an estimate of 37% for P and so a figure of 13.2% (i.e.(1-0.37) x 4.9) for the predicted proportion of 5-HT profiles forming synapses in the motor nucleus. The question prompted is why should the incidence of 5-HT profiles forming synapses
148 be so much lower than that of unlabelled profiles? Our suggestion is that the difference arises in part from an unusually large disparity between active zone length and bouton diameter for 5-HT-labelled terminals. The ratio for 29 serially reconstructed unlabelled boutons in the rat motor nucleus s is 0.63 (mean active zone
to a specific structure and so any masking may be less pronounced. The indirect peroxidase method may also provide a more sensitive method of localising 5-HT in terminals and this may explain the failure of Schaffar et al. 1° to identify either axo-axonic or axo-somatic synapses formed by 5-HT terminals.
length = 0.96/~m, mean b o u t o n diameter = 1.52/~m),
The arguments above lead to the conclusion that ex-
and 0.61 for 15 serially reconstructed G A B A - l a b e l l e d boutons 8 (active zone length = 0.76 # m , mean terminal
amination of serial sections of 5-HT-IR terminals provides the most reliable means of estimating the incidence
diameter = 1.25/~m). For comparison Schaffar et al.'s 1° data suggests a ratio of 0.44 for 5-HT profiles in the mo-
of 5-HT-labelled profiles forming synapses. Papadopoulos et al. 7 have performed the only exhaustive study to
tor nucleus (assumed active zone length of 0.4/~m, var-
date. They systematically examined 5-HT-labelled profiles in 50 uninterrupted series of serial sections from the visual cortex (rats) and found that 9{).2% (550/600) of
icosity diameter = 0.9 a m ) , while a ratio of 0.33 can be calculated from Beaudet and Sotelo's 1 data from the cerebellum (active zone length = 0.2/~m, varicosity diameter = 0.6 a m ) . A n implication of the smaller ratio for
labelled profiles formed synapses. O u r estimate that at
5-HT-labelled profiles is that the probability of observ-
cleus form synapses must be regarded as a m i n i m u m as
ing synapses for 5-HT structures in single sections must be much less than for either unlabelled or G A B A - l a -
this was based on examination of single sections. Nevertheless, the implication is that, as in the visual cortex,
belled structures. A n additional complication largely inherent in the E M autoradiographic method is that the
only a minority of 5-HT terminals fail to form synapses. Therefore the hypothesis 2 that most 5-HT actions are
tendency for silver grains to extend outside the labelled profile may result in masking of active zones and so fur-
exerted as a result of extrasynaptic release of transmitter clearly does not apply either to the visual cortex or trigeminal motor nucleus.
ther reduce the apparent incidence of synapses formed by labelled profiles. Immunohistochemical methods are, in principle, far superior at localising a reaction product
1 Beaudet, A. and Sotelo, C., Synaptic remodelling of serotonin axon terminals in rat agranular cerebellum, Brain Research, 206 (1981) 305-329. 2 Beaudet, A. and Descarries, L., The monoamine innervation of rat cerebral cortex: synaptic and non synaptic axon terminals, Neuroscience, 3 (1978) 851-860. 3 Buijs, R.M., Pool, C.W., Van Heerkhuize, J.J., Sluiter, A.A., Van der Sluis, P.J., Ramkema, M., Van der Woude, T.P. and Van der Beek, E., Antibodies to small transmitter molecules and peptides: production and application of antibodies to dopamine, serotonin, GABA, vasopressin, vasoactive intestinal peptides, neuropeptide Y, somatostatin and substance P, Biomed. Res., l0 Suppl. 3 (1983) 213-221. 4 Cropper, E.C., Eisenman, J.S. and Azmitia, E.C., 5-HT-immunoreactive fibers in the trigeminal nuclear complex of the rat, Exp. Brain Res., 55 (1984) 515-522. 5 Fort, P., Luppi, P-H, Salvert, D. and Jouvet, M., Nuclei of origin of monoaminoergic peptidergic and cholinergic afferents to the cat trigeminal motor nucleus: a double labelling study with cholera-toxin as a retrograde tracer, J. Comp. Neurol., 301 (1990) 262-275. 6 Fritsch, J-M., Lyons, W.E., Molliver, M.E. and Grzanna, R., Neurotoxic effects of p-chloroamphetamine on the serotoninergic innervation of the trigeminal motor nucleus: a retrograde transport study, Brain Research, 473 (1988) 261-270. 7 Papadopoulos, G.C., Parnavelas, J.G. and Buijs, R., Monoam-
least 74% of 5-HT profiles in the trigeminal motor nu-
This work was supported by the MRC.
8
9
10 11
12 13
inergic fibers form conventional synapses in the cerebral cortex, Neurosci. Lea., 76 (1987) 275-279. Saha, S., Appenteng, K. and Batten, T.EC., Quantitative analysis and postsynaptie targets of GABA-immunoreactive boutons within the rat trigeminal motor nucleus, Brain Research, in press. Saha, S., Appenteng, K. and Batten, T.EC., Distribution of 5-hydroxtryptamine-immunoreactive (5-HT-IR) synaptic boutons within the rat V motor nucleus as determined by EM, immunohistochemistry, J. Physiol., 438 (1991) 281P. Schaffar, N., Jean, A. and Calas, A., Radioautographic study of serotoninergic axon terminals in the rat trigeminal nucleus, Neurosci. Lett., 44 (1984) 31-36. Siaud, P., Denoroy, L., Assenmacher, I. and Alonso, G., Comparative immunocytochemical study of the catecholaminergic and peptidergic afferent innervation to the innervation to the dorsal vagal complex in rat and guinea pig, J. Comp. Neurol., 290 (1989) 323-335. Steinbusch, H.W.M., Distribution of serotonin immunoreactivity in the central nervous system of the rat-cell bodies and terminals, Neuroscience, 6 (1981) 557-618. Takeuchi, Y., Kojima, M., Matsuura, T. and Sano, Y., Serotonergic innervation of the motoneurons in the mammalian brainstem: light and electron microscopic immunohistoehemistry, Anat. Embryol., 167 (1983) 321-333.
145
BRES 24839
Light and electron microscopical Iocalisation of 5-HT-immunoreactive boutons in the rat trigeminal motor nucleus Sikha Saha 1, Kwabena Appenteng I and Trevor F.C. Batten 2 Departments of JPhysiology and 2Cardiovascular Studies, University of Leeds, Leeds (U. K.) (Accepted 18 June 1991)
Key words: Trigeminal; Serotonin; Electron microscopy; Synapse; Motoneuron
We have used pre-embedding EM immunohistoehemical methods to obtain quantitative data on the frequency and post-synaptic targets of 5-hydroxtryptamine-immunoreactive (5-HT-IR) boutons within the rat V motor nucleus. Thirteen percent (69/531) of all synaptic contacts in the motor nucleus involved 5-HT-IR boutons. Seventy-four percent of 5-HT-IR boutons made axo-dendritic contacts, 20% axo-somatie contaets, and 6% axo-axonic contacts. We conclude that a significant fraction of boutons in the motor nucleus are 5-HT-IR and most contribute to postsynaptie rather than presynaptie effects on trigeminal motoneurones.
LM studies have indicated that in common with other somatic motoneurone pools, the trigeminal motor nucleus receives a dense innervation from serotonin (5HT)-IR fibres 4"6'1°'12'13. The input arises from the raphe nuclei and the ventrolateral medullary reticular formation 5'6, among other sites, and has been suggested to provide a means of altering the gain or sensitivity of trigeminal motoneurones 6. However, there is little quantitative data on either the postsynaptic targets or the proportion of synaptic contacts involving 5-HT-IR boutons and so it has not been possiblie to obtain any assessment of the overall importance of 5-HT-IR fibres in the control of trigeminal motoneurone behaviour. We therefore set out to obtain quantitative data on the distribution, frequency and postsynaptic targets of 5-HT-IR boutons within the trigeminal motor nucleus. A preliminary abstract of this work has been communicated to the Physiological Society 9. The experiments were performed on 6 rats which were anaesthetised with pentobarbitone (dose = 60 mg/kg, i.v.) and perfused with 5% glutaraldehyde in 0.1 M phosphate buffer (pH = 7.4). The brains were removed and 1-2 mm blocks cut and post-fixed in the same fixative for 1-2 h at 4 °C. LM immunohistochemistry. Serial sections were cut on a vibratome (50 gm) and collected in phosphate-buffered saline (PBS, pH = 7.4). They were pre-incubated in normal (goat) serum for 30 min and then incubated overnight in a rabbit antiserum to glutaraldehyde-conjugated 5-HT 3, diluted 1/2000 in 0.01 M Tris-buffered sa-
line (TBS), containing 0.2% bovine serum albumin. To improve penetration of antibodies into the vibratome section 0.1% saponin was added to the dilution buffer 11. Bound antibody was visualised by incubating for 24 h at 4 °C with Fab-fragments of anti-rabbit IgG labelled with horseradish peroxidase (Sera Lab), diluted 1/300 in TBSsaponin, and the reaction visualised using DAB as the chromogen. EM immunohistochemistry. The brains of 4 rats were used for this phase of the study. 5-HT-immunostained vibratome sections were postflxed in 1% osmium tetroxide containing 1.5% potassium ferricyanide, dehydrated through 50% ethanol, 70% ethanol (containing 5% uranyl acetate), 90%, 100% (x2) ethanols, and flat-embedded on glass slides under plactic coverslips in Spurr's resin. The areas containing the trigeminal motor nucleus were cut out and re-embedded onto blocks. Only ultrathin sections taken from the surface 10 gm of each block were utilised for the ultrastructural analysis. EM data analysis. Somata were identified by the presence of dense stacks of rough endoplasmic reticulum and a nucleus. Axons were differentiated from dendrites by the absence of endoplasmic reticulum and ribosomes, and boutons identified by the presence of densely packed vesicles. The criteria used for the identification of synapses were the presence of an electron dense thickening associated with presynaptic vesicles. The proportion of 5-HT-IR boutons in the trigeminal motor nucleus was determined from a survey of 10 grids containing immunostained sections from 4 different animals. The grids
Correspondence: K. Appenteng, Department of Physiology, University of Leeds, Leeds LS2 9NQ, U.K. Fax: (44) (532) 334381.
146
e
D
f
D
Fig. 1. LM views of 5-HT-IR cells and fibres. A and B show lowand high-power views of labelling obtained in the nucleus raphe magnus. C and D show labelling obtained in the trigeminal motor nucleus. Note the apparent termination of 5-HT-IR fibres around neurones (marked by stars). E and F show locations of 5-HT-IR structures at the level of the trigeminal motor nucleus (E) and at the level of the genu of the VII nerve (F). The reconstructions in E and F based on serial sections taken from two animals. Reconstructions in E derived from sections spanning the entire length of the trigeminal motor nucleus (approximately 800/~m), while reconstructions in F derived from sections starting at the caudal border of the trigeminal motor nucleus and extending approximately 500 /~m caudal of this. Circles in E and F identify locations of dusters of neurones and crosses mark location of clusters of fibres. Bars in A and B = 100 #m, in C and D = 50/zm.
were r a n d o m l y selected prior to initial examination of the material under E M . E a c h grid was then placed into the E M and the b e a m r a n d o m l y focused o n t o two squares in the grid containing sections from the trigeminal motor nucleus. A l l b o u t o n s forming synaptic contacts within these squares were then surveyed and the numbers of 5 - H T - I R and non-immunoreactive boutons counted. In c o m m o n with others (e.g. ref. 4) we found, at the L M level, a profusion of 5 - H T - I R fibres within the trigeminal m o t o r nucleus (Fig. 1 C - E ) . Some of these a p p e a r e d to form b o u t o n s en passant and terminal boutons a r o u n d neurones in the m o t o r nucleus (Fig. 1C,D). 5 - H T - I R fibres were also o b s e r v e d in the nearby trigeminal main sensory nucleus and trigeminal nucleus oralis but the innervation of these a r e a s was very sparse in comparison to the m o t o r nucleus. N o 5 - H T - I R neurones were o b s e r v e d in any of these areas but clearly labelled neurones were observed in the raphe complex (raphe magnus and pontis) and the v e n t r o - m e d i a l portion of the
Fig. 2. Synaptic contacts formed by 5-HT-IR boutons within the trigeminal motor nucleus. A and B show two examples of axo-somatic contacts (S, soma), and C and D show two examples of axodendritic contacts (D, dendrite). Open arrowheads in A-D mark points of synaptic contact. Note the presence of dense-cored vesicles in all the 5-HT-IR boutons. The gold particles over the soma (S) in B were deposited in the course of an attempt to combine 5-HT immunostaining with post-embedding immunohistochemistry. All bars = 0.5/zm.
adjacent medullary reticular formation (Fig. 1 A - B , E F). All these observations are entirely in accord with previous reports from a n u m b e r of authors. A t E M level, 5 - H T - I R boutons within the trigeminal m o t o r nucleus were observed to m a k e axo-somatic (Fig. 2 A , B ) , axo-dendritic (Fig. 2 C , D ) , and axo-axonic (Fig. 3) synaptic contacts. We could not specifically identify the postsynaptic targets involved in axo-somatic and axodendritic synapses, as we did not a t t e m p t to combine neuroanatomical tracing of neurones in the m o t o r nucleus with 5-HT immunohistochemistry. H o w e v e r , o u r assumption is that the majority of such neurones are likely to have b e e n m o t o n e u r o n e s , as few interneurones have b e e n r e p o r t e d in the m o t o r nucleus. A consistent ultrastructural feature of 5 - H T - I R boutons was the presence of dense-cored vesicles (see Figs. 2 and 3). Such vesicles were also o b s e r v e d in 33% of a sample of seri-
147
Fig. 3. An example of an axo-axonic contact. A filled arrowhead marks the point of axo-axonic contact between a 5-HT-IR bouton and an urdabelled bouton while an open arrowhead marks the point of contact between unlabelled bouton and dendrite. Ax, axon; D, dendrite. Bar = 0.5/~m.
ally reconstructed ~-aminobutyric acid ( G A B A ) - I R boutons forming axo-somatic or axo-dendritic contacts within the motor nucleus s. They were also observed in G A B A - I R boutons forming axo-axonic contacts but no serial reconstructions of these were made so the true incidence of dense-cored vesicles could not be ascertained. The proportion of boutons immunolabelled for 5-HT and which all formed synaptic contacts was assessed from a random sample of 531 boutons. Overall 69 (i.e. 13% of total) boutons were 5-HT-IR and of these 74% made axo-dendritic contacts, 20% axo-somatic contacts, and 6% axo-axonic contacts. The incidence of 5-HT-IR boutons in the different animals was 9.1% (16/175), 12.3% (18/146), 14.9% (69/462), 23.5% (21/89), and 26.9% (14/52) of the total number of boutons. The variation in these figures between animals suggests that the overall figure given for the incidence of 5-HT-IR boutons should be treated as a minimum estimate because of the possibility that not all fibres are consistently labelled by pre-embedding immunohistochemistry, though we believe that the error may have been partly mitigated by examining only sections from the surface 10/~m of each block (see methods). Nevertheless it is clear that a significant fraction of synaptic boutons in the trigeminal motor nucleus may contain 5-HT. For comparison we have estimated, using the post-embedding immunogold labelling method, that 28% (i.e. 106/375) of boutons in the trigeminal motor nucleus contain G A B A and that 14% of G A B A - I R boutons formed axo-axonic synapses s. Two conclusions may be drawn from this; first 5-HT synaptic contacts in the trigeminal motor nucleus may be less prevalent than GABA-mediated synaptic contacts, and second any presynaptic 5-HT interactions will be much rarer than presynaptic G A B A actions.
In an earlier EM radioautographic study of 5-HT-IR varicosities in the rat trigeminal motor nucleus, Schaffar et al 1° reported that, overall, only 13-14% of 5-HT-IR varicosities formed synaptic contacts within the motor nucleus and that these contacts were exclusively axodendritic. Their failure to observe any axo-axonic axosomatic contacts represents an important difference from our data, but perhaps a more significant difference concerns their report of a lower incidence of synapses formed by 5-HT labelled terminals. We attempted to assess the possible incidence of non-synaptic release of 5-HT in a separate survey of 46 5-HT-IR profiles from our sample of four animals. The 5-HT-IR profiles all had the ultrastructural characteristics of terminals, namely the presence of densely packed vesicles coupled with the absence of endoplasmic reticulum. Thirty-four formed synaptic contacts but 12 did not. This may indicate the potential for non-synaptic release of transmitter but the absence of synapses can only be confirmed by serial sections. However this data does differ from that reported by Schaffar et a1.1° in that it suggests any non-synaptic release of 5-HT would only involve at best 26% of boutons as opposed to the figure of 86-87% suggested by Schaffar et al.l°. Their estimate of the incidence of 5-HT-labelled terminals was not based on a survey of boutons forming synapses, but simply on a survey of axonal profiles. Their observations were that 4.9% (91182) of 5-HT-labelled profiles formed clear synapses whereas 36.2% (1271351) of unlabelled profiles formed synapses. They used an empirical, and as yet an unproven formula originally developed by Beaudet and Sotelo 1 to estimate the incidence of 5-HT-labelled varicosities which would be expected to exhibit an active zone in random thin sections, assuming that all such varicosities formed a single synaptic contact. The probability (P) of observing active zones was given by d
2
P=--x--+-D ~
co D
where d is the mean active zone length, D the mean diameter of the varicosities or axonal profiles, and ~o the section thickness. The mean diameter of labelled varicosities in the motor nucleus was determined as 0.9 #m (range = 0.3-2), the mean active zone length was assumed to be 0.4 # m though no information was given as to how this was arrived at, and the section thickness taken as 0.08 #m. This gave an estimate of 37% for P and so a figure of 13.2% (i.e.(1-0.37) x 4.9) for the predicted proportion of 5-HT profiles forming synapses in the motor nucleus. The question prompted is why should the incidence of 5-HT profiles forming synapses
148 be so much lower than that of unlabelled profiles? Our suggestion is that the difference arises in part from an unusually large disparity between active zone length and bouton diameter for 5-HT-labelled terminals. The ratio for 29 serially reconstructed unlabelled boutons in the rat motor nucleus s is 0.63 (mean active zone
to a specific structure and so any masking may be less pronounced. The indirect peroxidase method may also provide a more sensitive method of localising 5-HT in terminals and this may explain the failure of Schaffar et al. 1° to identify either axo-axonic or axo-somatic synapses formed by 5-HT terminals.
length = 0.96/~m, mean b o u t o n diameter = 1.52/~m),
The arguments above lead to the conclusion that ex-
and 0.61 for 15 serially reconstructed G A B A - l a b e l l e d boutons 8 (active zone length = 0.76 # m , mean terminal
amination of serial sections of 5-HT-IR terminals provides the most reliable means of estimating the incidence
diameter = 1.25/~m). For comparison Schaffar et al.'s 1° data suggests a ratio of 0.44 for 5-HT profiles in the mo-
of 5-HT-labelled profiles forming synapses. Papadopoulos et al. 7 have performed the only exhaustive study to
tor nucleus (assumed active zone length of 0.4/~m, var-
date. They systematically examined 5-HT-labelled profiles in 50 uninterrupted series of serial sections from the visual cortex (rats) and found that 9{).2% (550/600) of
icosity diameter = 0.9 a m ) , while a ratio of 0.33 can be calculated from Beaudet and Sotelo's 1 data from the cerebellum (active zone length = 0.2/~m, varicosity diameter = 0.6 a m ) . A n implication of the smaller ratio for
labelled profiles formed synapses. O u r estimate that at
5-HT-labelled profiles is that the probability of observ-
cleus form synapses must be regarded as a m i n i m u m as
ing synapses for 5-HT structures in single sections must be much less than for either unlabelled or G A B A - l a -
this was based on examination of single sections. Nevertheless, the implication is that, as in the visual cortex,
belled structures. A n additional complication largely inherent in the E M autoradiographic method is that the
only a minority of 5-HT terminals fail to form synapses. Therefore the hypothesis 2 that most 5-HT actions are
tendency for silver grains to extend outside the labelled profile may result in masking of active zones and so fur-
exerted as a result of extrasynaptic release of transmitter clearly does not apply either to the visual cortex or trigeminal motor nucleus.
ther reduce the apparent incidence of synapses formed by labelled profiles. Immunohistochemical methods are, in principle, far superior at localising a reaction product
1 Beaudet, A. and Sotelo, C., Synaptic remodelling of serotonin axon terminals in rat agranular cerebellum, Brain Research, 206 (1981) 305-329. 2 Beaudet, A. and Descarries, L., The monoamine innervation of rat cerebral cortex: synaptic and non synaptic axon terminals, Neuroscience, 3 (1978) 851-860. 3 Buijs, R.M., Pool, C.W., Van Heerkhuize, J.J., Sluiter, A.A., Van der Sluis, P.J., Ramkema, M., Van der Woude, T.P. and Van der Beek, E., Antibodies to small transmitter molecules and peptides: production and application of antibodies to dopamine, serotonin, GABA, vasopressin, vasoactive intestinal peptides, neuropeptide Y, somatostatin and substance P, Biomed. Res., l0 Suppl. 3 (1983) 213-221. 4 Cropper, E.C., Eisenman, J.S. and Azmitia, E.C., 5-HT-immunoreactive fibers in the trigeminal nuclear complex of the rat, Exp. Brain Res., 55 (1984) 515-522. 5 Fort, P., Luppi, P-H, Salvert, D. and Jouvet, M., Nuclei of origin of monoaminoergic peptidergic and cholinergic afferents to the cat trigeminal motor nucleus: a double labelling study with cholera-toxin as a retrograde tracer, J. Comp. Neurol., 301 (1990) 262-275. 6 Fritsch, J-M., Lyons, W.E., Molliver, M.E. and Grzanna, R., Neurotoxic effects of p-chloroamphetamine on the serotoninergic innervation of the trigeminal motor nucleus: a retrograde transport study, Brain Research, 473 (1988) 261-270. 7 Papadopoulos, G.C., Parnavelas, J.G. and Buijs, R., Monoam-
least 74% of 5-HT profiles in the trigeminal motor nu-
This work was supported by the MRC.
8
9
10 11
12 13
inergic fibers form conventional synapses in the cerebral cortex, Neurosci. Lea., 76 (1987) 275-279. Saha, S., Appenteng, K. and Batten, T.EC., Quantitative analysis and postsynaptie targets of GABA-immunoreactive boutons within the rat trigeminal motor nucleus, Brain Research, in press. Saha, S., Appenteng, K. and Batten, T.EC., Distribution of 5-hydroxtryptamine-immunoreactive (5-HT-IR) synaptic boutons within the rat V motor nucleus as determined by EM, immunohistochemistry, J. Physiol., 438 (1991) 281P. Schaffar, N., Jean, A. and Calas, A., Radioautographic study of serotoninergic axon terminals in the rat trigeminal nucleus, Neurosci. Lett., 44 (1984) 31-36. Siaud, P., Denoroy, L., Assenmacher, I. and Alonso, G., Comparative immunocytochemical study of the catecholaminergic and peptidergic afferent innervation to the innervation to the dorsal vagal complex in rat and guinea pig, J. Comp. Neurol., 290 (1989) 323-335. Steinbusch, H.W.M., Distribution of serotonin immunoreactivity in the central nervous system of the rat-cell bodies and terminals, Neuroscience, 6 (1981) 557-618. Takeuchi, Y., Kojima, M., Matsuura, T. and Sano, Y., Serotonergic innervation of the motoneurons in the mammalian brainstem: light and electron microscopic immunohistoehemistry, Anat. Embryol., 167 (1983) 321-333.
