Neuropathology and Applied Neurobiology 1992,18,213-23 1
Ultrastructure of pre-synaptic input to motor neurons in Onuf‘s nucleus: controls and motor neuron disease A. H. PULLEN*, J. E. M A R T I N ? A N D M. SWASH? *Sobell Department of Neurophysiology, Institute of Neurology, Queen Square, London and ?Departments of Morbid Anatomy & Neurology, The Royal London Hospital, Whitechapel, London
PULLEN A. H., MARTIN J. E. & SWASH M. (1992) Neuropathology and Applied Neurobiology 18, 213-231 Ultrastructure of pre-synaptic input to motor neurons in Onuf’s nucleus: controls and motor neuron disease
Ultrastructural analyses of sphincteric motoneurons in Onuf‘s nucleus at S2 were undertaken in human spinal cord obtained 3-6 h post-mortem from three subjects with no neurological disease (‘controls’)and five in which death was due to motor neuron disease (MND). Neurons in specified locations within Onuf‘s nucleus of control subjects ranged between 17.8 and 71.7 pm diameter (mean 38.6 pm). Analyses of synaptology revealed five ultrastructural classes of presynaptic terminal synapsing with the neuronal surface membrane. When classified by size, vesicle morphology, and synaptic site structure these conformed to the S , F, T, M and Cterminals defining somatic motoneurons. No terminals characteristic of autonomic motoneurons were found. In MND subjects, neurons in Onuf‘s nucleus at S2 were preserved despite a paucity of neurons in medial and lateral motor nuclei and were of similar size range to those in control subjects. The morphological classes of pre-synaptic terminal found in controls, also characterized sphincteric motoneurons in MND subjects, including the C-type terminal. The presence of C-terminals indicates (i) that sphincteric motoneurons are somatic alphamotoneurons, and (ii) that hypotheses explaining the survival of sphincteric motoneurons in MND on the basis of Onuf’s nucleus being an extension of the pre-ganglionic parasympathetic nucleus, or having intrinsic autonomic properties are incorrect. Keywords: motor neuron, short post-mortem delay, ultrastructure, motor neuron disease, amyotrophic lateral sclerosis, Onuf’s nr :leus
INTRODUCTION Sphincteric motoneurons in Onuf‘s nucleus were originally classified ‘somatic’ (Onufrowicz, 1890). Curiously, sphincteric motoneurons are selectively resistant to degeneration in motor neuron disease (MND), which kills ‘somatic ’ motoneurons in other spinal segments (Mannen et al., 1977;Toyokura, 1977). The reason for their survival is unknown. Studies in other species have led to the suggestion that Onuf’s nucleus is not somatic but an anatomical extension of the Correspondence to: Dr A. H. Pullen, Sobell Department of Neurophysiology, Institute of Neurology, Queen Square, London WClN3BG.
214
A . H. Pullen. J. E. Martin & M . Swash
parasympathetic pre-ganglionic nucleus and therefore possesses autonomic properties (Rexed, 1954; Holstege & Tan, 1987). On the basis of its peptidergic input, others suggest that Onuf‘s nucleus has intrinsic autonomic-like properties (Schroder, 1984; Erdman et al., 1984; Katagiri et al., 1986; Kawatani, Nagel & de Groat, 1986; Gibson et al., 1988; Tashiro et al., 1989), or that sphincteric neurons are hybrid (Gibson et al., 1988). Clinical findings showing that Onuf’s nucleus co-degenerates with autonomic nuclei in the Shy-Drager form of multiple system atrophy tends to support these suggestions (Sung, Mastri & Segal, 1979; Mannen et al., 1982; Konno et al., 1986; Chalmers & Swash, 1987). However, confirmatory evidence of the identity of motoneurons in Onuf’s nucleus remains indirect and equivocal. To gain more insight into the general nature of sphincteric motoneurons Pullen (1988) adopted an alternative approach by investigating the ultrastructural characteristics of axon terminals pre-synaptic to sphincteric motoneurons in Onuf’s nucleus in cat. The premise was that the complement of morphological classes of terminals pre-synaptic to somatic motoneurons is different from that innervating pre-ganglionic parasympathetic motoneurons (Conradi, 1969a; Mawe et al., 1986). Sphincteric motoneurons labelled by intramuscular injection of HRP into the external anal sphincter, exhibited the S, F, T, M-type terminals associated with somatic motoneurons. More importantly, the presence of ‘C’-type axon terminals was demonstrated thus unequivocally characterizing feline sphincteric motoneurons as alphamotoneurons. No terminals dominated by the small dense-cored vesicles characteristic of adrenergic neurons were found, or the mixture of pleomorphic and 100 nm dense-cored vesicles defining inputs to autonomic motoneurons (Mawe et al., 1986; Leedy et al., 1988; Kawatani et al., 1989). The synaptology of human sphincteric motoneurons is unknown, but is an important indicator of their identity and offers an alternative approach to investigating possible mechanisms responsible for the differential response of sacral motoneurons in neurodegenerative disease. This study examines the ultrastructural characteristics of axon terminals pre-synaptic to neurons in Onuf‘s nucleus in spinal cord obtained 3-6 h post-mortem from patients in which death was due to MND/ALS or non-neurological disease. MATERIALS AND METHODS Subject details
Examinations focused on S2 in spinal cord obtained post-mortem from eight subjects. Death in five was due to MND/ALS. There was no neurological disease in the remaining three (Table I). Patient selection
In each case permission for autopsy was obtained from the family. In many of the patients with MND/ALS permission was obtained prior to death in a discussion arising from the patient’s concern that more knowledge about MND/ALS should be made available. Autopsies were carried out by J.E.M. as soon as possible after death had occurred. Tissue preparation
Spinal cord was approached by laminectomy and cervical, thoracic, lumbar and sacral segments were excised. Ultrastructural preservation was optimized by (i) minimizing post-mortem delay
Synaptology of Onuf’s nucleus, normals and MND
2 15
Table 1. Subject details and methodological data Case no.
Sex
Cause of death
Age (years) ~~
Post-mortem lh)
Fixative
Segments examined Cervical Lumbar Sacral Lumbar Sacral Sacral
~~
PM 100/89
F
65
Chronic renal failure
5
2.5% glut. in PBS
PM 182189
M
61
Lung carcinoma
3
PM206/89
F
45
Lung carcinoma
6
1Yo glut. + 2% paraform. in PBS 2.5% glut. in PBS
PM50/89 PM200/89
M M
79. 75
MND MND
6 3.5
PM219/89 PM 12/90
M M
58 38
MND
7 3
PM220/89 PM233/89
F F
70 63
MND MND
14.5
20
2.5% glut. in PBS 1% glut.+ 2% paraform. in PBS 2.5% glut. 2.5% glut. in PBS
Sacral Sacral
2.5% glut. in PBS 2.5% glut. in PBS
Sacral Sacral
Sacral Sacral
glut., glutaraldehyde. paraform., paraformaldehyde. PBS, phosphate buffered saline.
(3-6 h), and (ii) rapid immersion-fixation of 3-5 mm slices of cord for 24-48 h at 4°C in 2.5% glutaraldehyde or 1% glutaraldehyde + 2.0% paraformaldehyde prepared in 50 mM phosphate buffered saline (PBS) pH 7.3. After primary fixation, 1 mm slices were cut from each of the two transverse surfaces of the 5 mm slices and placed in fresh fixative for a further 24 h. Subsequent processing utilized a protocol developed originally to identify and examine HRP-labelled motoneurons in cat cord (Johnson, Pullen & Sears, 1985; Pullen, 1988). The 1 mm cord slices were cut into left and right halves. Each half was trimmed to isolate the anterior horn, and sectioned at 70 pm using a vibrating blade microtome. The 70 pm sections were processed for routine electronmicroscopy and flat-embedded in epoxy resin between two PTFE-coated glass microscope slides. Ventral horns cut from the wafer of resin-embedded section were examined using the specific protocols explained in context with the relevant sections of the ‘Results’. RESULTS A N D OBSERVATIONS Onuf’s nucleus Location Composite maps of neuronal locations obtained from 70 pm sections of S2 from control subjects confirmed that four major groups of neurons exist in the ventral horn (Figure 1). These consisted of an extreme lateral group on a level with the central canal (identified as the intermedio-lateral nucleus (Petras & Cummings, 1972; Konno et al., 1986)), a ventro-lateral group (the lateral motor nucleus (Toyokura, 1977; Konno er al., 1986)),a ventro-medialgroup (Konno et al., 1986), and a discrete ventral group of neurons sited adjacent to the ventral border of the grey matter between the ventro-lateral and ventro-medial nuclei. The ventral group neurons conformed in their location to descriptions of Onuf’snucleus in the human cord (Onufrowicz, 1890; Tokoyura, 1977; Mannen er al., 1977, 1982), and to the location of Onuf’s nucleus in the
216
A . H . Pullen. J . E. Martin 13M . Swash
Figure 1. Major neuronal groups found in S2 of control and MND subjects. Circles denote positions of neurons identified in two representative 70 pm sections through S2 from cases PM206/89 (control) and PM50/89 (MND). Open circles are neurons in the medial (VM) and lateral nuclei (LM); filled circles are sphincteric neurons in Onuf's nucleus (ON). Fewer neurons were found in LM and VM nuclei in the MND subject, relative to nuclei of the control, while neurons in ON were equally numerous in control and MND material.
Table 2. a, Diameters (in pm) of neurons in S2 of control and MND subjects. Measured in 0.5 pm resin sections. Lateral and medial nuclei
Onuf's nucleus n
Controls MND
80 57
Range
Mean
s.d.
s.e.
n
Range
Mean
s.d.
s.e.
17.8-71.1 16.1-77.8
38.6 45.0
12.8 15.8
1.44 2.1
80 42
14.2-70.8 16.9-75.9
39.9 47.7
13.4 14.3
1.5 2.2
Table 2. b, Diameters (in pm) of neurons in lumbar cords of control and MND subjects Controls MND
88 115
16.1-75.9 12.0-75.1
40.2 41.1
13.6 16.4
1.4 1.5
Synaptology of Onuf's nucleus, normals and MND
217
Neuron diameter ( p m )
-.-3m 40 = P c
0
50 r
4-
I
4
0
a
s
Neuron diameter (pm) Figure 2. Distribution profiles of neuronal diameters in Onuf's nucleus (-) and the combined lateral and medial nuclei (---) in a, controls and b, MND subjects. Ranges of size are similar for Onuf's and (medial and lateral nuclei) in control and MND subjects, and no difference occurs in the relative ranges of diameters found in control and MND subjects. Neurons in sections of lumbar cord, c, show more large neurons than do sacral segments (-, controls; ---, MND).
cat and rhesus monkey (Sato, Mizuno & Konishi, 1978; Roppolo, Nadelhaft & DeGroat, 1985). In confirmation of previous reports (Toyokura, 1977; Mannen et al., 1977, 1982) the ventrolateral and ventro-medial motoneurons were less prominent in specimens from MND/ALS patients than in those from controls (Figure 1) specimens. In contrast, neurons in Onuf's nucleus were clearly recognizable at S 2 in 70 pm sections from both controls and MND/ALS patients, confirming that in MND Onuf's nucleus is preserved. Structure In control and MND/ALS material, sections for examination were taken from the mid-segment of S2. In consequence, no 'bridge' of neurons was found between Onuf's nucleus and the preganglionic parasympathetic neurons in the intermedio-lateral nucleus (Rexed, 1954; Konno e f al., 1986).In confirmation of previous observations (Mannen et al., 1982), Onuf's nucleus in control and MND/ALS patients formed a single unified group of neurons and was not subdivided by a prominent diagonal bundle of myelinated axons as in cat and macacque (Pullen, 1988; Roppolo et al., 1985).
Synaptology of Onuf's nucleus, normals and MND
219
Identification of neurons in Onuf's nucleus To ensure that neurons for ultrastructural examination derived from Onuf's nucleus, rather than from a neighbouring group of cells, a procedure was adopted that was originally devised to focus attention on single cytochemically labelled neurons in a defined area of Onuf's nucleus in cat (Pullen, 1988). Areas containing Onuf's nucleus were cut from the 70 pm 'plastic' sections and mounted with cyanoacrylate glue on one end of a polished resin cylinder. The cylinder was illuminated through its opposite end and viewed with a dissecting microscope thus revealing the position of each neuron. The positions were noted and checked at three stages of sectioning: (i) when the specimen was first mounted in the ultramicrotome; (ii) in the toluidine blue stained 0.5 pm semi-thin section, and (iii) in the immediately adjacent ultra-thin section viewed in the electron microscope. Sizes of neurons in Onuf's nucleus Diameters of neurons in S2 of control and MND/ALS material were measured on 0.5 pm resin sections (Table 2a). In sections from control patients, neurons in Onuf's nucleus ranged from 17.8 pm to 71.1 pm (mean 38.6pmk 12.9pm, n=80). While the overall range of size in the ventro-medial and ventro-lateral nuclei was similar to that for Onuf's nucleus (Table 2a), the distribution of sizes within the range was different (Figure 2a). In Onuf's nucleus 42% of neurons were >35 pm diameter, but in the ventro-medial and ventro-lateral nuclei the proportion >35 pm was greater (64%). By their size and position within the ventral tip of the anterior horn, neurons > 35 pm diameter were classified as alpha-motoneurons. Smaller neurons were considered to be a mixture of principally local interneurons, and some gamma motoneurons (Johnson, 1986). In sections from MND patients, neurons in Onuf's nucleus ranged from 16.1 pm to 77.8 pm (mean 45.0f 15.8 pm, n = 57) (Table 2a). T-tests showed these values to be not significantly different from those obtained from control material (P>0.1). Analyses of distribution (Figure 2b) revealed that the proportion of sphincteric neurons >35 pm diameter (45%) was also similar to that found in controls. Qualitative examinations showed that in MND material neurons in the ventro-lateral and ventro-medial nuclei were not as frequently encountered as in control material, indicating a probable loss of neurons from these nuclei in MND. (Relative counts were not performed.) While residual neurons in the ventro-lateral and ventro-medial nuclei of MND patients exhibited a similar range of size to those in the nuclei of controls (Table 2a), they showed a different distribution of sizes (Figure 2b), with 78% possessing diameters > 35 pm. To determine differences between the range and distribution of sizes in Onuf's nucleus and the lumbar cord, neurons in the ventral tip and intermediate region of the anterior horn were measured in 0.5 pm resin sections of lumbar cord from control and MND patients prepared using methods identical to those applied to the sacral cord. Results are summarized in Table 2b,
Figure 3. Representative motoneurons in Onuf's nucleus of a control (a,& case PM206/89) and MND patient (c,d; Case PM50/89) illustrated from 0.5 pm resin sections stained with toluidine blue (scale bars in a and c = 100 pm, and in b and d = 50 pm). Representative neurons in sections of lumbar cord from a control (PM 182/89) and MND (PM219/89) are shown in e and f (scale bars in each case = 50 iun).
220
A . H . Pullen, J. E. Martin & M . Swash
Synaptology of Onuf's nucleus, normals and MND
22 1
Figure 2c, and representative sacral and lumbar neurons are illustrated in Figure 3. There was no significant difference between the ranges of size exhibited by sacral and lumbar neurons in either control or MND material. This finding suggested the loss of motoneurons from lumbar cord in MND patients which was detected in the qualitative examinations, had occurred across the entire spectrum of sizes. This suggestion was confirmed by the distribution profiles of neuronal diameters (Figure 2c) which showed that in both control and MND patients 55% of neurons in the anterior horn of the lumbar cord were > 35 pm diameter. Ultrastructural preservation of sphincteric neurons Previous studies of perfusion-fixed cat spinal cord have shown that the morphology of Nissl bodies, Golgi and rER cisterns, mitochondria, and synaptic vesicles are reliable indicators of tissue preservation (Pullen & Sears, 1978, 1983; Johnson, Pullen & Sears, 1985). Since these ultrastructural features are sensitive to poor access of fixative, delays in perfusion, and physiological stresses such as respiratory arrest prior to perfusion, they have been used to assess the quality of ultrastructural preservation in immersion-fixed human autopsy material. As shown in Figure 4a, in perfusion-fixed cat spinal cord, sphincteric motoneurons in Onuf's nucleus display intact cell and nucleur membranes and intact mitochondria. The highly structured Nissl bodies comprise a multilamellated stack of rER lamellae alternating with linear arrays of polyribosomes. This multilamellated structure is particularly well delineated in the Nissl body post-synaptic to the C-type synapse (Figure 4b). Normal cat motoneurons exhibit no cytoplasmic vacuolation, and rER and Golgi cisterns possess lumen widths ranging between 10-40nm(mean21.2nm+5.2nm). Neurons in Onuf's nucleus examined in autopsied, immersion-fixed human spinal cord obtained 3-6 h after death (Figure 5a) exhibited intact cell and nuclear membranes. However, while most mitochondria possessed double membranes with unbroken contours, the matrix was generally electron-lucent and cristae were disrupted or missing (Figure 5). Nissl body structure was highly preserved and showed the same multilamellated structure as found in perfusionfixed cat cord (Figure 5b). Neurons in autopsied material showed variable cytoplasmic vacuolation which mostly arose from swelling of ER and mitochondria. Lumen widths ofrER cisterns were 10-50 nm wide (mean 22.5 nmk7.7 nm). The quality of ultrastructural preservation of neurons in Onuf's nucleus of M N D material (Figure 6) obtained 3-6 h post-mortem was similar to that in controls, but the ultrastructure of Nissl bodies and Golgi cisterns deteriorated with longer post-mortem delays. Thus shortening the post-mortem delay improved the ultrastructure of most organelles sufficiently to provide adequate cellular detail for direct comparison with animal tissue. However, mitochondria1 preservation is always poor in autopsied tissue indicating either that further shortening of post-mortem delay is required for their preservation, or that factors other than post-mortem autolysis contribute to their destruction.
Figure 4. Normal cat motoneuron in Onuf's nucleus obtained from perfusion-fixed spinal cord. a, The cytoplasm is unvacuolated, and contains numerous intact mitochondria and highly structured Nissl bodies. b, The multilamellated structure of Nissl bodies is particularly well demonstrated by the Nissl body found post-synaptic to the 'C'-type pre-synaptic terminal which characterizes alpha-motoneurons.
222
A . H . Pullen, J. E. Martin & M . Swash
Synaptology of Onurs nucleus, normals and MND
223
S ynaptology Control material On the basis of size, synaptic vesicle shape and diameter, and synaptic site ultrastructure five morphological classes of nerve terminal were found presynaptic to control sphincteric motoneurons. Their relative characteristics were: 1 Approximate appositional length 2-3 pm; round electron-lucent synaptic vesicles ranging from 30-50 nm diameter, several focal and symmetric presynaptic densities. The synaptic region of such a terminal is shown magnified in Figure 7, top left. These terminals conformed to the S-type terminals found on spinal motoneurons of the cat and primate (Conradi, 1969a: Bodian, 1975). 2 Length 2-3 pm; ‘flattened’ or pleomorphic electron-lucent vesicles, 25-30 nm x 8-10 nm, paired focal synaptic densities organized asymmetrically. The synaptic region of this class of terminal is illustrated in Figure 7, middle left. Such terminals were consistent with the F-type terminals found on spinal motoneurons. 3 Length, 2 pm; electron-lucent round vesicles approximately 30 nm diameter, paired focal synaptic densities, but exhibiting a short row of 3-6 round electron-dense bodies subjacent to the post-synaptic density, consistent with Taxi bodies (Conradi, 1969a). The synaptic region is shown in Figure 7, bottom left. The presence of post-synaptic Taxi bodies identified these terminals as T-type terminals. 4 Length 3 pm; electron-lucent round vesicles 40-50 nm diameter; one or two paired synaptic densities longer in length than those in the other classes. The main characteristic of this class was the presence of a small tetminal containing synaptic vesicles which synapsed with a larger terminal residing on the neuronal surface membrane (Figure 7, top right). This smaller presynaptic P-type terminal identified the larger terminal as an M-type terminal (see Conradi, 1969a,b). 5 Length 3-6 pm, the largest of the terminals found on control pudendal motoneurons. This pre-synaptic terminal was characterized by densely packed 50 nm electron-lucent vesicles, and by prominent mitochondria (Figure 7, bottom right). The pre-synaptic membrane exhibited occasional electron-densities which projected into the matrix of the terminal. No corresponding focal post-synaptic densities were found. Immediately subjacent to the post-synaptic membrane, and extending the length of the pre-synaptic terminal was a narrow 10-15 nm closed cistern. This was topographically associated with a multi-lamellated Nissl body comprising alternating lamellae of rER and complex arrays of polyribosomes.
The sub-synaptic cistern with its associated Nissl body characterizes this terminal as a ‘C’-type. The C-type terminal characterized its post-synaptic neuron as a somatic alpha-motoneuron. Pre-synaptic terminals containing predominantly small diameter or large diameter densecore vesicles of the type recognized in cat pre-synaptic parasympathetic neurons (Mawe, Bresnahan & Beattie, 1986; Leedy et al., 1988) were not found associated with sphincteric motoneurons in the cord of control patients.
Figure 5. Control motoneuron in Onuf‘s nucleus from case 206/89 (6 h delay). a, The soma, nucleus and nucleolus are well preserved, and individual Nissl bodies well delineated in the cytoplasm. b, The multilamellate structure of the Nissl body is illustrated by an example from a different control subject in the inset (case 100/89, 5 h delay). Scale bar for (a) is 10 pm.
224
A . H.Pullen, J. E. Martin & M . Swash
Synaptology of Onuf’s nucleus, normals and MND
225
M N D material All the classes of pre-synaptic terminal recognized in control material were also identified on motoneurons in Onuf‘s nucleus examined in the material from MND patients (Figure 8). M-type terminals, however, were few in number, and no example is illustrated. Failure to identify more M-terminals in Onuf‘s nucleus in MND material could have been due to sampling error since in the cat M terminals are not uniformly distributed on pudendal motoneurons. Of particular significance however, was the occurrence of ‘C’-type terminals on neurons in Onuf’s nucleus in patients with MND (Figure 8, RHS), an observation which identified the presence of alpha-motoneurons in Onuf‘s nucleus of these patients. DISCUSSION Our ultrastructural analysis of human post-mortem spinal cord has shown that neurons in specified locations within Onuf’s nucleus from controls without neurological disease were approximately 40 pm in diameter, and exhibited five morphological classes of presynaptic terminal on their surface membrane. These synaptic classes conformed to those defining somatic motoneurons in mammalian spinal cord, and included the ‘C’-type synaptic terminal which is a unique morphological characteristic of alpha-motoneurons. Similar examination of Onuf‘s nucleus in post-mortem material from MND/ALS patients revealed the same five classes of terminal including the ‘C’ type, although the M-type terminals were few in number. Methodologicalconsiderations
Rather than employing 10% formalin in saline as fixative, this study specifically aimed at improving fixation of post-mortem material by introducing procedures devised for ultrastructural studies of mammalian spinal cord (Pullen & Sears, 1978; Johnson et al., 1975). Crucial to good tissue preservation is the use of low ionic strength buffers (< 50 mM) and glutaraldehyde concentrations < 3%, generally 1.5-2.5%. It is unlikely that the motoneurons examined derive from nuclei other than Onuf‘s nucleus since Onuf‘s nucleus was identified using a procedure first devised to identify single cytochemically labelled motoneurons in a given spinal nucleus initially by light-microscopy, and subsequently by electron-microscopy. The same procedure has previously been applied to Onuf’s nucleus in cat (Pullen, 1988). In the present study the positions of motoneurons in Onuf‘s nucleus were carefully noted prior to sectioning, and repeatedly checked thereafter in order to ensure each motoneuron examined by electronmicroscopy derived from a known position on Onuf‘s nucleus. Organizationof Onuf’s nucleus
The motoneurons examined formed a single ventrolateral group medial to the lateral motoneuron group which conformed to the position of Onuf‘s nucleus as described by Toyokura et al. (1977) and Mannen et al. (1982), and the posterolateral group of motoneurons of Konno et al. (1986) and Yamamoto et al. (1 986)
Figure 6. Motoneuron in Onuf‘s nucleus from MND subject 200/89 (3.5 h delay). a, The general ultrastructure is no different from that in controls. b, Nissl bodies are generally highly structured with normal spatial relationships between rER, rER-bound ribosomes and inter-lamellae polyribosomes (case 50/89,6 h delay). Scale in (a) is 10 pm.
226
A . H. Pullen, J. E. Martin & M . Swash
Synaptology of Onuf’s nucleus, normals and M N D
227
Motoneuron morphology and size We have measured motoneuron size in Onuf‘s nucleus of the human cord in both control and MND material. Mean diameters of sphincteric motoneurons were approximately 40 pm (controls) and 45 pm (MND), indicating that overall, there was no reduction in motoneuron size in the Onuf‘s nucleus in MND. Onufrowicz (1890) noted that motoneurons in nucleus X were ‘smaller’ than other sacral motoneurons, but no measurements were reported. Unfortunately, this description is often merely repeated in subsequent studies of Onuf‘s nucleus, but without supporting measurements. Two reports, however, suggest they may not be ‘small’ (as for example interneurons; < 30 pm diameter). Firstly, Mannen et al. (1977) described motoneurons in Onuf‘s nucleus as ‘medium’ sized, and secondly Konno et al. (1986) reported them to be ‘polygonal medium to large’ in diameter. Our work suggests that a major proportion are as large as those found elsewhere in the sacral cord, and some are similar in size to lumbar motoneurons. Possible reasons for the apparent discrepancy between our findings and previous reports are several. One is that most previous pathological material is fixed in 10% formol-saline and embedded in paraffin wax. This procedure imparts considerable tissue shrinkage resulting in an overall underestimate of cell size. Shrinkage is further compounded by subsequent staining procedures. By contrast, the present ultrastructural study minimizes osmotic effects by using buffer of low ionic strength (20-50 mM) and low fixative concentrations ( < 3%) with minimal fixation times. This procedure is known to give ‘good’ fixation in experimental studies, and our study shows it to give adequate preservation of human autopsy material immersion-fixed 3-6 h after death. Another possible factor is sexual dimorphism. In rat, dimorphism is known to express itself as differences of size and viability in pudendal neurons of male and female animals (McKenna & Nadelhaft, 1986). One factor affecting neuronal size in the sacral ventro-medial and ventro-lateral nuclei is the duration of the disease. Since this is progressive, at any given time the neuronal population will express the net balance between neurons subject to atrophy and those surviving but showing compensatory changes of size in response to enlargement of their associated motor unit territories (i.e. collateral reinnervation). Such a consideration would not apply to Onuf‘s nucleus which is spared the disease. Morphology of pre-synaptic terminals in ‘controls’ In mammals, somatic motoneurons exhibit five morphological classes of pre-synaptic terminal on their surface membrane. While initially classified in the monkey by Bodian (1975), the full definitive classification now in common use derives from work in the cat (Conradi, 1969a; McLaughlin, 1972a,b) and rat (Bernstein & Bernstein, 1976). The five classes are termed S - , F-, T-, M-, and C. This paper presents the first description of the same five morphologically classified classes of axon terminals pre-synaptic to the human spinal motoneuron. Criteria for classification include size, shape of synaptic vesicle and ultrastructure of the synaptic region. While vesicle shape alone distinguishes S- from F-terminals, a post-synaptic row of electrondense particles (eponymously named Taxi-bodies) identified the T-terminal. The M-terminal
Figure 7.Synaptic regions showing the distinguishing features of axon terminals found pre-synaptic to rnotoneurons in Onuf‘s nucleus from control subjects. The five classes are S, F, T, M, and C. Spherical vesicles of the S-type contrast with the flattened vesicles of the F-type. Post-synaptic Taxi bodies identify the T-type. The P-terminal synapsing (at arrows) with the larger terminal presynaptic to the motoneuron identifies the M-type terminal. The sub-synaptic cistern (labelled) and associated Nissl body (labelled) identifies a C-type synapse. Scale bars = 0.5pm.
228
A . H.Pullen. J,E. Martin & M . Swash
Figure 8. Four of the five classes of axon terminal pre-synapticto Onuf‘s motoneurons in MND subjects are illustrated using the features of the different synaptic regions. The ‘(2’-type terminal characterizesthe post-synaptic neuron as an alpha-motoneuron.a, S-type (case 200/89). b, F-type (case 200/89). c, T-type (case 50/89). d, C-type (case 200/89). Scale bars =0.5 Fm.
Synaptology of Onuf’s nucleus, normals and MND
229
derives its name from its assumed monosynaptic input to the motoneuron. M-terminals selectively degenerate following dorsal root section (Conradi, 1969b; McLaughlin, 1972a). The most distinctive terminal is the C-type with it high packing density of 50 LUII lucent spherical synaptic vesicles, 15-20 nm sub-synaptic cistern and subjacent multilamellated Nissl body. C-terminals survive dorsal root section, and near or remote spinal hemisection (Conradi, 1969b; McLaughlin, 1972a,b; Bodian, 1975; Pullen & Sears, 1978, 1983). Degeneration studies indicated that C-terminals derive from short-axon propnospinal interneurons (Matsushita & Ikeda, 1973). The C-terminal uniquely characterizes alpha-motoneurons and is not found associated with either gamma motoneurons or Renshaw cells (Conradi, 1969a; Lagerback & Ronnevi, 1982). As yet, no direct association between specific neurons and S , F, or T terminals has been reported. None of the pre-synaptic terminals found on somatic motoneurons are dominated by dense-core synaptic vesicles, although ‘S’ and ‘C’ terminals may exhibit the occasional 100 nm diameter dense-core vesicle. In direct contrast, large and small dense-core vesicles predominate in a majority of terminals pre-synaptic to pre-ganglionicparasympathetic neurons (Mawe et al., 1986;Leedy et al., 1988). Autonomic neurons exhibit three main classes of synaptic terminal; those with small spherical vesicles, those with pleomorphic vesicles, and those with either spherical or pleomorphic vesicles accompanied by numerous dense-cored vesicles. Post-synaptic ultrastructure is variable and may present either symmetric or asymmetricsynapticsites, or post-synaptic dense bodies. The dense bodies are smaller than Taxi bodies seen in T-terminals on somatic motoneurons. In the present study of human sphincteric motoneurons no terminals were encountered that contained mainly dense-core vesicles, or equated with terminals found on autonomic neurons. The overall complement of terminals unequivocally corresponded to that described for somatic motoneurons. The present of the C-type axon terminal identified the sphincteric motoneurons examined here as alpha-motoneurons., This classification is compatible with their innervation of striated skeletal muscle. Implicationsof the occurrence of pre-synaptic terminals in Onuf’s nucleus in MND The identification of S , F, T and C-terminals on sphincteric motoneurons of MND patients indicates that at least some sphinctericmotoneurons survive in this disease. It is not possible to test whether any neurons die in a study designed to examine morphology, as apposed to numerical representation. The presence of C-terminals indicates that sphincteric motoneurons are somatic alpha- motoneurons, and that hypotheses relating the survival of sphincteric motoneurons in MND to an intrinsic autonomic property are unfounded. It remains to be seen what cellular mechanisms or neuronal characteristics enable the selective survival of Onuf‘s neurons in MND, but one possibility is that differences in type and source of interneuronal input, in conjunction with corresponding differences in post-synaptic membrane receptor response, may provide a framework for the observed segmental variability in response to neurodegenerative disease. ACKNOWLEDGEMENTS This work was supported by project grants from the Motor Neuron Disease Association (M.S. & J.E.M., A.H.P.). J.E.M. holds an MRC Training Fellowship and a Neuropathology and Applied Neurobiology Journal Bursary for the collection of short-delay post-mortem material. We wish especially to thank the patients, relatives and staff of St Christopher’s Hospice for their kind help.
230
A . H. Pullen, J , E. Martin & M . Swash
REFERENCES Bernstein J.J. & Bernstein M.E. (1976) Ventral horn synaptology in the rat. Journalof Neurocytology 5, 109-123 Bodian D. (1975) Origin of synaptic types in the motoneurone neuropil of the monkey. Journal of Comparative Neurology 159,225-244 Chalmers D. & Swash M. (1987) Selectivevulnerability of urinary Onuf motoneurons in Shy-Drager syndrome. Journal of Neurology 234,259-260 Conradi S . (1969a) Ultrastructure and distribution of neuronal and glial elements on the motoneurone surface in the lumbosacral spinal cord of the adult cat. Acta Physiologica Scandinavica Supplement 3385-48 Conradi S . (1969b) Ultrastructure of dorsal root boutons on lumbosacral motoneurones of adult cat revealed by dorsal root section. Acta Physiologica Scandinavica Supplement 332,85-111 Erdman S.L.,Kawatani M., Thor K.B., Eskay R. & DeGroat W.C. (1984) Identification of neuropeptides in Onuf's nucleus in the cat. Society of Neuroscience Abstracts 10,432 Gibson S.L.,Polak J.M., Katagiri T., Su H., Weller R.O., Brownell D.B., Holland S., Hughes J.T., Kikuyama S., Ball J., Bloom S.R.& Clifford-Rose F. (1988) A comparison of the distribution of eight peptides in spinal cord from normal controls and cases of motor neurone disease with special reference to Onuf's nucleus. Brain Research 474,255-278
Holstege G. & Tan J. (1987) Supraspinal control of motoneurons innervating the striated muscles of the pelvic floor including urethral and anal sphincters in the cat. Brain 110, 1323-1344 Johnson I.P. (1986) A quantitative ultrastructural comparison of alpha- and gamma-motoneurones in the thoracic region of the spinal cord of the adult cat. Journal of Anatomy 147,55-72 Johnson I.P., Pullen A.H. & Sears T.A. (1985) Target-dependence of Nissl body ultrastructure in cat thoracic motoneurones. Neuroscience Letters 61,201-205 Katagiri T., Gibson S.J., Su H.C. & Polak J.M. (1986) Composition and central projections of the pudendal nerve in the rat investigated by combined peptide immunocytochemistry and retrograde fluorescent labelling. Brain Research 372,313-322 Kawatani M., Nagel J. & deGroat W.C. (1986) Identification of neuropeptides in pelvic and pudendal afferent pathways to the sacral spinal cord of the cat. Journal of Comparative Neurology 249,117-132 Kawatani M., Shioda S., Nakai Y., Takeshige C. & deGroat W.C. (1989) Ultrastructural analysis of enkephalinergic terminals in parasympathetic ganglia innervating the urinary bladder of the cat. Journal of Comparative Neurology 288,s 1-9 1 Konno H., Yamamoto T., Iwasaki Y. & Iizuka H. (1986) Shy-Drager syndrome and amyotrophic lateral sclerosis: cytoarchitectonic and morphometric studies of sacral autonomic neurones. Journal of Neurological Sciences 73, 193-204
Leedy M.G., Bresnahan J.C., Mawe G.M. & Beattie M.S. (1988) Differences in synaptic inputs to preganglionic neurons in the dorsal and lateral band divisions of the cat sacral parasympathetic nucleus. Journal of Comparative Neurology 268,84-90 Lagerback P.A. & Ronnevi P.O. (1982) An ultrastructural study of serially sectioned Renshaw cells 11. Synaptic types. Brain Research 246,18 1-192 Mannen T., Iwata M. & Toyokura M. (1977) Preservation of a certain motoneurone group of the sacral cord in amyotrophic lateral sclerosis; its clinical significance. Journal of Neurology, Neurosurgery and Psychiatry 40, 464-469
Mannen T., Iwata M., Toyokura Y. & Nagashima K. (1982) The Onuf's nucleus and the external anal sphincter muscles in amyotrophic lateral sclerosis and Shy-Drager syndrome. Acta Neuropathologica 58,255-260 Matsushita M. & Ikeda M. (1973) Propriospinal fiber connections of the cervical motor nuclei in the cat. A light and electronmicroscope study. Journal of Comparative Neurology 150, 1-32 Mawe G.M., Bresnahan J. & Beattie M.S. (1986) A light and electronmicroscope analysis of the sacral parasympathetic nucleus after labelling primary afferent and efferent elements with HRP. Journal of Comparative Neurology 250, 33-57
McKenna K.E. & Nadelhaft I. (1986) The organisation of the pudendal nerve in the male and female rat. Journal of Comparative Neurology 248,532-549 McLaughlin B.J. (1972a) Dorsal root projections to the motor nuclei in the cat spinal cord. Journal of Comparative Neurology 144,461-474 McLaughlin B.J.(1972b) Propriospinal and supraspinal projections to the motor nuclei in the cat spinal cord. Journalof Comparative Neurology 144,474-500
Synaptology of Onuf’s nucleus, normals and MND
23 1
Onufrowicz B. (1890) On the arrangement and function of the cell groups of the sacral region of the spinal cord in man. Archives of Neurology and Psychopathology 3,387-412 Pertras J.M. &Cummings J.F. (1972) Autonomic neurons in the spinal cord of the rhesus monkey. A correlation of the findings of cyto-architectonics and sympathectomy with fiber degeneration following dorsal rhizotomy. Journal of Comparative Neurology 146,189-2 18 Pullen A.H. (1988) Quantitative synaptology of feline motoneurones to external anal sphincter muscle. Journal of Comparative Neurology 269,414-424 Pullen A.H. & Sears T.A. (1978) Modification of ‘C’-synapses following partial central deafferentation of thoracic motoneurones. Brain Research 145,141-146 Pullen A.H. & Sears T.A. (1983) Trophism between C-type axon terminals and thoracic motoneurones in the cat. Journalof Physiology (London) 337,373-388 Rexed B.A. (1954) A cytoarchitechtonic atlas on the spinal cord in the cat. Journal of Cornpararive Neurology 100, 297-379 Roppolo J.R., Nadelhaft I. & DeGroat, W.C. (1985) The organisation of pudendal motoneurones and primary afferent projections in the spinal cord of the rhesus monkey revealed by horseradish peroxidase. Journal of Comparative Neurology 234,475-488 Sato M., Mizuno N. & Konishi A. (1978)Localisation ofmotoneurones innervating peroneal muscles: an HRPstudy in cat. Brain Research 140, 149-154 Schroder H.D (1984) Somatostatin in the caudal spinal cord. An immuno-histochemical study of the spinal centers involved in the innervation of the pelvic organs. Journalof Comparative Neurology 223,400-414 Sung, J.H., Mastri A.E. & Segal E. (1979) Pathology of Shy-Drager syndrome. Journal of Neuropathology and Experimental Neurology 38,353-368 Tashiro T., Sadota T., Matsushima R. & Mizuno N. (1989) Convergence of serotonin-. enkephalin- and substance P-like immunoreactive afferent fibres on single pudendal motoneurones in Onuf‘s nucleus of the cat: a light microscope study combining the triple immunocytochemical staining technique with the retrograde HRP-tracing method. Brain Research 481,392-398 Toyokura Y. (1977) Amyotrophic lateral sclerosis: a clinical and pathological study of the ‘negative features’ of the disease. Japanese Journal of Medicine 16,269-273
Received 5 August 1991 Accepted 5 November 1991
Ultrastructure of pre-synaptic input to motor neurons in Onuf‘s nucleus: controls and motor neuron disease A. H. PULLEN*, J. E. M A R T I N ? A N D M. SWASH? *Sobell Department of Neurophysiology, Institute of Neurology, Queen Square, London and ?Departments of Morbid Anatomy & Neurology, The Royal London Hospital, Whitechapel, London
PULLEN A. H., MARTIN J. E. & SWASH M. (1992) Neuropathology and Applied Neurobiology 18, 213-231 Ultrastructure of pre-synaptic input to motor neurons in Onuf’s nucleus: controls and motor neuron disease
Ultrastructural analyses of sphincteric motoneurons in Onuf‘s nucleus at S2 were undertaken in human spinal cord obtained 3-6 h post-mortem from three subjects with no neurological disease (‘controls’)and five in which death was due to motor neuron disease (MND). Neurons in specified locations within Onuf‘s nucleus of control subjects ranged between 17.8 and 71.7 pm diameter (mean 38.6 pm). Analyses of synaptology revealed five ultrastructural classes of presynaptic terminal synapsing with the neuronal surface membrane. When classified by size, vesicle morphology, and synaptic site structure these conformed to the S , F, T, M and Cterminals defining somatic motoneurons. No terminals characteristic of autonomic motoneurons were found. In MND subjects, neurons in Onuf‘s nucleus at S2 were preserved despite a paucity of neurons in medial and lateral motor nuclei and were of similar size range to those in control subjects. The morphological classes of pre-synaptic terminal found in controls, also characterized sphincteric motoneurons in MND subjects, including the C-type terminal. The presence of C-terminals indicates (i) that sphincteric motoneurons are somatic alphamotoneurons, and (ii) that hypotheses explaining the survival of sphincteric motoneurons in MND on the basis of Onuf’s nucleus being an extension of the pre-ganglionic parasympathetic nucleus, or having intrinsic autonomic properties are incorrect. Keywords: motor neuron, short post-mortem delay, ultrastructure, motor neuron disease, amyotrophic lateral sclerosis, Onuf’s nr :leus
INTRODUCTION Sphincteric motoneurons in Onuf‘s nucleus were originally classified ‘somatic’ (Onufrowicz, 1890). Curiously, sphincteric motoneurons are selectively resistant to degeneration in motor neuron disease (MND), which kills ‘somatic ’ motoneurons in other spinal segments (Mannen et al., 1977;Toyokura, 1977). The reason for their survival is unknown. Studies in other species have led to the suggestion that Onuf’s nucleus is not somatic but an anatomical extension of the Correspondence to: Dr A. H. Pullen, Sobell Department of Neurophysiology, Institute of Neurology, Queen Square, London WClN3BG.
214
A . H. Pullen. J. E. Martin & M . Swash
parasympathetic pre-ganglionic nucleus and therefore possesses autonomic properties (Rexed, 1954; Holstege & Tan, 1987). On the basis of its peptidergic input, others suggest that Onuf‘s nucleus has intrinsic autonomic-like properties (Schroder, 1984; Erdman et al., 1984; Katagiri et al., 1986; Kawatani, Nagel & de Groat, 1986; Gibson et al., 1988; Tashiro et al., 1989), or that sphincteric neurons are hybrid (Gibson et al., 1988). Clinical findings showing that Onuf’s nucleus co-degenerates with autonomic nuclei in the Shy-Drager form of multiple system atrophy tends to support these suggestions (Sung, Mastri & Segal, 1979; Mannen et al., 1982; Konno et al., 1986; Chalmers & Swash, 1987). However, confirmatory evidence of the identity of motoneurons in Onuf’s nucleus remains indirect and equivocal. To gain more insight into the general nature of sphincteric motoneurons Pullen (1988) adopted an alternative approach by investigating the ultrastructural characteristics of axon terminals pre-synaptic to sphincteric motoneurons in Onuf’s nucleus in cat. The premise was that the complement of morphological classes of terminals pre-synaptic to somatic motoneurons is different from that innervating pre-ganglionic parasympathetic motoneurons (Conradi, 1969a; Mawe et al., 1986). Sphincteric motoneurons labelled by intramuscular injection of HRP into the external anal sphincter, exhibited the S, F, T, M-type terminals associated with somatic motoneurons. More importantly, the presence of ‘C’-type axon terminals was demonstrated thus unequivocally characterizing feline sphincteric motoneurons as alphamotoneurons. No terminals dominated by the small dense-cored vesicles characteristic of adrenergic neurons were found, or the mixture of pleomorphic and 100 nm dense-cored vesicles defining inputs to autonomic motoneurons (Mawe et al., 1986; Leedy et al., 1988; Kawatani et al., 1989). The synaptology of human sphincteric motoneurons is unknown, but is an important indicator of their identity and offers an alternative approach to investigating possible mechanisms responsible for the differential response of sacral motoneurons in neurodegenerative disease. This study examines the ultrastructural characteristics of axon terminals pre-synaptic to neurons in Onuf‘s nucleus in spinal cord obtained 3-6 h post-mortem from patients in which death was due to MND/ALS or non-neurological disease. MATERIALS AND METHODS Subject details
Examinations focused on S2 in spinal cord obtained post-mortem from eight subjects. Death in five was due to MND/ALS. There was no neurological disease in the remaining three (Table I). Patient selection
In each case permission for autopsy was obtained from the family. In many of the patients with MND/ALS permission was obtained prior to death in a discussion arising from the patient’s concern that more knowledge about MND/ALS should be made available. Autopsies were carried out by J.E.M. as soon as possible after death had occurred. Tissue preparation
Spinal cord was approached by laminectomy and cervical, thoracic, lumbar and sacral segments were excised. Ultrastructural preservation was optimized by (i) minimizing post-mortem delay
Synaptology of Onuf’s nucleus, normals and MND
2 15
Table 1. Subject details and methodological data Case no.
Sex
Cause of death
Age (years) ~~
Post-mortem lh)
Fixative
Segments examined Cervical Lumbar Sacral Lumbar Sacral Sacral
~~
PM 100/89
F
65
Chronic renal failure
5
2.5% glut. in PBS
PM 182189
M
61
Lung carcinoma
3
PM206/89
F
45
Lung carcinoma
6
1Yo glut. + 2% paraform. in PBS 2.5% glut. in PBS
PM50/89 PM200/89
M M
79. 75
MND MND
6 3.5
PM219/89 PM 12/90
M M
58 38
MND
7 3
PM220/89 PM233/89
F F
70 63
MND MND
14.5
20
2.5% glut. in PBS 1% glut.+ 2% paraform. in PBS 2.5% glut. 2.5% glut. in PBS
Sacral Sacral
2.5% glut. in PBS 2.5% glut. in PBS
Sacral Sacral
Sacral Sacral
glut., glutaraldehyde. paraform., paraformaldehyde. PBS, phosphate buffered saline.
(3-6 h), and (ii) rapid immersion-fixation of 3-5 mm slices of cord for 24-48 h at 4°C in 2.5% glutaraldehyde or 1% glutaraldehyde + 2.0% paraformaldehyde prepared in 50 mM phosphate buffered saline (PBS) pH 7.3. After primary fixation, 1 mm slices were cut from each of the two transverse surfaces of the 5 mm slices and placed in fresh fixative for a further 24 h. Subsequent processing utilized a protocol developed originally to identify and examine HRP-labelled motoneurons in cat cord (Johnson, Pullen & Sears, 1985; Pullen, 1988). The 1 mm cord slices were cut into left and right halves. Each half was trimmed to isolate the anterior horn, and sectioned at 70 pm using a vibrating blade microtome. The 70 pm sections were processed for routine electronmicroscopy and flat-embedded in epoxy resin between two PTFE-coated glass microscope slides. Ventral horns cut from the wafer of resin-embedded section were examined using the specific protocols explained in context with the relevant sections of the ‘Results’. RESULTS A N D OBSERVATIONS Onuf’s nucleus Location Composite maps of neuronal locations obtained from 70 pm sections of S2 from control subjects confirmed that four major groups of neurons exist in the ventral horn (Figure 1). These consisted of an extreme lateral group on a level with the central canal (identified as the intermedio-lateral nucleus (Petras & Cummings, 1972; Konno et al., 1986)), a ventro-lateral group (the lateral motor nucleus (Toyokura, 1977; Konno er al., 1986)),a ventro-medialgroup (Konno et al., 1986), and a discrete ventral group of neurons sited adjacent to the ventral border of the grey matter between the ventro-lateral and ventro-medial nuclei. The ventral group neurons conformed in their location to descriptions of Onuf’snucleus in the human cord (Onufrowicz, 1890; Tokoyura, 1977; Mannen er al., 1977, 1982), and to the location of Onuf’s nucleus in the
216
A . H . Pullen. J . E. Martin 13M . Swash
Figure 1. Major neuronal groups found in S2 of control and MND subjects. Circles denote positions of neurons identified in two representative 70 pm sections through S2 from cases PM206/89 (control) and PM50/89 (MND). Open circles are neurons in the medial (VM) and lateral nuclei (LM); filled circles are sphincteric neurons in Onuf's nucleus (ON). Fewer neurons were found in LM and VM nuclei in the MND subject, relative to nuclei of the control, while neurons in ON were equally numerous in control and MND material.
Table 2. a, Diameters (in pm) of neurons in S2 of control and MND subjects. Measured in 0.5 pm resin sections. Lateral and medial nuclei
Onuf's nucleus n
Controls MND
80 57
Range
Mean
s.d.
s.e.
n
Range
Mean
s.d.
s.e.
17.8-71.1 16.1-77.8
38.6 45.0
12.8 15.8
1.44 2.1
80 42
14.2-70.8 16.9-75.9
39.9 47.7
13.4 14.3
1.5 2.2
Table 2. b, Diameters (in pm) of neurons in lumbar cords of control and MND subjects Controls MND
88 115
16.1-75.9 12.0-75.1
40.2 41.1
13.6 16.4
1.4 1.5
Synaptology of Onuf's nucleus, normals and MND
217
Neuron diameter ( p m )
-.-3m 40 = P c
0
50 r
4-
I
4
0
a
s
Neuron diameter (pm) Figure 2. Distribution profiles of neuronal diameters in Onuf's nucleus (-) and the combined lateral and medial nuclei (---) in a, controls and b, MND subjects. Ranges of size are similar for Onuf's and (medial and lateral nuclei) in control and MND subjects, and no difference occurs in the relative ranges of diameters found in control and MND subjects. Neurons in sections of lumbar cord, c, show more large neurons than do sacral segments (-, controls; ---, MND).
cat and rhesus monkey (Sato, Mizuno & Konishi, 1978; Roppolo, Nadelhaft & DeGroat, 1985). In confirmation of previous reports (Toyokura, 1977; Mannen et al., 1977, 1982) the ventrolateral and ventro-medial motoneurons were less prominent in specimens from MND/ALS patients than in those from controls (Figure 1) specimens. In contrast, neurons in Onuf's nucleus were clearly recognizable at S 2 in 70 pm sections from both controls and MND/ALS patients, confirming that in MND Onuf's nucleus is preserved. Structure In control and MND/ALS material, sections for examination were taken from the mid-segment of S2. In consequence, no 'bridge' of neurons was found between Onuf's nucleus and the preganglionic parasympathetic neurons in the intermedio-lateral nucleus (Rexed, 1954; Konno e f al., 1986).In confirmation of previous observations (Mannen et al., 1982), Onuf's nucleus in control and MND/ALS patients formed a single unified group of neurons and was not subdivided by a prominent diagonal bundle of myelinated axons as in cat and macacque (Pullen, 1988; Roppolo et al., 1985).
Synaptology of Onuf's nucleus, normals and MND
219
Identification of neurons in Onuf's nucleus To ensure that neurons for ultrastructural examination derived from Onuf's nucleus, rather than from a neighbouring group of cells, a procedure was adopted that was originally devised to focus attention on single cytochemically labelled neurons in a defined area of Onuf's nucleus in cat (Pullen, 1988). Areas containing Onuf's nucleus were cut from the 70 pm 'plastic' sections and mounted with cyanoacrylate glue on one end of a polished resin cylinder. The cylinder was illuminated through its opposite end and viewed with a dissecting microscope thus revealing the position of each neuron. The positions were noted and checked at three stages of sectioning: (i) when the specimen was first mounted in the ultramicrotome; (ii) in the toluidine blue stained 0.5 pm semi-thin section, and (iii) in the immediately adjacent ultra-thin section viewed in the electron microscope. Sizes of neurons in Onuf's nucleus Diameters of neurons in S2 of control and MND/ALS material were measured on 0.5 pm resin sections (Table 2a). In sections from control patients, neurons in Onuf's nucleus ranged from 17.8 pm to 71.1 pm (mean 38.6pmk 12.9pm, n=80). While the overall range of size in the ventro-medial and ventro-lateral nuclei was similar to that for Onuf's nucleus (Table 2a), the distribution of sizes within the range was different (Figure 2a). In Onuf's nucleus 42% of neurons were >35 pm diameter, but in the ventro-medial and ventro-lateral nuclei the proportion >35 pm was greater (64%). By their size and position within the ventral tip of the anterior horn, neurons > 35 pm diameter were classified as alpha-motoneurons. Smaller neurons were considered to be a mixture of principally local interneurons, and some gamma motoneurons (Johnson, 1986). In sections from MND patients, neurons in Onuf's nucleus ranged from 16.1 pm to 77.8 pm (mean 45.0f 15.8 pm, n = 57) (Table 2a). T-tests showed these values to be not significantly different from those obtained from control material (P>0.1). Analyses of distribution (Figure 2b) revealed that the proportion of sphincteric neurons >35 pm diameter (45%) was also similar to that found in controls. Qualitative examinations showed that in MND material neurons in the ventro-lateral and ventro-medial nuclei were not as frequently encountered as in control material, indicating a probable loss of neurons from these nuclei in MND. (Relative counts were not performed.) While residual neurons in the ventro-lateral and ventro-medial nuclei of MND patients exhibited a similar range of size to those in the nuclei of controls (Table 2a), they showed a different distribution of sizes (Figure 2b), with 78% possessing diameters > 35 pm. To determine differences between the range and distribution of sizes in Onuf's nucleus and the lumbar cord, neurons in the ventral tip and intermediate region of the anterior horn were measured in 0.5 pm resin sections of lumbar cord from control and MND patients prepared using methods identical to those applied to the sacral cord. Results are summarized in Table 2b,
Figure 3. Representative motoneurons in Onuf's nucleus of a control (a,& case PM206/89) and MND patient (c,d; Case PM50/89) illustrated from 0.5 pm resin sections stained with toluidine blue (scale bars in a and c = 100 pm, and in b and d = 50 pm). Representative neurons in sections of lumbar cord from a control (PM 182/89) and MND (PM219/89) are shown in e and f (scale bars in each case = 50 iun).
220
A . H . Pullen, J. E. Martin & M . Swash
Synaptology of Onuf's nucleus, normals and MND
22 1
Figure 2c, and representative sacral and lumbar neurons are illustrated in Figure 3. There was no significant difference between the ranges of size exhibited by sacral and lumbar neurons in either control or MND material. This finding suggested the loss of motoneurons from lumbar cord in MND patients which was detected in the qualitative examinations, had occurred across the entire spectrum of sizes. This suggestion was confirmed by the distribution profiles of neuronal diameters (Figure 2c) which showed that in both control and MND patients 55% of neurons in the anterior horn of the lumbar cord were > 35 pm diameter. Ultrastructural preservation of sphincteric neurons Previous studies of perfusion-fixed cat spinal cord have shown that the morphology of Nissl bodies, Golgi and rER cisterns, mitochondria, and synaptic vesicles are reliable indicators of tissue preservation (Pullen & Sears, 1978, 1983; Johnson, Pullen & Sears, 1985). Since these ultrastructural features are sensitive to poor access of fixative, delays in perfusion, and physiological stresses such as respiratory arrest prior to perfusion, they have been used to assess the quality of ultrastructural preservation in immersion-fixed human autopsy material. As shown in Figure 4a, in perfusion-fixed cat spinal cord, sphincteric motoneurons in Onuf's nucleus display intact cell and nucleur membranes and intact mitochondria. The highly structured Nissl bodies comprise a multilamellated stack of rER lamellae alternating with linear arrays of polyribosomes. This multilamellated structure is particularly well delineated in the Nissl body post-synaptic to the C-type synapse (Figure 4b). Normal cat motoneurons exhibit no cytoplasmic vacuolation, and rER and Golgi cisterns possess lumen widths ranging between 10-40nm(mean21.2nm+5.2nm). Neurons in Onuf's nucleus examined in autopsied, immersion-fixed human spinal cord obtained 3-6 h after death (Figure 5a) exhibited intact cell and nuclear membranes. However, while most mitochondria possessed double membranes with unbroken contours, the matrix was generally electron-lucent and cristae were disrupted or missing (Figure 5). Nissl body structure was highly preserved and showed the same multilamellated structure as found in perfusionfixed cat cord (Figure 5b). Neurons in autopsied material showed variable cytoplasmic vacuolation which mostly arose from swelling of ER and mitochondria. Lumen widths ofrER cisterns were 10-50 nm wide (mean 22.5 nmk7.7 nm). The quality of ultrastructural preservation of neurons in Onuf's nucleus of M N D material (Figure 6) obtained 3-6 h post-mortem was similar to that in controls, but the ultrastructure of Nissl bodies and Golgi cisterns deteriorated with longer post-mortem delays. Thus shortening the post-mortem delay improved the ultrastructure of most organelles sufficiently to provide adequate cellular detail for direct comparison with animal tissue. However, mitochondria1 preservation is always poor in autopsied tissue indicating either that further shortening of post-mortem delay is required for their preservation, or that factors other than post-mortem autolysis contribute to their destruction.
Figure 4. Normal cat motoneuron in Onuf's nucleus obtained from perfusion-fixed spinal cord. a, The cytoplasm is unvacuolated, and contains numerous intact mitochondria and highly structured Nissl bodies. b, The multilamellated structure of Nissl bodies is particularly well demonstrated by the Nissl body found post-synaptic to the 'C'-type pre-synaptic terminal which characterizes alpha-motoneurons.
222
A . H . Pullen, J. E. Martin & M . Swash
Synaptology of Onurs nucleus, normals and MND
223
S ynaptology Control material On the basis of size, synaptic vesicle shape and diameter, and synaptic site ultrastructure five morphological classes of nerve terminal were found presynaptic to control sphincteric motoneurons. Their relative characteristics were: 1 Approximate appositional length 2-3 pm; round electron-lucent synaptic vesicles ranging from 30-50 nm diameter, several focal and symmetric presynaptic densities. The synaptic region of such a terminal is shown magnified in Figure 7, top left. These terminals conformed to the S-type terminals found on spinal motoneurons of the cat and primate (Conradi, 1969a: Bodian, 1975). 2 Length 2-3 pm; ‘flattened’ or pleomorphic electron-lucent vesicles, 25-30 nm x 8-10 nm, paired focal synaptic densities organized asymmetrically. The synaptic region of this class of terminal is illustrated in Figure 7, middle left. Such terminals were consistent with the F-type terminals found on spinal motoneurons. 3 Length, 2 pm; electron-lucent round vesicles approximately 30 nm diameter, paired focal synaptic densities, but exhibiting a short row of 3-6 round electron-dense bodies subjacent to the post-synaptic density, consistent with Taxi bodies (Conradi, 1969a). The synaptic region is shown in Figure 7, bottom left. The presence of post-synaptic Taxi bodies identified these terminals as T-type terminals. 4 Length 3 pm; electron-lucent round vesicles 40-50 nm diameter; one or two paired synaptic densities longer in length than those in the other classes. The main characteristic of this class was the presence of a small tetminal containing synaptic vesicles which synapsed with a larger terminal residing on the neuronal surface membrane (Figure 7, top right). This smaller presynaptic P-type terminal identified the larger terminal as an M-type terminal (see Conradi, 1969a,b). 5 Length 3-6 pm, the largest of the terminals found on control pudendal motoneurons. This pre-synaptic terminal was characterized by densely packed 50 nm electron-lucent vesicles, and by prominent mitochondria (Figure 7, bottom right). The pre-synaptic membrane exhibited occasional electron-densities which projected into the matrix of the terminal. No corresponding focal post-synaptic densities were found. Immediately subjacent to the post-synaptic membrane, and extending the length of the pre-synaptic terminal was a narrow 10-15 nm closed cistern. This was topographically associated with a multi-lamellated Nissl body comprising alternating lamellae of rER and complex arrays of polyribosomes.
The sub-synaptic cistern with its associated Nissl body characterizes this terminal as a ‘C’-type. The C-type terminal characterized its post-synaptic neuron as a somatic alpha-motoneuron. Pre-synaptic terminals containing predominantly small diameter or large diameter densecore vesicles of the type recognized in cat pre-synaptic parasympathetic neurons (Mawe, Bresnahan & Beattie, 1986; Leedy et al., 1988) were not found associated with sphincteric motoneurons in the cord of control patients.
Figure 5. Control motoneuron in Onuf‘s nucleus from case 206/89 (6 h delay). a, The soma, nucleus and nucleolus are well preserved, and individual Nissl bodies well delineated in the cytoplasm. b, The multilamellate structure of the Nissl body is illustrated by an example from a different control subject in the inset (case 100/89, 5 h delay). Scale bar for (a) is 10 pm.
224
A . H.Pullen, J. E. Martin & M . Swash
Synaptology of Onuf’s nucleus, normals and MND
225
M N D material All the classes of pre-synaptic terminal recognized in control material were also identified on motoneurons in Onuf‘s nucleus examined in the material from MND patients (Figure 8). M-type terminals, however, were few in number, and no example is illustrated. Failure to identify more M-terminals in Onuf‘s nucleus in MND material could have been due to sampling error since in the cat M terminals are not uniformly distributed on pudendal motoneurons. Of particular significance however, was the occurrence of ‘C’-type terminals on neurons in Onuf’s nucleus in patients with MND (Figure 8, RHS), an observation which identified the presence of alpha-motoneurons in Onuf‘s nucleus of these patients. DISCUSSION Our ultrastructural analysis of human post-mortem spinal cord has shown that neurons in specified locations within Onuf’s nucleus from controls without neurological disease were approximately 40 pm in diameter, and exhibited five morphological classes of presynaptic terminal on their surface membrane. These synaptic classes conformed to those defining somatic motoneurons in mammalian spinal cord, and included the ‘C’-type synaptic terminal which is a unique morphological characteristic of alpha-motoneurons. Similar examination of Onuf‘s nucleus in post-mortem material from MND/ALS patients revealed the same five classes of terminal including the ‘C’ type, although the M-type terminals were few in number. Methodologicalconsiderations
Rather than employing 10% formalin in saline as fixative, this study specifically aimed at improving fixation of post-mortem material by introducing procedures devised for ultrastructural studies of mammalian spinal cord (Pullen & Sears, 1978; Johnson et al., 1975). Crucial to good tissue preservation is the use of low ionic strength buffers (< 50 mM) and glutaraldehyde concentrations < 3%, generally 1.5-2.5%. It is unlikely that the motoneurons examined derive from nuclei other than Onuf‘s nucleus since Onuf‘s nucleus was identified using a procedure first devised to identify single cytochemically labelled motoneurons in a given spinal nucleus initially by light-microscopy, and subsequently by electron-microscopy. The same procedure has previously been applied to Onuf’s nucleus in cat (Pullen, 1988). In the present study the positions of motoneurons in Onuf‘s nucleus were carefully noted prior to sectioning, and repeatedly checked thereafter in order to ensure each motoneuron examined by electronmicroscopy derived from a known position on Onuf‘s nucleus. Organizationof Onuf’s nucleus
The motoneurons examined formed a single ventrolateral group medial to the lateral motoneuron group which conformed to the position of Onuf‘s nucleus as described by Toyokura et al. (1977) and Mannen et al. (1982), and the posterolateral group of motoneurons of Konno et al. (1986) and Yamamoto et al. (1 986)
Figure 6. Motoneuron in Onuf‘s nucleus from MND subject 200/89 (3.5 h delay). a, The general ultrastructure is no different from that in controls. b, Nissl bodies are generally highly structured with normal spatial relationships between rER, rER-bound ribosomes and inter-lamellae polyribosomes (case 50/89,6 h delay). Scale in (a) is 10 pm.
226
A . H. Pullen, J. E. Martin & M . Swash
Synaptology of Onuf’s nucleus, normals and M N D
227
Motoneuron morphology and size We have measured motoneuron size in Onuf‘s nucleus of the human cord in both control and MND material. Mean diameters of sphincteric motoneurons were approximately 40 pm (controls) and 45 pm (MND), indicating that overall, there was no reduction in motoneuron size in the Onuf‘s nucleus in MND. Onufrowicz (1890) noted that motoneurons in nucleus X were ‘smaller’ than other sacral motoneurons, but no measurements were reported. Unfortunately, this description is often merely repeated in subsequent studies of Onuf‘s nucleus, but without supporting measurements. Two reports, however, suggest they may not be ‘small’ (as for example interneurons; < 30 pm diameter). Firstly, Mannen et al. (1977) described motoneurons in Onuf‘s nucleus as ‘medium’ sized, and secondly Konno et al. (1986) reported them to be ‘polygonal medium to large’ in diameter. Our work suggests that a major proportion are as large as those found elsewhere in the sacral cord, and some are similar in size to lumbar motoneurons. Possible reasons for the apparent discrepancy between our findings and previous reports are several. One is that most previous pathological material is fixed in 10% formol-saline and embedded in paraffin wax. This procedure imparts considerable tissue shrinkage resulting in an overall underestimate of cell size. Shrinkage is further compounded by subsequent staining procedures. By contrast, the present ultrastructural study minimizes osmotic effects by using buffer of low ionic strength (20-50 mM) and low fixative concentrations ( < 3%) with minimal fixation times. This procedure is known to give ‘good’ fixation in experimental studies, and our study shows it to give adequate preservation of human autopsy material immersion-fixed 3-6 h after death. Another possible factor is sexual dimorphism. In rat, dimorphism is known to express itself as differences of size and viability in pudendal neurons of male and female animals (McKenna & Nadelhaft, 1986). One factor affecting neuronal size in the sacral ventro-medial and ventro-lateral nuclei is the duration of the disease. Since this is progressive, at any given time the neuronal population will express the net balance between neurons subject to atrophy and those surviving but showing compensatory changes of size in response to enlargement of their associated motor unit territories (i.e. collateral reinnervation). Such a consideration would not apply to Onuf‘s nucleus which is spared the disease. Morphology of pre-synaptic terminals in ‘controls’ In mammals, somatic motoneurons exhibit five morphological classes of pre-synaptic terminal on their surface membrane. While initially classified in the monkey by Bodian (1975), the full definitive classification now in common use derives from work in the cat (Conradi, 1969a; McLaughlin, 1972a,b) and rat (Bernstein & Bernstein, 1976). The five classes are termed S - , F-, T-, M-, and C. This paper presents the first description of the same five morphologically classified classes of axon terminals pre-synaptic to the human spinal motoneuron. Criteria for classification include size, shape of synaptic vesicle and ultrastructure of the synaptic region. While vesicle shape alone distinguishes S- from F-terminals, a post-synaptic row of electrondense particles (eponymously named Taxi-bodies) identified the T-terminal. The M-terminal
Figure 7.Synaptic regions showing the distinguishing features of axon terminals found pre-synaptic to rnotoneurons in Onuf‘s nucleus from control subjects. The five classes are S, F, T, M, and C. Spherical vesicles of the S-type contrast with the flattened vesicles of the F-type. Post-synaptic Taxi bodies identify the T-type. The P-terminal synapsing (at arrows) with the larger terminal presynaptic to the motoneuron identifies the M-type terminal. The sub-synaptic cistern (labelled) and associated Nissl body (labelled) identifies a C-type synapse. Scale bars = 0.5pm.
228
A . H.Pullen. J,E. Martin & M . Swash
Figure 8. Four of the five classes of axon terminal pre-synapticto Onuf‘s motoneurons in MND subjects are illustrated using the features of the different synaptic regions. The ‘(2’-type terminal characterizesthe post-synaptic neuron as an alpha-motoneuron.a, S-type (case 200/89). b, F-type (case 200/89). c, T-type (case 50/89). d, C-type (case 200/89). Scale bars =0.5 Fm.
Synaptology of Onuf’s nucleus, normals and MND
229
derives its name from its assumed monosynaptic input to the motoneuron. M-terminals selectively degenerate following dorsal root section (Conradi, 1969b; McLaughlin, 1972a). The most distinctive terminal is the C-type with it high packing density of 50 LUII lucent spherical synaptic vesicles, 15-20 nm sub-synaptic cistern and subjacent multilamellated Nissl body. C-terminals survive dorsal root section, and near or remote spinal hemisection (Conradi, 1969b; McLaughlin, 1972a,b; Bodian, 1975; Pullen & Sears, 1978, 1983). Degeneration studies indicated that C-terminals derive from short-axon propnospinal interneurons (Matsushita & Ikeda, 1973). The C-terminal uniquely characterizes alpha-motoneurons and is not found associated with either gamma motoneurons or Renshaw cells (Conradi, 1969a; Lagerback & Ronnevi, 1982). As yet, no direct association between specific neurons and S , F, or T terminals has been reported. None of the pre-synaptic terminals found on somatic motoneurons are dominated by dense-core synaptic vesicles, although ‘S’ and ‘C’ terminals may exhibit the occasional 100 nm diameter dense-core vesicle. In direct contrast, large and small dense-core vesicles predominate in a majority of terminals pre-synaptic to pre-ganglionicparasympathetic neurons (Mawe et al., 1986;Leedy et al., 1988). Autonomic neurons exhibit three main classes of synaptic terminal; those with small spherical vesicles, those with pleomorphic vesicles, and those with either spherical or pleomorphic vesicles accompanied by numerous dense-cored vesicles. Post-synaptic ultrastructure is variable and may present either symmetric or asymmetricsynapticsites, or post-synaptic dense bodies. The dense bodies are smaller than Taxi bodies seen in T-terminals on somatic motoneurons. In the present study of human sphincteric motoneurons no terminals were encountered that contained mainly dense-core vesicles, or equated with terminals found on autonomic neurons. The overall complement of terminals unequivocally corresponded to that described for somatic motoneurons. The present of the C-type axon terminal identified the sphincteric motoneurons examined here as alpha-motoneurons., This classification is compatible with their innervation of striated skeletal muscle. Implicationsof the occurrence of pre-synaptic terminals in Onuf’s nucleus in MND The identification of S , F, T and C-terminals on sphincteric motoneurons of MND patients indicates that at least some sphinctericmotoneurons survive in this disease. It is not possible to test whether any neurons die in a study designed to examine morphology, as apposed to numerical representation. The presence of C-terminals indicates that sphincteric motoneurons are somatic alpha- motoneurons, and that hypotheses relating the survival of sphincteric motoneurons in MND to an intrinsic autonomic property are unfounded. It remains to be seen what cellular mechanisms or neuronal characteristics enable the selective survival of Onuf‘s neurons in MND, but one possibility is that differences in type and source of interneuronal input, in conjunction with corresponding differences in post-synaptic membrane receptor response, may provide a framework for the observed segmental variability in response to neurodegenerative disease. ACKNOWLEDGEMENTS This work was supported by project grants from the Motor Neuron Disease Association (M.S. & J.E.M., A.H.P.). J.E.M. holds an MRC Training Fellowship and a Neuropathology and Applied Neurobiology Journal Bursary for the collection of short-delay post-mortem material. We wish especially to thank the patients, relatives and staff of St Christopher’s Hospice for their kind help.
230
A . H. Pullen, J , E. Martin & M . Swash
REFERENCES Bernstein J.J. & Bernstein M.E. (1976) Ventral horn synaptology in the rat. Journalof Neurocytology 5, 109-123 Bodian D. (1975) Origin of synaptic types in the motoneurone neuropil of the monkey. Journal of Comparative Neurology 159,225-244 Chalmers D. & Swash M. (1987) Selectivevulnerability of urinary Onuf motoneurons in Shy-Drager syndrome. Journal of Neurology 234,259-260 Conradi S . (1969a) Ultrastructure and distribution of neuronal and glial elements on the motoneurone surface in the lumbosacral spinal cord of the adult cat. Acta Physiologica Scandinavica Supplement 3385-48 Conradi S . (1969b) Ultrastructure of dorsal root boutons on lumbosacral motoneurones of adult cat revealed by dorsal root section. Acta Physiologica Scandinavica Supplement 332,85-111 Erdman S.L.,Kawatani M., Thor K.B., Eskay R. & DeGroat W.C. (1984) Identification of neuropeptides in Onuf's nucleus in the cat. Society of Neuroscience Abstracts 10,432 Gibson S.L.,Polak J.M., Katagiri T., Su H., Weller R.O., Brownell D.B., Holland S., Hughes J.T., Kikuyama S., Ball J., Bloom S.R.& Clifford-Rose F. (1988) A comparison of the distribution of eight peptides in spinal cord from normal controls and cases of motor neurone disease with special reference to Onuf's nucleus. Brain Research 474,255-278
Holstege G. & Tan J. (1987) Supraspinal control of motoneurons innervating the striated muscles of the pelvic floor including urethral and anal sphincters in the cat. Brain 110, 1323-1344 Johnson I.P. (1986) A quantitative ultrastructural comparison of alpha- and gamma-motoneurones in the thoracic region of the spinal cord of the adult cat. Journal of Anatomy 147,55-72 Johnson I.P., Pullen A.H. & Sears T.A. (1985) Target-dependence of Nissl body ultrastructure in cat thoracic motoneurones. Neuroscience Letters 61,201-205 Katagiri T., Gibson S.J., Su H.C. & Polak J.M. (1986) Composition and central projections of the pudendal nerve in the rat investigated by combined peptide immunocytochemistry and retrograde fluorescent labelling. Brain Research 372,313-322 Kawatani M., Nagel J. & deGroat W.C. (1986) Identification of neuropeptides in pelvic and pudendal afferent pathways to the sacral spinal cord of the cat. Journal of Comparative Neurology 249,117-132 Kawatani M., Shioda S., Nakai Y., Takeshige C. & deGroat W.C. (1989) Ultrastructural analysis of enkephalinergic terminals in parasympathetic ganglia innervating the urinary bladder of the cat. Journal of Comparative Neurology 288,s 1-9 1 Konno H., Yamamoto T., Iwasaki Y. & Iizuka H. (1986) Shy-Drager syndrome and amyotrophic lateral sclerosis: cytoarchitectonic and morphometric studies of sacral autonomic neurones. Journal of Neurological Sciences 73, 193-204
Leedy M.G., Bresnahan J.C., Mawe G.M. & Beattie M.S. (1988) Differences in synaptic inputs to preganglionic neurons in the dorsal and lateral band divisions of the cat sacral parasympathetic nucleus. Journal of Comparative Neurology 268,84-90 Lagerback P.A. & Ronnevi P.O. (1982) An ultrastructural study of serially sectioned Renshaw cells 11. Synaptic types. Brain Research 246,18 1-192 Mannen T., Iwata M. & Toyokura M. (1977) Preservation of a certain motoneurone group of the sacral cord in amyotrophic lateral sclerosis; its clinical significance. Journal of Neurology, Neurosurgery and Psychiatry 40, 464-469
Mannen T., Iwata M., Toyokura Y. & Nagashima K. (1982) The Onuf's nucleus and the external anal sphincter muscles in amyotrophic lateral sclerosis and Shy-Drager syndrome. Acta Neuropathologica 58,255-260 Matsushita M. & Ikeda M. (1973) Propriospinal fiber connections of the cervical motor nuclei in the cat. A light and electronmicroscope study. Journal of Comparative Neurology 150, 1-32 Mawe G.M., Bresnahan J. & Beattie M.S. (1986) A light and electronmicroscope analysis of the sacral parasympathetic nucleus after labelling primary afferent and efferent elements with HRP. Journal of Comparative Neurology 250, 33-57
McKenna K.E. & Nadelhaft I. (1986) The organisation of the pudendal nerve in the male and female rat. Journal of Comparative Neurology 248,532-549 McLaughlin B.J. (1972a) Dorsal root projections to the motor nuclei in the cat spinal cord. Journal of Comparative Neurology 144,461-474 McLaughlin B.J.(1972b) Propriospinal and supraspinal projections to the motor nuclei in the cat spinal cord. Journalof Comparative Neurology 144,474-500
Synaptology of Onuf’s nucleus, normals and MND
23 1
Onufrowicz B. (1890) On the arrangement and function of the cell groups of the sacral region of the spinal cord in man. Archives of Neurology and Psychopathology 3,387-412 Pertras J.M. &Cummings J.F. (1972) Autonomic neurons in the spinal cord of the rhesus monkey. A correlation of the findings of cyto-architectonics and sympathectomy with fiber degeneration following dorsal rhizotomy. Journal of Comparative Neurology 146,189-2 18 Pullen A.H. (1988) Quantitative synaptology of feline motoneurones to external anal sphincter muscle. Journal of Comparative Neurology 269,414-424 Pullen A.H. & Sears T.A. (1978) Modification of ‘C’-synapses following partial central deafferentation of thoracic motoneurones. Brain Research 145,141-146 Pullen A.H. & Sears T.A. (1983) Trophism between C-type axon terminals and thoracic motoneurones in the cat. Journalof Physiology (London) 337,373-388 Rexed B.A. (1954) A cytoarchitechtonic atlas on the spinal cord in the cat. Journal of Cornpararive Neurology 100, 297-379 Roppolo J.R., Nadelhaft I. & DeGroat, W.C. (1985) The organisation of pudendal motoneurones and primary afferent projections in the spinal cord of the rhesus monkey revealed by horseradish peroxidase. Journal of Comparative Neurology 234,475-488 Sato M., Mizuno N. & Konishi A. (1978)Localisation ofmotoneurones innervating peroneal muscles: an HRPstudy in cat. Brain Research 140, 149-154 Schroder H.D (1984) Somatostatin in the caudal spinal cord. An immuno-histochemical study of the spinal centers involved in the innervation of the pelvic organs. Journalof Comparative Neurology 223,400-414 Sung, J.H., Mastri A.E. & Segal E. (1979) Pathology of Shy-Drager syndrome. Journal of Neuropathology and Experimental Neurology 38,353-368 Tashiro T., Sadota T., Matsushima R. & Mizuno N. (1989) Convergence of serotonin-. enkephalin- and substance P-like immunoreactive afferent fibres on single pudendal motoneurones in Onuf‘s nucleus of the cat: a light microscope study combining the triple immunocytochemical staining technique with the retrograde HRP-tracing method. Brain Research 481,392-398 Toyokura Y. (1977) Amyotrophic lateral sclerosis: a clinical and pathological study of the ‘negative features’ of the disease. Japanese Journal of Medicine 16,269-273
Received 5 August 1991 Accepted 5 November 1991
