Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 1999 Mar 15;515(Pt 3):787–797. doi: 10.1111/j.1469-7793.1999.787ab.x

Distribution of cholinergic contacts on Renshaw cells in the rat spinal cord: a light microscopic study

F J Alvarez 1, D E Dewey 1, P McMillin 1, R E W Fyffe 1
PMCID: PMC2269191  PMID: 10066905

Abstract

  1. Cholinergic terminals in the rat spinal cord were revealed by immunohistochemical detection of the vesicular acetycholine transporter (VAChT). In order to determine the relationships of these terminals to Renshaw cells, we used dual immunolabelling with antibodies against gephyrin or calbindin D28k to provide immunohistochemical identification of Renshaw cells in lamina VII of the ventral horn.

  2. A total of 50 Renshaw cells were analysed quantitatively using a computer-aided reconstruction system to provide accurate localization of contact sites and determination of somatic and dendritic surface area. Dendrites could be traced for up to 413 μm from the soma in calbindin D28k-identified Renshaw cells and up to 184 μm in gephyrin-identified cells.

  3. A total of 3330 cholinergic terminals were observed on 50 Renshaw cells, with a range of 21–138 terminal appositions per cell (mean 66.6 ± 25.56 contacts per cell). The vast majority (83.5%) of the terminals were apposed to dendrites rather than the soma. The overall density of cholinergic contacts increased from a little above 1 per 100 μm2 on the soma and initial 25 μm of proximal dendrites to 4–5 per 100 μm2 on the surface of dendritic segments located 50–250 μm from the soma. Single presynaptic fibres frequently formed multiple contacts with the soma and/or dendrites of individual Renshaw cells.

  4. VAChT-immunoreactive terminals apposed to Renshaw cells varied in size from 0.6 to 6.9 μm in diameter (mean 2.26 ± 0.94; n= 986) and were on average smaller than the cholinergic C-terminals apposed to motoneurones, but larger than VAChT-immunoreactive terminals contacting other ventral horn interneurones.

  5. The high density and relatively large size of many cholinergic terminals on Renshaw cells presumably correlates with the strong synaptic connection between motoneurones and Renshaw cells. The fact that the majority of contacts are distributed over the dendrites makes the motoneurone axon collateral input susceptible to inhibition by the prominent glycinergic inhibitory synapses located on the soma and proximal dendrites. The relative positions and structural features of the excitatory cholinergic and inhibitory glycinergic synapses may explain why Renshaw cells, although capable of firing at very high frequency following motor axon stimulation, appear to fire at relatively low rates during locomotor activity.

A number of different classes of segmental interneurones serve to control motoneurone excitability. One form of postsynaptic inhibition, recurrent inhibition, has been intensively studied over the years and is known to be mediated by Renshaw cells (Eccles et al. 1954). These interneurones use glycine and/or GABA as their neurotransmitter (Cullheim & Kellerth, 1981; Fyffe, 1991a; Schneider & Fyffe, 1992), and make contact predominantly on the dendrites of their target motoneurones (Fyffe, 1991b). Although the role of recurrent inhibition in motor control is still not well defined (see e.g. Binder et al. 1996), it remains important to understand how Renshaw cells themselves might be regulated. Recent work suggests that Renshaw cells are subject to strong glycinergic inhibitory inputs distributed over their cell bodies and proximal dendrites (Alvarez et al. 1997). In addition, a number of lines of evidence point to the presence of a powerful cholinergic input to Renshaw cells, a significant proportion of which is probably mediated by postsynaptic nicotinic acetylcholine receptors. The most prominent cholinergic input comes via the recurrent collaterals of motor axons; Renshaw cells display a characteristic high frequency burst of action potentials following antidromic stimulation of motor axons, and a single impulse in a single motor axon is sufficient to elicit action potentials in a Renshaw cell (van Kuelen, 1981) or to modulate its firing rate (Ross et al. 1975). The EPSP underlying motor axon collateral activation of Renshaw cells has a monophasic time course lasting longer than 50 ms (Walmsley & Tracey, 1981), and even at low levels of synaptic drive frequently has an action potential(s) superimposed on its early rising phase. The synaptic activation of Renshaw cells by motor axon collaterals (and the subsequent recurrent IPSPs in motoneurones) can be blocked, almost completely, by nicotinic antagonists such as d-tubocurarine, dihydro-β-erythroidine, or mecamylamine (Eccles et al. 1954; Curtis & Ryall, 1966a, b, c; Noga et al. 1987; Schneider & Fyffe, 1992).

It is not known which structural/functional features of motor axon collateral inputs on Renshaw cells underlie the genesis of such a strong synaptic connection, because little is known about the number and size, or the spatial organization, of cholinergic inputs on Renshaw cells. Furthermore, it is not known how many motoneurones converge on a single Renshaw cell. Data from intracellular staining studies provided evidence of considerable variability in the size of motor axon terminals (Lagerbäck et al. 1978), and indicated that some recurrent contacts were located on the soma of presumed Renshaw cells (e.g. Lagerbäck et al. 1981; Lagerbäck & Ronnevi, 1982). These studies also indicated that a single motoneurone usually made more than one contact on each of its target cells. Unfortunately these previous studies of motor axon terminals in contact with presumed Renshaw cells were constrained to examination of the soma and most proximal dendrites of the postsynaptic neurone, and could not provide information about more distal synaptic contacts. Recently, localization of the vesicular acetylcholine transporter (VAChT) by immunohistochemistry has been shown to be an effective method for visualizing cholinergic neurones and axon terminals (Gilmor et al. 1996; Arvidsson et al. 1997). The present study used VAChT immunoreactivity in combination with two immunohistochemical criteria to identify Renshaw cells. One approach was based on the characteristic membrane covering of gephyrin-immunoreactive patches on Renshaw cells (Alvarez et al. 1997). Gephyrin is a protein that is required for clustering glycine receptors in the postsynaptic region (Kirsch et al. 1993) and in Renshaw cells maps the size and shape of postsynaptic glycine receptor clusters (Alvarez et al. 1997). A second approach identified Renshaw cells by their expression of calbindin D28k immunoreactivity (Arvidsson et al. 1992; Carr et al. 1998). These techniques permitted us to sample a large number of Renshaw cells and visualize substantial parts of their dendritic trees. It was found that the vast majority of cholinergic terminals are distributed, at high density, along the proximal and mid-order dendrites. These findings are discussed in relation to the distribution of other defined classes of synapses on Renshaw cells.

METHODS

Tissue preparation

Adult rats of both sexes (200–300 g Sprague-Dawley) were deeply anaesthetized with pentobarbital (85 mg kg−1i.p.) and perfused transcardially with 4 % paraformaldehyde in 0.1 M phosphate buffer (pH 7.3; PB). Spinal cord blocks from L4 and L5 segments were removed and postfixed for 1–2 h. Tissue blocks were stored overnight in 30 % sucrose in 0.1 M PB. The next day, 40 μm thick frozen transverse sections were obtained and collected in 0.01 M phosphate buffer (pH adjusted to 7.1–7.3 with 0.09 % saline; PBS).

Immunocytochemistry

The purpose of the immunolabelling procedure was to reveal VAChT-immunoreactive (VAChT-IR) terminals on Renshaw cells identified by their gephyrin or calbindin D28k immunolabelling characteristics (Alvarez et al. 1997; Carr et al. 1998). Preparations that could be analysed under bright-field light microscopy conditions were sought in order to obtain accurate morphometric data on the location of contacts in the somatodendritic tree and dimensions of the labelled structures (see below).

A number of different immunohistochemical procedures were used. Sections were blocked for 30 min to 1 h with normal horse serum (NHS; Vector Laboratories, Burlingame, CA, USA) 1:10 in PBS with 0.1 % Triton X-100 (PBS-TX) and then placed in one of the following primary antibodies: mouse anti-gephyrin antiserum (Boehringer Mannheim; dilution 1:100), rabbit anti-VAChT antiserum (Gilmor et al. 1996; dilution 1:200–500), goat anti-VACht antiserum (Chemicon, Temecula, CA, USA; dilution 1:10000–1:20000) or rabbit anti-calbindin D28k antiserum (Swant, Bellinzona, Switzerland; dilution 1:20000), diluted in PBS-TX and incubated for 2 days at 4°C. Immunoreactive sites were then revealed using standard avidin-biotin complex (ABC)-peroxidase protocols (mouse, rabbit or goat ABC kits, Vector Laboratories). Peroxidase histochemistry was performed using diaminobenzidine (DAB) as substrate (0.02 % DAB and 0.01 % H2O2 diluted in 0.05 M Tris buffer, pH 7.4). In many experiments VAChT-IR was revealed with silver intensification of the DAB reaction product (Gallyas et al. 1982; Liposits et al. 1984). After several thorough washes in PBS the sections were placed for a further 2 days incubation at 4°C in a second primary antibody complementary to the first immunoreaction. The peroxidase labelling of antigenic sites from the second immunoreaction was developed with either Vector SG chromogen- H2O2 (SG; Vector Laboratories) or DAB with no silver intensification. All the sections were dehydrated and mounted in DPX (BDH Laboratory Supplies, Poole, UK). The DAB reaction product appeared brown, SG revealed immunoreactive sites in blue-black, and DAB followed by silver intensification resulted in a black signal. The three colours were easily distinguished under bright-field microscopy. Each marker labelled different structures specific for each antibody, indicating lack of cross-talk between the first and second immunostainings.

Several variations of the combination of antibodies, the order of immunostaining, and the chromogens were tested. Three combinations provided the best labelling and were used for all the quantitative analysis reported in Results. The protocols were (1) mouse anti-gephyrin-DAB followed by rabbit or goat anti-VAChT-SG, (2) silver-intensified goat anti-VAChT-DAB followed by rabbit anti-calbindin D28k-SG, and (3) silver-intensified goat anti-VAChT-DAB followed by rabbit anti-calbindin D28k-DAB. Illustrations were prepared from bright-field or DIC images obtained with an Olympus BX60 microscope and captured with a digital colour camera (Spot2 camera, Diagnostic Instruments, Sterling Heights, MI, USA) at 24 bits per pixel. Camera gains were adjusted automatically for each image. Composite figures were prepared and labelled using CorelDraw (version 3.0).

Some sections were immunolabelled with dual colour fluorescence to reveal at high magnification the spatial relations between VAChT-IR boutons and gephyrin-IR clusters in the ventral horn neuropil. After blocking with NHS the sections were incubated at 4°C for 2 days in a combination of primary antibodies including mouse anti-gephyrin (Boehringer Mannheim; dilution 1:100) and goat anti-VAChT antisera (Chemicon; dilution 1:2000) in PBS-TX. Immunoreactive sites were revealed with donkey anti-mouse IgG coupled to Cy3 and donkey anti-goat IgG coupled to fluorescein isothiocyanate (FITC; Jackson Labs, West Grove, PA, USA; dilution 1:25 in PBS-TX) or tetramethylrhodamine isothiocyanate (TRITC; Jackson Labs; dilution 1:25 in PBS-TX). The sections were then washed, and mounted in Vectashield (Vector Laboratories), and analysed with confocal microscopy (Olympus Fluoview; ×60 oil immersion objective lens, NA 1.4).

Quantitative analysis

Sections were analysed under bright-field illumination using a ×100 objective lens. Only neurones that had extensive dendritic labelling within the plane of a single section were analysed. As previously described (Alvarez et al. 1998), we use the term ‘contact’ for close appositions between immunolabelled presynaptic axonal varicosities (boutons) and immunolabelled postsynaptic structures, in this case the soma or dendrites of immunohistochemically identified Renshaw cells; interpretation of this type of data is fully described elsewhere (Fyffe, 1991a; Alvarez et al. 1998). Cell somas and dendritic outlines were traced using a computer-aided neurone tracing system (Eutectic, Raleigh, NC, USA) and the locations of all contacts plotted on these reconstructions. Dendritic surface area was calculated from dendritic lengths and diameters. Somatic surface area was estimated as the surface of an ellipsoid with major and minor diameters equal to the major and minor cross-sectional diameters of the cell soma. Dendrites were divided in bin segments at 25 μm increments of distance from the cell soma. The density of VAChT-IR contacts was expressed as the number of contacts per 100 μm2 of surface membrane in each dendritic bin or soma.

For measurement of bouton dimensions, VAChT-IR terminals were drawn at ×100 using a camera lucida. For this analysis we used DAB-silver-intensified VAChT-IR terminals since these gave the best definition of the bouton contours. The drawings were scanned (Hewlett-Packard 4c) into a computer and their major diameters calculated using Image Pro-Plus Analysis software (Media Cybernetics, Silver Spring, MD, USA). Statistical differences were tested using a one-way ANOVA (significance was set at P < 0.001), or Kruskal-Wallis one-way ANOVA when the samples did not distribute normally. Pairwise comparisons were performed using Dunn's test (significance set at P < 0.05). All statistical tests were performed in Sigma Stat version 2.0 (Jandel, San Rafael, CA, USA).

RESULTS

VAChT staining in the spinal cord

The anti-VAChT antibodies revealed a plexus of immunostained axons and varicosities throughout the grey matter of the spinal cord L4 and L5 segments (Fig. 1). The most intense labelling was observed in the ventral horn, particularly in and around the motor pools of lamina IX. Within lamina IX, the cell bodies of large cells, presumed motoneurones, contained immunoreactivity in their Nissl and Golgi network and were surrounded by large VAChT-IR terminals (Fig. 1). The terminals apposed to motoneurones were the largest VAChT-IR terminals in the spinal cord (see below). Other neurones in and around the motoneurone pools displayed fewer (if any) juxtasomatic appositions with VAChT-IR terminals (see Fig. 1) and the terminals were usually of smaller size. The overall labelling pattern was essentially the same regardless of the particular antibody or chromogen used. Gephyrin and calbindin labelling patterns were as described previously (Alvarez et. al. 1997; Carr et al. 1998).

Figure 1. Distribution of VAChT-IR terminals and calbindin D28k-IR cells in the ventral horn of the rat spinal cord.

Figure 1

Open in a new tab

A, low power image of the ventral horn of the rat spinal cord L5 segment immunostained with antibodies against VAChT (black-brown reaction deposits; silver-intensified DAB) and calbindin D28k (blue-black SG reaction deposits). B and C show the regions enclosed in A at higher magnification and visualized using DIC optics. VAChT immunoreactivity (brown-black) is present inside the cell bodies of presumed motoneurones (arrowheads) and in axon terminals throughout the grey matter of the ventral horn. The density of VAChT immunoreactivity inside cell bodies is relatively low, in contrast to the more intense VAChT immunoreactivity inside the terminals; the latter structures thus contain a more prominent (black) silver deposit after intensification. Small VAChT-IR terminals are present at low density throughout the grey matter of the ventral horn. Larger VAChT-IR terminals form two distinct types of arrangements in the ventral horn. The most prominent VAChT-IR terminals (corresponding to C-terminals) surround the cell bodies and proximal dendrites of motoneurones. In B, a motoneurone (brown cytoplasm) appears covered by VAChT-IR C-terminals (black). The final group of VAChT-IR terminals in the ventral horn form a plexus in ventral lamina VII. In this region, VAChT-IR terminals are predominantly associated with calbindin D28k-IR dendrites (blue-black). This area is shown at higher magnification in C. Calbindin D28k-IR cells in this area have been shown to correspond with Renshaw cells according to immunocytochemical and electrophysiological criteria (Alvarez et al. 1997; Carr et al. 1998). Other, larger, ventral horn interneurones located more dorsally also express calbindin D28k immunoreactivity (see B, blue-black neurone) but receive very little input from VAChT-IR terminals. These calbindin D28k-IR neurones were found not to display Renshaw cell characteristics (Carr et al. 1998). Scale bars are 500 μm in A and 100 μm in C. B is at the same magnification as C.

Identification of Renshaw cells

A total of 50 lamina VII interneurones that fulfilled the immunohistochemical identification criteria for Renshaw cells were analysed quantitatively. Twenty Renshaw cells were labelled with non-intensified brown DAB-gephyrin immunoreactivity and blue SG-VAChT-IR contacts, and 30 Renshaw cells were identified with calbindin D28k immunoreactivity (7 revealed with blue-black SG reaction product and 23 with brown DAB) and associated VAChT-IR terminals labelled with silver-intensified DAB.

The extent to which dendrites could be traced differed between the two markers. Gephyrin immunoreactivity revealed the dendrites up to a distance of about 180 μm from the soma, with a mean dendritic length stained of 127.1 ± 39.98 μm (mean ±s.d.). Calbindin D28k revealed more distal dendrites, to a distance of 413 μm from the soma (mean ±s.d., 253.45 ± 62.13 μm). The densities of VAChT-IR terminals were first calculated independently in each set. The density of contacts in the first 175 μm of dendrite in Renshaw cells identified with gephyrin immunoreactivity was always slightly larger than when calbindin D28k immunoreactivity was used for identification (perhaps because the exact limits of dendritic profiles are harder to determine when following the outline of gephyrin-IR clusters and therefore dendritic diameters might have been slightly underestimated in this material). Nevertheless, Kruskal- Wallis one-way ANOVA demonstrated that this difference did not reach statistical significance (P > 0.05). Hence, data from all 50 Renshaw cells were pooled for subsequent analysis.

VAChT-IR terminals associated with Renshaw cells

Double labelling revealed numerous VAChT-IR terminals (n= 3330) in contact with Renshaw cells, with contact sites including the cell bodies and dendritic trunks, as well as higher-order dendritic branches to the limits of dendrite visualization (Fig 1, Fig 2 and Fig 3). The number of contacts per Renshaw cell ranged from 21 to 138, with a mean of 66.6 ± 25.56 (s.d.) per cell. Calbindin D28k-labelled cells had more contacts (mean ±s.d., 72.17 ± 27.79 per cell) than gephyrin-labelled cells (mean ±s.d., 58.25 ± 19.59 per cell) because of the greater length of dendrite available for analysis. Figure 4 shows reconstructions of four Renshaw cells (labelled with antibodies against calbindin D28k) and the location of VAChT-IR terminals found in contact with the cells. Overall, about 16.5 % (n= 549) of the VAChT-IR contacts on Renshaw cells were apposed to the somatic membrane, with the majority (82.5 %; n= 2781) being on the dendrites. In many cases multiple boutons from a single preterminal axon contacted a single neurone (Fig 2 and Fig 3). The initial part of the Renshaw cell axon could be identified in a few cases and did not receive any VAChT-IR contacts.

Figure 2. High magnification images showing examples of VAChT-IR terminals contacting calbindin D28k-IR Renshaw cells.

Figure 2

Open in a new tab

A, a Renshaw cell identified according to its prominent gephyrin immunoreactivity (brown clusters covering the somatodendritic membrane) and contacted by VAChT-IR terminals (blue-black SG reaction product, arrows). B, a Renshaw cell identified by its calbindin D28k-immunoreactivity (blue-black SG reaction product) contacted by VAChT-IR terminals (black reaction product; silver-intensified DAB; arrows). C, a Renshaw cell identified by its calbindin D28k immunoreactivity (brown reaction product, DAB) contacted by VAChT-IR terminals (black reaction product; silver-intensified DAB; arrows). In all three examples, although there are some juxtasomatic contacts (arrowheads in C), most VAChT-IR contacts are located on the dendrites. The inset in C shows the diversity in sizes of VAChT-IR terminals in contact with Renshaw cell calbindin D28k-IR dendrites. A and B are at the same magnification. Scale bars in B and C, 50 μm. Scale bar in inset, 5 μm.

Figure 3. High magnification reconstruction of axon terminal collaterals displaying multiple VAChT-IR boutons (black) in contact with the somatodendritic membrane of a Renshaw cell (brown).

Figure 3

Open in a new tab

A series of optical sections was obtained under DIC optics and the various optical planes were superimposed to render a reconstruction of several Renshaw cell dendrites and their complement of VAChT-IR contacts. It is frequently possible to follow single VAChT-IR collaterals making multiple en passent contacts with a segment of dendrite (arrows). Scale bar, 50 μm.

Figure 4. Camera lucida reconstructions of 4 Renshaw cells from the sample identified by their calbindin D28k immunoreactivity.

Figure 4

Open in a new tab

VAChT-IR terminals in contact with the Renshaw cells are indicated by the irregular black spots. Intervaricose portions of the presynaptic axons were omitted for clarity. VAChT-IR terminals are distributed over the whole extent of the stained dendrites, especially at distances beyond the most proximal regions (< 25 μm) of each dendrite. Scale bar, 50 μm.

The density of contacts was measured on the soma and dendrites (Fig. 5). Contact density was relatively low (about 1.1 per 100 μm2) on the soma, changed abruptly to a higher level on the dendritic trunks (within 25 μm of the soma) and reached maximal levels (usually about 4–5 per 100 μm2) about 50 μm from the soma. The contact density remained relatively constant over the range 50–250 μm. Only one dendrite was sampled beyond 300 μm and it displayed a lower density of contacts. It is not known whether additional cholinergic contacts are made on the most distal dendrites, which may extend out to about 800 μm (see e.g. Lagerbäck & Kellerth, 1985b; Alvarez et al. 1997).

Figure 5. Histogram of the average density of VAChT-IR contacts at different distances from the cell soma.

Figure 5

Open in a new tab

Numbers in parentheses indicate the number of Renshaw cells that displayed dendritic segments at each distance (total number of cells = 50). We were able to follow dendrites beyond a distance of 200 μm from the cell body in less than half of the Renshaw cells. Error bars indicate s.e.m.

Overall, the density of dendritic cholinergic contacts over the somatodendritic membrane ranged from about one to four per 100 μm2 in different cells. The inter-cell variability in contact density was not due to the method used (gephyrin or calbindin D28k) to identify the Renshaw cells, and was not correlated with soma size or shape, or the length of dendrite that was analysed. In gephyrin-identified Renshaw cells, contact densities varied from 1.45 to 3.89 contacts per 100 μm2 (mean ±s.d., 2.45 ± 0.61 μm). In calbindin D28k-identified cells the range was 1.29–4.16 contacts per 100 μm2 (mean ±s.d., 2.68 ± 0.78 μm).

The VAChT-IR terminals apposed to Renshaw cells were of variable size (Fig 1, Fig 2, Fig 3 and Fig 5). The inset in Fig. 2C shows three adjacent contacts on a Renshaw cell dendrite, including a small, medium and large terminal in apposition. VAChT-IR boutons on Renshaw cells ranged in size from 0.6 to 6.9 μm in their longest axis (mean ±s.d., 2.26 ± 0.94 μm; n= 986). Figure 6A shows drawings of a subset of contacts on Renshaw cells, for comparison with samples representing the terminals in contact with presumed motoneurone cell bodies (Fig. 6C) and terminals in contact with calbindin D28k-IR non-Renshaw cell interneurones in dorsal lamina VII (Fig. 6E). The size distributions are illustrated in the corresponding histograms (Fig. 6B, D and F).

Figure 6. Size distribution of VAChT-IR terminals in relation to Renshaw cells (A and B), motoneurones (C and D) or other calbindin D28k-IR cells in the dorsal aspects of LVII (E and F).

Figure 6

Open in a new tab

A, C and E show examples of camera lucida drawings of each terminal population. Histograms B, D and F show the distribution of the maximun diameter of individual terminals. Scale bar in E applies to panels A and C, 10 μm.

Compared with the VAChT-IR terminals on Renshaw cells, those contacting the proximal dendrites and somas of lamina IX motoneurones were larger (range 1.09–8.4 μm; mean ±s.d., 3.76 ± 1.21 μm; n= 441). In contrast, those contacting calbindin D28k-IR interneurones in dorsal lamina VII were smaller (range 0.21–3.98 μm; mean ±s.d. 1.68 ± 0.57 μm; n= 444). One-way ANOVA of these bouton samples showed that the differences in size were statistically significant among the three populations of VAChT-IR terminals (P < 0.001).

The combination of gephyrin and VAChT immunolabelling also provided the opportunity to examine the relationships of cholinergic boutons to glycinergic postsynaptic sites, using confocal microscopy (Fig. 7). VAChT-IR terminals were never seen to directly overlie the patches of gephyrin immunoreactivity but were apposed to membrane interposed between patches of gephyrin staining. Both classes of synapse, cholinergic and glycinergic, were closely intermingled on the dendrites out to about 175 μm from the soma. In the cell soma and initial 50 μm of proximal dendrite glycinergic synaptic sites greatly outnumber VAChT-IR boutons (Fig. 7).

Figure 7. Confocal images of a dual colour immunofluorescence preparation showing VAChT-IR terminals (green, FITC) and gephyrin-IR clusters (red, TRITC).

Figure 7

Open in a new tab

VAChT-IR terminals (arrows) were never associated with postsynaptic gephyrin clusters. The neurone illustrated (A, 15 focal planes; B, 4 focal planes) was identified as a Renshaw cell based on the size and density of the complement of gephyrin-IR clusters in the soma and proximal dendrites. Scale bar, 10 μm.

DISCUSSION

The major conclusion from our study is that the cholinergic input to Renshaw cells is predominantly distributed, at high density, on the dendrites. Although this type of spatial distribution of synapses should be expected because there is much more membrane surface available on the dendrites, it is striking that such a powerful synaptic input as that from motor axon collaterals is largely removed from the soma and juxtasomatic region. However, the large size of the boutons and the high density of cholinergic synapses contacting Renshaw cell dendrites very likely constitutes the structural basis for the strength and efficacy characteristic of synaptic input from motor axon collaterals to Renshaw cells. On the other hand, the higher density of glycinergic synapses in the soma and juxtasomatic region, relative to this cholinergic input, suggests an orderly spatial organization by which the cholinergic input could be strongly modulated by glycinergic inhibitory mechanisms. A modulation of this sort might explain observations that, in certain types of fictive locomotion, Renshaw cells fire at relatively low rates even though they are capable of high frequency firing (e.g. McCrea et al. 1980).

Immunostaining with anti-VAChT, anti-gephyrin and anti-calbindin D28k antisera in the ventral horn

Previous studies have demonstrated that antibodies against the vesicular acetylcholine transporter provide reliable visualization of cholinergic terminals (Gilmor et al. 1996; Arvidsson et al. 1997). Thus, we are confident that VAChT immunostaining specifically labels cholinergic terminals. The antibody labels fibres and axonal varicosities of variable size. Some of the largest VAChT-IR terminals are apposed to the soma and proximal dendrites of presumed motoneurones, and are therefore likely to be the C-terminals, whose origin remains uncertain (Conradi, 1969; Nagy et al. 1993; Li et al. 1995). In ventral lamina VII there is great variability in the size of cholinergic terminals, and those in contact with Renshaw cells are larger, on average, than those in contact with other lamina VII interneurones. Cholinergic synapses onto Renshaw cells comprise a type of VAChT-IR bouton that could be distinguished by their size from both VAChT-IR C-terminals and other VAChT-IR terminals scattered in the ventral horn neuropil and contacting non-Renshaw cell interneurones. Although there was some overlap in the distribution of bouton sizes between the three populations, statistical analysis indicates that they form three distinct populations, with the origin of cholinergic boutons contacting Renshaw cells most probably being the recurrent collaterals of motor axons.

The Renshaw cells labelled by gephyrin or calbindin D28k expression are entirely consistent with earlier descriptions of their size, location, dendritic morphology and number (Fyffe, 1990; Alvarez et al. 1997; Carr et al. 1998).

The major problem concerning this method of identifying Renshaw cells is that the immunostaining does not reveal the full extent of the dendrites. In addition, we only sampled dendrites which could be followed for a considerable distance within a single section. The data therefore seriously underestimate the absolute number of cholinergic synapses on Renshaw cell dendrites. It is difficult to estimate just how much of the dendrites are missed in this analysis, because there are no morphometric data on intracellularly stained Renshaw cells in the rat, as there are for the cat (Lagerbäck & Kellerth, 1985; Alvarez et al. 1997), although our qualitative impression is that Renshaw cells in the rat are smaller than those in the cat. The same problem also causes an overestimate of the proportion of cholinergic synapses formed on the soma versus the dendrites. However, the methods do permit evaluation of the density of contacts on the soma and proximal dendrites (up to about 300 μm from the soma), and allowed us to study a much larger sample of neurones than would have been possible using intracellular staining techniques to visualize dendritic trees in their entirety. The tissue sections included several hundred other Renshaw cells which displayed identical patterns of staining to the 50 that were used for quantitative analysis, suggesting that our sample provides a good representation of Renshaw cells in the lumbar region of the rat spinal cord. It remains to be determined whether there are species differences in the distribution of cholinergic inputs, or if there are differences between Renshaw cells located at different segmental levels.

The criteria we used for defining a contact at the light microscopic level are well established (Brown & Fyffe, 1981; Fyffe, 1991a; Alvarez et al. 1998). Although electron microscopic examination is required to establish the synaptic nature of the contacts, the light microscopic approach has given reliable estimates of contact number in many similar studies. The determination of contacts over the soma of stained neurones is always somewhat problematic. This problem was not as much of an issue in the present study as in other recent studies (e.g. Alvarez et al. 1998) because different chromogens were used to label the pre- and postsynaptic structures, and the different colours were usually easily differentiated. Nor was penetration of antibody or other reagents into the tissue likely to have affected the results to a great extent, because there appeared to be relatively uniform staining throughout the depth of each section. Thus we conclude that the distribution patterns described here are a reasonably accurate indication of the location and density of contacts produced by cholinergic axons on the soma and proximal dendrites of Renshaw cells.

Cholinergic contacts on Renshaw cells

The density of dendritic cholinergic contacts is several orders of magnitude higher than that of serotoninergic inputs to Renshaw cells (P. A. Carr, J. C. Pearson and R. E. W. Fyffe, unpublished observations) and several times higher than the density of serotoninergic and I a afferent inputs to motoneurones (e.g. Burke & Glenn, 1996; Alvarez et al. 1998). Although the density of glycinergic contacts has not been calculated over the complete neuronal surface, ultrastructural analysis indicates that glycinergic inhibitory synapses comprise around 80 % of the somatic synapses on Renshaw cells, and about 75 % of synapses on the proximal dendrites (D. A. Harrington, F. J. Alvarez & R. E. W. Fyffe, 1994; unpublished observations); S-type synapses, which presumably include the population of cholinergic synapses, comprise about 10 % and 20 %, respectively. Thus the vast majority of synapses on the soma and proximal dendrites of Renshaw cells are either glycinergic or cholinergic. The increase in the proportion of S-type synapses from the soma to the proximal dendrites correlates well with the increase in cholinergic contact density seen in the present study over that same region.

It is likely that a significant proportion of the cholinergic inputs are formed by the axon collaterals of motoneurones and provide the substrate for the powerful recurrent input to Renshaw cells. The average size of VAChT-IR boutons that contact Renshaw cells was significantly different to that of other VAChT-IR boutons in the ventral horn neuropil. Nevertheless, their size distribution displayed considerable variability from terminal to terminal (see also Lagerbäck et al. 1978, 1981). This size variability may ultimately be reflected in variability in synaptic transmission at different contact sites. Overall, however, the involvement of large, densely packed boutons and the fact that chains of boutons often make contact along a single postsynaptic dendrite, presumably reflects a structural basis for effective synaptic transmission (e.g. Walmsley et al. 1998). Using minimal (possibly a single input axon) stimulation of motor axons it is possible to reveal the underlying EPSP (B. Walmsley & D. J. Tracey, 1981; unpublished observations), but since action potentials are frequently superimposed on the early part of the EPSP, measurement of the EPSP rise time is difficult. On the other hand, the EPSPs have a long time course (Walmsley & Tracey, 1981), and it would be reasonable to propose that the time course reflects the dendritic distribution of synapses rather than the time course of transmitter action or channel properties (Sargent, 1993).

It would be of interest to know the degree of divergence and convergence in the recurrent inhibitory pathway, and the present results provide information about another variable that is necessary for understanding the organization of the segmental motor system. Previous work has shown that the boutons generated by single α-motoneurone axons are distributed within a rostrocaudal distance of less than 1 mm around the position of the parent cell bodies and that each α-motoneurone, depending upon its functional type, generates on average about 60 axon collateral swellings (Cullheim & Kellerth, 1978), the great majority of which contact only one postsynaptic profile (soma or dendrite; Lagerbäck et al. 1981). Within a 1 mm length of spinal cord, there are about 100 Renshaw cells in each ventral horn (Carr et al. 1998). Thus, the maximum number of Renshaw cells that could be contacted by a single motoneurone (i.e. if it made only a single contact with each postsynaptic Renshaw cell) is in the range of about 50–100. This range is likely to be an overestimate because several boutons from a single collateral may contact the same Renshaw cell (Lagerbäck et al. 1981; present observations). Because the distal dendrites were not labelled we lack a measurement of the total number of cholinergic contacts on a single cell. However, we do know the density of contacts over the proximal dendrites. Extrapolating to the total Renshaw cell surface area (data for cat Renshaw cells, e.g. Lagerbäck & Kellerth, 1985; R. E. W. Fyffe, unpublished observations), we estimate that there may be a maximum of 400–500 cholinergic contacts in total on each Renshaw cell. Since each cholinergic axon collateral appears to contribute multiple inputs (Lagerbäck et al. 1981; present observations), the total contacts probably derive from, at most, 50–100 motoneurones. However, this indirect estimate requires confirmation by future analysis of contacts on intracellularly stained Renshaw cells to more accurately determine the number of contacts formed by collaterals from a single motor axon on a single Renshaw cell.

Acknowledgments

We are grateful to Dr Allan Levey for the generous gift of rabbit anti-VAChT antibodies. We thank Dr Patrick Carr for advice and Paula Somohano and Leslie Spencer for technical assistance. This work was supported by National Institutes of Health grants R01 NS 25547 to R. E. W. Fyffe and R29 NS 33555 to F. J. Alvarez.

References

  1. Alvarez FJ, Harrington DA, Dewey DE, Fyffe REW. Cell-type specific organization of glycine receptor clusters in the mammalian spinal cord. Journal of Comparative Neurology. 1997;379:150–170. 10.1002/(SICI)1096-9861(19970303)379:1<150::AID-CNE10>3.3.CO;2-M. [PubMed] [Google Scholar]
  2. Alvarez FJ, Pearson JC, Harrington DA, Dewey DE, Torbeck L, Fyffe REW. Distribution of 5-hydroxytryptamine-immunoreactive boutons on α-motoneurons in the lumbar spinal cord of adult cats. Journal of Comparative Neurology. 1998;393:69–83. 10.1002/(SICI)1096-9861(19980330)393:1<69::AID-CNE7>3.3.CO;2-X. [PubMed] [Google Scholar]
  3. Arvidsson U, Reidl M, Elde R, Meister B. Vesicular acetylcholine transporter (VAChT) protein: A novel and unique marker for cholinergic neurons in the central and peripheral nervous systems. Journal of Comparative Neurology. 1997;378:454–467. 10.1002/(SICI)1096-9861(19970224)378:4<454::AID-CNE2>3.3.CO;2-5. [PubMed] [Google Scholar]
  4. Arvidsson U, Ulfhake B, Cullheim S, Ramírez V, Shupliakov O, Hökfelt T. Distribution of calbindin D28k-like immunoreactivity (LI) in the monkey ventral horn: do Renshaw cells contain calbindin D28k-LI? Journal of Neuroscience. 1992;12:718–728. doi: 10.1523/JNEUROSCI.12-03-00718.1992. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Binder MD, Heckman CJ, Powers RK. The physiological control of motoneuron activity. In: Rowell LB, Shepherd JT, editors. Handbook of Physiology, section 12, Exercise: Regulation and Integration of Multiple Systems. New York, Oxford: American Physiological Society; 1996. pp. 3–53. [Google Scholar]
  6. Brown AG, Fyffe REW. Direct observations on the contacts made between I a afferent fibres and motoneurones in the cat's lumbosacral spinal cord. The Journal of Physiology. 1981;313:121–140. doi: 10.1113/jphysiol.1981.sp013654. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Burke RE, Glenn LL. Horseradish peroxidase study of the spatial and electrotonic distribution of group Ia synapses on type-identified ankle extensor motoneurons in the cat. Journal of Comparative Neurology. 1996;372:465–485. doi: 10.1002/(SICI)1096-9861(19960826)372:3<465::AID-CNE9>3.0.CO;2-0. [DOI] [PubMed] [Google Scholar]
  8. Carr PA, Alvarez FJ, Leman EA, Fyffe REW. Calbindin D28k expression in immunohistochemically-identified Renshaw cells. NeuroReport. 1998;9:2657–2661. doi: 10.1097/00001756-199808030-00043. [DOI] [PubMed] [Google Scholar]
  9. Conradi S. Ultrastructure and distribution of neuronal and glial elements on the motoneuron surface in the lumbosacral spinal cord of the adult cat. Acta Physiologica Scandanavica Supplementum. 1969;332:5–48. [PubMed] [Google Scholar]
  10. Cullheim S, Kellerth J-O. A morphological study of the axons and recurrent axon collaterals of cat α-motoneurones supplying different functional types of muscle unit. The Journal of Physiology. 1978;281:301–313. doi: 10.1113/jphysiol.1978.sp012423. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Cullheim S, Kellerth J-O. Two kinds of recurrent inhibition of cat spinal α-motoneurones as differentiated pharmacologically. The Journal of Physiology. 1981;312:209–224. doi: 10.1113/jphysiol.1981.sp013624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Curtis DR, Ryall RW. The excitation of Renshaw cells by cholinomimetics. Experimental Brain Research. 1966a;2:49–65. doi: 10.1007/BF00234360. [DOI] [PubMed] [Google Scholar]
  13. Curtis DR, Ryall RW. The acetylcholine receptors of Renshaw cells. Experimental Brain Research. 1966b;2:66–80. doi: 10.1007/BF00234361. [DOI] [PubMed] [Google Scholar]
  14. Curtis DR, Ryall RW. The synaptic excitation of Renshaw cells. Experimental Brain Research. 1966c;2:81–96. doi: 10.1007/BF00234362. [DOI] [PubMed] [Google Scholar]
  15. Eccles JC, Fatt P, Koketsu K. Cholinergic and inhibitory synapses in a pathway from motor axon collaterals to motoneurones. The Journal of Physiology. 1954;126:524–562. doi: 10.1113/jphysiol.1954.sp005226. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Fyffe REW. Evidence for different morphological classes of Renshaw cells in the cat's spinal cord. Brain Research. 1990;536:301–304. doi: 10.1016/0006-8993(90)90038-d. 10.1016/0006-8993(90)90038-D. [DOI] [PubMed] [Google Scholar]
  17. Fyffe REW. Glycine-like immunoreactivity in synaptic terminals of identified inhibitory interneurons in the cat's spinal cord. Brain Research. 1991a;547:175–179. doi: 10.1016/0006-8993(91)90590-r. 10.1016/0006-8993(91)90590-R. [DOI] [PubMed] [Google Scholar]
  18. Fyffe REW. The spatial distribution of recurrent inhibitory synapses on cat spinal motoneurons. Journal of Neurophysiology. 1991b;65:1134–1149. doi: 10.1152/jn.1991.65.5.1134. [DOI] [PubMed] [Google Scholar]
  19. Gallyas F, Gorcs T, Merchenthaler I. High-grade intensification of the end-product of the diaminobenzidine reaction for peroxidase histochemistry. Journal of Histochemistry and Cytochemistry. 1982;30:183–184. doi: 10.1177/30.2.7061820. [DOI] [PubMed] [Google Scholar]
  20. Gilmor ML, Nash NR, Roghani A, Edwards RH, Yi H, Hersch SM, Levey AI. Expression of the putative vesicular acetylcholine transporter in rat brain and localization in cholinergic synaptic vesicles. Journal of Neuroscience. 1996;16:2179–2190. doi: 10.1523/JNEUROSCI.16-07-02179.1996. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Harrington DA, Alvarez FJ, Fyffe REW. Somatic membrane covering by glycinergic terminals and glycine receptors of α-motoneurons and Renshaw cells in cat spinal cord. Society for Neuroscience Abstracts. 1994;20:1588. [Google Scholar]
  22. Kirsch J, Wolters I, Triller A, Betz H. Gephyrin antisense oligonucleotides prevent glycine receptor clustering in spinal neurons. Nature. 1993;266:745–748. doi: 10.1038/366745a0. 10.1038/366745a0. [DOI] [PubMed] [Google Scholar]
  23. Lagerbäck P-Å, Kellerth J-O. Light microscopic observations on cat Renshaw cells after intracellular staining with horseradish peroxidase. II. The cell bodies and dendrites. Journal of Comparative Neurology. 1985;240:368–376. doi: 10.1002/cne.902400405. [DOI] [PubMed] [Google Scholar]
  24. Lagerbäck P-Å, Ronnevi L-O. An ultrastructural study of serially sectioned Renshaw cells. I. Architecture of the cell body, axon hillock, initial axon segment and proximal dendrites. Brain Research. 1982;235:1–15. doi: 10.1016/0006-8993(82)90192-5. 10.1016/0006-8993(82)90192-5. [DOI] [PubMed] [Google Scholar]
  25. Lagerbäck P-Å, Ronnevi L-O, Cullheim S, Kellerth J-O. Ultrastructural characteristics of a central cholinergic synapse in the cat. Brain Research. 1978;148:197–201. doi: 10.1016/0006-8993(78)90389-x. 10.1016/0006-8993(78)90389-X. [DOI] [PubMed] [Google Scholar]
  26. Lagerbäck P-Å, Ronnevi L-O, Cullheim S, Kellerth J-O. An ultrastructural study of the synaptic contacts of α-motoneuron axon collaterals. II. Contacts in lamina VII. Brain Research. 1981;222:29–41. doi: 10.1016/0006-8993(81)90938-0. 10.1016/0006-8993(81)90938-0. [DOI] [PubMed] [Google Scholar]
  27. Li W, Ochalski PAY, Brimijoin S, Jordan LM, Nagy JI. C-terminals on motoneurons: electron microscope localization of cholinergic markers in adult rats and antibody-induced depletion in neonates. Neuroscience. 1995;65:879–891. doi: 10.1016/0306-4522(94)00511-3. 10.1016/0306-4522(94)00511-3. [DOI] [PubMed] [Google Scholar]
  28. Liposits Zs, Setalo Gy, Flerko B. Application of the silver-gold intensified 3,3-diaminobenzidine chromogen to the light and electron microscopic detection of the luteinizing hormone-releasing system of the rat brain. Neuroscience. 1984;13:513–525. doi: 10.1016/0306-4522(84)90245-8. 10.1016/0306-4522(84)90245-8. [DOI] [PubMed] [Google Scholar]
  29. McCrea DA, Pratt CA, Jordan LM. Renshaw cell activity and recurrent effects on motoneurons during fictive locomotion. Journal of Neurophysiology. 1980;44:475–488. doi: 10.1152/jn.1980.44.3.475. [DOI] [PubMed] [Google Scholar]
  30. Nagy JI, Yamamoto T, Jordan LM. Evidence for the cholinergic nature of C-terminals associated with subsurface cisterns in α-motoneurons of rat. Synapse. 1993;15:17–32. doi: 10.1002/syn.890150103. [DOI] [PubMed] [Google Scholar]
  31. Noga BR, Shefchyk SJ, Jamal J, Jordan LM. The role of Renshaw cells in locomotion: antagonism of their excitation from motor axon collaterals with intravenous mecamylamine. Experimental Brain Research. 1987;66:99–105. doi: 10.1007/BF00236206. [DOI] [PubMed] [Google Scholar]
  32. Ross H-G, Cleveland S, Haase J. Contribution of single motoneurons to Renshaw cell activity. Neuroscience Letters. 1975;1:105–108. doi: 10.1016/0304-3940(75)90053-1. 10.1016/0304-3940(75)90053-1-1">10.1016/0304-3940(75)90053-1. [DOI] [PubMed] [Google Scholar]
  33. Sargent PB. The diversity of neuronal nicotinic acetylcholine receptors. Annual Review of Neuroscience. 1993;16:403–443. doi: 10.1146/annurev.ne.16.030193.002155. 10.1146/annurev.ne.16.030193.002155. [DOI] [PubMed] [Google Scholar]
  34. Schneider SP, Fyffe REW. Involvement of GABA and glycine in recurrent inhibition of spinal motoneurons. Journal of Neurophysiology. 1992;68:397–406. doi: 10.1152/jn.1992.68.2.397. [DOI] [PubMed] [Google Scholar]
  35. van Keulen LCM. Autogenetic recurrent inhibition of individual spinal motoneurons of the cat. Neuroscience Letters. 1981;21:297–300. doi: 10.1016/0304-3940(81)90220-2. 10.1016/0304-3940(81)90220-2. [DOI] [PubMed] [Google Scholar]
  36. Walmsley B, Alvarez FJ, Fyffe REW. Diversity of structure and function at mammalian central synapses. Trends in Neurosciences. 1998;21:81–88. doi: 10.1016/s0166-2236(97)01170-3. 10.1016/S0166-2236(97)01170-3. [DOI] [PubMed] [Google Scholar]
  37. Walmsley B, Tracey DJ. An intracellular study of Renshaw cells. Brain Research. 1981;223:170–175. doi: 10.1016/0006-8993(81)90818-0. 10.1016/0006-8993(81)90818-0. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES