1. Top of page
  2. Abstract
  6. Acknowledgements
  • 1
    Noradrenaline is known to suppress transmission from group II muscle afferents when locally applied to γ-motoneurones, and serotonin (5-HT) facilitates the transmission. The purpose of this investigation was to search for evidence of monoaminergic innervation of γ-motoneurones.
  • 2
    Eight γ-motoneurones were labelled with rhodamine-dextran, and 50 μm thick sagittal sections of the spinal cord containing them were exposed to antibodies against dopamine β-hydroxylase (DBH) and 5-HT. All the cells were directly and/or indirectly excited by muscle group II afferents from the muscle they innervated and/or other muscles.
  • 3
    Appositions between monoaminergic fibres and the labelled somata and dendrites were located with three-colour confocal laser scanning microscopy by examining series of optical sections at 1 or 0.5 μm intervals.
  • 4
    DBH and 5-HT varicosities formed appositions with the somata and dendrites of all the γ-motoneurones. The mean packing densities for 5-HT (1.12 ± 0.11 appositions per 100 μm2 for somata and 0.91 ± 0.07 per 100 μm2 for dendrites) were similar to the densities of contacts reported for α-motoneurones. Monoaminergic varicosities in apposition to dendrites greatly outnumbered those on the somata.
  • 5
    The density of DBH appositions was consistently lower – corresponding means were 53 % and 62 % of those for 5-HT on the somata and dendrites, respectively.
  • 6
    It is concluded from an analysis of the distribution and density of varicosities in apposition to the γ-motoneurones compared with the density in the immediate surround of the dendrites that there is indeed both a serotoninergic and noradrenergic innervation of γ-motoneurones.

It has been known for some time that the activity of γ-motoneurones can be modulated by serotonin (5-HT) and noradrenaline (NA), and by their precursors and agonists. 5-HT excited γ-motoneurones, as did the precursor 5-hydroxytryptophan and agonist lysergic acid diethylamide (LSD) (Ahlman et al. 1971; Ellaway & Trott, 1975). The effects of noradrenaline and related compounds are more complex. Systemic L-3,4-dihydroxyphenylalanine (L-DOPA), the NA precursor, had a differential effect on the background discharges of dynamic and static γ-motoneurones in flexors and extensors; it was assumed that L-DOPA acted by releasing NA from noradrenergic axons (Grillner et al. 1967; Bergmans & Grillner, 1968; Grillner, 1969). However, clonidine, an α2 noradrenergic agonist, diminished or stopped background discharges of γ-motoneurones (Bennett et al. 1996).

It was recognised that the monoamines were not necessarily acting directly on γ-motoneurones, but could be acting on components of pathways impinging on them. Alternative sites of action could be in descending or spinal pathways, or even peripherally in the muscle spindles (see for example Ellaway & Trott, 1975; Bennett et al. 1996). Various types of interneurones are now known to respond to locally applied NA and 5-HT (Jankowska et al. 2000), among them intermediate zone interneurones which are premotor to γ-motoneurones. However, NA and 5-HT do indeed have actions at the level of γ-motoneurones. Ionophoretic application of NA and its α2 agonist tizanidine depressed responses evoked by stimulating group II muscle afferents, while 5-HT enhanced the resting activity of γ-motoneurones and facilitated activation by these afferents (Jankowska et al. 1998). This suggested a serotoninergic innervation, but the effects of 5-HT and NA on synaptic transmission could have been presynaptic. Amines may also modulate the activity of neurones when they are released as far as 20 μm away from the cell (see Bunin & Wightman, 1998); in fact, in primitive vertebrates serotoninergic fibres do not form synapses with motoneurones (see Jacobs & Azmitia, 1992). No morphological information was available about any monoaminergic innervation of γ-motoneurones. Accordingly we labelled some cells during the experiments reported (Jankowska et al. 1998) with rhodamine-dextran, and processed the spinal cords to reveal the varicosities of noradrenergic and serotoninergic axons in the immediate vicinity of the cells by immunofluorescence.

Confocal microscopy was employed to identify the monoaminergic varicosities and define their relation to the labelled cells. Actual contacts between varicosities and cells cannot be determined with complete confidence by confocal microscopy (or non-confocal light microscopy) because it is not possible to exclude the presence of very thin cell processes between the fluorescences of the varicosity and cell. However, whereas any intervening processes are readily identified by electron microscopy, it would be unrealistic to use this method to assess the distribution of varicosities in relation to the γ-motoneurones; electron microscopy would be prohibitively time consuming. With confocal microscopy we were able to map instances of close apposition of monoaminergic varicosities. Comparison with the density of varicosities in the immediate surround of the cells suggested that the instances of close apposition were more frequent than would be expected had they been randomly distributed spatially. A preliminary abstract on this work has been published (Sahal et al. 1998).


  1. Top of page
  2. Abstract
  6. Acknowledgements


Eight hindlimb γ-motoneurones were labelled in two deeply anaesthetised cats. Anaesthesia was induced with pentobarbital sodium (45 mg kg−1i.p.), and supplemented with α-chloralose (about 5 mg kg−1 h−1 i.v). During recording the animals were paralysed with pancuronium bromide (Pavulon i.v.), with initial doses of 0.4 mg supplemented by similar doses every 2–3 h. Regularly repeated tests were made to ensure that the pupils remained constricted to the same extent throughout the experiments and that the animals did not respond with an increase in either blood pressure or heart rate to any stimuli after they had been paralysed. The care of the preparation and the general experimental procedures were as described previously (Jankowska & Riddell, 1994). Briefly, the blood pressure was kept above 100 mmHg, and the end-tidal CO2 about 4 %, by adjusting the volume of the artificial respiration and by a continuous infusion of a bicarbonate buffer solution. The core temperature was kept at 37–38°C, and the temperature in the hindlimb oil pool between 33 and 36°C by heating lamps. A number of hindlimb nerves were dissected and stimulated in a mineral oil pool (separate muscle and skin branches of the sciatic nerve: posterior biceps and semitendinosus, PBST; anterior biceps and semimembranosus, ABSM; medial gastrocnemius, MG; lateral gastrocnemius and soleus, LGS; plantaris, Pl; flexor digitorum longus, FDL; deep peroneal, DP (including anterior tibial, TA, and extensor digitorum longus, EDL). The laminectomy exposed the spinal cord from L4 to the sacral segments, the dura was opened and the L7 and sacral dorsal roots were reflected to provide access to the lateral funiculus at the L7 and S1 levels.

γ-Motoneurones were searched for in the L7 segment and in the rostral part of the S1 segment. Their responses were first recorded extracellularly. Antidromic responses were identified by the following criteria: (i) constant latencies of the responses evoked by near threshold and stronger stimuli, and by the first and later stimuli in a train, (ii) latencies exceeding the latencies of activation of α-motoneurones by 1.5–9.2 ms and compatible with conduction velocities of 16–49 m s−1, (iii) stimulus thresholds exceeding the thresholds of activation for α-motoneurones (2–13 times the thresholds of group Ia afferents in a given muscle nerve); (iv) collision with synaptically or ‘spontaneously’ evoked responses at an appropriate critical interval (twice the latency of the antidromic responses plus about 0.7 ms, as illustrated in Fig. 2E and G).


Figure 2. Distribution of 5-HT- and DBH-immunoreactive varicosities in apposition to the cell bodies and proximal parts of the dendrites of two γ-motoneurones

A–D, plots of varicosities found in apposition with the two neurones (A and C: no. 1; B and D: no. 2) in a number of optical planes in single 50 μm sections. These have been projected on 2-D reconstructions of the largest contours of the soma and initial parts of the dendrites. Another part of the soma of γ-motoneurone no. 1 which was found in the neighbouring section is shown in Fig. 1A and B. The three sizes of filled circles indicate the presence of small, medium or large varicosities. E–H, examples of responses of the γ-motoneurone illustrated in A and C. E and G, records just before penetration of the neurone illustrating collision of the antidromically evoked responses by background discharges. F and H, PSPs evoked polysynaptically by double stimuli applied to the flexor digitorum longus and medial gastrocnemius nerves at a strength sufficient to recruit group II muscle afferents. Stimulus strength is indicated as a multiple of T, the threshold for the recruitment of Ia afferents. I and J, antidromically evoked blocked spikes and monosynaptically evoked EPSPs (arrow) in the γ-motoneurone illustrated in B and D. In each pair of records the upper ones are from the γ-motoneurone, and the lower are from the surface of the spinal cord and indicate the arrival of the afferent volleys following the nerve stimulation. Time calibrations: 5 ms for E–H, 2 ms for I and 1 ms for J.

Download figure to PowerPoint

All experiments were carried out according to Göteborg Universityguidelines. The cats were killed at the end of the experiments with an overdose of pentobarbital and formalin perfusion.

Labelling and confocal microscopy

The neurones were labelled by ionophoretically injecting rhodamine-dextran from microelectrodes containing a 2 % solution of tetramethylrhodamine-dextran (3000 MW, anionic, lysine fixable; Molecular Probes Inc.) in saline (pH 6; tips of about 2.5 μm; resistance 13–18 MΩ) following the procedure of Carr et al. (1994). Rhodamine was injected by passing a constant positive current of 5 nA, totally of 20–56 nA min. Neurones to be labelled were searched for at distances of 1–2 mm. Animals were perfused with fixative through the descending aorta. Initially they were perfused with a rinsing solution which consisted of 200 ml of 0.1 M phosphate-buffered saline (PBS) and subsequently with a fixative solution which consisted of 2 l of 4 % paraformaldehyde in PBS at pH 7.4. Segments of spinal cord 5 mm and 6.5 mm long containing the labelled neurones were removed and placed in the same fixative for up to 8 h. Sections (50 μm) were cut in the sagittal plane with a Vibratome, collected in serial order, mounted in Vectashield (Vector Laboratories, Peterborough, UK) and examined with a fluorescence microscope to identify labelled neurones. Short series of sections containing labelled neurones were processed for immunocytochemistry in serial order. Initially sections were preincubated in 10 % normal donkey serum in PBS for 30 min. They were then transferred to a solution containing a mixture of rat anti-5-HT antiserum (1:200; Affiniti Research Products Ltd) and rabbit anti-dopamine β-hydroxylase (DBH) antiserum (1:500; Affiniti Research Products Ltd) in a standard diluting solution consisting of PBS with 5 % normal donkey serum and 0.3 % Triton X-100, and incubated at room temperature for 48 h. Sections were washed in PBS and placed in a solution containing secondary antisera which consisted of fluorescein isothiocyanate (FITC)-donkey anti-rabbit (1:100) and cyanine 5.18 (Cy-5)-donkey anti-rat antisera (1:100) for 1 h at room temperature. These were obtained from Jackson Immunochemicals and were diluted in the standard solution. Finally, sections were washed again in PBS and mounted in Vectashield and stored at −20°C.

Sections were examined with a three-channel confocal laser-scanning microscope (BioRad MRC1024). Series of images were gathered sequentially from single optical sections at 0.5 or 1 μm intervals (11–30 planes per series) by using the 488, 568 and 647 nm laser lines to excite FITC, rhodamine-dextran and Cy-5, respectively. In control experiments, where primary antisera were omitted, there was no evidence of specific immunoreactions for 5-HT or DBH. The majority of images were obtained by using a ×60 oil-immersion lens (numerical aperture of 1.4); however, large structures such as cell bodies were examined with a ×20 dry lens and, occasionally, large dendrites were examined with a ×40 oil-immersion lens (numerical aperture of 1.3). After the confocal examination was completed, the dendritic trees of the neurones were reconstructed with the Lucivid/Neurolucida reconstruction program (MicroBrightField, Inc., Colchester, VT, USA). The locations of structures examined with the confocal microscope were identified from the reconstructions and the sites of the appositions related to the distances from the soma.


Series of images from single optical sections were combined, by using Confocal Assistant (version 3.10). Confocal images of the γ-motoneurone were visualised consecutively. Chains of varicosities could be clearly distinguished. Each noradrenergic and serotoninergic profile in the vicinity of the γ-motoneurone was followed, frame by frame, in order to find any appositions between NA- or 5-HT-immunoreactive varicosities and the motoneurone membrane. Images from all relevant optical sections were then superimposed, creating a projected image. Varicosities were taken to be in apposition with the γ-motoneurone when there were no intervening pixels between them and/or when there was an overlap between their contours. Dendrites were almost always orientated obliquely to the plane of the optical sections so estimates of appositions were made from lateral relationships. Lateral resolution with the ×60 lens should be about 0.17 μm (theoretical resolution at wavelength of 500 nm). The varicosities were classified into three categories: small (s), less than 1.5 μm in diameter; medium (m), between 1.5 and 3 μm in diameter, and large (l), which were 3 μm or over in diameter. These three categories were marked on the projected images of the optical sections in which they were seen by using different symbols (see Fig. 2). The larger varicosities were present in a number of successive optical sections, so care was taken to avoid counting them more than once. The smallest varicosities were distinguished from any background fluorescence by checking that the fluorescence appeared in exactly the same location in more than one optical section. The surface areas of the analysed parts of somata of all eight neurones (in planes in which they were surrounded by immunoreactive fibres) were calculated from the formulae for the surface areas of the corresponding 3-dimensional geometric shapes, or combinations of geometric shapes – equivalent cylinder, segment of a sphere cut by a single plane, or by two parallel planes. By relating the number of appositions to these areas, the density of the appositions was expressed as the number of appositions per 100 μm2. In some cases parts of the somata were damaged by the sectioning, or the immunoreaction was unsatisfactory at some depths, so the whole surface of the soma could not be analysed. In order to estimate the total number of appositions per soma, the area of the surface of the soma explored was first related to the surface of an equivalent sphere that had a radius equal to the mean radius of the soma; the number of appositions that had actually been observed was increased by the same factor. In addition to counting varicosities in apposition, varicosities within 5 μm of the γ-motoneurone soma and dendrites were counted in order to quantify the density of monoaminergic innervation in the immediate vicinity of the motoneurones.

Appositions with proximal dendrites (within about 100 μm of the soma) were searched for in the sections containing the soma for all the γ-motoneurones. Appositions with more distal parts of the dendrites could sometimes be analysed in the same sections but usually required a reconstruction of the dendritic tree from several sections using the Neurolucida reconstruction program. The distances of the segments of dendrites analysed were evaluated from these reconstructions. However, the distribution of the appositions as a function of the distance was only analysed for the two most strongly labelled motoneurones. The segments of dendrites selected for analysis were those that were sufficiently long in one plane and were surrounded by immunoreactive fibres. The total length of the analysed segments amounted to 7159 μm.

The dendritic segments analysed were allocated to four compartments, up to 100, 100–200, 200–300 and >300 μm from the soma. The density (number per 100 μm2) of appositions on all suitable dendritic segments in these compartments was calculated from the total numbers of appositions, the total lengths and the mean diameter of dendritic segments in each compartment. The segments of dendrites analysed belonging to the two neurones with the best-labelled dendritic trees are indicated as thickened regions in Fig. 3A and B. Note that the dendrites were divided into compartments by path length, rather than by radial distance from the soma. The densities for individual segments of dendrite were also calculated. In addition, in order to compare with data from premotor interneurones (Maxwell et al. 2000) the number of appositions with dendrites was calculated per 100 μm length. For comparison with data for α-motoneurones (Alvarez et al. 1998) the total number of appositions with dendrites was estimated by multiplying the value for unit length by the mean minimal and maximal total dendritic length of γ-motoneurones reported previously (Moschovakis et al. 1991). Data are presented as means ±s.e.m. Statistical significance was determined at P < 0.05 using Student's t test.


Figure 3. Comparison of the density of appositions of 5-HT- and DBH-immunoreactive fibres with somata and with dendrites at different distances from the somata

A and B, line drawings of the dendritic trees of the neurones illustrated in Fig. 2 (cells 1 and 2, respectively); the segments of dendrites analysed are indicated as thickened regions. The circles indicate 100 and 300 μm distances from the soma. C and D, density of appositions of 5-HT- and DBH-immunoreactive fibres with the somata of the cells shown in A and B, respectively, expressed as the number per 100 μm2. The proportions of small (▪), medium ( inline image) and large (□) varicosities are indicated. E and F, packing densities of 5-HT- and DBH-immunoreactive fibres with distance from the soma for cell 1 (▪) and for cell 2 ( inline image).

Download figure to PowerPoint


  1. Top of page
  2. Abstract
  6. Acknowledgements

Of eleven γ-motoneurones injected with rhodamine-dextran, eight were successfully labelled, three in one cat and five in the second cat. These eight γ-motoneurones included motoneurones innervating muscle spindles of MG (nos 1–5, see Table 1), LGS (no. 6) and FDL (nos 7 and 8). Their axons conducted at 23–34 m s−1 (see Table 1). Group II afferents provided input to all of these neurones. Two neurones (cells 2 and 3) were excited by group II afferents at latencies compatible with monosynaptically evoked actions, while the remaining neurones were excited and/or inhibited at longer latencies.

Table 1.  Comparison of densities of 5HT and NA fibre appositions on somata and dendrites of eight γ-motoneurones
 γ-Motoneurone no.12345678  
 Conduction velocity (m s−1)2829342523312426Means.e.m.
  1. Values are densities of appositions per 100 μm2. The muscles innervated by the motoneurones and the conduction velocities of the γmotor axons are indicated above the columns. The distance of the compartment of dendrite from the soma is given.

 Dendrites, 0–100 μm1.
 Dendrites,100–300 μm1.
 Dendrites,>300 μm0.580.590.58
 Dendrites,>300 μm0.260.360.31

Records from the two neurones which were most extensively studied (cells 1 and 2) are shown in Fig. 2E–J. Figure 2E and G illustrates the collision technique used to differentiate the antidromically evoked spike potentials from those evoked synaptically. The antidomically propagated spike seen in E was blocked in G on three occasions by orthodromically propagating background discharges. After penetrating cell 1, double stimulus pulses at short intervals were applied to the FDL and MG muscle nerves (F and H) at strengths sufficient to recruit group II muscle afferents. The latencies of the postsynaptic potentials (PSPs) evoked by these stimuli indicated that polysynaptic pathways were involved. Note that the response in H was from the homonymous muscle (for further details see Gladden et al. 1998). The intracellular records shown in Fig. 2I and J were from γ-motoneurone no. 2. Figure 2I shows the blocked antidromically propagated spike recorded when the MG muscle nerve was stimulated. In Fig. 2J stimulation of the hamstring nerve at group II muscle strength elicited PSPs (arrow) at latencies too short for an interneurone to be interposed in the pathway. These responses, and those of the other labelled cells, were representative of the responses shown by the whole population of 76 γ-motoneurones and already reported.

Aminergic fibres

Immunoreactive serotoninergic fibres were abundant in the ventral horns in both cats. These are seen in Fig. 1 as red strings and points. Both thin and thick terminal branches, with small, medium and large varicosities (see Methods) were seen in the vicinity of the labelled γ-motoneurones. There were fewer DBH-stained fibres (the punctate green areas in Fig. 1).


Figure 1. Examples of appositions between 5-HT- (red) and DBH- (green) immunoreactive fibre terminals and rhodamine-dextran-labelled γ-motoneurones

A, a projected image (21 optical sections) showing part of the soma and the most proximal parts of dendrites of a γ-motoneurone (no. 1) in planes across as well as just above and just below the neurone. B, appositions between 5-HT varicosities (arrows) and the soma shown in a single optical section across the neurone from the projected series illustrated in A. C, a projected image of an intermediate part of one of the dendrites of the same γ-motoneurone (200–300 μm from the soma; 15 optical sections). D, one of the optical sections from the projected series in C showing appositions between this dendrite and 5-HT axons (arrows) and a single small DBH varicosity (arrowhead). E, a projected image of a distal part of the same dendrite as in C (>300 μm from the soma; 18 optical sections). F, a single optical section from the projected series in E showing appositions between 5-HT axons (arrows). G, a projected image of a dendrite of another γ-motoneurone (no. 2) 200–300 μm from the soma (20 optical sections). H, a single optical section from the series shown in G. It illustrates appositions between large DBH varicosities and dendrites (arrowhead). Scale bars represent 50 μm for A and B, and 20 μm for C, D, E, F, G and H.

Download figure to PowerPoint

Appositions with cell bodies

Serotoninergic and DBH-immunoreactive nerve terminals were found in apposition to cell bodies of all of the labelled γ-motoneurones. Examples of their distribution are shown in Fig. 2A–D. Appositions found in series of optical sections of the two most intensively studied γ-motoneurones are indicated on projected reconstructions of those parts of the somata and dendritic trees present in each series. Serotoninergic appositions present in a single optical section of cell 1 are illustrated in Fig. 1B (arrows). 5-HT-immunoreactive appositions were not evenly distributed across the surface and they appeared to be less abundant than varicosites around some neighbouring α-motoneurones. The axons often followed the surface contour, providing a string of varicosities in apposition to the surface, but in no case was there any suggestion of a ring of contacts as is typical for α-motoneurones. In contrast, only single DBH-immunoreactive varicosities were apposed to the somata of the γ-motoneurones, although they could sometimes be connected to strings of varicosities that were frequently seen in the surround. This was also in contrast to prominent rings of varicosities that were seen around a number of neighbouring α-motoneurones.

Most of the varicosities found in apposition to γ-motoneurones were small (diameters < 5 μm). Over 60 % of the 5-HT-immunoreactive varicosities were small on 5 out of the 8 cells, and over 60 % of DBH-immunoreactive varicosities were small on all the cells. Cell 1 had an unusually high proportion of large 5-HT-immunoreactive varicosities (Fig 2A and Fig 3C), and cell 2 had the highest proportion of large DBH-immunoreactive varicosities (Fig 2D and Fig 3D). The sizes of varicosities in apposition to the cells did not reflect the proportions of the different sizes in the immediate vicinity of the somata.

The mean density of serotoninergic appositions was 1.12 ± 0.11 per 100 μm2, and of noradrenergic appositions 0.6 ± 0.10 per 100 μm2 (see Table 1). There was surprisingly little variation between cells. The highest and lowest densities differed by a factor of only two; the only exception was the very low density of DBH-immunoreactive appositions on cell 6 from the LGS motor nucleus, which was ten times less than that of the cell with the highest density. The mean number of serotoninergic appositions found on the somata was 18.7 ± 4.0, range 7–38 per cell. However, since the entire surface area could not be reliably analysed in every case (see Methods) a count of appositions actually seen underestimates the total number. For some cells estimates of the total number of appositions were therefore made from the number of contacts per 100 μm2 and the total surface areas calculated from the mean diameters of the γ-motoneurones. The estimated mean total number of serotoninergic appositions per soma was 31.6 ± 3.25 (range 22–44, see Table 2). The mean number of DBH-immunoreactive appositions was 9.3 ± 3.0, range 1–25 per cell and the estimated mean total number was 15.1 ± 3.17 (range 2–25).

Table 2.  Comparison of the distribution of contacts between monoaminergic fibres on γ-motoneurones, α-motoneurones and premotor group II interneurones
FibreSite of contactγ-Motoneuronesα-MotoneuronesGroup II premotor interneurones
  1. Data for γ-motoneurones are from this study, for α-motoneurones from Alvarez et al. (1998) and for the interneurones from Maxwell et al. (2000). Proximal dendrites are taken to be within 100 μm of the soma, intermediate at distances 100–200 μm, and distal, >300 μm away. <*These values were obtained by relating the mean number of appositions per soma (52 per cell) found by Alvarez et al. (1998) to the mean surface areas of 19 113 and 4657 μm2 calculated for extremes of mean soma diameters (78 and 38.5 μm) of α-motoneurones according to Burke et al. (1977). **These values were obtained by scaling the actual observed numbers by the ratio between the explored area and the calculated total area of the soma.

5HTSoma, total number18.7 (31.6 **)526.2
 Soma, per 100 μm21.12 ± 0.1110.27–1.12 *0.130
 Dendrites (< 500 μm), per 100 μm20.97 ± 0.0840.81 ± 0.250.29 ± 0.05
 Proximal dendrites, per 100 μm14.53.1
 Intermediate dendrites, per 100 μm11.34.0
 Distal dendrites, per 100 μm4.83.8
NASoma, total number9.4(21.7 **)2.0
 Soma, per 100 μm20.590.052
 Dendrites (< 500 μm), per 100 μm20.6 ± 0.010.082 ± 0.03
 Proximal dendrites, per 100 μm9.20.87
 Intermediate dendrites, per 100 μm7.61.05
 Distal dendrites, per 100 μm2.30.81


Noradrenergic and serotoninergic appositions occurred with the dendrites of all cells studied but there was no obvious relationship between them. In some dendritic segments both types of apposition were present in close proximity, as in Fig. 1C and D. In other areas there were serotoninergic appositions almost exclusively, as in Fig. 1E and F, and more rarely there were predominantly DBH-immunoreactive appositions, as in Fig. 1G and H. A clustering of appositions is also illustrated by the mapping of appositions on the dendrites to the right of the reconstruction of cell 1 (Fig. 2A and C). The lower right dendrite has dense serotoninergic but sparse noradrenergic appositions, but the reverse is true of the distal part of the upper right dendrite. This tendency for clustering of appositions resulted in a marked variability in the densities of appositions for individual segments of dendrite. The densities of 5-HT- and DBH-immunoreactive terminals in apposition to the two best-filled cells remained fairly constant within 300 μm distance from the somata, but fell away further than 300 μm from the soma (Fig. 3E and F, and Table 1). In both cases the fall was statistically significant. For the whole sample of neurones the mean density of 5-HT appositions with dendrites was 0.91 ± 0.07 per 100 μm2, and the value for NA was 0.56 ± 0.04, i.e. 62 % of the mean for 5-HT appositions. The densities of appositions for each of the eight neurones studied are given in Table 1. The total numbers of 5-HT-immunoreactive varicosities designated as appositions in each of the compartments studied (up to 100, 100–200, 200–300 and >300 μm from the soma) were 280, 189, 123 and 126. The corresponding numbers of DBH appositions were 178, 140, 47 and 54.

The majority of 5-HT-immunoreactive varicosities in contact with dendrites were small, as was the case with those in apposition to the somata. When the percentages of the different sizes of varicosity at the various distances from the somata were compared for all the cells it was found that, in the case of 5-HT, never less than 50 % were small, and for NA never less than 65 %. The percentage of small appositions tended to rise with distance from the cell, so that over 300 μm from the soma practically all appositions were small.

The total number of 5-HT varicosities in apposition to dendrites was estimated to be 157–843. This was calculated from the mean number of appositions over the first 300 μm (11.8 per 100 μm), and mean total lengths of 13296 and 71421 μm of dendrites of small and large γ-motoneurones (according to Moschovakis et al. 1991). The total number of DBH-immunoreactive varicosities in apposition was estimated in the same way to be 103–552, taking into account the mean number of appositions over the first 300 μm of 7.73 per 100 μm.

Varicosities in the surround of dendrites

Figure 1 shows that in addition to 5-HT- and DBH-immunoreactive varicosities in apposition to the somata and dendrites, many varicosities were present within a few micrometres of the surface of the labelled γ-motoneurones. In order to estimate whether the density of appositions with dendrites merely reflected the density in the surrounding area a comparison was made with the density of varicosities in a 5 μm-wide volume immediately surrounding each dendrite. If the monoaminergic varicosities were distributed evenly, and the axons reached the dendritic surfaces by chance, the numbers on the dendritic surface would reflect the numbers that would have occurred in the space occupied by the dendrites had the axons not been impeded by the dendritic surfaces. For this comparison the numbers of varicosities in apposition with dendrites were expressed in terms of their volume rather than surface area. In every case had the varicosities on the surface actually been present in the volume occupied by the dendrites (Fig. 4A and B, filled symbols) their density would have been greater than that of varicosities in the immediate surround (open symbols). The densities for the dendrites were increased by a mean factor of 13.8 ± 0.4 for 5-HT and 10.6 ± 1.5 for NA. Factors for individual sections of dendrites are shown in Fig. 4C and D. Figure 4A and B also shows that the density of serotoninergic and noradrenergic varicosities in the immediate surround of the dendrites was similar in the three motor nuclei (MG, FDL and LGS, indicated by different symbols). Taking together the volumes of the 5 μm surround of all the dendrites investigated from all the γ-motoneurones, and the total numbers of 5-HT- and DBH-immunoreactive varicosities in that space, their densities were estimated to be 798 and 585 varicosities mm−3.


Figure 4. Comparison of the densities of 5-HT- and DBH-immunoreactive varicosities in the immediate surround of the dendrites and those in apposition with distance from the soma

A and B, open symbols: the density of varicosities in the 5 μm-wide surround of the dendrites; filled symbols: the density of varicosities in apposition expressed as if they were distributed throughout the volume of the dendrite. Squares: MG γ-motoneurones: large filled square, cell 1; large hatched square, cell 2; large open square, cell 1 surround; large dotted square, cell 2 surround. Diamonds: FDL γ-motoneurones. Circles: LGS γ-motoneurone. Note that the density in the surround was always less. C and D, ratios of the densities of varicosities shown in A and B– an indication of the affinity of the dendrites for serotoninergic (C) and noradrenergic (D) axons.

Download figure to PowerPoint

Spines, beads and thread-like processes

Occasionally spines, beads and thread-like processes, as described by Ulfhake & Cullheim (1981), were encountered on the dendrites of γ-motoneurones. No appositions were observed with spines or thread-like processes, but both 5-HT and DBH appositions were found on beads.


  1. Top of page
  2. Abstract
  6. Acknowledgements

Although we cannot prove that the appositions between the monoaminergic varicosities and γ-motoneurones are synaptic contacts, the proposal that there is a serotoninergic and noradrenergic innervation of γ-motoneurones is strongly supported by analysis of their distribution. Firstly, the appositions were unevenly distributed. Clustering of appositions of both types was evident especially on proximal dendrites. This clustering did not necessarily mirror the density of varicosities in the immediate surround because when these were mapped on the outline of the cells the pattern of distribution was different. On more distal dendrites the uneven distribution was less obvious to the eye, but was indicated by large variations in density of appositions on different segments of dendrites, even though the mean densities remained fairly constant with distance from the soma (Fig. 3E and F). Secondly, the distribution of varicosities of different sizes changed with distance from the soma, there being scarcely any medium sized, and no large appositions with distal dendrites, although larger appositions were present in the surrounding areas. Thirdly, if the varicosities judged to be on the surface of the dendrites were distributed throughout the volume occupied by the dendrites they would be considerably more densely packed than varicosities found in a 5 μm-wide shell around the dendrites. It was estimated that the density would be 14 times greater, on average, for serotoninergic varicosities, and 11 times greater for noradrenergic varicosities. This suggests that the dendrites of γ-motoneurones have an affinity for monoaminergic axons.

Comparison of the density and distribution of serotoninergic and noradrenergic varicosities

Somata. The ratio of the densities of serotoninergic to noradrenergic varicosities in apposition to the somata of the γ-motoneurones was very variable (from 1:1 to 17:1, for values of densities see Table 1), and in the immediate surround of the somata it was even more variable. Some of this variation may be explained by sampling problems because not all areas of the somata could be accessed. However, there may be actual differences. For the two cells with the most fully explored surfaces the density of DBH appositions was 68 and 36 % of the density of 5-HT appositions, and the corresponding percentages in the immediate surround were 48 and 19 %. These wide variations in the proportions of serotoninergic and noradrenergic appositions may reflect functional differences seen in the records from individual γ-motoneurones. When the effects of locally applied noradrenaline were tested on γ-motoneurones (Jankowska et al. 1998) the effects on background discharges were variable, although noradrenaline depressed transmission from group II afferents to all γ-motoneurones tested. It should also be remembered that the γ-motoneurones in this study could be static or dynamic – tests to identify the functional types could not be carried out because it was necessary for the animals to be paralysed for intracellular recording. It is likely that most of them will have been static because the proportion of dynamic γ-motoneurones is usually less than 30 %, but even the static category are not functionally homogeneous (Dickson et al. 1993; Taylor et al. 1998).

Dendrites. The overall lower density of DBH- compared with 5-HT-immunoreactive varicosities in the motor nuclei studied was obvious (see Fig. 1). The density of NA fibres in motor nuclei is generally found to be lower than that of 5-HT fibres although in some nuclei it has been reported to be similar (Zhan et al. 1989). In the 5 μm-wide shell around the dendrites the density of DBH varicosities was 66 % of the density of 5-HT, but the overall percentage for appositions with the dendrites was lower – 56 %. The mean densities of monoaminergic appositions calculated for the whole sample of γ-motoneurones fell with distance from the somata (Table 1). However, there was a trend in some cells for the densities of 5-HT or DBH appositions to be higher on intermediate dendrites than on proximal dendrites (Fig. 3E and F, and Table 1). It is tempting to speculate that the monoaminergic appositions should be strategically placed to modulate transmission in specific pathways, and that regional accumulations could indicate the location of incoming traffic. Since 5-HT and NA both modulate the synaptic actions of group II muscle afferents the location of direct connections of these afferents on γ-motoneurones, and indirect connections via group II interneurones, may be primarily on the intermediate and distal dendrites of the γ-motoneurones. Thus it is interesting that relatively few contacts from group Ia axons were found on or near the somata of α-motoneurones (Burke & Glenn, 1996). It is unlikely that the monoaminergic axons make presynaptic connections with group II afferents because these axons have not been found to make axo-axonic contacts with fibres synapsing with α-motoneurones (Ulfake et al. 1987), or other types of spinal neurones (for references see Maxwell et al. 2000).

Comparison of distribution and density of contacts of serotoninergic fibres on α- and γ-motoneurones

Synaptic contacts of 5-HT-immunoreactive fibres with α-motoneurones have been demonstrated by EM in several studies (Takeuchi et al. 1983; Ulfhake et al. 1987; Holtman et al. 1990; Alvarez et al. 1998). As with γ-motoneurones 5-HT has facilitatory effects on α-motoneurones, and it is also known that the ability of α-motoneurones to generate plateau potentials depends on continued activity in the serotoninergic pathway from the midbrain (Hounsgaard et al. 1988). The distribution over both the soma and dendrites of α-motoneurones was investigated systematically by Alvarez et al. (1998). It is interesting, therefore, to compare the present results with theirs. However, they counted the total number of the apparent 5-HT contacts with the somata of α-motoneurones by light (non-confocal) microscopy. In order to accommodate the data derived by different approaches it was necessary to make some assumptions (see legend to Table 2). Nevertheless the mean density and the number of serotoninergic varicosities found in apposition to the somata of γ-motoneurones, whether actual or estimated, fall within the values published by Alvarez et al. (1998). They noted a wide variation in the numbers of varicosities contacting the somata – their range was 11–211. The packing density of appositions with dendrites of γ-motoneurones (about 1 per 100 μm2) is also within their range of values (see Table 2). It is interesting that Alvarez et al. (1998) estimated the number of serotoninergic contacts with α-motoneurones to be in excess of the number of synapses by Ia afferents.

Comparison with noradrenergic contacts on α-motoneurones

Synaptic contacts formed by noradrenergic fibres have been identified in some motor nuclei by EM (Holstege & Bongers, 1991; Rajaofetra et al. 1992), but there has not been a systematic analysis of the distribution. It is therefore not possible to make any comparisons of the distribution of noradrenergic contacts on α- and γ-motoneurones, except to point out that noradrenergic fibres appear to contact both somata and dendrites in both types of neurone. However, Rosin et al. (1996) reported α2C-adrenergic receptor-like immunoreactivity in the ventral horn at all spinal levels in rats, and stressed that the labelling was restricted to a fraction of large motoneurones. It may be possible to demonstrate more significant structural correlates of the different effects of NA on the responsiveness of α- and γ-motoneurones once the various receptor subtypes can be localised more reliably.

Possible explanation for the richer monoaminergic innervation of γ-motoneurones compared with premotor interneurones with input from group II afferents

In the last column of Table 2 are data for premotor interneurones (Maxwell et al. 2000) which are likely to be interposed in pathways from group II muscle afferents to both α- and γ-motoneurones. The actions of monoamines on these interneurones are similar to their actions on γ-motoneurones. As with γ-motoneurones NA powerfully depressed transmission from group II afferents to these interneurones, but 5-HT depressed transmission in about half and facilitated it in the rest (Jankowska et al. 2000). Table 2 shows that there are some common features in the coupling of monoaminergic fibres with these interneurones and with γ-motoneurones: a higher density of 5-HT than NA appositions, and a greater number of these appositions with dendrites than with somata. However, the density of serotoninergic and noradrenergic appositions is strikingly less on somata and dendrites than on γ- or indeed α-motoneurones. These intermediate zone interneurones are nevertheless more richly endowed than some types of dorsal horn interneurones which have scarcely any monoaminergic contacts (Stewart & Maxwell, 1999).

The premotor interneurones are smaller cells with less extensive dendritic trees than γ-motoneurones, and far less than those of α-motoneurones. This difference in scale cannot explain the markedly lower density of contacts. However, it could be an indication of fewer interaction sites along their dendrites. It could also be related to their input. In anaesthetised preparations group II interneurones respond readily to muscle stretch but have hardly any resting discharges (in contrast to non-anaesthetised decerebrate preparations, see Shefchyk et al. 1990). On the other hand very deep anaesthetic levels are required to silence all γ-motoneurones, and they are likely to be activated by a greater number of descending systems. A larger number of contacts between the modulatory NA- and 5-HT-releasing neurones and γ-motoneurones may thus be needed to adjust the rate of their discharge.

Fusimotion contributing to integrated responses?

γ-Motoneurones become excited in arousal and stress (see for example the responses described recently in humans by Ribot-Ciscar et al. 2000), and in these situations activation of the serotoninergic projection to the motor nuclei from the raphe nuclei is expected. The increased γ-activity recorded in these conditions may in fact originate from the raphe nuclei that lie in the medulla (the raphe pallidus and obscurus nuclei). If this is correct the increased fusimotor activity would be part of an integrated alerting response which includes lowering the threshold for activating α-motoneurones, autonomic changes and analgesia (see Lovick, 1997). Not only could a randomly activated fusimotor system enhance the information content of the signals from the muscle spindles (see Cordo et al. 1996), but also the increased frequency of discharge of Ia afferents would have a priming effect in the multitude of pathways to which Ia afferents provide an input.

  • Ahlman, H., Grillner, S. & Udo, M. (1971). The effect of 5-HTP on the static fusimotor activity of an extensor muscle. Brain Research 27, 393396.
  • Alvarez, F. J., Pearson, J. C., Harrington, D., Dewey, D., Torbeck, L. & Fyffe, R. E. (1998). Distribution of 5-hydroxytryptamine-immunoreactive boutons on alpha-motoneurons in the lumbar spinal cord of adult cats. Journal of Comparative Neurology 393, 6983.
  • Bennett, D. J., De Serres, S. J. & Stein, R. B. (1996). Regulation of soleus muscle spindle sensitivity in decerebrate and spinal cats during postural and locomotor activities. The Journal of Physiology 495, 835850.
  • Bergmans, J. & Grillner, S. (1968). Changes in dynamic sensitivity of primary endings of muscle spindle afferents induced by DOPA. Acta Physiologica Scandinavica 74, 629636.
  • Bunin, M. A. & Wightman, R. M. (1998). Quantitative evaluation of 5-hydroxytryptamine (serotonin) neuronal release and uptake: an investigation of extrasynaptic transmission. Journal of Neuroscience 18, 48544860.
  • Burke, R. E. & Glenn, L. L. (1996). Horseradish peroxidase study of the spacial and electrotonic distribution of group Ia synapses on type-identified ankle extensor motoneurones in the cat. Journal of Comparative Neurology 372, 465485.
  • Burke, R. E., Strick, P. L., Kanda, K., Kim, C. C. & Walmesley, B. (1977). Anatomy of medial gastrocnemius and soleus motor nuclei in cat spinal cord. Journal of Neurophysiology 40, 667680.
  • Carr, P. A., Noga, B. R., Nance, D. M. & Jordan, L. M. (1994). Intracellular labeling of cat spinal neurons using a tetramethylrhodamine-dextran amine conjugate. Brain Research Bulletin 34, 447451.
  • Cordo, P., Inglis, J. T., Vershueren, S., Collins, J. J., Merfeld, D. M., Rosenblum, S., Buckley, S. & Moss, F. (1996). Noise in human muscle spindles. Nature 383, 769770.
  • Dickson, M., Emonet-Dénand, F., Gladden, M. H., Petit, J. & Ward, J. (1993). Incidence of non-driving excitation of Ia afferents using ramp frequency stimulation of static γ-axons in cat hindlimbs. The Journal of Physiology 460, 657673.
  • Ellaway, P. H. & Trott, J. R. (1975). The mode of action of 5-hydroxytryptophan in facilitating a stretch reflex in the spinal cat. Experimental Brain Research 22, 145162.
  • Gladden, M. H., Jankowska, E. & Czarkowska-Bauch, J. (1998). New observations on coupling between group II muscle afferents and feline γ-motoneurones. The Journal of Physiology 512, 507520.
  • Grillner, S. (1969). Supraspinal and segmental control of static and dynamic gamma-motoneurones in the cat. Acta Physiologica Scandinavica, suppl. 327, 134.
  • Grillner, S., Hongo, T. & Lundberg, A. (1967). The effect of DOPA on the spinal cord. 7. Reflex activation of static gamma-motoneurones from the flexor reflex afferents. Acta Physiologica Scandinavica 70, 403411.
  • Holstege, J. C. & Bongers, C. M. (1991). Ultrastructural aspects of the coeruleo-spinal projection. Progress in Brain Research 88, 143156.
  • Holtman, J. R. Jr, Vascik, D. S. & Maley, B. E. (1990). Ultrastructural evidence for serotonin-immunoreactive terminals contacting phrenic motoneurons in the cat. Experimental Neurology 109, 269272.
  • Hounsgaard, J., Hultborn, H., Jesperson, B. & Kien, O. (1988). Bistability of α-motoneurones in the decerebrate cat and in the acute spinal cat after intravenous 5-hydroxytryptophan. The Journal of Physiology 405, 345367.
  • Jacobs, B. L. & Azmitia, E. C. (1992). Structure and function of the brain serotonin system. Physiological Reviews 72, 165229.
  • Jankowska, E., Gladden, M. H. & Czarkowska-Bauch, J. (1998). Modulation of responses of feline γ-motoneurones by noradrenaline, tizanidine and clonidine. The Journal of Physiology 512, 521531.
  • Jankowska, E., Hammar, I., Chojnicka, B. & Hedén, C. H. (2000). Effects of monoamines on interneurones in four spinal reflex pathways from group I and/or group II muscle afferents. European Journal of Neuroscience 12, 701714.
  • Jankowska, E. & Riddell, J. S. (1994). Interneurones in pathways from group II muscle afferents in sacral segments of the feline spinal cord. The Journal of Physiology 475, 455468.
  • Lovick, T. A. (1997). The medullary raphe: a system for integration and gain control in autonomic and somatomotor responsiveness? Experimental Physiology 82, 3141.
  • Maxwell, D. J., Riddell, J. S. & Jankowska, E. (2000). Serotoninergic and noradrenergic axonal contacts associated with premotor interneurons in spinal pathways from group II muscle afferents. European Journal of Neuroscience 12, 12711280.
  • Moschovakis, A. K., Burke, R. E. & Fyffe, R. E. (1991). The size and dendritic structure of HRP-labeled gamma motoneurons in the cat spinal cord. Journal of Comparative Neurology 311, 531545.
  • Rajaofetra, N., Ridet, J.-L., Poulat, P., Marlier, L., Sandillon, F., Geffard, M. & Privat, A. (1992). Immunocytochemical mapping of noradrenergic projections to the rat spinal cord with an antiserum against noradrenaline. Journal of Neurocytology 21, 481494.
  • Ribot-Ciscar, E., Rossi-Durand, C. & Roll, J.-P. (2000). Increased muscle spindle sensitivity to movement during reinforcement manoeuvres in relaxed human subjects. The Journal of Physiology 523, 271282.
  • Rosin, D. L., Talley, E. M., Lee, A., Stornetta, R. L., Gaylinn, B. D., Guyenet, P. G. & Lynch, K. R. (1996). Distribution of α2C-adrenergic receptor-like immunoreactivity in the rat central nervous system. Journal of Comparative Neurology 372, 135165.
  • Sahal, A., Maxwell, D. J., Jankowska, E. & Gladden, M. H. (1998). Relationships between noradrenaline- and serotonin-immunoreactive axons and γ-motoneurones. The Journal of Physiology 509.P, 172P..
  • Shefchyk, S., McCrea, D., Kriellaars, D., Fortier, P. & Jordan, L. (1990). Activity of interneurons within the L4 spinal segment of the cat during brainstem-evoked fictive locomotion. Experimental Brain Research 80, 290295.
  • Stewart, W. & Maxwell, D. J. (1999). A population of lamina III/IV neurokinin-1-immunoreactive cells are targeted by serotonin-immunoreactive axons in the rat dorsal horn. The Journal of Physiology 518.P, 166P..
  • Takeuchi, Y., Kojima, M., Matsuura, T. & Sano, Y. (1983). Serotonergic innervation on the motoneurons in the mammalian brainstem. Light and electron microscopic immunohistochemistry. Anatomy and Embryology 167, 321333.
  • Taylor, A., Ellaway, P. H. & Durbaba, R. (1998). Physiological signs of the activation of bag2 and chain intrafusal muscle fibres of gastrocnemius muscle spindles in the cat. Journal of Neurophysiology 80, 130142.
  • Ulfhake, B., Arvidsson, U., Cullheim, S., Hökfelt, T., Brodin, E., Verhofstad, A. & Visser, T. (1987). An ultrastructural study of 5-hydroxytryptamine-, thyrotropin-releasing hormone- and substance P-immunoreactive axonal boutons in the motor nucleus of spinal cord segments L7–S1 in the adult cat. Neuroscience 23, 917929.
  • Ulfhake, B. & Cullheim, S. (1981). A quantitative light microscopic study of the dendrites of cat spinal γ-motoneurones after intracellular staining with horseradish peroxidase. Journal of Comparative Neurology 202, 585596.
  • Zhan, W. Z., Ellenberger, H. H. & Feldman, J. L. (1989). Monoaminergic and GABAergic terminations in phrenic nucleus of rat identified by immunohistochemical labeling. Neuroscience 31, 105113.


  1. Top of page
  2. Abstract
  6. Acknowledgements

The study was supported by grants from the Wellcome Trust to D.J.M. (University Award No. 039925) and M.H.G., from Glasgow University to A.S., and from the Swedish Medical Research Council (No. 05648) to E.J. Our warmest thanks are due to Mrs R. Larsson for her assistance in the experiments, and to Mr R. Kerr for excellent technical assistance with immunocytochemical processing.