| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Área de Fisiología, Facultad de Medicina, Universidad de Cádiz, Cádiz, Spain
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionReferences |
|---|
(Received 18 December 2003;
accepted after revision 13 April 2004;
first published online 16 April 2004)
Corresponding author B. Moreno-López: Área de Fisiología, Facultad de Medicina, Plaza Falla, 9, 11003 Cádiz, Spain. Email: bernardo.moreno{at}uca.es
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Although much of this previous work has been directed towards defining the response of motoneurones to muscle disconnection, emphasis has usually been placed on the electrical and synaptic effects in synaptically silent preparations. Little is known about the discharge activity and rate modulation of axotomized motoneurones under normal afferent synaptic drive. Related work on this issue has been reported for axotomized abducens motoneurones recorded in chronic cat preparations (Delgado-García et al. 1988). However, as noticed by these authors, excitability is reduced in axotomized abducens motoneurones, and so their results differ from those described in axotomized spinal and HMNs in which excitability is increased. Likewise, in such studies contralateral eye position was taken as a reference index for the control afferent synaptic drive, despite the possibility that compensatory motor adaptations in the non-treated side can occur (Moreno-López et al. 1997). Knowledge of the discharge activity and modulation of axotomized motoneurones in response to physiological patterns of afferent activation has a crucial importance since (a) axotomy-induced increases in membrane excitability and decreases in synaptic efficacy would be expected to have an opposite influence on firing activity, and (b) it would determine the levels of electrical activity in regenerating axotomized motoneurones, which in turn have been shown to affect both the speed and the precision for target re-innervation (Al-Majed et al. 2000; Brushart et al. 2002).
The tongue participates in diverse essential motor tasks including respiration-related movements; this function is very important for keeping open upper airways during breathing. These breathing-related movements rely on the coordinated actions of protruder (genioglossus) and retractor (hypoglossus and styloglossus) muscles (Sawczuk & Mosier, 2001). In the hypoglossal nucleus (HN), most motoneurones discharge bursts of action potentials synchronized with respiratory phases, even though the great majority of hypoglossal respiratory activity is inspiratory (Hwang et al. 1983a). During inspiration, HMNs are synaptically excited by pre-motor neurones located within the medulla and pons (Ono et al. 1994; Peever et al. 2002). A great advantage for our purpose is that, while most oro-facial motor drive is suppressed following decerebration, respiratory afferent activity on HMNs persists (Wang & Nims, 1948; Hwang et al. 1983a,b). Additionally, in in vivo decerebrated, and lightly anaesthetized preparations, respiratory activity can be modulated by changes in arterial blood CO2 and O2 pressures (Cohen, 1968; Hwang et al. 1983a,b).
In this work we report that (1) axonal injury induces a depression in firing activity of HMNs, (2) motoneurone sensitivity to afferent drive modulated by changes in end-tidal CO2 was also reduced following XIIth nerve crushing, (3) synaptophysin immunoreactivity was reduced after lesion, and (4) the morphofunctional alterations were reversible, and recovered in parallel with the compound muscle action potential. We conclude that previously reported increases in motoneurone excitability do not completely compensate for the effects of axotomy-induced deafferentation, which accounts for the observation that the inspiratory discharge of HMNs is reduced both under basal conditions and conditions of altered chemoreceptor drive. Possible implications in the speed and selectivity of regenerative processes following peripheral nerve injury are discussed.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Adult male Wistar rats weighing 250–400 g, obtained from an authorized supplier (Animal Supply Services, University of Cádiz, Spain), were used in this study. Animals were individually housed in cages with water and food pellets available ad libitum, under temperature-controlled conditions at 21 ± 1°C, with a cycle of 12 h light and 12 h darkness. Experiments were performed in accordance with the European Union directive 609/86/CEE and with Spanish legislation (RD 233/89) on the use and care of laboratory animals, and approved by the local Animal Care and Ethics Committee. All surgical procedures were carried out under aseptic conditions.
XIIth nerve injury
Animals were anaesthetized with chloral hydrate (0.5 g kg–1; I.P.). A sufficient depth of anaesthesia was judged from the absence of withdrawal reflexes. The right XIIth nerve was isolated from surrounding tissue and the common nerve trunk was then thoroughly crushed with microdissecting tweezers, applied for 30 s. The lesion was made just proximal to the bifurcation into lateral and medial branches (Fig. 1). The incision was sutured, cleaned with an aseptic solution (povidone–iodine) and the animals were allowed to survive 1, 3, 7, 15, 22, 30 or 45 days after the crushing. In the control group this surgery was absent, and in sham-operated animals the nerve was dissected but crushing was omitted. All animals received one post-operative injection of penicillin (20 000 i.u. kg–1; I.M.) in order to prevent infection. Pirazolone (0.1 mg kg–1; I.M.) was given on awakening for post-operative analgesia. At least three animals per experimental condition were used.
|
To study the time course effects of nerve injury on compound muscle action potential (CMAP) and basal activity of the XIIth nerve, experimental animals were anaesthetized as indicated above, both XIIth nerves were dissected and Teflon-isolated silver bipolar electrodes were placed 3–4 mm proximal to their bifurcation (Fig. 1). Electrodes were electrically isolated from neighbouring tissue with Vaseline jelly and parafilm. CMAP was recorded by means of two stainless steel Teflon-coated hook electrodes implanted in the genioglossus muscle. The XIIth nerve was stimulated and the orthodromic field potential elicited in the muscle (Fig. 1A) was recorded, amplified, and low-pass filtered to 3 kHz. Stimuli were single cathodic square pulses of 50 µs at 1 Hz and the current intensity applied was three times the threshold current for CMAP detection. In all animals tested at 1, 3 and 7 days, threshold could not be determined, in which case stimulation intensities fivefold higher than threshold measured on the control side (< 0.1 mA) were tested.
Electroneurographic activity was recorded using the same electrodes implanted in the XIIth nerves for stimulation purposes (Fig. 1B, upper trace). Spontaneous activity from both nerves in anaesthetized animals was recorded in monopolar mode, AC-coupled, amplified and filtered (10 Hz to 10 kHz). The signals were visualized, transferred and stored in a computer. Animals used in these experiments were finally perfused, as indicated below, and the brainstems removed and stored for later histological processing.
Extracellular recordings of HMNs
Rats were anaesthetized as previously indicated. Atropine sulphate (0.05 mg) and dexamethasone sodium phosphate (0.2 mg) dissolved in saline were administered intramuscularly to minimize airway fluid secretion and brain oedema, respectively. A ventral approach was used to cannulate trachea, bladder, and femoral artery and vein. Bipolar stimulating electrodes were bilaterally implanted on XIIth nerves (Fig. 1). Vagotomy was bilaterally performed to eliminate the effect of pulmonary afferent signals on central respiratory drive. Then each animal was placed in a stereotaxic frame, injected with the neuromuscular blocking agent gallamine triethiodide (20 mg kg–1I.V. initially and supplemented with 4 mg kg–1 as needed) and mechanically ventilated. Expired CO2 and O2 were continuously monitored (Eliza duo, Gambro Engström, Bromma, Sweden), and data were stored in a computer. The end-tidal CO2(ETCO2) was changed (3–3.5 to 7–7.5%) as required by adjusting ventilation parameters (tidal volume and/or respiratory rate). In all experiments, expired O2 was always between 19% and 14%, and so was higher than those values below which hypoxia-induced alterations have been reported (Hwang et al. 1983a). Femoral arterial blood pressure was monitored and maintained between 80 and 110 mmHg (Fig. 1E). The heart rate was similar in all experimental conditions (370–445 beats min–1) as measured from the pulse pressure wave. When necessary, a modified Ringer solution (2 ml 1 M NaHCO3 and 10 ml 5% glucose in 38 ml Ringer solution) was infused through the femoral vein. Rectal temperature was monitored and maintained at 37 ± 1°C by placing the animal on a temperature-controlled heating pad. Precollicular decerebration, confirmed by post-mortem examination, was performed after craniotomy. Some anaesthetics, including chloral hydrate, can affect respiratory parameters (Hwang et al. 1983b). Therefore, discontinuation of anaesthesia could progressively modify respiratory function. For this reason, although decerebration rendered the animals insentient, low supplemental doses of chloral hydrate (0.05 g kg–1 h–1) were given. This procedure allowed the basal burst rate (BR) to be kept steady during the recording session. Before the beginning of neuronal recording, animals were allowed to stabilize for 30 min after decerebration, maintaining ETCO2 at 4.8–5.2%.
Extracellular recordings were carried out with glass micropipettes bevelled to a resistance of 1–3 M
and filled with a conductive solution of 2 M NaCl. Micropipettes were advanced with a three-axis micromanipulator and visually guided to reach the brainstem at the level of the obex. Then the micromanipulator was laterally displaced 0.2–0.5 mm and advanced through the brainstem down to the HN. The characteristic inspiratory pattern of the HN was heard in the audio monitor and its localization was confirmed by the presence of the antidromic field potential produced by electrical stimulation of the ipsilateral XIIth nerve (Fig. 1C). The stimuli consisted of cathodic square pulses (50 µs, < 0.1 mA, 1 Hz). HMNs were positively identified by their antidromic activation from the XIIth nerve and by the collision test between the orthodromic and antidromic action potentials (Fig. 1C, upper trace). The extracellular neuronal activity was recorded with the aid of a high-impedance circuit located in a head-stage close to the preparation. The electrical signals were amplified and filtered at a bandwidth of 10 Hz to 10 kHz for display and digitization purposes. Responses of HMNs were recorded initially at normocapnia (ETCO2= 4.8–5.2%), and then in response to a change in ETCO2 from hypocapnic (
3%) to hypercapnic (7.5%) conditions. From basal conditions, motoneurones were recorded during a slow increase in ETCO2 up to 7–7.5% followed by a slow decrease down to 3–3.5% and finally recovery to 4.8–5.2%. For each analysed motoneurone, the test was completed in 5 min. Motoneurones remained well isolated during the entire test. Only motoneurones discharging at basal conditions were considered in this study. During extracellular recordings, bursts of enhanced activity of the HN were clearly discriminated on the background noise (Fig. 6A and B). Since HN activity has been shown to be mostly inspiratory (Hwang et al. 1983a,b), bursts were used as indicative of the inspiratory phase. Then, we classified HMNs as inspiratory, expiratory or non-respiratory tonic units according to their relationship with the inspiratory burst activity of the HN. Only inspiratory HMNs were used in this study.
|
General. Electromyographic, electroneurographic and unitary discharge signals, percentages of expired CO2 and O2, and arterial pressure recordings were amplified, filtered, transferred and stored in a computer through a PowerLab/8SP A/D interface (ADInstruments, Castle Hill, Australia). Data acquisition was performed using the Scope and Chart software for CMAP (100 kHz sampling rate) and the remaining signals (40 kHz sampling rate), respectively. The same software was used for off-line analysis of data. Values are always expressed as mean ±S.E.M.
Electromyogram and electroneurogram.
To assess muscle de-innervation and study the time course of re-innervation, CMAPs evoked by XIIth nerve stimulation were analysed at different times after crushing. Peak latency and amplitude of CMAP were measured with respect to the stimulation artifact and baseline, respectively, in control and experimental animals. The electroneurographic signals were integrated (Fig. 1B, lower trace; 
= 20 ms) and measures were taken using a parabola automatically fitted to the integrated burst activity provided by the software (insets in Fig. 4B–D). Several burst parameters including area, duration (BW; burst width) and 10–90% slope (slope10–90) from the baseline were obtained by this method (Fig. 4B–D). Data were averaged from 10 CMAPs and 30 bursts for electromyographic and electroneurographic recordings, respectively, and expressed relative to the contralateral control side. Comparison between experimental and control values, at different post-injury times, was performed using the non-parametric Mann-Whitney U test.
|
0.8 and P < 0.001 were considered for statistical analysis. One-way ANOVA and post hoc comparison tests at a level of significance of P < 0.05 were used for the comparison of the mean parameters relative to different experimental conditions.
|
Animals used for electroneurographic and electromyographic recordings received a supplemental dose of chloral hydrate (0.1 g kg–1; I.P.), and were transcardially perfused first with phosphate-buffered saline (PBS), followed by 4% paraformaldehyde in 0.1 M phosphate buffer (PB), pH 7.4, at 4°C. The brains were removed, post-fixed for 2 h in the same fixative solution, and cryoprotected by overnight immersion in 30% sucrose in PB at 4°C. Coronal sections from the brainstem, 40 µm thick, were obtained using a freezing microtome. Brainstem sections containing the HNs were first incubated in a solution containing 2% (v/v) hydrogen peroxide and 60% (v/v) methanol in PBS, to block endogenous peroxidase activity. Then the tissue was rinsed in PBS and immersed in 2.5% (w/v) bovine serum albumin, 0.25% (w/v) sodium azide and 0.1% (v/v) Triton X-100 in PBS for 30 min, followed by overnight incubation at room temperature with specific anti-synaptophysin rabbit polyclonal antibody (1: 100; Zymed Laboratories, San Francisco, CA, USA). Subsequently, the tissue was rinsed with PBS and incubated for 1 h at room temperature with a biotinylated anti-rabbit IgG (1: 800; Sigma) as secondary antibody. Biotin was detected by means of the avidin–biotin–peroxidase system (Pierce, Rockford, IL, USA) using 3,3'-diaminobenzidine tetrahydrochloride as chromogen. Sections were mounted on slides, dehydrated, covered with DePeX and visualized under light microscopy. Omission of primary antibody resulted in no detectable staining (Fig. 9G).
|
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
In anaesthetized animals, orthodromic CMAPs evoked by single-pulse stimulation of XIIth nerve were recorded to evaluate the functionality of neuromuscular junctions. The amplitude and latency of the resulting wave were taken as an index of lesion and recovery degree of the motor nerve. In controls, CMAP was biphasic, with a large negative wave followed by a small positive wave (Fig. 3A). The negative wave had 2.5 ± 0.04 ms of latency and 9.5 ± 2.0 mV of amplitude. The recorded potential was postsynaptic, since it was eliminated in acute experiments by the blockade of acetylcholine receptors with a neuromuscular blocking agent (gallamine triethiodide; 3 mg kg–1, I.M.) injected in the genioglossus muscle. In all experiments performed 1, 3 or 7 days after crushing, CMAP was completely eliminated, even when supramaximal (< 0.1 mA) nerve stimulation was applied. No signs of recovery were observed during the first week (Fig. 3B and E). These results indicated that our method of nerve crush effectively disconnected most HMNs from their targets.
|
Depression of XIIth nerve activity after crushing
We investigated possible alterations in HMN pool activity by means of nerve recordings carried out proximal to the lesion at 1, 3, 7, 15, 22, 30 and 45 days after nerve crushing in the same animals as those used for electromyographic studies. Activity was expressed as the ratio between crushed and intact sides and compared with data from control animals. Recordings were integrated and area, BW and slope10–90 were measured to characterize nerve burst activity. In control animals, electrical activity of both XIIth nerves was characterized by a rhythmic inspiratory discharge (Fig. 4A). After nerve injury, the inspiratory-related activity was reduced in a time-dependent manner (Fig. 4A–C). Significant reductions in nerve activity appeared as soon as 1 day post-lesion and declined thereafter, reaching a minimum at 15 days post-crushing (Fig. 4B). The integrated area of XIIth nerve bursts was reduced by 23.8 ± 7.6%, 41.1 ± 14.6% and 51.6 ± 8.8% at 1, 3 and 7 days post-injury, respectively. Maximum significant depression occurred 15 days after crushing (62.2%) and a rapid increase towards control values was noticed 7 days later (93.6 ± 8.9% of control; Fig. 4). Integrated activity was maintained at control levels when measured on subsequent days (Fig. 4B).
Besides integrated burst area, other burst parameters were also modified after crushing. This is the case for the slope10–90 of burst, a parameter that was reduced by 49.0 ± 4.9% at 7 days, and whose time course of recovery was parallel to that indicated for mean integrated area (Fig. 4C). Mean values of BW were not significantly affected by nerve crushing at any time post-injury (Fig. 4D). In the sham-operated group, differences with respect to control were not observed (Fig. 4). Therefore, reduced basal XIIth nerve activity was transitory and persisted as long as the CMAP was absent, suggesting that motor nerve and muscle were disconnected. Early after re-innervation was initiated at the second week, a fast and complete re-establishment of normal nerve activity occurred (Figs 3 and 4). Three rats per experimental condition were used in this study, since electroneurographic changes were very homogeneous; results from these experiments were mainly used to select the time points for unitary extracellular HMN recordings.
Slowing of conduction velocity in proximal motor axons
The HN was located by means of the antidromic field potential produced by the electrical stimulation of the ipsilateral XIIth cranial nerve (Figs 1C and 5A, arrowheads). The antidromic spike of HMNs appeared summated on the antidromic field, and in some cases it was somewhat difficult to distinguish the neuronal spike from the antidromic field. To ascertain that the recorded neurone was the same as that activated antidromically, the collision test was used. An orthodromic spike was fed through the window discriminator and the generated pulse was used to trigger, at a certain latency, the antidromic stimulus (Fig. 1C). By shortening the interval between the spontaneous spike and the antidromic stimulus it was possible to occlude the antidromic spike, revealing only the antidromic field potential and demonstrating unequivocally that the recorded neurone was a HMN (Figs 1 and 5A). Only HMNs meeting this criterion and showing inspiratory-related activity were included in this study. The antidromic latency of control HMNs followed a bell-shaped distribution (P < 0.01, F-Snedecor test) of mean 1.1 ± 0.02 ms (Fig. 5B and C).
|
Basal activity of HMNs
The effects of nerve injury on neurogram activity could be explained by the blockade of impulse conduction along axons and/or diminished activity of the motoneurone pool. To evaluate this issue, unitary basal firing activities of HMNs in basal conditions (ETCO2= 4.8–5.2%) were recorded in decerebrated and vagotomized animals under neuromuscular blockade. Burst parameters were measured and averaged using a minimum of 30 bursts per motoneurone to characterize the basal activity of the nucleus at different time points after lesion. Control inspiratory HMNs showed a characteristic respiratory pattern that consisted of bursts of action potentials coincident with the inspiratory phase of breathing, clearly demonstrated by the background nucleus activity (Fig. 6A). This pattern was characterized in the control stage by a BR of 43.7 ± 0.9 bursts min–1 and a mFR of 36.6 ± 2.5 spikes s–1, with a DB of 361.9 ± 14.0 ms (Table 1). All parameters measured are presented in Table 1. At the earliest time point measured, 3 days after crushing, burst parameters underwent a significant reduction that persisted at 7 and 15 days but recovered to control values within 1 month (Fig. 6B–D; Table 1). The mFR was reduced by 35.8 ± 6.3% and 53.6 ± 4.9%, relative to control, at 7 and 15 days after crushing, respectively. However, at these times, BRs were similar to control, indicating that although injury induced severe changes in motoneurone basal activity, structures involved in the rhythmogenesis of breathing remained similar to the non-injured condition (Table 1). The BR was slightly decreased 3 days after injury and in the sham condition. This alteration was not associated with changes in other burst parameters in the sham group, indicating that both effects were independent (Table 1; Fig. 6D). Different anaesthesia levels could explain BR reduction in these two groups.
|
Discharge activity of HMNs in response to end-tidal CO2 changes
The modulation of discharge activity of inspiratory HMNs was studied in relation to ETCO2 variations; it has been well established that these variations modify respiratory parameters. The discharge pattern of control motoneurones consisted of bursts of action potentials during the inspiratory phase of breathing, and the properties of these bursts were generally modified in response to ETCO2 changes (Fig. 7A). The burst activity of motoneurones was increased when ETCO2 rose and was reduced when ETCO2 decreased (Fig. 7A, C and D). For the ETCO2 range used in our experiments, relationships obtained between mFR, pFR, SB, DB or BR and ETCO2 were linear (r 0.8, P < 0.001). Data were firstly grouped and averaged in 0.2% intervals of ETCO2 and then plotted and analysed (Fig. 7C and D). Statistically significant relationships between mFR and ETCO2 were observed in 94% of control motoneurones. The slope of the regression line represented the sensitivity or gain of mFR to ETCO2 that in control averaged 8.1 ± 0.7 spikes s–1%–1 (SmFR in Fig. 7D and E and Table 2). Examples of scatter plots obtained for a control motoneurone are shown in Fig. 7C and D (•). The mean gains of the remaining parameters that were linearly related to ETCO2 are presented in Table 2.
|
|
) together with that obtained for the control motoneurone shown in Fig. 7A. No parallel changes were found in SDB and SBR after the lesion (Table 2). The lack of effects on SDB and SBR is crucial for the validation of these results because these parameters are indicative of the integrity of chemosensors and pre-motor structures. Burst parameters were not affected in the sham condition. Thus, nerve lesion decreased the chemoreceptor-modulated responsiveness of HMNs to ETCO2 changes, with no effects on pre-motor respiratory modulation. Recruitment of the HMN pool
Threshold distribution and recruitment properties of control and lesioned HMNs were analysed to investigate further if they contributed to determining the functional alterations described above. Theoretical recruitment thresholds were the abscissa intercepts in the mFR–ETCO2 regression lines. For a great proportion of units, recruitment thresholds were negative (48.7% of control motoneurones) and therefore outside the physiological ETCO2 range. The mean threshold in control motoneurones was –2.2 ± 0.9% and was not significantly affected at 3 days (–1.9 ± 1.8%), 7 days (–5.1 ± 1.2%), 15 days (–0.1 ± 0.6%) or 30 days (–0.1 ± 0.7%) after nerve crushing. However, although mean values were not altered by axonal lesion, we noticed changes in threshold distribution and in the relation between this and other functional parameters (Fig. 8).
|
To evaluate to what extent changes in recruitment properties and rate modulation could contribute to depressing the electrical activity in the HN after nerve lesion, we calculated the whole nucleus firing activity according to experimental data obtained from the control, lesioned (at 7 and 15 days after crushing) and recovered (at 30 days after crushing) motor pools along the 0–5%ETCO2 interval (Fig. 8; see legend). Predicted curves showed that, despite the higher percentage of recruited units at 0% in the crushed groups (Fig. 8F), the discharge activity of experimental groups never exceeded that calculated in the control. Likewise, mean activities estimated at 7 and 15 days in the 4.5–5% interval were reduced by 57.1% and 60.0% compared with the control, respectively, which was in agreement with experimental observations. These results indicate that, in addition to firing rate modulation, recruitment properties were altered by nerve lesion in the HN. However, although changes in threshold distributions led to a higher proportion of early recruited units in the crushed groups, this was not sufficient to compensate for reductions in firing rates and modulation.
Changes in synaptophysin immunoreactivity in the HN
All the effects reported in relation to firing properties were long lasting. Changes appeared early during the first days, reached a plateau between days 7 and 15, and finally recovered over the third and fourth weeks. We further investigated whether changes in firing properties were accompanied by parallel morphological alterations affecting the density of synaptic boutons. To investigate this hypothesis we carried out immunostaining against synaptophysin, a marker of synaptic terminals. Optical density of immunolabelling in the hypoglossal neuropil was measured at different time points after nerve crushing. Synaptophysin immunolabelling was more dense in the HN than in the surrounding reticular formation (Fig. 9A–D) and appeared as a punctate-like staining when visualized at high magnification (Fig. 9E and F). No significant differences in the optical density between the two sides were observed at 1 day after nerve crushing (Figs 9A and B, and 10). However, at 7 and 15 days after injury, the optical density of synaptophysin immunoreactivity was reduced by 17.1 ± 2.7% and 21.0 ± 5.4% in the experimental side compared with the control side (Figs 9C–F and 10). Synaptic density in the crushed side was similar to the control side at 1 month after injury (Fig. 10). Inspection of the immunostained material at higher magnification demonstrated a reduction in synaptophysin-immunoreactive terminals in the surrounding neuronal neuropil 2 weeks after nerve injury (Fig. 9E and F). These observations indicate that crushing the XIIth nerve induced a decrease in the number of synaptic terminals in the HN that was recovered with muscle re-innervation.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionReferences |
|---|
Remarks about the methodological approach
Although significant efforts have been made to study membrane and synaptic responses to motor nerve axotomy, little is known about how these alterations affect responses of motoneurones to physiological inputs. Chronic studies in the spinal motor system have shown a reversible decline in motoneurone activity during locomotion following axotomy of several mixed nerves (Gordon et al. 1980). A major disadvantage in the interpretation of previous results is that loss of proprioceptive signals on the lesioned side would contribute to a reduced homonymous motor pool activity. In our experimental approach, where animals were vagotomized, decerebrated and a neuromuscular blocking agent administered, it would be expected that pulmonary, descending and proprioceptive afferents were similarly suppressed in control and lesioned sides. Eliminating the influence of these afferent sources that could be directly unbalanced by peripheral lesions and restricting the modalities of afferent information would permit more unambiguous interpretations.
Since most previous studies on the effects of muscle disconnection on motoneurones have used complete nerve transection, it is important to stress distinctions between different forms of nerve lesion. Regenerative and survival capacities could be differentially compromised in adult motoneurones following transection or avulsion of the peripheral nerve (Li et al. 1998). Crushing lesion causes complete axotomy but, unlike nerve transection, it preserves the endoneural tube, providing neurotrophic support and a physical guide for the proximal axonal ends (De Medinaceli, 1988). Consequently, motor nerve crush involves a shorter delay in achieving muscle re-innervation without significant neuronal loss (Rende et al. 1995; Tanaka et al. 1998). According to our observations, crushing the XIIth nerve immediately induced a complete suppression of CMAP, which was fully restored 30 days later in parallel with the integrated burst area of neurographic recordings. The complete recovery of these parameters one month after injury probably reflects that the number of functional motor units in both control and re-innervated conditions is similar. Therefore, the reduced nerve activity we observed after nerve lesion would be determined by functional axonal and central alterations without neuronal loss.
Peripheral response to XIIth nerve crushing
Following the XIIth nerve crushing, we found peripheral electrophysiological alterations that appeared on the proximal motor axons and dennervated muscle. The slowed proximal impulse conduction and the increased threshold to evoke antidromic spikes have been also reported in peripheral and sensory myelinated fibres after nerve transection (Milner & Stein, 1981; Foehring et al. 1986). Several factors have been suggested as underlying these axonal disturbances, such as decreases in the thickness of the myelin sheaths and the axonal diameter (Hoffman et al. 1987; Titmus & Faber, 1990). The distribution of antidromic latencies was progressively widened and skewed to longer values from 3 to 15 days post-injury. This latter time point coincided with initial phases of muscle re-innervation as indicated by the re-appearance of CMAPs. The fact that they were reduced in amplitude and polyphasic probably reflects increased asynchrony in the arrival of motor impulses. Thereafter, antidromic latency returned to normal in parallel with most of the CMAP parameters, which is in agreement with a target-derived trophic control of axonal properties.
Despite proximal recovery at 30 days, CMAP latency remained slightly longer than the control, perhaps due to an incomplete recovery in fibre size, myelinization or both, in regenerated distal axons. It has been suggested that functional and structural axonal properties of fast conducting units (with greater axon calibre) would be preferentially affected over slow units following motor or sensory nerve axotomy (Milner & Stein, 1981; Titmus & Faber, 1990). Consequently, a larger and faster decline in fibre size and conduction velocity and a longer delay in recovery would be expected in large myelinated motor fibres (Titmus & Faber, 1990). The longer latency observed one month after XIIth nerve crushing could indicate a partial prevalence of the structural and functional effects on faster motor units.
Depression in firing activity of HMNs following XIIth nerve crushing
The main alterations observed in the firing pattern of HMNs following XIIth nerve crushing were an overall reduction in firing rates and an almost complete loss of modulation by chemosensory afferents. Factors contributing to these alterations could include changes either in pre-motor afferent activity, in motoneurone excitability and/or in the strength of the synaptic inputs to motoneurones. Modifications of central respiratory patterns induced by peripheral crushing are discounted since (1) although mechanosensory afferent signals could differentially modulate respiratory output in crushed versus intact pools, these peripheral inputs were mostly and similarly reduced in both; (2) although permanent motor or sensory deprivation can induce changes in the activity and topography of higher order central structures (Kaas et al. 1999; Chen et al. 2002), such extension for axotomy-induced retrograde effects has never been reported following reversible peripheral nerve lesions; and (3) the modulation in BR and DB by ETCO2 changes remained similar in control and lesioned animals, suggesting that pre-motor responsiveness was largely preserved.
Membrane excitability is up-regulated in hypoglossal and spinal motoneurones (Eccles et al. 1958; Kuno & Llinás, 1970a,b; Gustaffson, 1979; Takata et al. 1980) following axotomy. Accordingly, it could be expected that membrane alterations would make axotomized HMNs more responsive to synaptic inputs. However, our results appeared to indicate reduced sensitivity to afferent drive. Therefore, although membrane alterations could contribute to the modification of some functional properties such as recruitment distribution (see below), they are not likely to underlie the changes in discharge activity and modulation shown here. As suggested by parallel reductions in synaptophysin immunolabelling, these firing alterations would be better explained by loss of afferent signals.
In relation to the present results, previous studies have revealed that firing sensitivity to afferent signals could be linked to the degree of afferent synaptic innervation (Moreno-López et al. 1997; González-Forero et al. 2002, 2003). Likewise, axotomized motoneurones show attenuated synaptic potentials that have been associated with massive synaptic stripping (Kuno & Llinás, 1970b; Brännström & Kellerth, 1998). Such synaptic detachment may differentially affect excitatory and inhibitory terminals. In HMNs, axotomy mainly affects S-type boutons, forming excitatory synapses (Sumner, 1975). Because the inspiratory-related discharge in HMNs is determined by afferent synaptic excitation (Ono et al. 1994; Peever et al. 2002), and since that firing activity was reduced in experimental animals, we might conclude that excitatory synaptic transmission was depressed by lesion of the XIIth nerve.
Changes in recruitment scheme and motor output after XIIth nerve crushing
The output of a motor pool is regulated by both recruitment gradation and firing rate modulation. Since thresholds, firing rate gains, force production and, frequently, synaptic strengths vary for each unit in a motor pool, one would expect a certain degree of covariation between these variables in order to adjust force production to motor requirements. Although recruitment order is primarily determined by the size principle (Henneman et al. 1965; Gustaffson & Pinter, 1985), the organization of synaptic inputs could ultimately determine the rank and gain for recruitment or even facilitate disruption of the size-determined scheme (Kernell & Hultborn, 1990; Heckman & Binder, 1993). According to the size principle, slower conducting phrenic motoneurones are recruited first during inspiration (Jodkowski et al. 1987). For HMNs this information is lacking, although it has been suggested that the recruitment pattern of this pool differs from that of phrenic units (Hwang et al. 1983a). In our study, no information about size-related parameters other than unitary antidromic latency could be obtained. However, we did not find good relationships (r < 0.5) between latency and threshold.
On the other hand, recruitment thresholds were exponentially correlated with SmFR. Motoneurones with lower SmFR were usually recruited earlier than those with higher gains. This recruitment scheme was not altered by XIIth nerve lesions, indicating that SmFR were good predictors of the recruitment sequence in both control and experimental HNs. Similar relationships between firing rate gains and threshold have been found in cat abducens motoneurones (Pastor & González-Forero, 2003) and would be expected from the work of Monster & Chan (1977) on human spinal motoneurones. Therefore, such distributions could represent a common principle for recruitment organization in motor pools. Likewise, the rate of recruitment (recruitment gain) increased with thresholds in parallel with firing rate gains. A non-homogeneous distribution of thresholds and SmFR over the motor range could be related to requirements for a gradation of force production. The poor firing rate modulation and reduced recruitment gain for low threshold units would be more suitable for sustaining a repetitive inspiratory discharge at low levels of force output. Since for these earlier recruited units, thresholds usually fell outside the 0–5%ETCO2 range, a minimal basal inspiratory activation would be always guaranteed in physiological conditions. This can be critical for maintaining upper airways patency during low activity states such as sleep. The absence of these always-active units could be associated with the symptomatic presence of the obstructive sleep apnoea syndrome. In contrast, a higher firing rate and recruitment gains beyond 0% of ETCO2 would make the motor pool more responsive to changes in afferent drive in physiological conditions. Therefore, the distribution of recruitment and rate gradations over the motor range would facilitate appropriate force modulation between 0 and 5% of ETCO2 superimposed on a minimal basal activation.
Since nerve lesion affects both intrinsic (membrane properties) and extrinsic (synaptic organization) properties, the recruitment alterations reported here would be complex to interpret. The elevation of input resistance and reduction of rheobase lead to a decrease in the current thresholds for impulse discharge in axotomized motoneurones (Gustaffson, 1979). On the other hand, synaptic input loss could differentially affect recruitment properties depending on afferent distribution over the motoneurone pool (Kernell & Hultborn, 1990; Heckman & Binder, 1993). We found that XIIth nerve lesion led to a higher proportion of low-sensitivity and early recruited units as well as to reductions in the recruitment gain across the physiological range. Therefore, it is possible that the recruitment alterations noticed here result from both increased excitability and loss of afferent signals. Increased input resistance could explain an elevated incidence of low threshold units, while changes in the recruitment gain would be better related to the reorganization of synaptic inputs. This suggestion is in agreement with present findings and previous reports on spinal and cranial motoneurones showing changes in the recruitment gain following selective or massive deafferentation (Powers & Rymer, 1988; Haftel et al. 2001; Pastor & González-Forero, 2003). After XIIth nerve crushing, a reduction in the recruitment gain could be associated with the loss of efficacy from inspiratory afferent inputs that scale in parallel to firing rate gains in the HMN pool, i.e. inducing higher synaptic responses in motoneurones with high sensitivities and thresholds. Alternatively, as noticed above, it is also possible that XIIth nerve lesions induce more profound functional and anatomical alterations in high threshold motoneurones (Milner & Stein, 1981; Giannini et al. 1989; Titmus & Faber, 1990). Hence, a more specific effect of XIIth nerve crushing on high-threshold HMNs, implying greater membrane and synaptic alterations, could contribute to reducing the proportion of motoneurones recruited in the 0–5%ETCO2 interval, and to increasing the number of those with lower thresholds and sensitivities.
As indicated by experimental and predicted data, an earlier recruitment was not associated with higher motor output in the lesioned HMN pool. Total firing activity in the experimental nucleus never exceeded that of the control pool. That would be obvious since the relationships between thresholds and SmFR persisted after crushing, and therefore the reduction of thresholds was also associated with gain decreases. These data indicate that alterations in the recruitment distribution did not contribute to the depression of electrical activity in the HN. We conclude that the deficient firing rate modulation in the HN is the primary cause of activity depression after peripheral nerve lesion.
Opposition between membrane and synaptic alterations following target disconnection. Compensatory response or functional de-differentiation?
A generalized finding for axotomized motoneurones is the increased membrane excitability (Titmus & Faber, 1990; González-Forero et al. 2004). Electrical changes implicate reductions in the rheobase and increases in the membrane time constant, input resistance and gain for the ratio between firing rate and current intensity (Heyer & Llinás, 1977; Gustaffson, 1979; Takata et al. 1980). Consequently, electrical changes would be expected to facilitate firing activity in response to synaptic inputs and reduce the threshold for discharge. However, despite this presumed facilitation, axotomized motoneurones usually show attenuated synaptic responses (Eccles et al. 1958; Kuno & Llinás, 1970b) and as reported here, have depressed firing activity, and do not respond to enhanced synaptic (chemoreceptor) drive. Therefore, we suggest that, under physiological patterns of activation, the effects of deafferentation prevail over those arising from membrane alterations in axotomized HMNs.
The functional significance of these obviously counterposed changes remains uncertain, although some authors have suggested that electrical alterations would imply a functional de-differentiation in axotomized motoneurones towards a postnatal-like stage (Kuno et al. 1974). Anatomical and electrical properties of neonatal motoneurones mature in parallel with muscle fibres after motor axons reach their target muscles (Navarrette & Vrbova, 1993). Therefore, the activation of regrowth programmes in adult motor axons following axotomy could require reversion towards an immature electrical phenotype. Another possibility is that increased excitability in axotomized motoneurones would constitute a compensatory response to counteract the deficiency in ongoing excitatory drive (Eccles et al. 1958: Kuno & Llinás, 1970a). In lesioned HMNs, reductions in firing activity and modulation were not completely compensated by possible membrane alterations since they persisted during muscle disconnection and recovered when neuromuscular transmission was restored. However, increased membrane excitability could sustain certain levels of electrical activity necessary to prevent atrophy or neuronal death, and favour axonal growth, synaptogenesis and differentiation in both immature and regenerating motoneurones.
Implications for nerve regeneration and functional recovery
Motor unit differentiation during development is modulated by patterns of electrical activity (Kalb & Hockfield, 1992; Navarrette & Vrbova, 1993). Because axotomized motoneurones exhibit electrical, histochemical and anatomical properties that resemble those observed in neonatal motoneurones (Kuno et al. 1974), it is possible that the differential patterns of electrical activity shown by regenerating motoneurones, at least to a certain degree, play a role in instructing axonal regrowth and target re-innervation. Chronic electrical stimulation increases the speed and accuracy of motor axon regeneration following peripheral nerve transection in adulthood (Al-Majed et al. 2000). Therefore, we can speculate that reduced afferent synaptic and firing activity in HMNs after XIIth nerve lesions would condition a low rate of regeneration. Additionally, firing depression in the HMN pool could also contribute to the loss of specificity during regeneration that usually leads to aberrant muscle re-innervation and anomalous contraction patterns (Mizuno et al. 1980; Wilson et al. 1994). Our findings could be used as a reference for developing therapeutic approaches that compensate discharge depression after peripheral nerve injury. It remains to be determined if, as has been demonstrated for the effect of electrical stimulation, treatments that increase afferent synaptic drive and firing activity in motoneurones would promote a faster and more appropriate muscle re-innervation.
The findings reported here contribute to a better understanding of the physiological alterations derived from physical disconnection between motoneurones and muscle. We have demonstrated that lesion of the XIIth nerve led to reduced firing activity in the HN that, in combination with the marked reduction in synaptic inputs, resulted in a poor motor output across the whole motor range. Previous studies have reported similar reductions in the spontaneous activity displayed by sensory and motor neurones (Gordon et al. 1980; Delgado-García et al. 1988; Seburn et al. 1999). In conclusion,
) and sham (•) groups. Values represent means ±S.E.M.*Significant differences (P < 0.05; one-way ANOVA; post hoc Dunnett's method) relative to the control group. E, plot of mFR versus neurogram area (both expressed as percentage of control) including data points corresponding to the control, sham, and 3, 7, 15 and 30 days post-lesion groups. For each point, the average mFR was calculated from the whole sample of units recorded (n= 18–53) in decerebrated and vagotomized animals under neuromuscular blockade, and the neurogram area as the mean value for the neurographic recordings (n= 3) in anaesthetized animals in the same conditions. This relationship was fitted by the regression line: y=–14.0 + 1.12x (r= 0.98; P < 0.0001).
) of the antidromic activation latencies obtained in the control (n= 112), and at 7 (n= 55), 15 (n= 65) and 30 (n= 55) days post-lesion. C, time course of changes in the antidromic activation latency (means ±S.E.M.) after XIIth nerve crushing (
