Reflex wind-up in early chronic spinal injury: plasticity of motor outputs

This research is the first to assess temporal summation, also called wind-up, of muscle reflexes during the 1-mo recovery period following spinal injury. Our results show that two types of muscle reflex activity are differentially modulated 1 mo after spinal cord injury (SCI) and that spinal reflexes are altered in a muscle-specific manner during this critical period. This postinjury plasticity likely plays an important role in spasticity experienced by individuals with SCI.

Abstract

In this study we evaluate temporal summation (wind-up) of reflexes in select distal and proximal hindlimb muscles in response to repeated stimuli of the distal tibial or superficial peroneal nerves in cats 1 mo after complete spinal transection. This report is a continuation of our previous paper on reflex wind-up in the intact and acutely spinalized cat. To evaluate reflex wind-up in both studies, we recorded electromyographic signals from the following left hindlimb muscles: lateral gastrocnemius (LG), tibialis anterior (TA), semitendinosus (ST), and sartorius (Srt), in response to 10 electrical pulses to the tibial or superficial peroneal nerves. Two distinct components of the reflex responses were considered, a short-latency compound action potential (CAP) and a longer duration bout of sustained activity (SA). These two response types were shown to be differentially modified by acute spinal injury in our previous work (Frigon A, Johnson MD, Heckman CJ. J Physiol 590: 973-989, 2012). We show that these responses exhibit continued plasticity during the 1-mo recovery period following acute spinalization. During this early chronic phase, wind-up of SA responses returned to preinjury levels in one muscle, the ST, but remained depressed in all other muscles tested. In contrast, CAP response amplitudes, which were initially potentiated following acute transection, returned to preinjury levels in all muscles except for Srt, which continued to show marked increase. These findings illustrate that spinal elements exhibit considerable plasticity during the recovery process following spinal injury and highlight the importance of considering SA and CAP responses as distinct phenomena with unique underlying neural mechanisms.

NEW & NOTEWORTHY This research is the first to assess temporal summation, also called wind-up, of muscle reflexes during the 1-mo recovery period following spinal injury. Our results show that two types of muscle reflex activity are differentially modulated 1 mo after spinal cord injury (SCI) and that spinal reflexes are altered in a muscle-specific manner during this critical period. This postinjury plasticity likely plays an important role in spasticity experienced by individuals with SCI.

temporal summation (wind-up) to repeated inputs is a common feature in spinal neural elements and highly dependent on descending neuromodulatory inputs from the brain stem. It has been known for over 50 years that spinal neurons display wind-up (Crone et al. 1988; Granit et al. 1957; Mendell and Wall 1965). Assessing wind-up of reflex responses can serve as a direct measure of spinal circuit function in response to spinal cord injury (SCI). Precise sensory processing by spinal neural circuits is critical for proper motor control and exhibits considerable plasticity during the recovery process following SCI (Johnson et al. 2013). Both acute and chronic SCI can significantly alter spinal sensory processing and resulting reflex activations. These alterations can include debilitating spasms and muscle cramps that may be triggered locally by activation of muscles at a single joint and then become amplified and spread throughout the entire limb. Wind-up of responses in spinal neurons can exacerbate spasticity and unintended muscle activations and contribute to the motor dysregulation associated with spinal injury.

Spinal injury not only disrupts the pathways that carry volitional motor commands from the cortex but also severely impairs descending pathways from the brain stem that provide neuromodulatory input to spinal neurons. Motoneuron (MN) persistent inward currents (PICs), which provide synaptic amplification and are critical for normal MN function, exhibit wind-up and are dependent on brain stem monoaminergic inputs that are eliminated following spinal transection. It is now known that the immediate loss of MN excitability following spinal transection is only temporary. In as little as 3 wk following acute spinal transection, the intrinsic properties underlying MN excitability start to return (Bennett et al. 2001). This is thought to be at least partly due to the emergence of constitutive activity in monoaminergic receptors on spinal MNs (Murray et al. 2010; Rank et al. 2011). Concurrent with the return of MN excitability is the emergence of hyperreflexia (Bennett et al. 2001) and disorganized reflex activations (Hyngstrom et al. 2008; Frigon et al. 2011; Frigon et al. 2012a; Frigon et al. 2012b). The activation of these aberrant spinal reflexes is likely a mechanism by which spasticity can develop and spread throughout the limb.

Wind-up, defined as a progressive increase in neuronal response in both duration and amplitude to a repeated stimulus, is a common feature of neurons in the spinal cord (Mendell and Wall 1965). Numerous studies have been done showing the involvement of N-methyl-d-aspartate (Davies and Lodge 1987), metabotropic glutamate (Russo et al. 1997), monoamines (Diaz et al. 1997; Herrero et al. 2000; Sullivan et al. 1992), and opioids in the production and maintenance of wind-up (Dickenson and Sullivan 1986). For the most part, these studies focused on recordings in dorsal horn interneurons or in dorsal or ventral roots. Spinal MN intrinsic properties include prominent PICs, which are responsible for the generation of plateau potentials and self-sustained firing (Hounsgaard and Kiehn 1985; Schwindt and Crill 1980). PICs are also subject to wind-up (Bennett et al. 1998; Svirskis and Hounsgaard 1997). However, few studies have been done on the role of wind-up of muscle activation or force production, the final output of the motor system. Fewer still have attempted to study wind-up in the context of spinal injury, although see Hornby et al. (2006). These researchers found temporal facilitation of stretch reflexes in human SCI subjects (8 motor complete, 5 motor incomplete) in ankle plantar flexor muscles looking at only one component of the electromyogram (EMG), the sustained (SA) gross EMG responses, and ankle torque. We have previously demonstrated that wind-up of two distinct reflex responses, a short-latency CAP, which consists of the synchronous activation of multiple motor units, and the SA response, as well as muscle force, is prominent in the intact mammalian spinal cord. Moreover, following acute spinal transection, force wind-up in many muscles is abolished and CAP and SA responses are differentially modulated. Specifically, SA responses tended to be drastically reduced and CAP responses increased, suggesting that these two signals are the result of distinct neuronal circuits. This provides the basis for a potentially valuable tool to assess changes in spinal motor circuit excitability in patients with spinal injury (Frigon et al. 2012a). Also, these studies have focused on muscles acting on a single joint. There is evidence that in the weeks following spinal injury, as spinal MN excitability reemerges, aberrant activation of motor pools throughout the limb occurs in response to focal sensory inputs (Johnson et al. 2013).

It has been shown that in decerebrate rat preparations, acute spinalization enhances wind-up responses compared with those in anesthetized animals (Gozariu et al. 1997). However, under specific circumstances (i.e., hyperalgesia and wind-up from large fiber input), spinalization can have the opposite effect (Herrero and Cervero 1996). Others have demonstrated that wind-up of reflex responses parallels the development of spasticity following spinal injury in rats spinalized at the sacral level (Bennett et al. 2001). In humans with spasticity associated with spinal injury, stretch reflexes of the ankle muscles exhibit significant wind-up with repeated joint rotations as well as prolongation of EMG responses in response to ramp and hold excursions, both of which are partially ameliorated by baclofen, a GABA agonist used to control spasticity (Hornby et al. 2003). The pathological spread of muscle activations following spinal injury has also been demonstrated in studies evaluating wind-up in lower limbs (Onushko et al. 2010; Onushko et al. 2011).

We hypothesize that with the return of motoneuron excitability (Bennett et al. 2001) and dysregulation of sensory processing following loss of descending brain stem inputs (Johnson et al. 2013), wind-up responses will reemerge in some motor pools. We have previously tested the effects of acute SCI on wind-up responses of select hindlimb motor pools. In the present study, we evaluate wind-up of reflex responses in these same motor pools during the recovery phase 1-mo post-SCI to better understand neural plasticity during the recovery process in the weeks following spinal injury.

METHODS

Ethical Information and Surgical Procedures

All procedures were approved by the Institutional Animal Care and Use Committee of Northwestern University. All animals were obtained from a designated breeding establishment for scientific research. Before the experiments, animals were housed and fed within designated areas, which were monitored daily by veterinary staff and trained personnel. The current data set is compiled from 7 adult cats weighing between 2.5 and 5.0 kg. The data presented are derived from experiments specifically designed to address multiple scientific questions related to the pathophysiology of spinal injury (Frigon et al. 2011; Frigon et al. 2012a; Frigon et al. 2012b). No additional animals were used to compile the current data set. This is part of our ongoing effort to maximize the scientific output from every animal experiment.

Survival surgery.

In 7 cats, the spinal cord was completely transected at low thoracic levels (T12–T13) ~1 mo before the terminal experiment. The spinalization was performed under aseptic conditions in an operating room with sterilized equipment. Before surgery, cats were sedated (Telazol, 3 mg/kg im; acepromazine, 0.05-0.1 mg/kg im; glycopyrrolate, 0.01 mg/kg sc). Anesthesia was induced with propofol (2–3 mg/kg iv). Once anesthetized, the cat was immediately intubated with a flexible endotracheal tube. Anesthesia was maintained by adjusting isoflurane concentration as needed (1.5–3%). The fur overlying the back was shaved with electric clippers, and loose hair was removed by vacuum. An intravenous line was placed in the cephalic vein of either the right or left forelimb for administration of fluids (warmed lactated ringer solution or saline + 2.5% dextrose) at a rate of 5–10 ml·kg⁻¹·h⁻¹ throughout the surgery to maintain hemodynamic stability. The level of anesthesia was confirmed and adjusted throughout the surgery by monitoring cardiac rate, respiratory rate, and jaw tone and by applying pressure to the paw to detect limb withdrawal. Body temperature was recorded using a rectal thermometer. Monitoring of the parameters was done on a continuous basis (i.e., cardiac and respiratory rates) and recorded every 15 min. Ophthalmic ointment was applied to the eyes.

A dorsal vertebral laminectomy was performed at T12–T13, a small incision was made in the dura, and, after local lidocaine application (Xylocaine; 2%), the spinal cord was completely transected with tenotomy scissors. Hemostatic material (Surgicel) was inserted within the resulting gap, and muscles and skin were closed in anatomic layers. A transdermal fentanyl patch (25 µg/h) was adhered to the back of the animal 2–3 cm from the base of the tail. At the beginning of surgery and ~7 h later, an analgesic (buprenorphine, 0.01 mg/kg) was administered subcutaneously. An antibiotic (Baytril; 5 mg/kg sc) was given once a day for 5 days after spinalization to prevent urinary infection. The bladder was manually emptied one to two times each day up to the acute experiment. The animals were monitored daily by veterinary personnel. The hindlimbs of the cats were frequently cleaned by placing the lower half of the body in a warm soapy bath.

Terminal experiment.

For the terminal experiment, anesthesia was induced in a clear plastic cylindrical chamber with 1.5–3% isoflurane in a 1:3 mixture of O₂ and NO₂. After ∼15 min, the animal was transferred to a surgical table and anesthesia was continued with a mask. A tracheotomy was performed and a permanent tracheal tube inserted through which anesthesia was delivered for the duration of initial procedures. The right common carotid artery and right jugular vein were cannulated to monitor blood pressure and administer intravenous fluids and drugs. Anesthesia level was adjusted in response to continuous monitoring of blood pressure and heart rate as well as responses to toe pinch. The animal was then transferred to a stereotaxic frame (Kopf) for further surgery. After a craniotomy, a precollicular decerebration was performed. At this point, animals were considered to have complete lack of sentience and anesthesia was discontinued (Silverman et al. 2005). A lethal injection of potassium chloride (2 mg/kg iv) was administered at the end of the experiment.

Experimental Design

A schematic illustration of the experimental setup is shown in Fig. 1A. Bipolar wire electrodes were inserted into the soleus (Sol; ankle extensor), lateral gastrocnemius (LG; ankle extensor/knee flexor), semitendinosus (ST; knee flexor/hip extensor), anterior sartorius (Srt; hip flexor/knee extensor), and tibialis anterior (TA; ankle flexor) of the left hindlimb and in the right ST for electromyography (EMG). The left knee was fixed with a custom-made clamp attached to the femoral epicondyles and both hind paws were held with clamps (not shown). Data from the left LG, TA, ST, and Srt are reported in this study. EMG signals were amplified (×1,000) with a multichannel amplifier (model 3500; AM Systems), bandpass filtered (300–3,000 Hz), and sampled at 10,000 Hz. Bipolar stimulating cuff electrodes were placed around the left distal tibial and superficial peroneal (SP) nerves near the ankle joint.

Fig. 1.

Schematic illustration of the experimental setup. Bipolar wire electrodes were inserted into the soleus (ankle extensor), lateral gastrocnemius (ankle extensor/knee flexor), semitendinosus (ST; knee flexor/hip extensor), anterior sartorius (hip flexor/knee extensor), and tibialis anterior (ankle flexor) of the left hindlimb and in the right ST for electromyography (EMG). The left knee was fixed with a custom-made clamp attached to the femoral epicondyles, and both hind paws were held with clamps (not shown). Bipolar stimulating cuff electrodes were placed around the left tibial and superficial peroneal nerves near the ankle joint (A). Sustained activity (SA) responses were calculated by rectifying and smoothing the EMG waveforms (B). Compound action potential (CAP) signals were measured as the peak-to-peak amplitude of the signal in a window from 0.005 to 0.031 s following each stimulus pulse (C). Tib n. stim, tibial nerve stimulation.

Experimental Protocol

Motor thresholds for each nerve stimulation were determined. For the tibial nerve, the motor threshold was the stimulation intensity required to evoke a small plantar flexion of the right toes; for the SP nerve, it was the stimulation intensity required to evoke a small flexion response at the knee. All stimulation intensities are expressed as multiples of these thresholds (T) and were chosen to predominantly activate large-diameter afferents with minimum activation of nociceptive fibers (Jack 1978). Our stimulation protocol consisted of a series of 10 pulses (0.2-ms pulse width) to the left tibial or SP nerves at 2T or 5T at a frequency of 1 or 2 Hz. Typically, wind-up of reflex responses is more evident at stimulation frequencies of 1 Hz or more (Mendell and Wall 1965). Generally, two trials were averaged for each stimulation parameter. Separate trials were performed at intervals of no less than 1 min.

Measurements

All EMG measurements were made with Spike2 version 6.0 (Cambridge Electronic Design). Two EMG measurements were made. The first measure was made to assess the presence of sustained activity (SA) following stimulation. The SA is an asynchronous signal that generally takes the form of a prolonged burst of activity following the stimulation (see Fig. 1B), at least in the intact state. To calculate the SA, the EMG waveforms were rectified and smoothed with a 0.03 time constant. Ten windows were placed following each pulse from 0.05 s to 1 or 0.5 s for 1- or 2-Hz stimulation frequencies, respectively. The SAs were calculated as the areas under the curve (modulus function in Spike2) within these 10 windows (in mV).

The second measure was the short-latency compound action potential (CAP) illustrated in Fig. 1C, a signal that primarily results from synchronous activation of multiple motor units. To calculate the short-latency CAP, 10 windows were created following each stimulus pulse from 0.006 to 0.031 s. The minimal and maximal values were determined within these windows and the CAP peak-to-peak amplitude was measured (in mV). A short-latency response is typically evoked at ~8–10 ms following stimulation of cutaneous nerves at rest or during locomotion in intact or spinalized cats (Frigon and Rossignol 2007; Frigon and Rossignol 2008a; Frigon and Rossignol 2008b).

To evaluate the presence of wind-up the slope of the responses evoked by the first five stimulations was calculated. To facilitate comparisons between states (i.e., intact, acute spinal, and chronic spinal) and responses (i.e., CAP and SA), the first five data points were normalized to the first response (see Frigon et al. 2012a; Hornby et al. 2003). Only the first five values were used because the responses tended to plateau around the fifth stimulation. Slopes and coefficients of determination (r²) were measured by linear regression analysis using Excel software (Microsoft). Slope calculations have been used to assess the presence of wind-up (Mitsuyo et al. 2006).

Statistical Analysis

Responses obtained from stimulating the left tibial and SP nerves were treated separately. For each nerve, responses obtained at 2T, 5T, 1 Hz, and 2Hz were pooled for statistical analysis. An analysis of variance (single-factor ANOVA) was done with the 10 stimulation pulses. The intact, acute spinal, and early chronic spinal states were considered the intersubject factor for the slopes of the linear regression analysis obtained for LG, TA, ST, and Srt EMG responses. Statistical significance was set at P ≤ 0.05. Data in the graphs are group averages (±SE) of each stimulus condition (frequency and intensity).

RESULTS

Our objective was to determine the effects of recovery time on wind-up of two distinct muscle responses, SA and CAPs, 1 mo following complete spinal transection. This transection was identical in extent and location to the injury imposed in our acute spinal injury protocol. Our experience with chronic full spinal transection has shown that the secondary effects of injury (necrosis, cyst, and glial scar formation) extend approximately one spinal segment rostral and caudal from the injury site. The low thoracic location of the transection site was chosen to be adequately far from the motor pools of the muscles in this study. In this study and in our previous work, we have observed that the nerve stimuli evoked muscle activation with minimal movement of the restrained limb. In our previous work, we examined the impact that acute spinal transection had on wind-up (Frigon et al. 2012a). In the current study, we test the effects of spinalization 1 mo into the recovery process following acute transection, which we termed “early chronic.”

The results from our previous work are included in the current figures. These describe responses in cats 1 mo into the recovery period following spinal transection to illustrate the effect of the three different states: intact, acute, and early chronic spinal transection on spinal reflexes. To simplify the graphs in this study, only the average response across both stimulus intensities and frequencies is shown for the intact (thick green lines) and acute spinal states (thick red lines). For the early chronic state, the average response (thick blue lines) as well as the responses to all conditions (thin blue lines) are shown. Initial analysis of the individual intensities and frequencies did not show statistical significance and were therefore pooled together for analysis. Our previous study showed that acute thoracic spinalization completely abolished wind-up and greatly reduced the amplitude of SA reflex responses in hip, knee, and ankle flexors and extensors (Frigon et al. 2012a). The effect on SA wind-up and SA amplitude was consistent across stimulus intensities and frequencies. The exception was the hip extensor/knee flexor ST, which at the higher stimulus intensity (5T) continued to exhibit some, but less, wind-up immediately after complete spinalization (see Figs. 3C and 4C).

Fig. 3.

Sustained activity (SA) slope group data with tibial nerve stimulation (Tib Stim). A–D: graphs show the slope (m) of the first 5 SA responses normalized to the first response for intact, acute spinal, and chronic spinal states. Slopes were calculated from the average across all stimulus parameters (1- and 2-Hz frequency and 2T and 5T thresholds). Wind-up of SA responses remains abolished 1 mo postspinal transection in lateral gastrocnemius (LG; A), tibialis anterior (TA; B), and sartorius muscles (Srt; D) but shows strong recovery in the semitendinosus muscle (ST; C).

Fig. 4.

Sustained activity (SA) amplitude group data with super peroneal nerve stimulation. A–D: graphs show SA reflex responses in lateral gastrocnemius (LG; A), tibialis anterior (TA; B), semitendinosus (ST; C), and sartorius muscles (Srt; D) evoked by a series of 10 stimulation pulses of the superficial peroneal nerve (SP stim). Each trace is the average (avg) response across 7 cats for 2 different thresholds (2T, 5T) and 2 different frequencies (1 Hz, 2 Hz). The thick lines are the averages across all stimulus parameters. Green, red, and blue traces indicate the 3 states (spinal intact, acute transection, and early chronic transection, respectively). As with the tibial nerve stimulus trials, the ST muscle EMG amplitudes in the early chronic state approach those in the preinjury state with superficial peroneal nerve stimulation (C).

Acute spinalization had the opposite effect on CAPs, which tended to show an increase in amplitude following acute transection in all muscles (Frigon et al. 2012a) (see also Figs. 6 and 7). Before spinalization, stimulus frequency and intensity had no significant effects on wind-up of SAs (see Frigon et al. 2012b) for further discussion).

Fig. 6.

Compound action potential (CAP) amplitude group data with tibial nerve stimulation. A–D: graphs show CAP responses in lateral gastrocnemius (LG; A), tibialis anterior (TA; B), semitendinosus (ST; C), and sartorius muscles (Srt; D) evoked by a series of 10 stimulation pulses of the distal tibial nerve (Tib stim). Each trace is the average (avg) response across 7 cats for 2 different thresholds (2T, 5T) and 2 different frequencies (1 Hz, 2 Hz). The thick lines are the averages across all stimulus parameters. Green, red, and blue traces indicate the 3 states (spinal intact, acute transection, and early chronic transection, respectively). CAP amplitudes are smaller in the early chronic state in all muscles except Srt, which shows larger amplitude CAP than in the already potentiated acute state (D; thick blue trace).

Fig. 7.

Compound action potential (CAP) amplitude group data with super peroneal nerve stimulation. A–D: graphs show CAP responses in lateral gastrocnemius (LG; A), tibialis anterior (TA; B), semitendinosus (ST; C), and sartorius muscles (Srt; D) evoked by a series of 10 stimulation pulses of the superficial peroneal nerve (SP stim). Each trace is the average (avg) response across 7 cats for 2 different thresholds (2T, 5T) and 2 different frequencies (1 Hz, 2 Hz). The thick lines are the averages across all stimulus parameters. Green, red, and blue traces indicate the 3 states (spinal intact, acute transection, and early chronic transection, respectively). As with the tibial nerve stimulation, Srt CAP responses to superficial peroneal nerve stimulation in the early chronic state start to show recovery toward preinjury levels (D; thick blue trace).

Effect of Early Chronic Spinalization on Reflexes Responses

Sustained activity reflex responses.

Figures 2 and 4 summarize SA reflex responses to stimulation of the distal tibial and SP nerves in all cats in the intact (uninjured), acutely spinalized, and early chronic spinalized (1 mo) states. Figures 3 and 5 characterize reflex wind-up in these three states and show the slope of a line fit to the first five responses to nerve stimulation. Overall, SA reflex responses in the intact state tended to be large and showed pronounced wind-up (Figs. 3 and 5, solid traces). These responses were greatly reduced in amplitude and showed significantly less wind-up (P < 0.005), as assessed by slope values, immediately following spinal transection (Figs. 2, 3, 4, and 5, dashed traces). When reflexes were tested 1 mo after transection, SA response amplitudes evoked by tibial nerve stimulation remained significantly depressed compared with the intact state and were not significantly different from the acute state in LG, Srt, and TA (Fig. 2). Amplitudes were generally smaller in these muscles in response to SP stimulation (Fig. 4, A–D), although this reduction in amplitude reached significance (P < 0.05) in TA only. However, SA amplitudes recovered to near preinjury levels in ST (Figs. 2C and 4C, blue lines). In the early chronic state, nearly all SA responses showed little to no wind-up and were not significantly different from responses in the acute state (Figs. 3 and 5). The exception was ST, which showed a strong tendency toward recovery of wind-up (Figs. 3C and 5C, blue lines).

Fig. 2.

Sustained activity (SA) amplitude group data with tibial nerve stimulation. A–D: graphs show SA reflex responses in lateral gastrocnemius (LG; A), tibialis anterior (TA; B), semitendinosus (ST; C), and sartorius muscles (Srt; D) evoked by a series of 10 stimulation pulses of the distal tibial nerve (Tib stim). Each trace is the average (avg) response across 7 cats for 2 different thresholds (2T, 5T) and 2 different frequencies (1 Hz, 2 Hz). The thick lines are the averages across all stimulus parameters. Green, red, and blue traces indicate the 3 states (spinal intact, acute transection, and early chronic transection, respectively). In the ST muscle, SA response amplitudes show a tendency to return to preinjury levels (C; thick blue trace).

Fig. 5.

Sustained activity (SA) slope group data with super peroneal nerve stimulation (SP Stim). A–D: graphs show the slope (m) of the first 5 SA responses normalized to the first response for intact, acute spinal, and early chronic spinal states in lateral gastrocnemius (LG; A), tibialis anterior (TA; B), semitendinosus (ST; C), and sartorius muscles (Srt; D). Slopes were calculated from the average across all stimulus parameters (1- and 2-Hz frequency and 2T and 5T thresholds). As with the tibial nerve stimulation, ST muscle responses with superficial peroneal nerve stimulation in the early chronic state start to show recovery toward preinjury levels (C).

Compound action potentials.

Figures 6 and 7 summarize the effects of nerve stimulus rate and intensity on CAP responses in the intact, acute spinal, and early chronic spinal states. Figure 6 shows results from stimulation of the distal tibial nerve, and Fig. 7 shows results from stimulation of the SP nerve. We previously showed that in the acute stage of spinal transection, TA, ST, and Srt CAP amplitudes were significantly larger than in the intact state (Frigon et al. 2012a). In the current study, 1 mo into the recovery period following spinal transection, CAP amplitudes in the Srt muscle continued to grow larger than their already potentiated amplitudes in the acute state (Figs. 6D and 7D, blue lines), whereas CAP amplitudes were significantly smaller than in the acute state for LG, TA, and ST muscles (P < 0.05) and smaller than preinjury for TA only (P < 0.05) (summary in Table 1).

Table 1.

Effects of acute and early chronic spinal transection on SA and CAP EMG responses

	Soleus	Lateral Gastrocnemius	Semitendinosus	Sartorius	Tibialis anterior
Acute spinal state
SA overall amplitude	↓	↓	↓	↓	↓
SA wind-up	―	―	↓	―	―
CAP overall amplitude	↓	↓	↑	↑	↑
Early chronic spinal state
SA overall amp	↓	↓	↑	↓	↓
SA wind-up	―	―	↑	―	―
CAP overall amp	↓	↓	↓	↑	↓

Summary data describe the effects of immediate (acute) and 1-mo after (early chronic) spinal transection on sustained activity (SA) and compound action potential (CAP) EMG responses. Acute spinal transection reduced (↓) the overall amplitude of EMG in all muscles tested and significantly reduced wind-up in the SA of semitendinosus muscle (―, no significant change). Acute spinal transection had mixed effects on the amplitude of CAP responses, specifically, reduction in soleus and lateral gastrocnemius and potentiation (↑) in semitendinosus, sartorius, and tibialis anterior. In the early chronic phase, SA amplitude and wind-up both began to show recovery toward preinjury level in semitendinosus only. The initial potentiation of CAP amplitude seen in semitendinosus and tibialis anterior following acute transection subsided and returned to preinjury levels.

DISCUSSION

In summary, this work reveals three findings: 1) SA reflex amplitudes remain low in the early chronic spinalized state, similar to the acutely spinalized state in all muscles except the hip extensor/knee flexor ST, which shows a strong tendency to revert to preinjury amplitudes; 2) wind-up of the ST SA responses also shows reversion to preinjury levels, whereas wind-up in the other muscles remains abolished; and 3) the facilitation of CAP amplitudes following acute spinal transection is short lived, because their amplitudes revert back to preinjury magnitudes or smaller in most muscles, with the exception of the hip flexor Srt, whose CAP amplitudes became even larger in the early chronic stage than they were in the already facilitated acute state.

The descending monoaminergic tracts of the locus coeruleus and caudal raphe brain stem nuclei innervate all segments and lamina of the spinal cord (Björklund 1982). The activity of these nuclei correlate with state of arousal, showing increased firing as the animal transitions from sleep to quiet rest to alertness and locomotion (Jacobs et al. 2002). Dialysis of the spinal ventral horn shows that there is a significant increase in 5-HT concentration following exercise (Gerin et al. 1995), and complete transection of the spinal cord, including these monoaminergic tracts, results in a 75% decrease of 5-HT (Shibuya and Anderson 1968). Immediately following complete spinal injury, spinal motoneuron (MN) intrinsic excitability is greatly reduced, primarily due to loss of descending monoaminergic (MA) inputs from the brain stem. This is shown in voltage-clamp recordings where MN excitatory postsynaptic currents (EPSCs) in acutely spinalized animals are roughly half the amplitude of EPSCs in the normal uninjured state (Hyngstrom et al. 2008), as well as in reflex activations as shown in our previous article (Frigon et al. 2012a). Complete spinalization and loss of descending neuromodulation not only decreases MN excitability but also rearranges MN output patterns (Hyngstrom et al. 2008). We have shown that ankle extensor MNs, which would normally only be responsive to rotations of the ankle, exhibit depolarizing EPSCs in response to passive hip rotations in spinalized cats (Hyngstrom et al. 2008). The normally focused and joint-specific Ia muscle spindle reflex pathways and reciprocal inhibitory pathways, which typically operate in agonist-antagonist pairs and with close synergists, lose joint specificity in spinalized preparations where descending MA projections are abolished. In the normal intact spinal cord, rotations of the ankle joint strongly activate currents in ankle extensor MNs and produce only weak EPSCs in hip MNs (Hyngstrom et al. 2008). Complete spinal transection, which abolishes the descending MA inputs, causes a widening of MN receptive fields to muscle stretch inputs, and hence a loss of the focused nature of the Ia system. Even though excitatory synaptic currents are overall smaller in this reduced MA state, the relative level of activation of hip MNs compared with ankle MNs in response to ankle rotation favors the hip cells, a pattern completely opposite that seen in the uninjured state. This receptive field widening could confound proper MN activation by spared descending commands in spinal injured animals and humans, as well as interfering with central pattern generation in spinalized animals.

Although loss of descending MA drive greatly reduces spinal excitability acutely, there is strong evidence that MN intrinsic excitability returns during the weeks following injury (Bennett et al. 2001; Eken et al. 1989; Murray et al. 2010). However, proper sensory input processing does not parallel this return in excitability; in fact, MN receptive fields remain in a widened disorganized state (Johnson et al. 2013). With the reemergence of MN excitability and a continuation of altered MN sensory input processing, spinal motor output patterns become severely altered. This loss of input specificity at the cellular level is manifest at the motor output stage in the form of inappropriate muscle activations (Johnson et al. 2013). Receptive field widening may also be the basis of aberrant muscle activations (spasticity) in humans and animals with SCI.

In the current study and in our previous work (Frigon et al. 2012a), we demonstrate that cutaneous reflex pathways show altered activation patterns immediately after and in the recovery process 1 mo following spinal transection. The reflex activations evoked likely involve mostly polysynaptic pathways. The role neuromodulation plays in synaptic processing by proprioceptive interneurons in the spinal cord is less understood, and neuromodulatory effects are complicated by the existence of multiple receptor subtypes and the ratios of these receptors on different classes of interneurons. Studies show that MA drive has varied effects on interneurons involved in different reflex pathways (Jankowska et al. 2000; Maxwell et al. 2000). However, it is clear from both single-cell recordings and reflex studies that immediately following complete spinal injury and loss of descending neuromodulation, sensory integration in the spinal cord is altered. The net effect of this alteration appears to bias inhibition of interneurons and activation of MNs driving more proximal muscles. As intrinsic excitability returns in the early chronic and chronic stages of SCI, these synaptic events, even those that were subthreshold initially in the acute spinal state, have the potential to initiate MN firing and muscle activation.

The potentiation of CAPs immediately following spinal transection did not persist into the early chronic stage (with the exception of the sartorius muscle). The initial short-lived hyperexcitability of CAPs, likely the activation of multiple excitatory INs which then activate MNs, could arise from several sources as multiple descending systems influence the behaviors of these cells (Eccles and Lundberg 1958; Hultborn 2001; Lundberg 1964), primary among them those that modulate presynaptic inhibition (Enríquez et al. 1996; Rudomin et al. 1986) (for further discussion, see Frigon et al. 2012a). Additionally, further studies to differentiate afferent type are required to identify the specific inputs activated by peripheral nerve stimulation. Potentiation of some peripheral sensory input modalities may be an adaptive response to the initial shock of spinal injury that then subside as the recovery process progresses.

The change in CAP amplitude from the intact state through the acute spinal to the early chronic spinal state demonstrates considerable plasticity in reflex modulation following SCI. These findings not only suggest that CAP responses arise from different sources than the SA reflex responses but also demonstrate that both CAP and SA measures have the potential to be used as a diagnostic and research tool to evaluate the mechanisms and time course of changes in multiple spinal systems in SCI.

GRANTS

This study was supported by Wings for Life and by National Institute of Neurological Disorders and Stroke Grant R01 NS089313.

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.D.J., A.F., and C.C. performed experiments; M.D.J., A.F., M.-F.H., and C.J.H. interpreted results of experiments; M.D.J. prepared figures; M.D.J. drafted manuscript; M.D.J. approved final version of manuscript; A.F. conceived and designed research; A.F., M.-F.H., C.C., and C.J.H. edited and revised manuscript; M.-F.H. analyzed data.