- Email: [email protected]
Spinal cord injury in neonates alters respiratory motor output via supraspinal mechanisms
Experimental Neurology 206 (2007) 137 – 145 www.elsevier.com/locate/yexnr
Spinal cord injury in neonates alters respiratory motor output via supraspinal mechanisms M. Beth Zimmer ⁎, Harry G. Goshgarian Wayne State University, Department of Anatomy and Cell Biology, 540 East Canfield, Detroit, MI 48201, USA Received 28 February 2007; revised 27 April 2007; accepted 2 May 2007 Available online 8 May 2007
Abstract Upper cervical spinal cord injury (SCI) alters respiratory output and results in a blunted respiratory response to pH/CO2. Many SCI studies have concentrated on respiratory changes in neural function caudal to injury; however few have examined whether neural plasticity occurs rostral to SCI. Golder et al. (2001a) showed that supraspinal changes occur to alter respiratory output after SCI. Furthermore, Brown et al. (2004) showed that neural receptors change rostral to a thoracic SCI. We hypothesized that SCI in neonates will alter supraspinal output, show a blunted response to pH and alter receptor protein levels in the medulla. On postnatal day 0/1, a C2 SCI surgery was performed. Two days later, neonates were anesthetized and brainstem-spinal cords removed. Respiratory-related activity was recorded using the in vitro brainstem-spinal cord preparation and the superfusate pH was changed (pH 7.2, 7.4 and 7.8). The respiratory-like frequency was significantly reduced in SCI rats indicating supraspinal plasticity. Increasing the pH decreased respiratory-like frequency and peak amplitude in injured and sham controls. Increasing the pH increased burst duration and area in sham controls, whereas in injured rats, the burst duration and area decreased. Western blot analysis demonstrated significant changes in glutamate receptor subunits (NR1, NR2B and GluR2), adenosine receptors (A1, A2A), glutamic acid decarboxylase (65) and neurokinin-1 receptors in medullary tissue ipsilateral and contralateral to injury. These data show that supraspinal plasticity in the respiratory system occurs after SCI in neonate rats. The mechanisms remain unknown, but may involve alterations in receptor proteins involved in neurotransmission. Published by Elsevier Inc. Keywords: Spinal cord injury; Control of breathing; pH chemosensitivity; Neonate; In vitro
Introduction Injury to the spinal cord interrupts both descending and ascending neural projections which results in paralysis and a loss of sensation caudal to the site of injury. With regard to the respiratory system, many researchers have shown that alterations in neural function occur caudal (Goshgarian, 2003; Fuller et al., 2003, Golder et al., 2003; Nantwi and Goshgarian, 2005) and rostral to the site of injury (Fuller et al., 2006; Golder et al., 2001a,b; Goshgarian et al., 1986). However, few studies have shown that neural plasticity of respiratory function occurs in supraspinal centers after spinal cord injury (Golder et al., 2001a) and the mechanisms underlying supraspinal neural plasticity have not been investigated. High cervical spinal cord hemisection (C2) interrupts the ipsilateral descending premotor ⁎ Corresponding author. Fax: +1 313 577 3125. E-mail address: [email protected] (M.B. Zimmer). 0014-4886/$ - see front matter. Published by Elsevier Inc. doi:10.1016/j.expneurol.2007.05.003
projections, effectively paralyzing the ipsilateral hemidiaphragm, and thus, reduces respiratory volume (Golder et al., 2003). This reduced volume is sensed by lung and chest wall receptors which feedback through vagal mechanisms to increase respiratory frequency; routinely observed in vivo (Fuller et al., 2006; Golder et al., 2001a, 2001b; Goshgarian et al., 1986). Below the site of injury, respiratory neural plasticity occurs within the phrenic motor nucleus (C3–C6 in rats) to allow for the expression of a latent respiratory motor pathway, termed the crossed phrenic pathway (see Goshgarian, 2003 for review). Activation of the latent crossed phrenic pathway results in the restoration of function to the previously paralyzed hemidiaphragm (see Goshgarian, 2003 for review). Thus, many studies have focused on the mechanisms underlying the regulation and activation of the crossed phrenic pathway in order to restore motor function to the paralyzed hemidiaphragm. Rostral to the site of injury within the medulla, however, the effect of spinal cord injury on respiratory neural function is less clear.
138
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
Recently, Golder et al. (2001a) showed that respiratory neural plasticity occurs in supraspinal respiratory centers after spinal cord injury in adult male rats. Using anesthetized, paralyzed, ventilated rats, 2 months after a C2 hemisection, the respiratory frequency was significantly decreased. In addition, examination of hypoglossal motor output, originating rostral to the site of the injury, revealed that the peak amplitude of the hypoglossal roots was significantly decreased (Golder et al., 2001a). Furthermore, rats with a C2 hemisection showed a blunted response to hypercapnia; the increase in amplitude of the hypoglossal burst was blunted in spinal cord injured rats compared to non-injured controls (Golder et al., 2001a). Together, the data indicate a decreased respiratory drive after chronic spinal cord injury due to supraspinal mechanisms. The underlying cause of this altered supraspinal mechanism is not clear, but could involve injuryinduced plasticity in supraspinal brain centers resulting from either the axotomy of respiratory premotor neurons located in the ventral lateral medulla and/or the axotomy of ascending spinobulbar projections that connect to respiratory neurons in the medulla. Therefore, this study was designed to examine the effect of spinal cord injury on brainstem neural output using the brainstem-spinal cord preparation of the neonate, such that no other inputs other than from within the neuraxis itself are present. In adult, whole animal models, attempts are made to eliminate and/or reduce the feedback from higher brain centers, peripheral chemoreceptors and muscle and lung receptors; however, it is difficult to eliminate all feedback, and the effect of anesthetic on neural function may obscure the results. This is not a problem in the present study. After spinal cord injury, specific receptor subunits involved in neurotransmission, such as AMPA receptor subunits (GluR) and NMDA receptor subunits (NR), change within the spinal cord both rostral and caudal to the site of injury (Park et al., 2003; Grossman et al., 2000; Brown et al., 2004; Grossman et al., 1999; Grossman et al., 2001; Fuller et al., 2005). Many studies have shown that decreases in specific receptors or receptor subunits occur caudal to the site of injury; however, Brown et al., 2004 found that using Western blot analysis, GluR1, GluR2, GluR4, NR2A and NR2B were all significantly decreased in spinal cord tissue both rostral and caudal to an acute T8 contusion injury in young rats. We predict that a high cervical spinal cord injury (around C2) in neonate rats will alter receptors rostral to the site of injury within the medulla, which could have a significant impact on the descending control mechanisms of many autonomic functions, including respiration. In the present study, we examined whether supraspinal neural plasticity occurs within the respiratory system as a result of an acute, high cervical spinal cord injury in neonate rats. After an acute spinal cord injury (2 days), the in vitro brainstem-spinal cord preparation was used to determine whether changes in medullary neural output occurred as a result of the injury. We hypothesized that supraspinal alterations in respiratory control mechanisms would manifest in neonates after an acute injury similar to the results of Golder et al. (2001a) in adult rats, since the neonate is still developing and highly plastic. The respiratory response to changes in pH was examined to determine if medullary chemosensing mechanisms are also altered as a result
of the spinal cord injury. Many studies have shown a reduced sensitivity to hypercapnia after spinal cord injury in both animal models (Teng et al., 2003) as well as human studies (Kelling et al., 1985; Manning et al., 1992); however, the underlying mechanism is not known. We hypothesized that modifications in supraspinal brain centers may be a part of the underlying mechanism behind this reduced sensitivity to hypercapnia. And finally, Western blot analysis was performed on medullary tissue both ipsilateral and contralateral to the injury to examine whether changes in receptors involved in neurotransmission are altered in response to an acute spinal cord injury. Like Brown et al. (2004), we hypothesized that many receptor subunits rostral to the injury, in this case, within the brainstem, would be altered by an upper cervical spinal cord injury. Methods Surgery All surgeries and protocols were reviewed and approved by the Animal Investigation Committee (AIC) at Wayne State University acting under the guidelines set up by the AAALAC prior to experimentation. Neonate rats (first day of birth or postnatal day (1) were removed, individually, from their mother and deeply anesthetized with cold. An incision was made on the dorsal aspect of the neck and the muscles were cut to expose the C1 vertebra. A laminectomy at the C1 vertebra was made followed by a lesion in the left C2 spinal cord using microscissors (n = 10). Attempts were made to completely hemisect the spinal cord, but all pups were used in this study, despite the lack of full hemisection in all pups. The skin was closed with tissue glue and the pups warmed on a heating pad until awake. They were immediately placed in the cage with the other pups and mother. At this time, another pup was removed from the litter for surgery. Sham surgeries (n = 9) were performed in the same manner as described above, except that the spinal cord was not cut. Hemisection and sham surgeries were alternated. The mothers received all pups and immediately fed them. Physiology experiments After 2 days (P3 or P4), the pups were removed and anesthetized deeply with halothane. The brainstem-spinal cord (pons removed; medulla-T5/6) was carefully dissected and the dura removed. The brainstem-spinal cord was placed in a recording chamber supplied with artificial cerebral spinal fluid (aCSF composed of 113.0 mM sodium chloride, 3.0 mM potassium chloride, 1.2 mM sodium phosphate, 1.5 mM calcium chloride, 1.0 mM magnesium chloride, 30.0 mM sodium bicarbonate and 30.0 mM dextrose) that was bubbled with 95% oxygen and 5% carbon dioxide. Suction electrodes were placed on both the right and left C4 ventral roots to record bilateral respiratory-related activity. Electrical signals were amplified and filtered (300–3000 Hz, gain 20,000) using Grass amplifiers (Model P511). Baseline recordings were taken for approximately 20 min at a pH of 7.4 and then the pH of the
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
superfusate was changed by altering the sodium bicarbonate and carbon dioxide in the solutions, first to a high pH (∼7.8) for another 20 min followed by low pH (∼ 7.2) for another 20 min. The data were recorded continuously on computer using Spike2 data acquisition software (CED, Cambridge UK) (sample rate 3000 Hz). Bilateral respiratory-like recordings were full-wave rectified and integrated using CED data analysis software (Cambridge UK). The peak amplitude, area and the duration of the respiratory-like bursts were analyzed for the last 5 min of each pH exposure and comparisons made between spinal cord injured and sham control rat pups. Western blot analysis Immediately after the physiological recordings were taken, the tissue was dissected and the right and left side of the medulla were frozen at − 80 °C until processing. Total protein was extracted from the medulla by sonication in a modified Glasgow protein extraction buffer (150 mM NaCl, 1.0 mM EDTA, 50 mM Tris, 1% NP40, 0.1% Triton X-100 and protease inhibitors) and protein concentrations were determined using a Bradford assay (Sigma, St. Louis, MO). Medullary protein homogenates (50 ìg) containing Laemmli sample buffer were loaded onto 7.5% Tris–HCl gels (BioRad, Hercules, CA) and separated. Each gel contained protein homogenate from control rats and spinal cord injured rats from one side of the medulla along with a Precision Plus Protein™ Standard (10–250 kDa, BioRad, Hercules, CA). Following separation, the proteins were transferred to a polyvinylidine difluoride (PVD) membrane (BioRad, Hercules, CA) overnight at 4 °C. The membranes were washed 5× with a Tris–Saline–-Tween buffer (TSTB), followed by non-specific blocking with 5% milk. The membranes were then incubated overnight at 4 °C in primary antibody (see below). The next morning, the membranes were washed (5 × 5 min) and incubated in peroxidase-conjugated secondary antibody diluted in a 1% TSTB solution for 1–2 h at room temperature. The membranes were washed again and incubated in an enhanced chemiluminescent peroxidase substrate (Chemicon, Temecula, CA) for 3 min. The membranes were apposed to film (30 s–10 min) and developed. The films were scanned with Image Scanner II (Amersham Biosciences, Piscataway, NJ) and the bands corresponding to the various proteins were quantified using the ImageQuant TL program (Amersham Biosciences, Piscataway, NJ). Protein bands were measured for total density. Spinal cord-injured values were compared to controls for quantitative and statistical analysis. â-actin controls (1:5000, Sigma-Aldrich, St. Louis, MO) were used as a loading control. Antibody solutions Western blots were probed using the following antibodies. AMPA receptor subunit protein expression levels were determined using rabbit anti-GluR1 (1:1000), mouse antiGluR2 (1:1000), mouse anti-GluR3 (1:1000) and rabbit antiGluR4 (1:1000). NMDA receptor subunit protein expression
139
levels were determined using mouse anti-NR1 (1:1000), rabbit anti-phospho-NR1 (Ser897) (1:000) (Upstate Cell Signaling Solutions, Lake Placid, NY), mouse anti-NR2A (1:500), rabbit anti-NR2B and rabbit anti-NR2B phosphoTyr1472 (1:1000). Adenosine receptor protein levels were determined with rabbit anti-A1 (1:1000) and rabbit anti-A2A (1:2500; Sigma). Mouse anti-glutamic acid decarboxylase (65) (1:1000), mouse antiGAD67 (1:5000), rabbit anti-glycine receptor (1:1000) and rabbit anti-substance P receptor (NK-1) (1:1000) were also used to determine protein level differences between non-injured and spinal cord-injured rats. All antibodies, except those already listed above, were purchased from Chemicon International (Temecula, CA). All secondary antibodies were purchased from Sigma-Aldrich (St. Louis, MO); peroxidase-conjugated goat anti-rabbit (1:10,000) and peroxidase-conjugated goat antimouse (1:2500). Statistics Baseline respiratory variables (frequency, burst peak amplitude, burst area and burst duration) of sham-operated control and injured rats were analyzed for 5 continuous minutes and the average value obtained for each rat. The differences between sham controls and spinal cord injured rats were examined at ‘normal’ pH 7.4 using an ANOVA. The responses of each respiratory variable to changes in pH (7.8 and 7.2) were assessed using a repeated measures ANOVA. The results obtained with Western blot analysis were analyzed using a t-test. p b 0.5 was considered significant. All results are expressed as means ± standard error (SE). Results Upper cervical spinal cord-injured neonates and neonates with a sham surgery (non-injured, control rats) gained a significant amount of weight over two days post surgery (Table 1) ( p = 0.001). There were no significant difference in weight between non-injured and spinal cord-injured rats ( p = 0.15). Baseline recordings of the right and left C4 ventral roots at normal pH (7.4) revealed that the respiratory-like frequency of the spinal cord-injured rats was significantly reduced ( p = 0.04) compared to the sham-operated controls (Fig. 1). In addition, there was a significant decrease in the peak amplitude of the burst ( p = 0.04), the burst area ( p = 0.001) and the burst duration ( p = 0.024) on the side ipsilateral to injury, but not on the contralateral, or uninjured side (Fig. 1).
Table 1 Body weights (grams) of neonate rats on the day of surgery, sham or incomplete spinal cord injured (P0–P1), and 2 days later on the day of physiology experiments and tissue collection for Western blot analysis (P3/4)
Control Spinal cord injured
Body weight (P0/P1)
Body weight (P3/4)
7.3 ± 0.3 7.2 ± 0.2
11.1 ± 0.7* 9.9 ± 0.5*
Both groups of rats significantly gained weight (*), but there was no difference between control and spinal cord injured rats.
140
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
Fig. 1. Upper cervical spinal cord injury (SCI) in neonate rats caused a significant decrease in the resting frequency of the respiratory-like bursts of the brainstem-spinal cord preparation (bars ± SE). Compared to control sham operated rats, spinal cord injury caused a significant reduction in the peak amplitude, the burst area and the burst duration of the C4 ventral root ipsilateral to the injury compared to contralateral to the injury during normal pH (7.4).
Changing the pH of the superfusate of the brainstem-spinal cord preparation resulted in significant effects on all respiratory variables examined (Fig. 2). The respiratory-like frequency of spinal cord-injured preparations was altered in response to pH ( p = 1.1 × 10− 4), but the response to pH was not different between injured and sham control preparations, i.e., the frequency decreased in response to high pH and increased in response to low pH in both non-injured and injured rats (Fig. 2A). The change in burst duration in response to pH was significantly different in absolute raw values, i.e., seconds (right nerve, p = 0.016; left nerve, p = 0.024, data not shown), but not in relative terms, i.e., percent difference between non-injured, sham controls and injured rats (Fig. 2B). There was, however, a significant interaction between changing the pH and spinal cord
injury when examining the relative changes in burst duration ( p = 0.026). This significant interaction between pH and injury was also evident in the total burst area ( p = 0.011) (Fig. 2D). This is most likely due to the opposite effect that pH had on burst duration and area; in non-injured, sham control brainstemspinal cord preparations, raising the pH of the superfusate (∼ 7.8) increased the burst duration and the area, whereas in the injured rats, raising the pH caused a decrease in burst duration and area (Figs. 2B, D). When pH was lowered to 7.2 in noninjured, sham controls, the burst duration did not change, but in spinal cord injured preparations, the burst duration and area increased. The relative change in peak amplitude in response to changing the pH was also significant ( p = 0.001), however, there were no differences between spinal cord injured and sham-
Fig. 2. Changing the pH of the superfusate from the control level of 7.4 significantly altered the frequency (A) and the peak amplitude (C) of the respiratory-like discharge of both non-injured and spinal cord injured neonate rats. There was a significant interaction between spinal cord injury and the response to pH when examining the burst duration (B) and area (D); the duration and area tended to change in an opposite manner in injured rats compared to non-injured rats. Spinal cord injury was made on the left side of the spinal cord.
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
operated control rats (Fig. 2C), i.e., peak amplitude increased in response to high pH and there was a slight decrease in response to low pH in both groups. Western blot analysis revealed that several key receptor subunits involved in neurotransmission were altered in the medulla on both the side contralateral and ipsilateral to the upper cervical spinal cord injury. Of the glutamate receptor subunits, the protein levels of the NR1 subunit of the NMDA receptor were significantly decreased rostral and ipsilateral to the injury (Fig. 3A) ( p = 0.027). However, the phosphorylated NR1 (pNR1) protein and the ratio of pNR1 relative to the total NR1 were not significantly changed. Contralateral to the injury, the NR1 protein levels were not different between non-injured, sham controls and spinal cord injured rats ( p = 0.08); however, there was a significant decrease in the amount of pNR1 protein ( p = 0.002). The ratio of pNR1 to total NR1 was also significantly reduced ( p = 0.03) indicating that although NR1 protein levels may not have significantly decreased after injury, there is a decrease in functional/phosphorylated NR1 protein.
141
The NR2B subunit protein levels were significantly increased ipsilateral to the spinal cord injury ( p = 0.03); however, there was no change in the amount of phosphorylated NR2B (pNR2B) protein levels (Fig. 3A). The ratio of pNR2B to total NR2B protein was significantly decreased ( p = 0.003) on the side ipsilateral to the injury, possibly indicating that although the number of NR2B subunits increased in quantity they were not being phosphorylated. On the side contralateral to the injury, NR2B protein levels did not change significantly; however, the quantity of pNR2B protein significantly decreased ( p = 0.02). This resulted in a significant decrease in the ratio of pNR2B to total NR2B protein levels on the right side (contralateral to the injury) of the medulla of spinal cord injured rats ( p = 0.007). The NR2A protein expression did not change rostral to the injury on either side of the medulla in response to spinal cord injury (Fig. 3B). Of the AMPA receptor subunits, there was a bilateral decrease in the protein levels of the GluR2 subunit in the medulla of spinal cord injured neonate rats (Fig. 3C) (right p = 0.02; left p = 0.03). No other subunits of
Fig. 3. Western blot analysis revealed that specific protein levels involved in glutamate receptor transmission are significantly altered in the medulla after C2 spinal cord injury in neonatal rats. The left side of the medulla is ipsilateral to the injury, whereas the right side is contralateral to the injury. The bar graphs show the amount of protein in medullary tissue both contralateral (right) and ipsilateral (left) to SCI (white bars) compared to non-injured (black bars) rats for the various receptor subunits of the NMDA receptor (NR1, 2A, 2B) and the AMPA receptor (GluR1, 2, 3, 4). Significant difference between non-injured and injured rats is marked with an asterisk (*).
142
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
Fig. 4. Western blot analysis revealed that other receptors and proteins involved in neurotransmission were altered rostral to spinal cord injury (left side) (white bars) in the medulla in neonates. Glutamic acid decarboxylase (GAD), the rate limiting enzyme in the production of GABA was slightly elevated ipsilateral to the injury (left side). Adenosine receptors (A1, A2A) and the neurokinin 1 (NK-1) receptor protein levels were also altered in response to acute spinal cord injury. Significant difference between control (black bars) and injured rats (white bars) is marked with an asterisk (*).The left side of the medulla is ipsilateral to the injury, while the right side is contralateral to the injury.
the AMPA receptor that were examined (GluR1, GluR3, GluR4) were altered in response to upper cervical spinal cord injury. Several other proteins that could potentially be involved in respiratory control were examined. Glutamic acid decarboxyl-
ase (GAD) is a key enzyme involved in the production of GABA, a major inhibitory neurotransmitter. One of the two isoforms, GAD65, was significantly elevated on the left side of the medulla ( p = 0.036) (Fig. 4), possibly indicating an increase in the production of GABA after spinal cord injury. Adenosine receptors, A1 and A2A, were also examined. After spinal cord injury, A1 protein levels were significantly elevated ( p = 0.016) and the A2A protein levels were significantly decreased ( p = 0.001) on the side contralateral to the injury (Fig. 4). Finally, the tachykinin receptor, NK-1, protein levels were examined, since they have been localized to neurons involved in respiratory rhythmogenesis. After spinal cord injury, levels of NK-1 protein levels were significantly decreased ipsilateral to the injury ( p = 0.02) and significantly increased contralateral to the injury (Fig. 4) ( p = 0.04). In summary, Western blot analysis revealed that there was a tendency for a downregulation of receptor subunits involved in excitatory neurotransmission and an upregulation of receptor subunits involved in inhibitory neurotransmission rostral to the injury in the medulla of upper spinal cord injured neonate rats (Fig. 5). Discussion
Fig. 5. A summary of the receptor protein levels that were significantly altered by a left C2 spinal cord injury as measured by Western blot analysis. The general trend was a decrease in protein levels that are associated with excitatory neurotransmission and an increase in protein levels that are associated with inhibitory neurotransmission.
The results from this study show that upper cervical spinal cord injury can induce neural plasticity and modulate supraspinal respiratory centers in as little as 2 days in the neonate. The data from this study are consistent with the results of Golder et al. (2001a) showing a decrease in central respiratory drive after chronic upper cervical spinal cord injury in adult rats. A decrease in central respiratory drive could be detrimental in a condition such as spinal cord injury where
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
respiratory motor function is already compromised. The underlying mechanism behind the decreased respiratory neural output is not known, but Western blot analysis demonstrated that many receptor subunits and the phosphorylation state of some subunits change in the medulla after injury. The general trend was a reduction in protein subunits that are involved in excitatory neurotransmission and an increase in those proteins involved in inhibitory neurotransmission, which is consistent with the findings of a reduced respiratory drive after injury. The significance or specific location of the receptor changes is not yet known, but knowing that medullary neural control is altered after upper cervical spinal cord injury could significantly alter our view of how spinal cord injury affects respiratory control, as well as other autonomic systems. Respiratory motor output after acute upper cervical spinal cord injury Upper cervical spinal cord injury (in the C2 region) caused a significant reduction in the respiratory frequency in neonates using an in vitro brainstem-spinal cord preparation to assess motor output. This is consistent with the observations by Golder et al. (2001a) who also found that the respiratory frequency was decreased after chronic spinal cord injury in anesthetized, paralyzed and ventilated rats. The underlying mechanism behind the reduced respiratory frequency is not known. Recall that in vivo, respiratory frequency is actually increased, not decreased after spinal cord injury. This is due to changes in receptor feedback from the paralysis of the hemidiaphragm and the subsequent decrease in tidal volume (Golder et al., 2001b). In adults, over time, the altered mechanoreceptor inputs may result in neural plasticity of respiratory centers such that when these inputs are removed, the resultant frequency is decreased compared to non-injured controls. In neonates, this may occur quickly due to the developing state of the respiratory system. Alternatively, the axotomy of ascending sensory or other modulating inputs or even specific alterations in neurotransmission from either direct or indirect mechanisms, including changes in receptor numbers, receptor composition, receptor function, as well as neurotransmitters after injury (see below, Western blot analysis) may be involved in the reduction of respiratory frequency that is observed using reduced preparations. Ipsilateral to the injury, the absolute phrenic motor output was also significantly decreased, but not absent. This is most likely due to the transection of descending premotor axons which project to the ipsilateral phrenic motor neurons through the lateral and ventral funiculi. In the present study, the primary objective was to lesion the cord on the left side in order to alter respiratory drive, not to make a complete hemisection. Thus, not all rats that were used in this study had a complete hemisection, as determined by visualization through a dissecting scope. Nonetheless, even in those rats with a complete hemisection, crossed phrenic activity is spontaneously present in neonates; this has been demonstrated both in vitro (Zimmer and Goshgarian, 2005), as well as in vivo (Huang, Y. and Goshgarian, H.G. unpublished data). Therefore, the amount of activity observed on the left C4 ventral roots was not due solely
143
to crossed phrenic pathways, although crossed phrenic pathways probably contributed to motor output in some cases. Respiratory response to pH Changes in pH/CO2 are sensed by both peripheral and central chemoreceptors. Central chemoreception has been localized to multiple sites within the brain which contain neurons that are responsive to changes in pH/CO2: the retrotrapezoid nucleus (RTN) (Mulkey et al., 2004), the nucleus of the solitary tract (Putnam, 2001), the cerebellum (Xu and Frazier, 2002), the locus coeruleus (Ballantyne and Scheid, 2001) and the raphe nucleus (Richerson, 2004). The in vitro brainstem-spinal cord preparation is a model system that others have used to examine the cellular and molecular mechanisms underlying the sensitivity to pH in brainstem regions (Ballantyne and Scheid, 2001; Putnam, 2001). Our data from noninjured rats are consistent with results from other labs and show that superfusion of aCSF with a high pH results in a decrease in respiratory frequency and a low pH causes an increase in frequency (Dubayle and Viala, 1998; Kawai et al., 1996; Oyamada et al., 1998). Spinal cord injury has been shown to alter the sensitivity to hypercapnia in vivo and many studies report a reduced or blunted response to hypercapnia (Kelling et al., 1985; Manning et al., 1992; Teng et al., 2003). The underlying mechanism behind this reduced sensitivity to CO2 is not known. The results presented in this study do not show a clear reduction in pH sensitivity, but there was a significant interaction between spinal cord injury and the sensitivity to pH when analyzing burst duration and burst area; that is, the injured rats responded in an opposite manner compared to the non-injured rats. Specifically, in non-injured rats, when pH was increased there tended to be a slight increase in burst duration and a decrease or no change when exposed to low pH bilaterally. The opposite was observed with spinal cord injured rats; a slight decrease in burst duration when exposed to high pH and an increase in burst duration during low pH. It is not clear what significance this has in reference to the whole animal, but it does indicate that upper cervical spinal cord injury alters supraspinal respiratory neural control causing a change in the respiratory response to pH. Therefore, it is possible that some of the reduced sensitivity to hypercapnia that has been observed after spinal cord injury may be due, in part, to alterations in supraspinal respiratory neural control. Western blot analysis Significant changes in receptor subunits involved in both excitatory and inhibitory neurotransmission were demonstrated rostral to the site of spinal cord injury using Western blot analysis, confirming the results of Brown et al., 2004. More importantly, upper cervical spinal cord injury induced changes in receptor subunits within the medulla. Many receptor subunits involved in excitatory neurotransmission, NR1, NR2B, GluR2, A2, were decreased after injury, while those involved in inhibition, GAD65, A1, were elevated. From these data it is hard to evaluate whether any specific functional or physiological
144
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
consequences result from these changes, but the changes are consistent with an overall decrease in respiratory drive observed after injury. The exact mechanisms behind a change in receptor protein levels are not known. Increased cell death may occur after spinal cord injury due to the axotomy of neurons. And glutamate-receptor mediated excitotoxicity has also been associated with the spinal cord injury site, whereby, glutamate spills out from the injury site and damages nearby neurons causing secondary cell death (Bittigau and Ikonomidou, 1997; Liu et al., 1999). This could result in a decrease in specific receptor protein concentrations relative to the total protein concentration; however, this does not seem likely since some receptor subtypes increased. It seems more likely that the changes in receptor quantity were due to modifications in various inputs that resulted in specific changes in receptor protein levels. This conclusion was also reached by other studies that showed that specific receptor changes occurred in specific cell types within specific regions of the spinal cord, gray or white matter after injury (Park et al., 2003; Grossman et al., 2000; Brown et al., 2004; Grossman et al., 1999; Grossman et al., 2001; Fuller et al., 2005). Not only were total protein levels changed after spinal cord injury, but also the amount of phosphorylated protein was changed. It appears that the phosphorylated state of NMDA channels decreased after injury; this was the case for both the NR1 and the NR2B subunit. Phosphorylation of the NR1 subunit has been associated with increasing the functional state of the receptor (Yun and Zhang, 2000), whereas, phosphorylation of the NR2B subunit has been associated with instability and internalization of receptors from the membrane. Therefore, spinal cord injury not only altered the total quantity of protein, but may also alter the functional state of the receptors. Neurokinin-1 receptors have been localized to a specific subset of neurons within the pre-Bötzinger complex, a site of respiratory rhythmogenesis within the ventral lateral medulla. Removal or destruction of these neurons results in a very unstable and non-rhythmic respiratory pattern, strongly indicating their role in generating the respiratory rhythm (Gray et al., 1999, 2001). Interestingly, neurons in the ventral medulla that express NK-1 receptors are also necessary for central and peripheral chemoreception (Nattie and Li, 2006). NK-1 receptor protein levels were altered in response to upper cervical spinal cord injury; the quantity of NK-1 protein increased contralateral to injury and decreased ipsilateral to injury. Although Western blot analysis does not give an indication of the specific location of protein changes, if changes in the relative amount of NK-1 receptor protein occur within the pre-Bötzinger complex and other regions involved in chemoreception, the respiratory frequency might be altered after spinal cord injury. Current studies in our lab are examining whether an alteration in preBötzinger NK1-protein expression directly affects respiratory rhythm transmission after upper cervical spinal cord injury. The mechanism(s) behind the changes in neural receptor composition and quantity after spinal cord injury are not known. The data here are consistent with the possibility that some of these changes in neurotransmission may be directly responsible
for the effects that are observed on the respiratory motor output after spinal cord injury. The in vitro brainstem spinal cord preparation The in vitro brainstem-spinal cord preparation is a useful model to examine respiratory neural control mechanisms and has been used for many years by respiratory neurobiologists. It is has both advantages and disadvantages. One problem is the time limitation; respiratory rhythm can only be examined within the first 4–6 days with many studies using postnatal day 4. However, in this particular case, we were able to examine the effect of spinal cord injury on brainstem neural output and the response to changes in pH/CO2 with no other inputs or feedback other than within the neuraxis itself to interfere with the motor output that is generated by brainstem mechanisms alone. In adult, whole animal models, attempts are made to eliminate, reduce, and/or equalize the feedback from higher brain centers, peripheral chemoreceptors, and muscle and lung receptors; however, it is difficult to eliminate all feedback. In addition, the effect of anesthetic on neural function may obscure the results. None of the above issues is a concern when using the in vitro brainstem-spinal cord preparation. Acknowledgment The work in this paper was supported by NIH grant HD31550 (Dr. H.G. Goshgarian). References Ballantyne, D., Scheid, P., 2001. Central chemosensitivity of respiration: a brief overview. Respir. Physiol. 129, 5–12. Bittigau, P., Ikonomidou, C., 1997. Glutamate in neurologic diseases. J. Child Neurol. 12 (8), 471–485. Brown, K.M., Wrathall, J.R., Yasuda, R.P., Wolfe, B.B., 2004. Glutamate receptor subunit expression after spinal cord injury in young rats. Dev. Brain Res. 152, 61–68. Dubayle, D., Viala, D., 1998. Effects of CO2 and pH on the spinal respiratory rhythm generator in vitro. Brain Res. Bull. 45 (1), 83–87. Fuller, D.D., Johnson, S.M., Olson Jr., E.B., Mitchell, G.S., 2003. Synaptic pathways to phrenic motorneurons are enhanced by chronic intermittent hypoxia after cervical spinal cord injury. J. Neurosci. 23 (7), 2993–3000. Fuller, D.D., Baker-Herman, T.L., Golder, F.J., Doperalski, N.J., Watters, J.J., Mitchell, G.S., 2005. Cervical spinal cord injury upregulates ventral spinal 5-HT2A receptors. J. Neurotrauma 22 (2), 203–213. Fuller, D.D., Golder, F.J., Olson Jr., E.B., Mitchell, G.S., 2006. Recovery of phrenic activity and ventilation after cervical spinal hemisection in rats. J. Appl. Physiol. 100, 800–806. Golder, F.J., Reier, P.J., Bolser, D.C., 2001a. Altered respiratory motor drive after spinal cord injury: supraspinal and bilateral effects of a unilateral lesion. J. Neurosci. 21 (21), 8680–8689. Golder, F.J., Reier, P.J., Davenport, P.W., Bolser, D.C., 2001b. Cervical spinal cord injury alters the pattern of breathing in anesthetized rats. J. Appl. Physiol. 91 (6), 2451–2458. Golder, F.J., Fuller, D.D., Davenport, P.W., Johnson, R.D., Reier, P.J., Bolser, D.C., 2003. Respiratory motor recovery after unilateral spinal cord injury: eliminating crossed phrenic activity decreases tidal volume and increases contralateral respiratory motor output, vol. 23(6), pp. 2494–2501. Goshgarian, H.G., 2003. The crossed phrenic phenomenon: a model for plasticity in the respiratory pathways following spinal cord injury. J. Appl. Physiol. 94 (2), 795–810.
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145 Goshgarian, H.G., Moran, M.F., Prcevski, P., 1986. Effect of cervical spinal cord hemisection and hemidiaphragmparalysis on artherial blood gases, pH and respiratory rate in the adult rat. Exp. Neurol. 93 (2), 440–445. Gray, P.A., Rekling, J.C., Bocchiaro, C.M., Feldman, J.L., 1999. Modulation of respiratory frequency by peptidergic input to rhythmogenic neurons in the preBotzinger complex. Science 286 (5444), 1566–1568. Gray, P.A., Janszewski, W.A., Mellen, N., McCrimmon, D.R., Feldman, J.L., 2001. Normal breathing requires preBotzinger complex neurokinin-1 receptor-expressing neurons. Nat. Neurosci. 4 (9), 927–930. Grossman, S.D., Wolfe, B.B., Yasuda, R.P., Wrathall, J.R., 1999. Alterations in AMPA receptor subunit expression after experimental spinal cord contusion injury. J. Neurosci. 19 (14), 5711–5720. Grossman, S.D., Wolfe, B.B., Yasuda, R.P., Wrathall, J.R., 2000. Changes in NMDA receptor subunit expression in response to contusive spinal cord injury. J. Neurochem. 75, 174–185. Grossman, S.D., Rosenberg, L.J., Wrathall, J.R., 2001. Relationship of altered glutamate receptor subunit mRNA expression to acute cell loss after spinal cord contusion. Exp. Neurol. 168, 283–289. Kawai, A., Ballantyne, D., Mückenhoff, K., Scheid, P., 1996. Chemosensitive medullary neurones in the brainstem-spinal cord preparation of the neonatal rat. J. Physiol. 492.1, 277–292. Kelling, J.S., DiMarco, A.F., Gottfried, S.B., Altose, M.D., 1985. Respiratory responses to ventilatory loading following low cervical spinal cord injury. J. Appl. Physiol. 59, 1752–1756. Liu, D., Xu, G.Y., Pan, E., McAdoo, D.J., 1999. Neurotoxicity of glutamate at the concentration released upon spinal cord injury. Neurosci. 93 (4), 1383–1389. Manning, H.L., Brown, R., Scharf, S.M., Leith, D.E., Weiss, J.W., Weinberger, S.E., Schwartzstein, R.M., 1992. Ventilatory and P0.1 response to hypercapnia in quadriplegia. Respir. Physiol. 89, 97–112.
145
Mulkey, D.K., Stornetta, R.L., Weston, M.C., Simmons, J.R., Parker, A., Bayliss, D.A., Guyenet, P.G., 2004. Respiratory control by ventral surface chemoreceptor neurons in rats. Nat. Neurosci. 7, 1360–1369. Nantwi, K.D., Goshgarian, H.G., 2005. Adensosinergic mechanisms underlying recovery of diaphragm motor function following upper cervical spinal cord injury: potential therapeutic implications. Neurol. Res. 27 (2), 195–205. Nattie, E., Li, A., 2006. Neurokinin-1 receptor-expressing neurons in the ventral medulla are essential for normal central and peripheral chemoreception in the conscious rat. J. Appl. Physiol. 101, 1596–1606. Oyamada, Y., Ballantyne, D., Mückenhoff, K., Scheid, P., 1998. Respirationmodulated membrane potential and chemosensitivity of locus coeruleus neurones in the in vitro brainstem-spinal cord of the neonatal rat. J. Physiol. 513.2, 381–398. Park, E., Liu, Y., Fehlings, M.G., 2003. Changes in glial cell white matter AMPA receptor expression after spinal cord injury and relationship to apoptotic cell death. Exp. Neurol. 182, 35–48. Putnam, R.W., 2001. Intracellular pH regulation of neurons in chemosensitive and nonchemosensitive areas of brain slices. Respir. Physiol. 129, 37–56. Richerson, G.B., 2004. Serotonergic neurons as carbon dioxide sensors that maintain pH homeostasis. Nat. Rev., Neurosci. 5 (6), 449–461. Teng, Y.D., Bingman, M., Taveira-DaSilva, A.M., Pace, P.P., Gillis, R.A., Wrathall, J.R., 2003. Serotonin 1A receptor agonists reverse respiratory abnormalities in spinal cord-injured rats. J. Neurosci. 23 (10), 4182–4189. Xu, F., Frazier, D.T., 2002. Role of the cerebellar deep nuclei in respiratory modulation. Cerebellum 1 (1), 35–40. Yun, L., Zhang, J., 2000. Recent development in NMDA receptors. Chin. Med. J. 113, 948–956. Zimmer, M.B., Goshgarian, H.G., 2005. Spontaneous crossed phrenic activity in the neonatal respiratory network. Exp. Neurol. 194 (2), 530–540.
Spinal cord injury in neonates alters respiratory motor output via supraspinal mechanisms M. Beth Zimmer ⁎, Harry G. Goshgarian Wayne State University, Department of Anatomy and Cell Biology, 540 East Canfield, Detroit, MI 48201, USA Received 28 February 2007; revised 27 April 2007; accepted 2 May 2007 Available online 8 May 2007
Abstract Upper cervical spinal cord injury (SCI) alters respiratory output and results in a blunted respiratory response to pH/CO2. Many SCI studies have concentrated on respiratory changes in neural function caudal to injury; however few have examined whether neural plasticity occurs rostral to SCI. Golder et al. (2001a) showed that supraspinal changes occur to alter respiratory output after SCI. Furthermore, Brown et al. (2004) showed that neural receptors change rostral to a thoracic SCI. We hypothesized that SCI in neonates will alter supraspinal output, show a blunted response to pH and alter receptor protein levels in the medulla. On postnatal day 0/1, a C2 SCI surgery was performed. Two days later, neonates were anesthetized and brainstem-spinal cords removed. Respiratory-related activity was recorded using the in vitro brainstem-spinal cord preparation and the superfusate pH was changed (pH 7.2, 7.4 and 7.8). The respiratory-like frequency was significantly reduced in SCI rats indicating supraspinal plasticity. Increasing the pH decreased respiratory-like frequency and peak amplitude in injured and sham controls. Increasing the pH increased burst duration and area in sham controls, whereas in injured rats, the burst duration and area decreased. Western blot analysis demonstrated significant changes in glutamate receptor subunits (NR1, NR2B and GluR2), adenosine receptors (A1, A2A), glutamic acid decarboxylase (65) and neurokinin-1 receptors in medullary tissue ipsilateral and contralateral to injury. These data show that supraspinal plasticity in the respiratory system occurs after SCI in neonate rats. The mechanisms remain unknown, but may involve alterations in receptor proteins involved in neurotransmission. Published by Elsevier Inc. Keywords: Spinal cord injury; Control of breathing; pH chemosensitivity; Neonate; In vitro
Introduction Injury to the spinal cord interrupts both descending and ascending neural projections which results in paralysis and a loss of sensation caudal to the site of injury. With regard to the respiratory system, many researchers have shown that alterations in neural function occur caudal (Goshgarian, 2003; Fuller et al., 2003, Golder et al., 2003; Nantwi and Goshgarian, 2005) and rostral to the site of injury (Fuller et al., 2006; Golder et al., 2001a,b; Goshgarian et al., 1986). However, few studies have shown that neural plasticity of respiratory function occurs in supraspinal centers after spinal cord injury (Golder et al., 2001a) and the mechanisms underlying supraspinal neural plasticity have not been investigated. High cervical spinal cord hemisection (C2) interrupts the ipsilateral descending premotor ⁎ Corresponding author. Fax: +1 313 577 3125. E-mail address: [email protected] (M.B. Zimmer). 0014-4886/$ - see front matter. Published by Elsevier Inc. doi:10.1016/j.expneurol.2007.05.003
projections, effectively paralyzing the ipsilateral hemidiaphragm, and thus, reduces respiratory volume (Golder et al., 2003). This reduced volume is sensed by lung and chest wall receptors which feedback through vagal mechanisms to increase respiratory frequency; routinely observed in vivo (Fuller et al., 2006; Golder et al., 2001a, 2001b; Goshgarian et al., 1986). Below the site of injury, respiratory neural plasticity occurs within the phrenic motor nucleus (C3–C6 in rats) to allow for the expression of a latent respiratory motor pathway, termed the crossed phrenic pathway (see Goshgarian, 2003 for review). Activation of the latent crossed phrenic pathway results in the restoration of function to the previously paralyzed hemidiaphragm (see Goshgarian, 2003 for review). Thus, many studies have focused on the mechanisms underlying the regulation and activation of the crossed phrenic pathway in order to restore motor function to the paralyzed hemidiaphragm. Rostral to the site of injury within the medulla, however, the effect of spinal cord injury on respiratory neural function is less clear.
138
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
Recently, Golder et al. (2001a) showed that respiratory neural plasticity occurs in supraspinal respiratory centers after spinal cord injury in adult male rats. Using anesthetized, paralyzed, ventilated rats, 2 months after a C2 hemisection, the respiratory frequency was significantly decreased. In addition, examination of hypoglossal motor output, originating rostral to the site of the injury, revealed that the peak amplitude of the hypoglossal roots was significantly decreased (Golder et al., 2001a). Furthermore, rats with a C2 hemisection showed a blunted response to hypercapnia; the increase in amplitude of the hypoglossal burst was blunted in spinal cord injured rats compared to non-injured controls (Golder et al., 2001a). Together, the data indicate a decreased respiratory drive after chronic spinal cord injury due to supraspinal mechanisms. The underlying cause of this altered supraspinal mechanism is not clear, but could involve injuryinduced plasticity in supraspinal brain centers resulting from either the axotomy of respiratory premotor neurons located in the ventral lateral medulla and/or the axotomy of ascending spinobulbar projections that connect to respiratory neurons in the medulla. Therefore, this study was designed to examine the effect of spinal cord injury on brainstem neural output using the brainstem-spinal cord preparation of the neonate, such that no other inputs other than from within the neuraxis itself are present. In adult, whole animal models, attempts are made to eliminate and/or reduce the feedback from higher brain centers, peripheral chemoreceptors and muscle and lung receptors; however, it is difficult to eliminate all feedback, and the effect of anesthetic on neural function may obscure the results. This is not a problem in the present study. After spinal cord injury, specific receptor subunits involved in neurotransmission, such as AMPA receptor subunits (GluR) and NMDA receptor subunits (NR), change within the spinal cord both rostral and caudal to the site of injury (Park et al., 2003; Grossman et al., 2000; Brown et al., 2004; Grossman et al., 1999; Grossman et al., 2001; Fuller et al., 2005). Many studies have shown that decreases in specific receptors or receptor subunits occur caudal to the site of injury; however, Brown et al., 2004 found that using Western blot analysis, GluR1, GluR2, GluR4, NR2A and NR2B were all significantly decreased in spinal cord tissue both rostral and caudal to an acute T8 contusion injury in young rats. We predict that a high cervical spinal cord injury (around C2) in neonate rats will alter receptors rostral to the site of injury within the medulla, which could have a significant impact on the descending control mechanisms of many autonomic functions, including respiration. In the present study, we examined whether supraspinal neural plasticity occurs within the respiratory system as a result of an acute, high cervical spinal cord injury in neonate rats. After an acute spinal cord injury (2 days), the in vitro brainstem-spinal cord preparation was used to determine whether changes in medullary neural output occurred as a result of the injury. We hypothesized that supraspinal alterations in respiratory control mechanisms would manifest in neonates after an acute injury similar to the results of Golder et al. (2001a) in adult rats, since the neonate is still developing and highly plastic. The respiratory response to changes in pH was examined to determine if medullary chemosensing mechanisms are also altered as a result
of the spinal cord injury. Many studies have shown a reduced sensitivity to hypercapnia after spinal cord injury in both animal models (Teng et al., 2003) as well as human studies (Kelling et al., 1985; Manning et al., 1992); however, the underlying mechanism is not known. We hypothesized that modifications in supraspinal brain centers may be a part of the underlying mechanism behind this reduced sensitivity to hypercapnia. And finally, Western blot analysis was performed on medullary tissue both ipsilateral and contralateral to the injury to examine whether changes in receptors involved in neurotransmission are altered in response to an acute spinal cord injury. Like Brown et al. (2004), we hypothesized that many receptor subunits rostral to the injury, in this case, within the brainstem, would be altered by an upper cervical spinal cord injury. Methods Surgery All surgeries and protocols were reviewed and approved by the Animal Investigation Committee (AIC) at Wayne State University acting under the guidelines set up by the AAALAC prior to experimentation. Neonate rats (first day of birth or postnatal day (1) were removed, individually, from their mother and deeply anesthetized with cold. An incision was made on the dorsal aspect of the neck and the muscles were cut to expose the C1 vertebra. A laminectomy at the C1 vertebra was made followed by a lesion in the left C2 spinal cord using microscissors (n = 10). Attempts were made to completely hemisect the spinal cord, but all pups were used in this study, despite the lack of full hemisection in all pups. The skin was closed with tissue glue and the pups warmed on a heating pad until awake. They were immediately placed in the cage with the other pups and mother. At this time, another pup was removed from the litter for surgery. Sham surgeries (n = 9) were performed in the same manner as described above, except that the spinal cord was not cut. Hemisection and sham surgeries were alternated. The mothers received all pups and immediately fed them. Physiology experiments After 2 days (P3 or P4), the pups were removed and anesthetized deeply with halothane. The brainstem-spinal cord (pons removed; medulla-T5/6) was carefully dissected and the dura removed. The brainstem-spinal cord was placed in a recording chamber supplied with artificial cerebral spinal fluid (aCSF composed of 113.0 mM sodium chloride, 3.0 mM potassium chloride, 1.2 mM sodium phosphate, 1.5 mM calcium chloride, 1.0 mM magnesium chloride, 30.0 mM sodium bicarbonate and 30.0 mM dextrose) that was bubbled with 95% oxygen and 5% carbon dioxide. Suction electrodes were placed on both the right and left C4 ventral roots to record bilateral respiratory-related activity. Electrical signals were amplified and filtered (300–3000 Hz, gain 20,000) using Grass amplifiers (Model P511). Baseline recordings were taken for approximately 20 min at a pH of 7.4 and then the pH of the
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
superfusate was changed by altering the sodium bicarbonate and carbon dioxide in the solutions, first to a high pH (∼7.8) for another 20 min followed by low pH (∼ 7.2) for another 20 min. The data were recorded continuously on computer using Spike2 data acquisition software (CED, Cambridge UK) (sample rate 3000 Hz). Bilateral respiratory-like recordings were full-wave rectified and integrated using CED data analysis software (Cambridge UK). The peak amplitude, area and the duration of the respiratory-like bursts were analyzed for the last 5 min of each pH exposure and comparisons made between spinal cord injured and sham control rat pups. Western blot analysis Immediately after the physiological recordings were taken, the tissue was dissected and the right and left side of the medulla were frozen at − 80 °C until processing. Total protein was extracted from the medulla by sonication in a modified Glasgow protein extraction buffer (150 mM NaCl, 1.0 mM EDTA, 50 mM Tris, 1% NP40, 0.1% Triton X-100 and protease inhibitors) and protein concentrations were determined using a Bradford assay (Sigma, St. Louis, MO). Medullary protein homogenates (50 ìg) containing Laemmli sample buffer were loaded onto 7.5% Tris–HCl gels (BioRad, Hercules, CA) and separated. Each gel contained protein homogenate from control rats and spinal cord injured rats from one side of the medulla along with a Precision Plus Protein™ Standard (10–250 kDa, BioRad, Hercules, CA). Following separation, the proteins were transferred to a polyvinylidine difluoride (PVD) membrane (BioRad, Hercules, CA) overnight at 4 °C. The membranes were washed 5× with a Tris–Saline–-Tween buffer (TSTB), followed by non-specific blocking with 5% milk. The membranes were then incubated overnight at 4 °C in primary antibody (see below). The next morning, the membranes were washed (5 × 5 min) and incubated in peroxidase-conjugated secondary antibody diluted in a 1% TSTB solution for 1–2 h at room temperature. The membranes were washed again and incubated in an enhanced chemiluminescent peroxidase substrate (Chemicon, Temecula, CA) for 3 min. The membranes were apposed to film (30 s–10 min) and developed. The films were scanned with Image Scanner II (Amersham Biosciences, Piscataway, NJ) and the bands corresponding to the various proteins were quantified using the ImageQuant TL program (Amersham Biosciences, Piscataway, NJ). Protein bands were measured for total density. Spinal cord-injured values were compared to controls for quantitative and statistical analysis. â-actin controls (1:5000, Sigma-Aldrich, St. Louis, MO) were used as a loading control. Antibody solutions Western blots were probed using the following antibodies. AMPA receptor subunit protein expression levels were determined using rabbit anti-GluR1 (1:1000), mouse antiGluR2 (1:1000), mouse anti-GluR3 (1:1000) and rabbit antiGluR4 (1:1000). NMDA receptor subunit protein expression
139
levels were determined using mouse anti-NR1 (1:1000), rabbit anti-phospho-NR1 (Ser897) (1:000) (Upstate Cell Signaling Solutions, Lake Placid, NY), mouse anti-NR2A (1:500), rabbit anti-NR2B and rabbit anti-NR2B phosphoTyr1472 (1:1000). Adenosine receptor protein levels were determined with rabbit anti-A1 (1:1000) and rabbit anti-A2A (1:2500; Sigma). Mouse anti-glutamic acid decarboxylase (65) (1:1000), mouse antiGAD67 (1:5000), rabbit anti-glycine receptor (1:1000) and rabbit anti-substance P receptor (NK-1) (1:1000) were also used to determine protein level differences between non-injured and spinal cord-injured rats. All antibodies, except those already listed above, were purchased from Chemicon International (Temecula, CA). All secondary antibodies were purchased from Sigma-Aldrich (St. Louis, MO); peroxidase-conjugated goat anti-rabbit (1:10,000) and peroxidase-conjugated goat antimouse (1:2500). Statistics Baseline respiratory variables (frequency, burst peak amplitude, burst area and burst duration) of sham-operated control and injured rats were analyzed for 5 continuous minutes and the average value obtained for each rat. The differences between sham controls and spinal cord injured rats were examined at ‘normal’ pH 7.4 using an ANOVA. The responses of each respiratory variable to changes in pH (7.8 and 7.2) were assessed using a repeated measures ANOVA. The results obtained with Western blot analysis were analyzed using a t-test. p b 0.5 was considered significant. All results are expressed as means ± standard error (SE). Results Upper cervical spinal cord-injured neonates and neonates with a sham surgery (non-injured, control rats) gained a significant amount of weight over two days post surgery (Table 1) ( p = 0.001). There were no significant difference in weight between non-injured and spinal cord-injured rats ( p = 0.15). Baseline recordings of the right and left C4 ventral roots at normal pH (7.4) revealed that the respiratory-like frequency of the spinal cord-injured rats was significantly reduced ( p = 0.04) compared to the sham-operated controls (Fig. 1). In addition, there was a significant decrease in the peak amplitude of the burst ( p = 0.04), the burst area ( p = 0.001) and the burst duration ( p = 0.024) on the side ipsilateral to injury, but not on the contralateral, or uninjured side (Fig. 1).
Table 1 Body weights (grams) of neonate rats on the day of surgery, sham or incomplete spinal cord injured (P0–P1), and 2 days later on the day of physiology experiments and tissue collection for Western blot analysis (P3/4)
Control Spinal cord injured
Body weight (P0/P1)
Body weight (P3/4)
7.3 ± 0.3 7.2 ± 0.2
11.1 ± 0.7* 9.9 ± 0.5*
Both groups of rats significantly gained weight (*), but there was no difference between control and spinal cord injured rats.
140
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
Fig. 1. Upper cervical spinal cord injury (SCI) in neonate rats caused a significant decrease in the resting frequency of the respiratory-like bursts of the brainstem-spinal cord preparation (bars ± SE). Compared to control sham operated rats, spinal cord injury caused a significant reduction in the peak amplitude, the burst area and the burst duration of the C4 ventral root ipsilateral to the injury compared to contralateral to the injury during normal pH (7.4).
Changing the pH of the superfusate of the brainstem-spinal cord preparation resulted in significant effects on all respiratory variables examined (Fig. 2). The respiratory-like frequency of spinal cord-injured preparations was altered in response to pH ( p = 1.1 × 10− 4), but the response to pH was not different between injured and sham control preparations, i.e., the frequency decreased in response to high pH and increased in response to low pH in both non-injured and injured rats (Fig. 2A). The change in burst duration in response to pH was significantly different in absolute raw values, i.e., seconds (right nerve, p = 0.016; left nerve, p = 0.024, data not shown), but not in relative terms, i.e., percent difference between non-injured, sham controls and injured rats (Fig. 2B). There was, however, a significant interaction between changing the pH and spinal cord
injury when examining the relative changes in burst duration ( p = 0.026). This significant interaction between pH and injury was also evident in the total burst area ( p = 0.011) (Fig. 2D). This is most likely due to the opposite effect that pH had on burst duration and area; in non-injured, sham control brainstemspinal cord preparations, raising the pH of the superfusate (∼ 7.8) increased the burst duration and the area, whereas in the injured rats, raising the pH caused a decrease in burst duration and area (Figs. 2B, D). When pH was lowered to 7.2 in noninjured, sham controls, the burst duration did not change, but in spinal cord injured preparations, the burst duration and area increased. The relative change in peak amplitude in response to changing the pH was also significant ( p = 0.001), however, there were no differences between spinal cord injured and sham-
Fig. 2. Changing the pH of the superfusate from the control level of 7.4 significantly altered the frequency (A) and the peak amplitude (C) of the respiratory-like discharge of both non-injured and spinal cord injured neonate rats. There was a significant interaction between spinal cord injury and the response to pH when examining the burst duration (B) and area (D); the duration and area tended to change in an opposite manner in injured rats compared to non-injured rats. Spinal cord injury was made on the left side of the spinal cord.
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
operated control rats (Fig. 2C), i.e., peak amplitude increased in response to high pH and there was a slight decrease in response to low pH in both groups. Western blot analysis revealed that several key receptor subunits involved in neurotransmission were altered in the medulla on both the side contralateral and ipsilateral to the upper cervical spinal cord injury. Of the glutamate receptor subunits, the protein levels of the NR1 subunit of the NMDA receptor were significantly decreased rostral and ipsilateral to the injury (Fig. 3A) ( p = 0.027). However, the phosphorylated NR1 (pNR1) protein and the ratio of pNR1 relative to the total NR1 were not significantly changed. Contralateral to the injury, the NR1 protein levels were not different between non-injured, sham controls and spinal cord injured rats ( p = 0.08); however, there was a significant decrease in the amount of pNR1 protein ( p = 0.002). The ratio of pNR1 to total NR1 was also significantly reduced ( p = 0.03) indicating that although NR1 protein levels may not have significantly decreased after injury, there is a decrease in functional/phosphorylated NR1 protein.
141
The NR2B subunit protein levels were significantly increased ipsilateral to the spinal cord injury ( p = 0.03); however, there was no change in the amount of phosphorylated NR2B (pNR2B) protein levels (Fig. 3A). The ratio of pNR2B to total NR2B protein was significantly decreased ( p = 0.003) on the side ipsilateral to the injury, possibly indicating that although the number of NR2B subunits increased in quantity they were not being phosphorylated. On the side contralateral to the injury, NR2B protein levels did not change significantly; however, the quantity of pNR2B protein significantly decreased ( p = 0.02). This resulted in a significant decrease in the ratio of pNR2B to total NR2B protein levels on the right side (contralateral to the injury) of the medulla of spinal cord injured rats ( p = 0.007). The NR2A protein expression did not change rostral to the injury on either side of the medulla in response to spinal cord injury (Fig. 3B). Of the AMPA receptor subunits, there was a bilateral decrease in the protein levels of the GluR2 subunit in the medulla of spinal cord injured neonate rats (Fig. 3C) (right p = 0.02; left p = 0.03). No other subunits of
Fig. 3. Western blot analysis revealed that specific protein levels involved in glutamate receptor transmission are significantly altered in the medulla after C2 spinal cord injury in neonatal rats. The left side of the medulla is ipsilateral to the injury, whereas the right side is contralateral to the injury. The bar graphs show the amount of protein in medullary tissue both contralateral (right) and ipsilateral (left) to SCI (white bars) compared to non-injured (black bars) rats for the various receptor subunits of the NMDA receptor (NR1, 2A, 2B) and the AMPA receptor (GluR1, 2, 3, 4). Significant difference between non-injured and injured rats is marked with an asterisk (*).
142
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
Fig. 4. Western blot analysis revealed that other receptors and proteins involved in neurotransmission were altered rostral to spinal cord injury (left side) (white bars) in the medulla in neonates. Glutamic acid decarboxylase (GAD), the rate limiting enzyme in the production of GABA was slightly elevated ipsilateral to the injury (left side). Adenosine receptors (A1, A2A) and the neurokinin 1 (NK-1) receptor protein levels were also altered in response to acute spinal cord injury. Significant difference between control (black bars) and injured rats (white bars) is marked with an asterisk (*).The left side of the medulla is ipsilateral to the injury, while the right side is contralateral to the injury.
the AMPA receptor that were examined (GluR1, GluR3, GluR4) were altered in response to upper cervical spinal cord injury. Several other proteins that could potentially be involved in respiratory control were examined. Glutamic acid decarboxyl-
ase (GAD) is a key enzyme involved in the production of GABA, a major inhibitory neurotransmitter. One of the two isoforms, GAD65, was significantly elevated on the left side of the medulla ( p = 0.036) (Fig. 4), possibly indicating an increase in the production of GABA after spinal cord injury. Adenosine receptors, A1 and A2A, were also examined. After spinal cord injury, A1 protein levels were significantly elevated ( p = 0.016) and the A2A protein levels were significantly decreased ( p = 0.001) on the side contralateral to the injury (Fig. 4). Finally, the tachykinin receptor, NK-1, protein levels were examined, since they have been localized to neurons involved in respiratory rhythmogenesis. After spinal cord injury, levels of NK-1 protein levels were significantly decreased ipsilateral to the injury ( p = 0.02) and significantly increased contralateral to the injury (Fig. 4) ( p = 0.04). In summary, Western blot analysis revealed that there was a tendency for a downregulation of receptor subunits involved in excitatory neurotransmission and an upregulation of receptor subunits involved in inhibitory neurotransmission rostral to the injury in the medulla of upper spinal cord injured neonate rats (Fig. 5). Discussion
Fig. 5. A summary of the receptor protein levels that were significantly altered by a left C2 spinal cord injury as measured by Western blot analysis. The general trend was a decrease in protein levels that are associated with excitatory neurotransmission and an increase in protein levels that are associated with inhibitory neurotransmission.
The results from this study show that upper cervical spinal cord injury can induce neural plasticity and modulate supraspinal respiratory centers in as little as 2 days in the neonate. The data from this study are consistent with the results of Golder et al. (2001a) showing a decrease in central respiratory drive after chronic upper cervical spinal cord injury in adult rats. A decrease in central respiratory drive could be detrimental in a condition such as spinal cord injury where
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
respiratory motor function is already compromised. The underlying mechanism behind the decreased respiratory neural output is not known, but Western blot analysis demonstrated that many receptor subunits and the phosphorylation state of some subunits change in the medulla after injury. The general trend was a reduction in protein subunits that are involved in excitatory neurotransmission and an increase in those proteins involved in inhibitory neurotransmission, which is consistent with the findings of a reduced respiratory drive after injury. The significance or specific location of the receptor changes is not yet known, but knowing that medullary neural control is altered after upper cervical spinal cord injury could significantly alter our view of how spinal cord injury affects respiratory control, as well as other autonomic systems. Respiratory motor output after acute upper cervical spinal cord injury Upper cervical spinal cord injury (in the C2 region) caused a significant reduction in the respiratory frequency in neonates using an in vitro brainstem-spinal cord preparation to assess motor output. This is consistent with the observations by Golder et al. (2001a) who also found that the respiratory frequency was decreased after chronic spinal cord injury in anesthetized, paralyzed and ventilated rats. The underlying mechanism behind the reduced respiratory frequency is not known. Recall that in vivo, respiratory frequency is actually increased, not decreased after spinal cord injury. This is due to changes in receptor feedback from the paralysis of the hemidiaphragm and the subsequent decrease in tidal volume (Golder et al., 2001b). In adults, over time, the altered mechanoreceptor inputs may result in neural plasticity of respiratory centers such that when these inputs are removed, the resultant frequency is decreased compared to non-injured controls. In neonates, this may occur quickly due to the developing state of the respiratory system. Alternatively, the axotomy of ascending sensory or other modulating inputs or even specific alterations in neurotransmission from either direct or indirect mechanisms, including changes in receptor numbers, receptor composition, receptor function, as well as neurotransmitters after injury (see below, Western blot analysis) may be involved in the reduction of respiratory frequency that is observed using reduced preparations. Ipsilateral to the injury, the absolute phrenic motor output was also significantly decreased, but not absent. This is most likely due to the transection of descending premotor axons which project to the ipsilateral phrenic motor neurons through the lateral and ventral funiculi. In the present study, the primary objective was to lesion the cord on the left side in order to alter respiratory drive, not to make a complete hemisection. Thus, not all rats that were used in this study had a complete hemisection, as determined by visualization through a dissecting scope. Nonetheless, even in those rats with a complete hemisection, crossed phrenic activity is spontaneously present in neonates; this has been demonstrated both in vitro (Zimmer and Goshgarian, 2005), as well as in vivo (Huang, Y. and Goshgarian, H.G. unpublished data). Therefore, the amount of activity observed on the left C4 ventral roots was not due solely
143
to crossed phrenic pathways, although crossed phrenic pathways probably contributed to motor output in some cases. Respiratory response to pH Changes in pH/CO2 are sensed by both peripheral and central chemoreceptors. Central chemoreception has been localized to multiple sites within the brain which contain neurons that are responsive to changes in pH/CO2: the retrotrapezoid nucleus (RTN) (Mulkey et al., 2004), the nucleus of the solitary tract (Putnam, 2001), the cerebellum (Xu and Frazier, 2002), the locus coeruleus (Ballantyne and Scheid, 2001) and the raphe nucleus (Richerson, 2004). The in vitro brainstem-spinal cord preparation is a model system that others have used to examine the cellular and molecular mechanisms underlying the sensitivity to pH in brainstem regions (Ballantyne and Scheid, 2001; Putnam, 2001). Our data from noninjured rats are consistent with results from other labs and show that superfusion of aCSF with a high pH results in a decrease in respiratory frequency and a low pH causes an increase in frequency (Dubayle and Viala, 1998; Kawai et al., 1996; Oyamada et al., 1998). Spinal cord injury has been shown to alter the sensitivity to hypercapnia in vivo and many studies report a reduced or blunted response to hypercapnia (Kelling et al., 1985; Manning et al., 1992; Teng et al., 2003). The underlying mechanism behind this reduced sensitivity to CO2 is not known. The results presented in this study do not show a clear reduction in pH sensitivity, but there was a significant interaction between spinal cord injury and the sensitivity to pH when analyzing burst duration and burst area; that is, the injured rats responded in an opposite manner compared to the non-injured rats. Specifically, in non-injured rats, when pH was increased there tended to be a slight increase in burst duration and a decrease or no change when exposed to low pH bilaterally. The opposite was observed with spinal cord injured rats; a slight decrease in burst duration when exposed to high pH and an increase in burst duration during low pH. It is not clear what significance this has in reference to the whole animal, but it does indicate that upper cervical spinal cord injury alters supraspinal respiratory neural control causing a change in the respiratory response to pH. Therefore, it is possible that some of the reduced sensitivity to hypercapnia that has been observed after spinal cord injury may be due, in part, to alterations in supraspinal respiratory neural control. Western blot analysis Significant changes in receptor subunits involved in both excitatory and inhibitory neurotransmission were demonstrated rostral to the site of spinal cord injury using Western blot analysis, confirming the results of Brown et al., 2004. More importantly, upper cervical spinal cord injury induced changes in receptor subunits within the medulla. Many receptor subunits involved in excitatory neurotransmission, NR1, NR2B, GluR2, A2, were decreased after injury, while those involved in inhibition, GAD65, A1, were elevated. From these data it is hard to evaluate whether any specific functional or physiological
144
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145
consequences result from these changes, but the changes are consistent with an overall decrease in respiratory drive observed after injury. The exact mechanisms behind a change in receptor protein levels are not known. Increased cell death may occur after spinal cord injury due to the axotomy of neurons. And glutamate-receptor mediated excitotoxicity has also been associated with the spinal cord injury site, whereby, glutamate spills out from the injury site and damages nearby neurons causing secondary cell death (Bittigau and Ikonomidou, 1997; Liu et al., 1999). This could result in a decrease in specific receptor protein concentrations relative to the total protein concentration; however, this does not seem likely since some receptor subtypes increased. It seems more likely that the changes in receptor quantity were due to modifications in various inputs that resulted in specific changes in receptor protein levels. This conclusion was also reached by other studies that showed that specific receptor changes occurred in specific cell types within specific regions of the spinal cord, gray or white matter after injury (Park et al., 2003; Grossman et al., 2000; Brown et al., 2004; Grossman et al., 1999; Grossman et al., 2001; Fuller et al., 2005). Not only were total protein levels changed after spinal cord injury, but also the amount of phosphorylated protein was changed. It appears that the phosphorylated state of NMDA channels decreased after injury; this was the case for both the NR1 and the NR2B subunit. Phosphorylation of the NR1 subunit has been associated with increasing the functional state of the receptor (Yun and Zhang, 2000), whereas, phosphorylation of the NR2B subunit has been associated with instability and internalization of receptors from the membrane. Therefore, spinal cord injury not only altered the total quantity of protein, but may also alter the functional state of the receptors. Neurokinin-1 receptors have been localized to a specific subset of neurons within the pre-Bötzinger complex, a site of respiratory rhythmogenesis within the ventral lateral medulla. Removal or destruction of these neurons results in a very unstable and non-rhythmic respiratory pattern, strongly indicating their role in generating the respiratory rhythm (Gray et al., 1999, 2001). Interestingly, neurons in the ventral medulla that express NK-1 receptors are also necessary for central and peripheral chemoreception (Nattie and Li, 2006). NK-1 receptor protein levels were altered in response to upper cervical spinal cord injury; the quantity of NK-1 protein increased contralateral to injury and decreased ipsilateral to injury. Although Western blot analysis does not give an indication of the specific location of protein changes, if changes in the relative amount of NK-1 receptor protein occur within the pre-Bötzinger complex and other regions involved in chemoreception, the respiratory frequency might be altered after spinal cord injury. Current studies in our lab are examining whether an alteration in preBötzinger NK1-protein expression directly affects respiratory rhythm transmission after upper cervical spinal cord injury. The mechanism(s) behind the changes in neural receptor composition and quantity after spinal cord injury are not known. The data here are consistent with the possibility that some of these changes in neurotransmission may be directly responsible
for the effects that are observed on the respiratory motor output after spinal cord injury. The in vitro brainstem spinal cord preparation The in vitro brainstem-spinal cord preparation is a useful model to examine respiratory neural control mechanisms and has been used for many years by respiratory neurobiologists. It is has both advantages and disadvantages. One problem is the time limitation; respiratory rhythm can only be examined within the first 4–6 days with many studies using postnatal day 4. However, in this particular case, we were able to examine the effect of spinal cord injury on brainstem neural output and the response to changes in pH/CO2 with no other inputs or feedback other than within the neuraxis itself to interfere with the motor output that is generated by brainstem mechanisms alone. In adult, whole animal models, attempts are made to eliminate, reduce, and/or equalize the feedback from higher brain centers, peripheral chemoreceptors, and muscle and lung receptors; however, it is difficult to eliminate all feedback. In addition, the effect of anesthetic on neural function may obscure the results. None of the above issues is a concern when using the in vitro brainstem-spinal cord preparation. Acknowledgment The work in this paper was supported by NIH grant HD31550 (Dr. H.G. Goshgarian). References Ballantyne, D., Scheid, P., 2001. Central chemosensitivity of respiration: a brief overview. Respir. Physiol. 129, 5–12. Bittigau, P., Ikonomidou, C., 1997. Glutamate in neurologic diseases. J. Child Neurol. 12 (8), 471–485. Brown, K.M., Wrathall, J.R., Yasuda, R.P., Wolfe, B.B., 2004. Glutamate receptor subunit expression after spinal cord injury in young rats. Dev. Brain Res. 152, 61–68. Dubayle, D., Viala, D., 1998. Effects of CO2 and pH on the spinal respiratory rhythm generator in vitro. Brain Res. Bull. 45 (1), 83–87. Fuller, D.D., Johnson, S.M., Olson Jr., E.B., Mitchell, G.S., 2003. Synaptic pathways to phrenic motorneurons are enhanced by chronic intermittent hypoxia after cervical spinal cord injury. J. Neurosci. 23 (7), 2993–3000. Fuller, D.D., Baker-Herman, T.L., Golder, F.J., Doperalski, N.J., Watters, J.J., Mitchell, G.S., 2005. Cervical spinal cord injury upregulates ventral spinal 5-HT2A receptors. J. Neurotrauma 22 (2), 203–213. Fuller, D.D., Golder, F.J., Olson Jr., E.B., Mitchell, G.S., 2006. Recovery of phrenic activity and ventilation after cervical spinal hemisection in rats. J. Appl. Physiol. 100, 800–806. Golder, F.J., Reier, P.J., Bolser, D.C., 2001a. Altered respiratory motor drive after spinal cord injury: supraspinal and bilateral effects of a unilateral lesion. J. Neurosci. 21 (21), 8680–8689. Golder, F.J., Reier, P.J., Davenport, P.W., Bolser, D.C., 2001b. Cervical spinal cord injury alters the pattern of breathing in anesthetized rats. J. Appl. Physiol. 91 (6), 2451–2458. Golder, F.J., Fuller, D.D., Davenport, P.W., Johnson, R.D., Reier, P.J., Bolser, D.C., 2003. Respiratory motor recovery after unilateral spinal cord injury: eliminating crossed phrenic activity decreases tidal volume and increases contralateral respiratory motor output, vol. 23(6), pp. 2494–2501. Goshgarian, H.G., 2003. The crossed phrenic phenomenon: a model for plasticity in the respiratory pathways following spinal cord injury. J. Appl. Physiol. 94 (2), 795–810.
M.B. Zimmer, H.G. Goshgarian / Experimental Neurology 206 (2007) 137–145 Goshgarian, H.G., Moran, M.F., Prcevski, P., 1986. Effect of cervical spinal cord hemisection and hemidiaphragmparalysis on artherial blood gases, pH and respiratory rate in the adult rat. Exp. Neurol. 93 (2), 440–445. Gray, P.A., Rekling, J.C., Bocchiaro, C.M., Feldman, J.L., 1999. Modulation of respiratory frequency by peptidergic input to rhythmogenic neurons in the preBotzinger complex. Science 286 (5444), 1566–1568. Gray, P.A., Janszewski, W.A., Mellen, N., McCrimmon, D.R., Feldman, J.L., 2001. Normal breathing requires preBotzinger complex neurokinin-1 receptor-expressing neurons. Nat. Neurosci. 4 (9), 927–930. Grossman, S.D., Wolfe, B.B., Yasuda, R.P., Wrathall, J.R., 1999. Alterations in AMPA receptor subunit expression after experimental spinal cord contusion injury. J. Neurosci. 19 (14), 5711–5720. Grossman, S.D., Wolfe, B.B., Yasuda, R.P., Wrathall, J.R., 2000. Changes in NMDA receptor subunit expression in response to contusive spinal cord injury. J. Neurochem. 75, 174–185. Grossman, S.D., Rosenberg, L.J., Wrathall, J.R., 2001. Relationship of altered glutamate receptor subunit mRNA expression to acute cell loss after spinal cord contusion. Exp. Neurol. 168, 283–289. Kawai, A., Ballantyne, D., Mückenhoff, K., Scheid, P., 1996. Chemosensitive medullary neurones in the brainstem-spinal cord preparation of the neonatal rat. J. Physiol. 492.1, 277–292. Kelling, J.S., DiMarco, A.F., Gottfried, S.B., Altose, M.D., 1985. Respiratory responses to ventilatory loading following low cervical spinal cord injury. J. Appl. Physiol. 59, 1752–1756. Liu, D., Xu, G.Y., Pan, E., McAdoo, D.J., 1999. Neurotoxicity of glutamate at the concentration released upon spinal cord injury. Neurosci. 93 (4), 1383–1389. Manning, H.L., Brown, R., Scharf, S.M., Leith, D.E., Weiss, J.W., Weinberger, S.E., Schwartzstein, R.M., 1992. Ventilatory and P0.1 response to hypercapnia in quadriplegia. Respir. Physiol. 89, 97–112.
145
Mulkey, D.K., Stornetta, R.L., Weston, M.C., Simmons, J.R., Parker, A., Bayliss, D.A., Guyenet, P.G., 2004. Respiratory control by ventral surface chemoreceptor neurons in rats. Nat. Neurosci. 7, 1360–1369. Nantwi, K.D., Goshgarian, H.G., 2005. Adensosinergic mechanisms underlying recovery of diaphragm motor function following upper cervical spinal cord injury: potential therapeutic implications. Neurol. Res. 27 (2), 195–205. Nattie, E., Li, A., 2006. Neurokinin-1 receptor-expressing neurons in the ventral medulla are essential for normal central and peripheral chemoreception in the conscious rat. J. Appl. Physiol. 101, 1596–1606. Oyamada, Y., Ballantyne, D., Mückenhoff, K., Scheid, P., 1998. Respirationmodulated membrane potential and chemosensitivity of locus coeruleus neurones in the in vitro brainstem-spinal cord of the neonatal rat. J. Physiol. 513.2, 381–398. Park, E., Liu, Y., Fehlings, M.G., 2003. Changes in glial cell white matter AMPA receptor expression after spinal cord injury and relationship to apoptotic cell death. Exp. Neurol. 182, 35–48. Putnam, R.W., 2001. Intracellular pH regulation of neurons in chemosensitive and nonchemosensitive areas of brain slices. Respir. Physiol. 129, 37–56. Richerson, G.B., 2004. Serotonergic neurons as carbon dioxide sensors that maintain pH homeostasis. Nat. Rev., Neurosci. 5 (6), 449–461. Teng, Y.D., Bingman, M., Taveira-DaSilva, A.M., Pace, P.P., Gillis, R.A., Wrathall, J.R., 2003. Serotonin 1A receptor agonists reverse respiratory abnormalities in spinal cord-injured rats. J. Neurosci. 23 (10), 4182–4189. Xu, F., Frazier, D.T., 2002. Role of the cerebellar deep nuclei in respiratory modulation. Cerebellum 1 (1), 35–40. Yun, L., Zhang, J., 2000. Recent development in NMDA receptors. Chin. Med. J. 113, 948–956. Zimmer, M.B., Goshgarian, H.G., 2005. Spontaneous crossed phrenic activity in the neonatal respiratory network. Exp. Neurol. 194 (2), 530–540.