Abstract

INTRODUCTION

The most important function of the plasma membrane is to protect the cell and maintain a functional intracellular environment. When the integrity of the neuronal membrane is disrupted, ions and other molecules leak in and out of the cell down their electrochemical gradients, resulting in cell death through various mechanisms (Lucas et al., 1985; Gross and Higgins, 1987; Fitzpatrick et al., 1998; Fiskum, 2000; Lewen et al., 2000; Ray et al., 2003). It is clear that membrane repair following physical insult is critical for survival and functional recovery in neurons, as it is for other cells. In the case of axons, membrane resealing is also required to prevent further degeneration of the severed fibers (Gross et al., 1983; Lucas et al., 1985; Yawo and Kuno, 1985; Shi et al., 1989; Xie and Barrett, 1991; Shi and Borgens, 1999; Shi et al., 2000; Shi and Borgens, 2000). Hence, the earlier the breach is closed, the better the prospects for maintaining the residual functions of the neuron and for potential regrowth and recovery of function through plasticity and regeneration.

Neuronal membrane disruption following mechanical insult has been documented extensively in vitro, ex vivo (Xie and Barrett, 1991; Spira et al., 1993; Krause et al., 1994; Eddleman et al., 1997; Shi et al., 2000; Shi and Borgens, 2000; Shi and Pryor, 2002) and in vivo (Pettus et al., 1994; Maxwell et al., 1999; Shi, 2002). Extended membrane disruption likely plays a key role in cell death, tissue loss, and neurological dysfunction. However, the environment following acute traumatic spinal cord injury in vivo is believed to be hostile for membrane repair (Stokes et al., 1983; Young and Koreh, 1986), and prolonged axonal membrane damage appears to exist following in vivo injury (Shi, 2004). Therefore, it is important to elucidate the mechanisms of membrane repair and to formulate effective pharmaceutical interventions that can promote membrane repair and enhance functional recovery.

It is well established that membrane resealing in mammalian nerve cells is calcium dependent (Xie and Barrett, 1991; Shi et al., 2000). Furthermore, it is believed that calcium influx following injury initiates many biochemical reactions that are critical for repair of damaged membranes (Yawo and Kuno, 1983, 1985; Terasaki et al., 1997). The most obvious avenue of calcium influx following mechanical insults is at the primary site of membrane disruption (Borgens et al., 1980; Strautman et al., 1990; Ziv and Spira, 1993; Krause et al., 1994; Eddleman et al., 2000). However, there is also a possible, and sometimes overlooked, pathway of calcium influx through depolarized voltage dependent calcium channels (VDCC) (Ziv and Spira, 1993; Sattler et al., 1996; Agrawal et al., 2000; Ouardouz et al., 2003; Ouardouz et al., 2005). It is commonly believed that calcium influx through the injury site is sufficient to trigger effective membrane resealing due largely to the assumption that this calcium influx is much greater than through individual calcium channels. However, the role of calcium influx through calcium channels has not been examined in the context of membrane resealing. Considering the fact that calcium channel blockers have been postulated as a potential treatment for traumatic CNS injury (Pointillart et al., 1993; Rhee et al., 1996; Imaizumi et al., 1999; Mason et al., 1999), it is particularly important to determine their role in influencing membrane repair, a critical factor in the recovery of CNS traumatic injury (Shi and Borgens, 1999; Shi et al., 2000; Shi and Borgens, 2000; Shi and Pryor, 2000).

The overall objective of this study was to examine the role of calcium influx through VDCC in axonal membrane resealing following two types of mechanical injuries, transection and compression. Two known calcium channel blockers, nimodipine, a specific L-type calcium channel blocker, and cadmium, a non-specific calcium channel blocker, were used to block calcium influx through Ca²⁺ channels. Finally, a specific L-type calcium channel agonist, BayK 8644, was used to study the effect of increased calcium channel activity on membrane resealing. This study was carried out using a well-established isolated spinal cord model where the membrane integrity can be assessed both functionally and anatomically (Shi et al., 2000; Shi and Borgens, 2000; Shi and Pryor, 2000; Peasley and Shi, 2002). The findings in this study support the hypothesis that calcium influx through VDCC likely plays an important role in membrane resealing after mechanical spinal cord injury.

EXPERIMENTAL PROCEDURES

The experimental protocols have been reviewed and approved by the Purdue University Animal Care and Use Committee (PACUC). All experiments were performed ex vivo on strips of spinal cord white matter, isolated from adult, female, Hartley guinea pigs weighing 300–500g. Histological studies were performed on the same tissue that was used for physiological analysis.

Recording chamber

Variations of the double sucrose gap recording chamber have been described. Further details can be found in our previous publications (Shi and Blight, 1996; Shi and Borgens, 1999; Shi et al., 2000; Shi and Pryor, 2002). Figure 1 shows a diagram of the recording chamber which consists of two outer wells containing isotonic potassium chloride (120 mM), and a center well containing oxygenated Krebs solution, separated by two middle gaps containing isotonic sucrose solution (320 mM). A strip of isolated spinal cord white matter, approximately 35 mm in length, was supported in the central compartment and continuously supplied with oxygenated Krebs’ solution (~2 mL/min). The ends of the tissue were drawn through the sucrose gap channels and ended in side compartments filled with isotonic potassium chloride (Fig. 1). The white matter strip was sealed on either side of the sucrose gap channels, using fragments of plastic cover-slip and a small amount of silicone grease to prevent any mixing of the solutions. Isotonic sucrose solution was continuously run through the gap channels at a rate of 1 mL/min. The temperature of the chamber was maintained at 37°C with a thermostatically controlled inline heater (Warner Instruments). Axons were stimulated and compound action potentials were recorded at opposite ends of the white matter with silver-silver chloride electrodes. The central bath was connected to an instrument ground. Stimuli approximately 0.1 ms in duration were delivered in the form of constant-current unipolar pulses through a stimulus isolation unit. Recordings were made with silver/silver chloride wire electrodes positioned within the side chambers and the central bath using a bridge amplifier. Subsequent analysis was performed using custom Labview^® software (National Instruments^™) on a Compaq PC^™.

Open in a separate window

Figure 1

Diagrams showing the method for guinea pig spinal cord white matter isolation (left) recording apparatus and injury technique (right). Once the spinal cord was removed from the vertebral column it was cut longitudinally down the center axis and then subdivided again to extract only ventral white matter strips. The isolated spinal cord tract is shown mounted in the middle of the center well, which is continuously perfused with oxygenated Krebs’ solution. The two end chambers containing the ends of the spinal cord segment contained isotonic KCl. These chambers were divided from the center chamber by two thinner gaps filled with flowing, isotonic sucrose solution. Stimuli were presented at the left with axonal activity recorded at the right. The two injury models used were a transection model in which the entire white matter strip was transected using micro scissors, or a crush model in which the white matter strips were injured using a pair of modified forceps.

Transection and compression injury

Transection injuries were produced by cutting the tissue strip at the face of the recording sucrose gap, using a micro-scissors (Shi et al., 2000; Shi and Pryor, 2000). Crush injuries were performed at the face of the recording sucrose gap using blunt forceps with a spacer placed between the tips. The severity of the crush injury was controlled by varying the width of the spacer so compression could be applied accordingly (Borgens and Shi, 2000; Luo et al., 2002). In the present study, the first group of crush injury (Fig. 4) was designed so the recovery of CAP was at approximately 45% of pre-crush levels determined at 60 min. post injury. The second group of crush injury (Fig. 6) was designed so the CAP recovery was approximately 10% of pre-crush levels examined at 60 min post injury.

Open in a separate window

Figure 4

The response of the compound action potential amplitude to compression injury in 3 groups of spinal cord strips. A: Examples of CAP changes in response to compression injury in control conditions or in the presence of cadmium or nimodipine. The compound action potential was extinguished initially after the crush injury, but recovered and reached a new plateau level. Note CAP recovery was noticeably less when cadmium or nimodipine was present compared to control conditions. B: Average CAP recovery after injury in control and cadmium or nimodipine treated injury groups. The recovery plateau amplitude was approximately 2 ½ times greater in the control solution (45.61±5.18% of pre-injury, n=9) compared to the solution containing cadmium (16.87±3.69% of pre-injury, n=9) or nimodipine treated group (14.48±5.28%, n=4). The recovered CAP amplitude in both cadmium and nimodipine treated groups are significantly lowered that that in control (P<0.01, ANOVA). However, there is no significant difference between cadmium and nimodipine treated group (P>0.05, ANOVA).

Open in a separate window

Figure 6

The response of the compound action potential amplitude to compression injury in 2 groups of spinal cord strips. A: Individual examples of CAP changes in response to compression injury in control conditions or in the presence of BayK 8644 (1 μM). The compound action potential was almost extinguished after the crush injury, but recovered and reached a new plateau level. Note CAP recovery was noticeably higher when BayK 8644 (1 μM) was present compared to control conditions (2 mM Ca²⁺ and 37°C). B: Average CAP recovery after injury in control and BayK 8644 treated injury groups. The recovery plateau amplitude was approximately 3 times greater in the compression injury bathed in the Krebs’ solution containing BayK 8644 (1 μM) (9 ± 1.67% of pre, N=8) than in the control solution (33.25 ± 3.41% of pre, N=8). (P<0.001).

RESULTS

Cadmium and nimodipine inhibit membrane repair following transection

It is known that transected guinea pig spinal cord axons are capable of resealing, both functionally and anatomically, within 60 minutes of injury (Shi et al., 2000). Successful resealing of the disrupted membrane could be achieved with normal extracellular calcium concentrations (1–2 mM), and normal physiological temperature (37°C) (Shi et al., 2000; Shi and Pryor, 2000). However, we have found that when calcium channel blockers, cadmium or nimodipine were introduced to the solution, recovery of resting membrane potential following transection (expressed as % pre-transection value) is significantly inhibited (Fig. 2). Specifically, in control (transection only) conditions (2 mM Ca²⁺, 37°C), the average membrane depolarization at 60 minutes post-injury was 5.6 ± 1.5% (N = 8). In the presence of cadmium, however, the membrane depolarization was 33.1 ± 5.0% (N = 8) at the same time point. This value is significantly higher than the transection only group (P < 0.001, ANOVA). Similarly, the application of nimodipine also significantly increased the membrane depolarization to 43 ± 5.0% at the same time point post injury (P < 0.001 when compared to control, ANOVA). However, there is no significant difference between the cadmium and nimodipine treated groups (P>0.05, ANOVA). Application of cadmium or nimodipine for up to 80 minutes did not induce significant membrane depolarization or reduction of CAP amplitude in control axons (data not shown).

Open in a separate window

Figure 2

Individual examples (A) and averaged membrane depolarization potential recordings (B) in response to transection under different calcium blocking conditions. A: There was a significant decrease in the rate of recovery of membrane repolarization when cadmium (100 μm) was added to the Krebs’ solution. Membrane repolarization was inhibited further when the specific L-type calcium channel blocker nimodipine (10 μm) was added to the Krebs’ solution. B: Average membrane depolarization at 60 minutes post injury. There was a significant increase in membrane depolarization when cadmium (33.13±5.04%, N=8) or nimodipine (43.13±4.8%, N=8) was added to the perfusion solution compared to Krebs’ alone (5.63±1.52%, n=8) (P < 0.001 when compared to control in both treatment groups, ANOVA). However, there is no significant difference between cadmium and nimodipine treated group (P>0.05, ANOVA).

In addition to monitoring membrane potential after axonal transection, we also examined the status of membrane integrity using a previously described HRP-exclusion assay (Shi et al., 2000; Shi and Borgens, 2000; Shi et al., 2001). The entry of HRP has been used to indicate the loss of membrane integrity in both in vitro (or ex vivo) and in vivo studies (Pettus et al., 1994; Shi et al., 2000; Shi and Borgens, 2000; Shi et al., 2001; Shi, 2004). Consistent with the membrane recovery profile, both cadmium and nimodipine significantly increased the number of axons labeled with HRP examined at 60 minutes post-injury (Fig. 3). Specifically, the density of axons labeled with HRP in control conditions was 34 ± 8.3 axons/mm² when examined 60 minutes post-transection. The presence of cadmium increased this value to 758.4 ± 83.9 axons/mm² while the application of nimodipine increased HRP density to 1153.8 ± 203.3 axons/mm² (P < 0.01 when compared to control in both treatment groups, ANOVA). However, there is no statistical difference between cadmium and nimodipine treated group (P>0.05, ANOVA).

Open in a separate window

Figure 3

Micrographs of vibratome sections showing HRP labeling (A, B, C) and the average density of labeled axons under different conditions (D). A: Micrographs of spinal cord strips stained for HRP at 60 minutes after transection in control conditions (2mM Ca²⁺ in Krebs’ solution, 37°C). B and C: Spinal cord strips stained with HRP at 60 minutes following transection and treated with either the general calcium channel blocker cadmium (B) or the L-type specific calcium channel blocker nimodipine (C). Notice the appearance of HRP-labeled axons in both conditions indicated by the open arrows while solid arrows indicate axons which have excluded HRP stain. D: Bar graph showing the quantification of HRP-labeled axons in different conditions. Note that there is a significant increase in the density of axons stained with HRP in axons treated with cadmium (758.4±83.92 axons/mm², N=5) as well as nimodipine (1153±203.31 axons/mm², n=5) compared to control conditions (34±8.32 axons/mm², N=5). (P<0.01 when cadmium treated groups were compared to control, P<0.001 when nimodipine treated groups were compared to control, ANOVA). However, there is no statistical difference between cadmium and nimodipine treated group (P>0.05, ANOVA). Scale bar indicates a distance of 10 μm.

Calcium channel agonist influences membrane resealing and CAP recovery

In order to examine whether enhanced calcium influx through VDCC promotes membrane resealing, we used BayK 8644, a calcium agonist that is known to promote calcium influx through VDCC. We have found that, at a concentration of 1 μM, BayK 8644 accelerates membrane resealing following transection compared to control (transection only) (Fig. 5). Specifically, the time required for the membrane to recover 95% of the pre-injury resting membrane potential was significantly less in the presence of BayK 8644 (1 μM) (43.1 ± 1.62 minutes) than control (55.4 ± 1.18 minutes, P < 0.001 student’s unpaired, two tail t-test). The time required for 50% repolarization was similar in both control (14.9 ± 0.53 minutes) and BayK 8644 (1 μM) treated groups (15.1 ± 1.05 minutes, P > 0.05, n=7 in both groups, student’s unpaired, two tail t-test).

Open in a separate window

Figure 5

Individual examples and averaged quantification of membrane potential recordings in response to transection injury in 2 different conditions. A: Individual examples of membrane potential recordings in response to transection injury. There was an obvious acceleration of membrane depolarization when BayK 8644 (1 μM) was added to the Krebs’ solution compared to the control condition. B: Examples of averaged quantification of membrane potential recordings in response to transection injury. The time it took for the membrane to reach 95% repolarization was significantly less in BayK 8644 (1 μM) treated groups compared to control rates (P<0.005, N=7 for both control and BayK8644 treated groups). Note that the time required for 50% of membrane repolarization was similar in both control and BayK 8644 (1 μM) treated (P>0.05).

Similarly, BayK 8644 at 1 μM also enhanced CAP recovery following compression. In this case, a more severe compression injury was chosen so that the average CAP recovery was approximately 10% of pre-crush levels (see method). Specifically, the CAP amplitude recovered to 9 ± 1.67% of pre-injury levels at 60 minutes after compression in the control (injury only) group. When BayK 8644 1 μM was introduced to the perfusion solution, the CAP amplitude at 60 minutes post compression was 33.25 ± 3.41% of pre-injury levels, which is significantly higher than control groups (P < 0.001, student’s unpaired, two tail t-test, Fig. 6).

The next group of experiments were designed to examine the effect of high level BayK 8644 (30 μM) on membrane resealing following transection. When BayK 8644 (30 μM) was added to the solution the resultant time required for 50% of membrane repolarization (19.73 ± 2.42 minutes) did not significantly differ from control (transection only, 14.9 ± 0.53 minutes) or the group incubated with 1 μM BayK 8644 (P > 0.05, student’s unpaired, two tail t-test). However, the time it took for the membrane to repolarize 95% in the BayK 8644 (30μM) treated group (59.8 ± 8.32 min) was significantly longer than the group treated with 1 μM BayK 8644 (43.1 ± 1.62 minutes, P < 0.05, student’s unpaired, two tail t-test, Fig. 7A). Interestingly, there is no significant difference of the time required for 95% repolarization in the group with BayK 8644 30μM and control (transection only) (p > 0.05, student’s unpaired, two tail t-test, Fig. 7A). This indicates that the accelerated membrane resealing in the presence of BayK 1 μM was largely eliminated when the concentration of BayK 8644 was elevated to 30 μM.

Open in a separate window

Figure 7

Averaged membrane potential recordings and average CAP recovery in response to transection (A) and crush injury (B) in control and in the presence of either 1 μm or 30 μm Bay K 8644. A: Data with Bay K 8644 1 μM was included here for ease of comparison. The time required for the membrane to repolarize to 50% was similar between the three groups. The time it took for the membrane to repolarize to 95% of pre injury levels was not significantly different between control conditions (55.4 ± 1.18min, N=9) and BayK 8644 (30 μM) treated groups (59.8 ± 8.2min, N=9). However, the repolarization rate was significantly faster when BayK 8644 1 μM was used compared to the other 2 groups. Note that the time required for 95% membrane resealing was similar to the time it took for control axons to reseal when BayK 8644 was increased from 1 μM to 30 μM. B: Average CAP recovery 60 minutes post crush injury was relatively equal when bathed in either the control Krebs’ solution (9 ± 1.67% of pre, N=5) or the BayK 8644 (30 μM) treated solution (9.53 ± 3.01% of pre, N=9, P>0.05). Note that the CAP recovery was reduced back to control levels when BayK 8644 was increased from 1 μM to 30 μM.

A similar result was also seen in compression injuries. Specifically, the average CAP recovery at 60 minutes post crush was 9 ± 1.67% of pre-crush, which was increased to 33.25 ± 3.41% when 1 μM BayK 8644 was added to the perfusion (Fig. 6, ​,7A).7A). However, this enhanced recovery of CAP was largely eliminated when BayK 8644 was increased to 30 μM (9.53 ± 3.01%, Fig. 7B). Therefore, even though BayK at 1 μM significantly enhanced CAP recovery, there was a reversal of CAP recovery to near control levels when BayK 8644 was increased from 1μM to 30μM.

DISCUSSION

The current study examines the role of VDCC in membrane resealing by blocking (nimodipine and cadmium) or enhancing (BayK 8644) VDCC in the context of axolemmal resealing following trauma. We have found that blocking VDCC significantly suppressed membrane resealing while moderate enhancement of VDCC was associated with significant acceleration of membrane resealing. These findings seem to suggest an important role of VDCC-mediated calcium influx in axonal resealing. We believe that this is the first observation to suggest that VDCC may play an important role in membrane resealing following traumatic injury.

We have found that there is no significant difference between nimodipine (a specific L-type channel blocker) and cadmium (non-specific calcium channel blocker) treated groups in membrane potential, action potential, or HRP labeling evaluation following trauma. Furthermore, membrane resealing is enhanced by 1 μM BayK 8644, an L-type calcium channel agonist. This suggests that L-type channels plays a significant role in mediating calcium influx and promoting membrane resealing while other types of calcium channels play a minor, if any role in this regard. One piece of evidence consistent with our data is that blocking L- type calcium channels has been shown to inhibit the regeneration of injured axons after transection injury (Kulbatski et al., 2004). One possible explanation is that the axons which received a calcium channel blocker failed to reseal the cut end properly and therefore show delayed or reduced regeneration.

Currently, it is not clear how the calcium influx through VDCC would contribute to the membrane resealing. One possibility is that this may relate to vesicle transport, a calcium-dependent process that is involved in membrane resealing (Steinhardt et al., 1994; Eddleman et al., 1997; Terasaki et al., 1997; Detrait et al., 2000; McNeil and Terasaki, 2001; McNeil and Steinhardt, 2003). It is possible that, in order for successful membrane resealing to occur, a critical number of vesicles must be produced and placed at the injury site. Membrane vesicles are likely formed not just near the focus of the disruption, but also at some distance along the axolemmal and subsequently are transported to the injury site. Indeed, it is well established vesicles and other cellular materials that are transported down the axon by fast and slow axonal transport are produced continuously at the soma, and in the case of axonal disruption, tend to accumulate at the site of blockage (Povlishock, 1992). It is likely that calcium influx through the injury site does not flow very far along the axon due to intrinsic cellular ability to impede diffusion of calcium (Ziv and Spira, 1993; Terasaki et al., 1997). Therefore, it is possible that calcium influx through VDCC ensures adequate calcium influx along the entire length of the damaged axon in response to axonal injury. Subsequently, the elevated calcium level along the axons will ensure the efficient transport of vesicles to the site of membrane disruption. One piece of evidence that supports this hypothesis is that calcium ionophores can enhance the overall number of vesicles near the membrane and induce considerable surface enlargement, both of which are known to enhance membrane resealing (Coorssen et al., 1996; Rodriguez et al., 1997; Cocucci et al., 2004).

Despite the importance of calcium influx in the axonal membrane resealing, it is understandable that excessive accumulation of intracellular calcium could be harmful. It is well established that uncontrolled high levels of calcium can trigger axonal degeneration (Schlaepfer and Bunge, 1973; Schlaepfer, 1974; Schanne et al., 1979; George et al., 1995; Ueda et al., 1997; Cano-Abad et al., 2001). This may explain why the enhanced axonal resealing in the presence of BayK 8644 at 1 μM was largely overcome when the concentration was increased to 30 μM. Interestingly, the biphasic effect of calcium influx through VDCC was also seen in a study done by Kulbatski et al. (Kulbatski et al., 2004). In their study, while 1 μM BayK8644 aided in neurite regeneration, this beneficial effect was largely eliminated or reversed when higher concentrations (up to 30 μM) were used.

Even though calcium influx through VDCC appears to be important for resealing, it does not mean that this will be beneficial in other types of injury. The overall effect of calcium influx through VDCC probably depends on whether mechanical membrane disruption is the primary cause of the injury. For example, unlike mechanical injury where membrane disruption is the key initial pathology, ischemic injury stems from the lack of oxygen and other life-supporting chemicals with little membrane disruption initially (Shi et al., 2000; Shi and Borgens, 2000; Shi and Pryor, 2000; Peasley and Shi, 2002). As a result, the axons depolarize slowly which causes prolonged calcium influx and degradation along an extensive length of axon from which no amount of resealing can protect them. Therefore, the net effect of calcium influx in ischemic insult is harmful which can elicit many pathological process leading to cell death. This may explain why calcium channel blockers are beneficial in ischemic injury (Fern et al., 1995; George et al., 1995; Sattler et al., 1996; Imaizumi et al., 1999; Agrawal et al., 2000), but not in transection injury where the membrane disruption is the primary injury mechanism and membrane resealing is the key to stop uncontrolled influx of calcium (Gross et al., 1983; Lucas et al., 1985; Yawo and Kuno, 1985; Shi et al., 1989; Xie and Barrett, 1991; Shi and Borgens, 1999; Shi et al., 2000; Shi and Borgens, 2000).

Another factor that supports the beneficial effects of Ca²⁺ following mechanical injuries is that intracellular Ca²⁺ activity is actually depressed following physical insults. For example, in traumatic injury, degradative processes are set in motion to a greater or lesser degree which result in binding most intracellular Ca²⁺. As elaborated by Young (Young, 1992) phospholipids are degraded after injury and bind to Ca²⁺, making intracellular Ca²⁺ very low. This could explain why adequate Ca²⁺ (sufficient to promote vesicle fusion with axon membranes but not so high as to exacerbate the degradative processes) is beneficial.

One complication of the current study regarding cadmium-induced impairment in membrane repair is the possible toxicity of cadmium in CNS axons as demonstrated by Fern and his colleagues (Fern et al., 1996). However, we believe this is unlikely. The main reason is that the application of cadmium for 80 minutes did not show any significant functional impairment in uninjured guinea pig spinal cord white matter (current study). In addition, the dosage of cadmium used in this study was only half of what was demonstrated to be harmful to rat optic nerve (Fern et al., 1996). Interestingly, cadmium at a dosage that is 60% of that used in the current study and applied at a duration that is 20 minutes longer than that in this study did not cause significant reduction of CAP amplitude in rat optic nerve (Fern et al., 1996). Therefore, it is more than likely that the effect of cadmium in this study was mainly due to its ability to block VDCC, and not the non-specific toxicity.

It is well documented that prolonged increases in intracellular calcium lead to neuronal degeneration in trauma and diseases (Schlaepfer and Bunge, 1973; Schlaepfer, 1974; Schanne et al., 1979; George et al., 1995; Ueda et al., 1997; Cano-Abad et al., 2001). Because of this, it has been a continuous effort to design therapies for spinal cord injury and traumatic brain injury that either block calcium from entering the cell, or work to maintain a low level of calcium at the injury site. Our study, along with many others, offers a balancing view regarding the role of calcium in overall response to mechanical trauma. It argues a temporal and spatial differential role of calcium in the cellular response to mechanical trauma. More specifically, calcium may be beneficial, or even crucial for membrane resealing following mechanical insults.

Based on our previous studies, it takes a minimum of 30–60 minutes to complete the resealing process in guinea pig spinal cord axons following trauma in vitro (in 2 mM extracellular Ca²⁺) (Shi et al., 2000). However, considering the fact that extracellular Ca²⁺ drops below the level that supports effective membrane resealing for several hours following spinal cord trauma (Young et al., 1982; Stokes et al., 1983), it may be considerably longer than 60 minutes under in vivo conditions for effective membrane resealing to occur. Therefore, we suggest that treatments designed to lower extracellular calcium concentrations or aimed to block calcium entry into the cell may be better suited if used an hour post injury, when the membrane integrity is presumably reestablished. This will avoid interfering with the resealing process while preventing prolonged elevation of cytosolic calcium that is known to be detrimental (Schlaepfer and Bunge, 1973; Schlaepfer, 1974; Schanne et al., 1979; George et al., 1995; Ueda et al., 1997; Cano-Abad et al., 2001). It is our desire in the current study to first stress the notion of the differential role of calcium in the recovery process and second, point out that it is in the period immediately following the insult that intracellular calcium is most critical. This knowledge may be important in designing new approaches or optimizing the effectiveness of existing methods in the treatment of CNS trauma.

Acknowledgments

Comprehensive list of abbreviations

Footnotes

References