General Anesthesia Blocks Pain-Induced Hemorrhage and Locomotor Deficits After Spinal Cord Injury in Rats

Abstract

Research has shown that engaging pain (nociceptive) pathways after spinal cord injury (SCI) aggravates secondary injury and undermines locomotor recovery. This is significant because SCI is commonly accompanied by additional tissue damage (polytrauma) that drives nociceptive activity. Cutting communication with the brain by means of a surgical transection, or pharmacologically transecting the cord by slowly infusing a sodium channel blocker (lidocaine) rostral to a thoracic contusion, blocks pain-induced hemorrhage. These observations suggest that the adverse effect of pain after SCI depends on supraspinal (brain) systems. We hypothesize that inhibiting brain activity using a general anesthetic (e.g., pentobarbital, isoflurane) should have a protective effect. The present study shows that placing rats in an anesthetic state with pentobarbital or isoflurane 24 h after a lower thoracic contusion injury blocks pain-induced intraspinal inflammation and hemorrhage when administered before pain. Pentobarbital also extends protective effects against locomotor deficits produced by noxious stimulation. Inducing anesthesia after noxious stimulation, however, has no effect. Similarly, subanesthetic dosages of pentobarbital were also ineffective at blocking pain-induced hemorrhage. Also examined were the hemodynamic impacts of both pain and anesthetic delivery after SCI. Peripheral pain-input induced an acute increase in systolic blood pressure; isoflurane and pentobarbital prevent this increase, which may contribute to the protective effect of anesthesia. The results suggest that placing patients with SCI in a state akin to a medically induced coma can have a protective effect that blocks the adverse effects of pain.

Introduction

The long-term consequences of spinal cord injury (SCI) are determined by both the tissue loss incurred at the time of injury (primary injury) and the processes that unfold over the hours to days afterward (secondary injury). Secondary injury is attributable to biological cascades and pathophysiological processes that enable the infiltration of blood (hemorrhage) at the site of injury, activate inflammatory processes, and activate signal pathways that drive cell death. These processes are initiated by the primary insult and can continue to unfold long after the initial injury.

Previous research has shown that engaging pain (nociceptive) pathways after injury can increase tissue loss, foster the development of spasticity and chronic pain, and impair locomotor recovery.^1-6 This is clinically important because the three most common sources of SCI in the United States are vehicle accidents (38.3% of SCI), falls (31.6% of SCI), and incidents of violence (13.8% of SCI).

In these cases, additional tissue damage (polytrauma) to the periphery is common. Indeed, up to 80% of injuries are incurred alongside some form of additional tissue damage⁸ and surgical interventions can introduce even more tissue damage. This polytrauma, whether traumatic or medical in nature, can serve as a source of peripheral pain input and impact tissue loss at the site of injury.

Our laboratory has explored the adverse effect of noxious stimulation on tissue loss and recovery using rats that have received a moderate thoracic (T12) contusion injury. Nociceptive fibers can be engaged using electrical stimulation (shock) or the peripheral application of capsaicin, which engages pain fibers that express the TRPV1 receptor. Just six minutes of intermittent noxious electrical stimulation to the tail or a single application of capsaicin to one hindpaw 24 h after injury is sufficient to drive increased inflammation and hemorrhage at the injury site.^3,5,9

Further work has shown that noxious stimulation caudal to the SCI leads to a breakdown of the blood-spinal cord barrier (BSCB) that enables the infiltration of blood. Because some bloodborne contents are neurotoxic,¹⁰ this hemorrhage fuels cell death, inflammation, and additional tissue loss. We have shown that blocking neural activity at the site of injury by applying the sodium-channel blocker lidocaine before noxious stimulation attenuates pain-induced hemorrhage and its adverse effect on recovery.¹¹

Other research has explored the effect of disrupting communication with the brain by means of a surgical transection at T2, rostral to the contusion injury. Because descending serotonergic fibers are known to quell overexcitation in response to noxious stimulation,² we hypothesized that a T2 transection would amplify nociception-induced hemorrhage. We found the opposite—a rostral transection blocked nociception-induced hemorrhage and inflammation.^6,12

Pharmacologically blocking ascending/descending neural activity at T2, by slowly infusing lidocaine before noxious stimulation, had a similar protective effect.¹¹ Importantly, infusing lidocaine rostral to injury also blocked the adverse effect noxious stimulation has on long-term recovery. These observations suggest that neural systems rostral to T2—presumably related to brain activity—play a critical role in driving pain-induced tissue loss after SCI.

If brain processes mediate pain-induced hemorrhage and the disruption in long-term recovery after SCI, then treatments that inhibit brain activity should have a protective effect. The present study shows that placing animals in a state akin to a medically induced coma, using the general anesthetic pentobarbital has a protective effect. A gas anesthetic (isoflurane) also blocked nociception-induced hemorrhage. Inducing an anesthetic state after noxious stimulation had no effect.

Methods

Subjects

Adult male Sprague-Dawley rats weighing 300-400 g were used for all experiments. Animals were maintained on a 12-h light-dark cycle. Before contusion surgery, rats were pair-housed and given ad libitum access to food and water. All rats were acclimated to handling and the behavior apparatus before testing. The experiments were performed in accordance with National Institute of Health (NIH) guidelines for care and use of laboratory animals and were approved by the Animal Care and Use Committee at Texas A&M University. Every effort was made to minimize the use and suffering of laboratory animal subjects.

Surgery

For all experiments, animals received a contusion injury at the T10-11 spinal level. To induce an anesthetic state, a 5% isoflurane/oxygen mixture was used. This anesthetic mixture was reduced to 2-3% during surgery. A 3 cm longitudinal incision was made through the skin along the midline.

Tissue was removed around the spinous processes of the T11-T12 area, exposing the spinal cord. The vertebral column was secured to a MASCIS device using clamps on either side of the surgical site.¹³ A 10 g weight was dropped directly onto the spinal cord from a 12.5 mm height. The surgical site was then closed with Michel clips. After surgery was complete, animals received supplemental injections of saline (3 mL) as well as penicillin (100,000 units/kg) to replace fluids lost during surgery and to prevent infection. Animals were placed in recovery housing overnight.

Noxious electrical stimulation

A day after the contusion injury, animals in Experiments 1-3 and 5-6 were given uncontrollable electrical noxious stimulation of the tail (for details, see Grau et al, 2004). Briefly, rats were secured in a Plexiglas tube and fitted with a modified fuse clip coated with electrode paste (Harvard Apparatus). A constant current AC shock source (660V AC) was used to generate intermittent electrical stimulation 100 msec in duration on a variable schedule (0.2-3.8 sec interstimulus interval) for a total of 6 min.

Capsaicin treatment

A day after injury, animals in Experiment 4 were secured in Plexiglas tubes with their hindlimbs exposed. They then received an intradermal injection of a 3% capsaicin solution (0.05 mL) or its vehicle to the dorsal surface of one hind paw. Injection side (left versus right hind paw) was counterbalanced across subjects. Rats were restrained in Plexiglas tubing for 6 min after injection.

Drug treatment

Pentobarbital

Rats in Experiments 1, 2, 4, and 5 were treated with 50 mg/kg pentobarbital or its vehicle through i.p. injection. In Experiment 3 animals received 0, 12.5, 25, or 50 mg/kg of pentobarbital i.p. In Experiments 1–5, pentobarbital was administered 30 min before pain stimulation to achieve an anesthetic state. In Experiment 5, half of the animals received pentobarbital (50 mg/kg) or its vehicle immediately after pain stimulation. Pentobarbital powder (Sigma-Aldrich; Saint Louis, MO) was mixed in a vehicle containing alcohol (10%), propylene glycol (40%), and water (50%).¹⁴

Isoflurane

Rats in Experiment 6 had an anesthetic state induced using a 5% isoflurane/oxygen mixture or just oxygen for 5 min immediately before pain stimulation. Once an anesthetic state was reached, the isoflurane concentration was dropped to 3% and maintained for the 6-min period.

Locomotor recovery

Locomotor performance was evaluated using the Basso, Beattie, and Bresnahan (BBB) locomotor rating scale.¹⁵ Animals were allowed to move freely throughout an open field while a pair of researchers evaluated movement. Intraobserver reliability was high (all r's > 0.94), and observers were blinded to treatment conditions. Scores were converted to a form more amenable to parametric analysis.¹⁶

In Experiment 1, locomotor performance was assessed once per day for the first week and on days 9, 11, 13, 15, 18, and 21. Coordination was assessed at the end of the recovery period using the beam walk¹⁷ and the ladder¹⁸ tests.

Tissue collection

At the end of all experimental procedures, rats were sacrificed with a 100 mg/kg i.p. injection of pentobarbital. In Experiments 2-6, spinal cord tissue enveloping the injury site (1 cm) was collected and immediately frozen in liquid nitrogen. The tissue was then processed, and protein was extracted using Qiazol lysis reagent according to manufacturer instructions. To ensure uniform protein concentration across samples, a Bradford assay was performed (BioRad, Hercules, CA), and samples were diluted in Laemmli sample buffer.

Spectrophotometry

Spectrophotometric analysis (Nanodrop, Thermo Scientific) was performed on the protein sample and absorbance of light was quantified at the wavelength (420 nm) associated with oxygenated hemoglobin.¹⁹

Immunoblotting

Protein samples diluted in Laemmli buffer were subjected to gel electrophoresis using SDS PAGE. Each sample was heated at 96 C° for 5 min and was centrifuged for 3-5 sec. Ten μL of each sample were then loaded into the wells of a 12% pre-cast tris-HCL gel (BioRad, Hercules, CA) as well as 5 μL of biotinylated ladder (Biorad). Gels were subjected to electrophoresis at 100 V for 2h.

After gel electrophoresis, samples were transferred from the gel onto a polyvinylidene difluoride membrane (PVDF; Millipore, Bedford, MA) according to manufacturer instructions. A gel-cassette complex was formed using fiber pads, filter paper, the gel, and the PVDF membrane. Afterward, the complex was placed within the transfer tank and proteins were transferred from the gel to the membrane at 100V for 1h.

After the transfer process was completed, membranes were blocked in 5% milk (BioRad, Hercules, CA) diluted in 1% Tris-buffered saline Tween-20(TBST) for 1h. Afterward, blots were allowed to incubate overnight in primary antibodies on an orbital shaker at 4 C°. Antibodies were used for hemoglobin alpha chain proteins (1:1000; Abcam [Cambridge, MA] ab92492, RRID: AB10561594), Il-1ß (1:1000; Novus Biologicals [Centennial, CO] NB600-633, RRID: AB10001060) and Il-18 (1:1000; Invitrogen [Carlsbad, CA] PA5-79482, RRID: AB2746598).

After primary antibody incubation, membranes were washed in TBST for 10 min, three times each, at room temperature. Each membrane was then incubated in secondary antibody compatible with the primary antibody they were initially incubated in: HRP-conjugated goat anti-rabbit secondary antibodies [1:5000; Sigma-Aldrich (Saint Louis, Missouri) ab258649, RRID: AB228341]. After incubating in secondary antibody for 1h, membranes were again washed in TBST for 10 min, three times each, at room temperature. Afterward, membranes were washed in 1% Tris-buffered saline (TBS) for 5 min. Membranes were then treated with clarity reagent electrochemiluminescence substrate kit (ECL; Pierce, Rockford, IL) for one min. Afterward, treatment with the reagent, membranes were imaged using a Fluorchem HD2 (ProteinSimple, Santa Clara, CA).

Experimental designs

Experiment 1

We first examined whether pentobarbital anesthesia blocks the adverse effect of noxious stimulation on locomotor recovery (Fig. 1A). Twenty-four hours after a lower thoracic spinal contusion, animals were assigned to one of four experimental groups (vehicle shock, vehicle unshock, pentobarbital shock, pentobarbital unshock) with baseline locomotor performance roughly matched across groups. Animals were then treated with anesthesia and noxious stimulation according to their grouping.

FIG. 1.

Pentobarbital anesthesia blocked the adverse effect of noxious stimulation on recovery. (A) Experimental design. Noxious electric stimulation impaired locomotor recovery in animals pre-treated with the vehicle (B) but not animals given pentobarbital (C). Likewise, shock disrupted the recovery of weight in animals given the vehicle (D) but not pentobarbital-treated rats (E). At the end of the recovery period, performance was disrupted on the ladder (F) and beam walk (G) tasks in the vehicle-treated shocked animals, but not in pentobarbital-treated rats. *Indicates statistical significance (p < 0.05). Error bars represent standard error of the mean (n = 10, 40 total animals).

Rats received either pentobarbital (50 mg/kg) or its vehicle through i.p. injection. Thirty minutes later, rats were placed in the restraining tubes and exposed to either uncontrollable electrical stimulation to the tail or nothing. Locomotor performance was then monitored over the course of 21 days. Motor coordination was assessed at the end of the recovery period using the beam walk and ladder tasks.

Experiment 2

We then tested whether pentobarbital administration attenuated nociception-induced hemorrhage in contused rats. Rats received a contusion injury and were allowed to recover for 24 h (Fig. 2A). Afterward, animals had their baseline blood pressure and locomotor scores evaluated.

FIG. 2.

Pentobarbital blocked shock-induced hemorrhage when administered before noxious shock stimulation. (A) Experimental design. (B) Exposure to noxious stimulation (shock) increased absorbance at the wavelength associated with hemoglobin (420 nm) in vehicle-treated rats, but not in those that received pentobarbital. (C) A similar pattern was observed when hemoglobin was assayed using Western blotting. (D) Nociceptive stimulation produced an increase in systolic blood pressure. Pentobarbital administration blocked this effect. Western blotting showed that exposure to noxious stimulation increased the expression of interleukin (IL)-1ß (E) and IL-18 (F) at the site of injury in vehicle, but not pentobarbital, treated rats. *Indicates statistical significance (p < 0.05). Error bars represent standard error of the mean (n = 8, 32 total animals).

Based on these scores, rats were assigned to four experimental groups (vehicle shock, vehicle unshock, pentobarbital shock, pentobarbital unshock) in a balanced manner. Animals were then treated with anesthesia and pain stimulation as described in Experiment 1. Immediately after noxious stimulation [time 0], and every hour after stimulation for three hours (1–3), non-invasive blood pressure recordings were taken. Rats were then sacrificed, and tissue was collected as described above.

Experiment 3

Next, we examined the effect of pentobarbital treatment on pain-induced hemorrhage across a range of doses, from 12.5 mg/kg (subanesthetic) to 50 mg/kg (deep anesthesia) (Fig. 3A). A day after spinal contusion injury, animals received 0, 12.5, 25, or 50 mg/kg of pentobarbital (i.p.). Thirty minutes later, half the animals were exposed to nociceptive stimulation. Three hours later rats were sacrificed, and tissue was extracted and processed.

FIG. 3.

A subanesthetic dose of pentobarbital did not protect against noxious shock- induced hemorrhage. (A) Experimental design. Pre-treatment with 25–50 mg/kg of pentobarbital attenuated shock-induced hemorrhage at the site of injury as measured by spectrophotometry (B) and Western blotting (C). *Indicates statistical significance (p < 0.05). Error bars represent standard error of the mean (n = 8, 32 total animals).

Experiment 4

We also examined whether pentobarbital treatment attenuates capsaicin-induced hemorrhage after SCI (Fig. 4A). As in previous experiments, rats were contused and allowed to recover. The next day, baseline blood pressure and locomotor performance were tested and rats were assigned to one of four different experimental groups (vehicle capsaicin, vehicle vehicle, pentobarbital capsaicin, pentobarbital vehicle). Rats were then administered pentobarbital (50 mg/kg) or its vehicle. After 30 min, rats received an injection of capsaicin or its vehicle into the hindpaw, with injection side counterbalanced across conditions. Rats remained in the restraint tube for 6 min. Blood pressure was then reassessed as described in Experiment 2. After the third hour, rats were sacrificed, and tissue was collected and processed.

FIG. 4.

Pre-treatment with pentobarbital blocked capsaicin-induced hemorrhage. (A) Experimental design. Both spectrophotometry (B) and Western blotting (C) showed that treatment with the irritant capsaicin increased hemorrhage at the site of injury in the vehicle-treated controls, but not in rats given pentobarbital. (D) Capsaicin induced a rise in systolic blood pressure in vehicle-treated rats. Pentobarbital attenuated this effect and generally lowered blood pressure in animals given the capsaicin vehicle (Cap-Vehicle). Western blotting showed that capsaicin treatment increased expression of interleukin (IL)-1β (E) and IL-18 (F) in vehicle-treated rats, but not in animals given pentobarbital. *Indicates statistical significance (p < 0.05). Error bars represent standard error of the mean (n = 8 Capsaicin groups, n = 10 Vehicle Groups, 36 total animals).

Experiment 5

We also tested whether inducing a state of anesthesia after painful stimulation attenuates the development of hemorrhage (Fig. 5A). Animals were given a contusion injury and allowed to recover over 24 h. Afterward, baseline blood pressure and locomotor performance were recorded, and rats were assigned to experimental groups (pentobarbital before shock, pentobarbital after shock, vehicle before shock, vehicle after shock). Animals were then treated with pentobarbital (50 mg/kg) or vehicle either 30 min before or immediately after noxious electrical stimulation. Blood pressure readings were recorded as described in previous experiments. After three hours, rats were sacrificed, and tissue was collected and processed.

FIG. 5.

Pentobarbital given before shock, but not after, attenuated hemorrhage. (A) Experimental design. (B) Spectrophotometry showed that rats given pentobarbital before shock treatment had less hemorrhage. This effect was not observed when pentobarbital was given after shock treatment. (C) A similar pattern was observed when hemoglobin was assessed using Western blotting. *Indicates statistical significance (p < 0.05). Error bars represent standard error of the mean (n = 8, 32 total animals).

Experiment 6

We concluded by testing the impact of an inhaled anesthetic drug (isoflurane) (Fig. 6A). A day after rats received a contusion injury, baseline blood pressure and locomotor readings were taken and rats were assigned to experimental groups (isoflurane shock, isoflurane unshock, vehicle shock, vehicle unshock). Rats were exposed to a 5% isoflurane/oxygen mix or oxygen alone for 5 min to induce an anesthetic state.

FIG. 6.

Isoflurane administration blocked shock-induced hemorrhage. (A) Experimental design. Both spectrophotometry (B) and Western blotting (C) showed that shock treatment induced hemorrhage in unanesthetized (Vehicle) animals. Inducing an anesthetic state with isoflurane blocked this effect. (D) Isoflurane attenuated systolic blood pressure. Exposure to shock increased expression of IL-1ß (E) and IL-18 (F) in unanesthetized (Vehicle) rats. Isoflurane anesthesia blocked this effect. *Indicates statistical significance (p < 0.05). Error bars represent standard error of the mean (n = 8, 32 total animals).

After induction, anesthetic/oxygen exposure was continued at 3% while rats were treated with noxious stimulation or an equivalent period of restraint. Blood pressure readings were taken over the subsequent 3h. At the end of the 3h period, rats were sacrificed, and tissue was collected and processed as described above.

Statistics

Data were analyzed with analysis of variance (ANOVA) and analysis of covariance (ANCOVA). Post hoc comparisons were performed using the Duncan New Multiple Range test. In all cases, a criterion of p < 0.05 was set as the threshold for statistical significance.

Results

Experiment 1: Pentobarbital blocks the induction of locomotor deficits after noxious stimulation

Peripheral pain input administered shortly after SCI has been shown to impair recovery. Disrupting brain-spinal cord communication by slowly infusing lidocaine rostral to the contusion injury blocks this effect.¹¹ What is not known is whether inhibiting brain systems would be sufficient to produce the same protective effect.

To explore this issue, the general anesthetic pentobarbital was administered to produce a state similar to a medically induced coma before noxious stimulation, and behavioral recovery was assessed (Fig. 1A). Animals were given a contusion injury as described above and were allowed to recover overnight. Animals were then given a dosage of pentobarbital or its vehicle 30 min before noxious electric stimulation. Functional recovery was monitored over the next 21 days.

Recovery of locomotor function was evaluated using the BBB scale of hindlimb movement as described above. In vehicle-treated animals, electrical stimulation resulted in a significant decrease in locomotor recovery relative to unshocked controls. Pentobarbital administration before shock blocked the expression of this deficit (Fig. 1B, C). An ANOVA performed on baseline BBB scores, before drug or shock treatment, confirmed that locomotor performance did not differ (All F's < 1.0, p > 0.05). An ANCOVA of BBB scores across days after treatment, using initial scores as a covariate, revealed a main effects of drug treatment [F (1,36) = 28.839, p < 0.01). In addition, the interaction term showed that the effect of shock depended on drug treatment (F (1,35) = 10.943, p < 0.01).

Post hoc comparisons of the group means confirmed that vehicle shocked group was lower than other three. In addition, the pentobarbital shocked group had significantly lower scores compared with vehicle unshocked animals (p < 0.05). Weight was monitored as a measure of overall animal health across time. Noxious stimulation undermined weight gain in the vehicle controls (Fig. 1D), but not pentobarbital treated rats (Fig. 1E).

An ANOVA performed on baseline weights, before shock/pentobarbital treatment revealed no significant effects (F < 1.0, p > 0.05). An ANCOVA performed on weights across days after treatment revealed a main effect of drug treatment (F (1,35) = 7.4237, p < 0.05). No other terms reach significance (all F's < 1.0, p > 0.05). Post hoc comparisons of the group means confirmed that saline-shocked group differed from all other groups (p < 0.05).

Additional assessments of locomotor performance were performed at the end of the recovery period using the ladder (Fig. 1F) and beam walk (Fig. 1G) assays of coordination and balance. In both tests, animals that received the vehicle before shock performed poorly relative to their unshocked counterparts. Pentobarbital blocked this effect (Fig. 1 C).

An ANOVA performed on scores from the beam walk (Fig.1F) assay showed a main effect of drug administration (F (1,36) = 6.366, p < 0.05) and an interaction between drug administration and shock treatment that approached significance (F (1,36) = 3.722, p < 0.068). No other term was significant (F (3,36) = 2.665, p > 0.05].

Post hoc comparison of group means confirmed that the vehicle shocked group differed from all other groups (p < 0.05). An ANOVA performed on ladder scores (Fig. 1G) revealed an interaction between drug administration and shock treatment (F (1,36) = 5.066, p < 0.05). No other terms were significant (all F's < 1.0, p > 0.05). Post hoc analysis confirmed that vehicle-shocked animals significantly differed from vehicle-unshocked animals (p < 0.05). In summary, pentobarbital ameliorated the adverse effect of acute pain on long-term recovery.

Experiment 2: Pentobarbital blocks the expression of shock-induced hemorrhage

Pentobarbital administered before nociceptive stimulation blocked its adverse effect on long-term recovery. Here, we explored whether pentobarbital administration before pain input also blocks nociception-induced hemorrhage at the injury site (Fig. 2A).

Light absorbance data at 420 nm demonstrated a protective effect of pentobarbital administration against electrical stimulation. Vehicle-treated animals given uncontrollable noxious stimulation showed increased hemorrhage at the site of injury relative to unshocked controls.

Pentobarbital treatment blocked the shock-induced increase in hemorrhage (Fig. 2B). An ANOVA conducted on the absorbance data revealed a main effect of drug administration and an interaction between shock treatment and drug administration (both F's < 6.917, p < 0.05). No other terms reached significance (F (3,28) = 1.884, p > 0.05). Post hoc comparisons of the group means confirmed that vehicle-shocked animals differed from all other groups (p < 0.05).

A similar pattern of results was obtained when hemoglobin was assessed using Western blotting. Animals treated with shock had higher levels of hemoglobin in protein samples when compared with unshocked controls. Pentobarbital blocked this effect (Fig. 2C). An ANOVA showed that the effect of shock depended on drug treatment, (F (1,28) = 4.441, p < 0.05). Post hoc comparisons revealed that animals treated with vehicle and shock differed from vehicle-unshock and pentobarbital-shock groups (p < 0.05).

Measurement of blood pressure over time after nociceptive stimulation demonstrated an increase in systolic blood pressure that persisted up to 3h after pain stimulation. Anesthetic administration blocked this increase (Fig. 2D). An ANOVA of baseline (before drug/shock treatment) blood pressure readings suggest no group difference (F < 1.0, p > 0.05).

An ANCOVA of the blood pressure data collapsed across post-stimulation time points revealed a main effect of drug treatment (F (1,28) = 8.889, p < 0.05). Post hoc comparison of group means revealed that vehicle-treated shocked animals differed from all other groups (p < 0.05).

Previous work has demonstrated that peripheral pain input can increase expression of inflammatory cytokines.^3,5,9 Given this, Western blotting was used to assess the expression of interleukin (IL)-1ß and IL-18 (Fig. 2E, F). Shock increased expression of IL-1β and pentobarbital blocked this effect. ANOVAs run on data sets from each cytokine assay revealed a main effect of drug administration and an interaction between drug and shock administration (all F's < 8.93, p < 0.05).

Post hoc comparisons of the group means showed that vehicle-shocked animals differed from all other groups for both assays (p < 0.05). While the overall pattern of results for IL-18 was similar, there were no statistically significant effects of shock or pentobarbital treatment (all F's < 2.95, p > 0.05).

Experiment 3: A subanesthetic dose of pentobarbital does not block the expression of shock-induced hemorrhage

Pentobarbital anesthesia attenuated the adverse effect of peripheral pain on hemorrhage and recovery after SCI. It is unknown, however, whether pentobarbital must be delivered at a dosage that induces a full anesthetic state. Experiment 3 explored this issue by evaluating the effect of pentobarbital treatment across a range of dosages (Fig. 3A). The doses included subanesthetic (12.5 mg/kg), low level anesthesia (25 mg/kg), and deep anesthesia (50 mg/kg).

Light absorbance data at 420 nm confirmed that pentobarbital had blocked noxious stimulation-induced increases in hemorrhage when delivered at either 50 or 25 mg/kg, but not at 12.5 mg/kg. As before, anesthetic treatment with pentobarbital before nociceptive stimulation resulted in reduced hemorrhage relative to vehicle controls (Fig. 3B).

An ANOVA conducted on absorbance data confirmed a significant difference in group means (F (3,28) = 3.810, p < 0.05). Post hoc comparisons showed that 50 mg/kg and 25 mg/kg pentobarbital-treated animals differed from the vehicle-treated counterparts. No other comparisons reached significance (p > 0.05).

Similar results were obtained when hemoglobin was assayed using Western blotting. Higher dosages (25–50 mg/kg), but not a lower dosage (12.5 mg/kg) of pentobarbital blocked pain-induced hemorrhage (Fig. 3C). An ANOVA showed a main effect of drug treatment, (F (1,28) = 3.730, p < 0.05). No other term was significant (p > 0.05). Post hoc comparisons demonstrated the 50 mg/kg and 25 mg/kg treated groups differed from the vehicle group (p < 0.05).

Experiment 4: Pentobarbital blocks the expression of capsaicin-induced hemorrhage

To evaluate the generality of our results, we tested whether pentobarbital anesthesia blocks hemorrhage elicited by another form of noxious stimulation elicited by applying the irritant capsaicin to one hindpaw (Fig. 4A). Capsaicin produced an increase in hemoglobin-associated absorbance. Pentobarbital administration blocked this effect (Fig. 4B). An ANOVA revealed a main effect of both drug and capsaicin treatment as well as an interaction effect between these variables (all F's > 5.268, p < 0.05). Post hoc comparisons showed that unanesthetized animals treated with capsaicin differed from all other groups (p < 0.05).

A similar pattern of results was obtained when hemoglobin was assessed using western blotting. Unanesthetized animals exposed to capsaicin had higher levels of hemoglobin in protein samples when compared to no pain controls. Pentobarbital blocked this effect (Fig. 4C). An ANOVA confirmed an interaction effect between capsaicin and drug treatment, (F (1,28) = 4.480, p < 0.05). No other terms reached significance, (both F's < 4.656, p > 0.05). Post hoc comparisons revealed that unanesthetized animals treated with capsaicin differed from all other groups (p < 0.05).

Capsaicin also induced an increase in blood pressure (Fig. 4D). As before, this effect was blocked by pentobarbital administration. An ANOVA found no differences in pre-treatment systolic blood pressure, (F < 1.0, p > 0.05). An ANCOVA of post-treatment systolic blood pressure collapsed across all time points demonstrated a main effect of capsaicin application as well as an interaction between drug administration and blood pressure time point, (Both F's > 4.575, p < 0.05). Post hoc comparisons showed that unanesthetized animals treated with capsaicin differed from animals treated with capsaicin and pentobarbital, and these groups differed from the unanesthetized group that did not receive capsaicin (p < 0.05).

Western blotting was performed to assess the expression of inflammatory cytokines. Capsaicin increased expression of IL-18 and IL-1β, while administration of pentobarbital blocked this effect (Fig. 4E, F). An ANOVA on the IL-1β results revealed an interaction between drug and capsaicin administration, (F (1,32) = 4.412, p < 0.05). The main effect of drug administration approached significance, (F (1,32) = 3.811, p = 0.0597). Post hoc comparisons showed that unanesthetized animals treated with capsaicin differed from all other groups (p < 0.05). Likewise, for IL-18 an ANOVA revealed a main effect of capsaicin administration, (F (1,32) = 5.800, p < 0.05). Post hoc comparisons showed that animals treated with capsaicin and vehicle differed from all vehicle-vehicle and pentobarbital-vehicle groups (p < 0.05). The overall pattern of inflammatory cytokine expression mirrors the findings reported above for shock treatment, with a more robust (and statistically significant) effect observed for IL-18.

Experiment 5: Pentobarbital before, but not after, noxious stimulation attenuates hemorrhage

We have shown that inducing an anesthetic state blocks pain-induced hemorrhage after SCI when administered before acute pain onset. Our next experiment tested whether inducing anesthesia immediately after noxious stimulation counters the development of hemorrhage (Fig. 5A).

Spectrophotometry showed that pentobarbital given before shock attenuated hemorrhage, relative to vehicle treated rats (Fig. 5B). Pentobarbital given after shock had no effect. An ANOVA revealed a main effect of injection time, (F (1,28) = 6.353, p < 0.05) and an interaction between drug and timing of administration, (both F's > 9.242, p < 0.05). The main effect of drug administration approached significance, (F (1,28) = 3.225, p = 0.0833). Post hoc comparisons demonstrated that the pentobarbital before shock group differed from the other three (p < 0.05).

Similar results were obtained when hemoglobin was assayed using Western blotting. Pentobarbital blocked the noxious stimulation-induced hemorrhage when administered before shock, but not afterward (Fig. 5C). An ANOVA showed a main effect of time of injection, (F (1,28) = 8.075, p < 0.05). The interaction between drug and timing of treatment approached significance, (F (1,28) = 3.830, p = 0.0604). No other term was significant (p > 0.05). Post hoc comparisons showed that the pentobarbital-before-shock group differed from animals that received pentobarbital or vehicle afterward (p < 0.05).

Experiment 6: Isoflurane blocked the expression of electric shock-induced hemorrhage

We have shown that inducing general anesthesia attenuates nociception-induced hemorrhage after SCI. It is unclear whether this protective effect is unique to barbiturates or if it extends to other general anesthetic drugs. Here, we explore this issue by testing whether the general anesthetic isoflurane has a similar protective effect,

As shown previously, shock produced an increase in hemoglobin-associated absorbance at the injury site. Isoflurane administration given before acute pain onset blocked this effect (Fig. 6B). An ANOVA revealed a main effect of shock and drug treatment, as well as an interaction effect between these variables, (all F's > 6.509, p < 0.05). Post hoc comparisons revealed that animals treated with vehicle and shock differed from all the other groups (p < 0.05).

Animals treated with shock had higher levels of hemoglobin in protein samples when compared with unshocked controls. Isoflurane blocked this effect (Fig. 6C). An ANOVA revealed a main effect of shock and drug treatment, as well as an interaction between these variables, (all F's > 4.534, p < 0.05). Post hoc comparisons revealed that animals treated with vehicle and shock differed from all the other groups (p < 0.05).

Isoflurane limited pain-induced alteration of blood pressure similar to pentobarbital. Noxious stimulation resulted in an acute increase in systolic blood pressure. Isoflurane administration was associated with lowered blood pressure (Fig. 6D). An ANOVA found no differences in pre-treatment systolic blood pressure (p > 0.05).

An ANCOVA of post-treatment systolic blood pressure collapsed across time revealed a main effect of drug administration, (F (1,27) = 15.314, p < 0.05). Post hoc comparisons showed that animals treated with shock and vehicle differed from groups treated with isoflurane given shock or no shock.

Additional Western blots targeted proinflammatory cytokines (IL-1ß, IL-18). Shock increased expression of these proteins, while administration of isoflurane blocked this effect (Fig. 6E, F). Independent ANOVAs for each cytokine revealed a main effect of shock (both F's < 5.553, p < 0.05). The IL-1β assay also yielded a main effect of drug administration, (F (1,28) = 4.250, p < 0.05). The interaction effect was significant for the IL-18 assay, (F (1,28) = 7.957, p < 0.05) and approached significance for IL-1β, (F (1,28) = 3.110, p = 0.0887). Post hoc comparisons of the group means showed that vehicle-shocked animals differed from all other groups for both assays (p < 0.05). The overall pattern of results suggests that noxious stimulation (both shock and capsaicin treatment) fosters proinflammatory cytokine expression and that this effect is blocked by anesthesia (both pentobarbital and isoflurane).

Discussion

Noxious stimulation after SCI expands the extent of tissue loss and inhibits locomotor recovery.^1-3,5,6 Previous work has shown that cutting brain-spinal cord communication by transecting the spinal cord rostral to injury (at T2) blocks nociception-induced hemorrhage. Likewise, pharmacologically blocking neural signaling at T2 with lidocaine blocked both pain-induced hemorrhage and its adverse effect on locomotor recovery.^6,11,12,20

The current results show that disrupting brain activity using a general anesthetic also has a protective effect. Pentobarbital given before noxious stimulation (capsaicin and electric shock) blocked nociception-induced hemorrhage and eliminated the disruption in locomotor recovery observed after exposure to noxious electrical stimulation. A subanesthetic (12.5 mg/kg) dose of pentobarbital did not have a protective effect, nor did application of pentobarbital after pain-stimulation had run its course. Importantly, inducing an anesthetic state with isoflurane also blocked nociception-induced hemorrhage, which suggests that the protective effect is not limited to barbiturates.

Previous work showed that disrupting spinal cord-brain signaling by applying lidocaine at T2 after noxious stimulation has no effect.²¹ Likewise, inducing an anesthetic state with pentobarbital after animals received noxious electrical stimulation did not attenuate nociception-induced hemorrhage. Because hemorrhage was induced by a brief period of stimulation (6 min of intermittent shock), these observations suggest that the processes that drive hemorrhage emerge quickly and are irreversible. This implies that anesthetic interventions targeting either the brain or spinal cord must be administered proactively to protect against pain-induced hemorrhage and inflammation.

Underlying processes and mechanisms

Previous studies suggest that inducing a state of anesthesia can attenuate secondary injury in other instances of central nervous system (CNS) injury (stroke, SCI, traumatic brain injury [TBI]).^22-24 By quelling neural excitability after injury, anesthetic drugs have been shown to reduce markers of various pathophysiological processes—such as neuroinflammation and oxidative stress—implying that the current results may generalize to a number of injury models and anesthetic regimens.

Some debate exists surrounding the clinical translatability of anesthetic neuroprotection. Like many proposed treatments of CNS injury, clinical studies have failed to produce the same neuroprotective effects observed in pre-clinical work. Anesthetic administration, however, to curb peripheral pain input is a novel proposed mechanism for anesthetic neuroprotection. Here, we have shown that general anesthetic drugs produce protective effects against nociception-induced increases in hemorrhage and inflammation after SCI.

Anesthetic drugs could yield a neuroprotective effect by impacting systemic factors, such as acute hypertension and stress-related processes engaged by peripheral pain stimulation. Supporting this, analysis of plasma from pain-stimulated rats demonstrates that pain produces a sustained increase in corticosterone observable as early as 24 h after shock treatment and maintained for one week after.²⁵ While the severance of brain-spinal cord communication using anesthetics is sufficient to block the effects of both pain and stress stimulation on recovery, treatments that target the systemic processes that fuel nociception-induced hemorrhage may also have a neuroprotective effect. Supporting this, administration of alpha-1 adrenergic receptor antagonist prazosin to reduce pain-induced hypertension yielded a reduction in hemorrhage at the site of injury.²⁶ This suggests that acute hypertension brought on by peripheral pain stimulation may, in part, drive hemorrhage.

Importantly, the timing of anesthetic intervention would need to overlap with the period of vulnerability of the spinal cord to pain-related changes. Previous data suggest that noxious stimulation does not induce hemorrhage or impair long-term recovery when given two weeks after SCI.²⁷ Because many indices of secondary injury have run their course by two weeks after injury, this establishes a window of therapeutic opportunity for anesthetic intervention.

A variety of pathophysiological processes contribute to the expansion of the initial lesion site shortly after injury. The destruction of blood vessels near the site of injury enables a leakage of blood into the spinal cord parenchyma. This hemorrhage causes cell death at site of infiltration from the neurotoxicity of red blood cells.²⁸ Thrombin, free iron, and even hemoglobin itself can induce cell death and neuroinflammation.¹⁰ In addition, the leakage of other neurotoxic vascular contents, such as leukocytes and albumin, contribute to the expansion of the lesion site shortly after injury.^10,29

Shortly after these vessels are ruptured, the release of vascular content after injury triggers an extensive inflammatory response. As systemic immune cells infiltrate the site of injury, partly from compromised blood vessels, resident immune cells also become active.^30-32 These activated cells produce pro-inflammatory cytokines, which initiate pyroptotic cell death pathways, fueling additional tissue loss. Proinflammatory cytokines, such a IL-1ß, are upregulated within minutes after injury.³¹ This cytokine expression can peak within hours to days after injury.^31,33,34

In addition, the neuroinflammatory response will further disrupt CNS blood barriers and promote immune cell infiltration.³⁵ Notably, the presence of proinflammatory cytokines activate surrounding immune cells, a feed-forward process that drives further cytokine release driving greater cell death.³⁶

Both tested general anesthetics and the local anesthetic lidocaine, delivered at the level of the spinal cord, may play a role in quieting pain-related secondary signals. Future studies will explore the importance of neuroanatomical features of both the brain and spinal cord in propagating pain-induced deficits.

The rationale for the present work stemmed from previous studies demonstrating that cutting communication with the brain blocks the adverse effect noxious stimulation has on acute tissue loss and long-term recovery after SCI. Given these observations, we hypothesized that treatments that induce a sedative state, which inhibits brain activity, would also have a protective effect.

While the present results lend some support for this concept, it must be acknowledged that general anesthetics have a broad spectrum of effects and can impact spinal function. In addition, it is also possible that spinal systems rostral to T2 play a pivotal role. To implicate brain systems, further work is needed to demonstrate that selective manipulations, which target particular brain areas, impact tissue loss after SCI. Our work suggests that inhibiting activity in these regions could have a protective effect while overactivity may fuel tissue.

We have described these effects in terms of nociception-induced hemorrhage and inflammation. This is consistent with previous work showing that electrical stimulation engages (C and A) nociceptive pathways and the fact that selectively engaging nociceptive-specific fibers with capsaicin induces hemorrhage and impairs recovery.^9,37

Interestingly, these effects are not attenuated by pre-treatment with the analgesic morphine, given at a dosage that blocks brain-dependent measures of pain.²⁰ This suggests that attenuating brain-mediated pain, or engaging mu opioid receptors within the spinal cord, is not sufficient to counter the adverse effects of noxious stimulation. This contrasts with the protective effect of curbing sensory transmission to the brain (via lidocaine at T2) and general anesthesia (induced by pentobarbital or isoflurane). The overall pattern of results implicates neural systems engaged by noxious stimulation that are not regulated by a widely used opiate analgesic. One interpretation of these observations is that tissue loss after injury is driven by the induction of heightened arousal or stress. Further work is being conducted to explore this possibility.

Hemodynamic changes after pain

Nociceptive input and stress may impact tissue loss by engaging the sympathetic nervous system to drive a rise in blood pressure. This is consistent with previous work showing that a period of hypertension after SCI is associated with poor recovery in an animal model.^38,39 Other work suggests that blood pressure dysregulation is predictive of diminished neurological recovery in humans.⁴⁰ Further, we have shown that noxious electrical stimulation induces a rise in blood pressure and that this effect is blocked by both a T2 transection and pentobarbital anesthesia.⁶ This acute sympathetic activity, particularly consequent hemodynamic changes, is dependent on rostral systems. The implication is that nociception-induced hemorrhage may be driven, at least in part, by a brain-dependent surge in blood pressure/blood flow.⁴¹

Other findings suggest that hemodynamic dysregulation may fuel tissue loss after SCI. Indeed, pharmacologically inducing a rise in systolic blood pressure through norepinephrine administration is sufficient to induce locomotor deficits after SCI in rats.⁴² Further, hypertension is commonly associated with increased CNS blood barrier permeability and hemorrhage after brain injuries.^43,44 An acute hypertensive period brought about by peripheral pain may put stress on injured capillaries at the site of injury, fueling further barrier permeability and greater blood infiltration.

It is also important to note that the maintenance of normotensive, rather than low, blood pressure is key in protecting the injured nervous system. Hypotension can be just as destructive as hypertension after CNS injury. Recent work suggests that deviation from optimal mean arterial pressure after SCI, whether hypertensive or hypotensive, produces poor recovery.⁴⁰ The present results suggest that peripheral pain input produces an acute hypertensive state and that a potential mechanism for anesthetic neuroprotection may be a consequent return to normal hemodynamic levels by drug administration.

While hemodynamic regulation is critical after SCI, new findings suggest that nociception-induced hypertension does not fully explain the consequent hemorrhage and lasting impairment. Indeed, while pharmacologically driving a rise in systolic blood pressure produced deficits in locomotor recovery, this did not result in the expression of increased hemorrhage at the site of injury.⁴² This suggests that changes in systolic blood pressure alone may not be sufficient to produce hemorrhage. Conversely, administration of prazosin around the time of peripheral pain stimulation to reduce blood pressure blocked the expression of acute hemorrhage but did not yield a protective effect that extended to locomotor recovery.²⁶

Summary

Independent of the mode of action, it is clear that nociceptive stimulation after injury has an adverse effect on recovery. In cases of polytrauma, the current findings suggest that placing patients in a medically induced coma would have therapeutic benefits. The results also have implications for treatments that engage nociceptive fibers. Indeed, others have shown extended stretching can engage pain fibers and disrupts locomotor performance in rats.⁴⁵ Likewise, application of capsaicin disrupts locomotor performance in humans undergoing step-training on a treadmill.⁴⁶

Finally, surgical intervention after injury will induce tissue damage and engage nociceptive fibers. Our work suggests that to achieve the benefit of these interventions, it may be necessary to maintain an extended anesthetic state or block communication of pain signals to and from the brain. Other recent work has revealed that noxious stimulation also increases the area of hemorrhage after TBI.⁴⁷ Further work is needed to determine whether the neuroprotective effect of anesthesia extends to this example of pain-induced hemorrhage.

Funding Information

Work on this project was supported by the Mary Tucker Currie Professorship and the National Institute of Neurological Disorders and Stroke (NS104422) and the Office of the Assistant Secretary of Defense for Health Affairs, through the Spinal Cord Injury Research Program under Award No. W81XWH-18-1-0807.

Author Disclosure Statement

No competing financial interests exist.