Spinal Cord Ischemic Injury
Development of a New Model in the Rat
Background and Purpose Spinal cord ischemic injury (SCII) with resulting paralysis is a major cause of morbidity after operations on the thoracic aorta. Since the vascular supply to the spinal cord is similar in rats and humans, the rat appears important for studies of mechanisms of injury and development of therapeutic strategies to avoid this complication.
Methods In group A rats, we induced SCII using a previously described method, by occluding the descending thoracic aorta for 15, 20, 24, or 30 minutes with the inflated balloon of a 2F Fogarty catheter inserted through the femoral artery. In group B, the catheter was inserted through the left common carotid artery, and the aorta was occluded just distal to the carotid origin for 20 minutes. In group C, in addition to the procedure described for group B, hypovolemia was induced during a 12-minute period of aortic occlusion by equilibrating the left femoral artery pressure to the atmospheric pressure. The motor function of the hind limbs and the associated spinal cord histopathology were studied.
Results At 96 hours, 9 of 10 rats in group C were paraplegic. This rate was significantly higher than that of group A (1 of 21, P=.00000) or group B (4 of 10, P<.03). In all groups, the histopathological changes became more severe from the rostral to the caudal direction along the spinal cord and from the peripheral to the central location in transverse sections.
Conclusions The combination of aortic arch occlusion with induced hypovolemia resulted in a reproducible model of SCII in rats.
Spinal cord ischemic injury is an uncommon but devastating entity in clinical practice.1 A substantial proportion of the cases occur after operations on the descending thoracic or thoracoabdominal aorta.2 Spinal cord ischemia may also represent an important secondary injury mechanism after spinal cord trauma.3 Stenonis, as quoted by Zivin and DeGirolami,4 was the first to describe the relationship between the interruption of the blood flow in the aorta and the development of paralysis in hind limbs of dogs more than 300 years ago. A rabbit model that was systematically characterized by Zivin and DeGirolami4 has been one of the most frequently used models for the study of SCII. Paraplegia can be reliably produced in the rabbit by temporarily interrupting the blood flow to the infrarenal aorta. Nevertheless, the spinal cord arterial system in the rabbit is almost purely segmental5 and therefore different from that in humans.6 In contrast, several studies have shown that the vascular anatomy of the rat and the human spinal cord are almost identical.7–9 SCII has been produced in the rat by temporarily clamping the aortic arch, which has been accessed through a thoracotomy.10 In addition, Coston et al11 described a minimally invasive rat model of SCII in 1983. However, none of the various rat models of SCII has gained significant popularity. The reasons for this are not entirely clear but may include the inability to reproduce previously reported results. The present study represents a renewed effort to develop a reliable SCII model in rats with the use of a minimal access surgical technique.
Materials and Methods
Animal Care and Surgical Technique
Seventeen SD and 26 LE outbred male rats from Harlan Sprague Dawley, Inc, weighing between 350 and 375 g were used in the study. All animals were allowed free access to laboratory chow and tap water in day/night-regulated quarters at 25°C. Animal care and experiments complied with the “Principles of Laboratory Animal Care” (Guide for the Care and Use of Laboratory Animals, NIH publication 86–23, 1985) and was approved by the Animal Studies Committee, Washington University School of Medicine (protocol No. 96261). The rats underwent surgery in a room kept at 24°C. Anesthesia was induced in a chamber containing 3% halothane and was maintained by inhalation through a facial mask of 1% to 2% halothane driven by oxygen flow of 2 L/min. Temperature was continuously monitored with a flexible probe inserted 3 cm into the rectum. During the surgical procedure, the temperature was maintained at 37±0.5°C with an underbody thermal pad and a heat lamp. No additional measures were taken during the period of aortic occlusion to compensate for the decrease in the rectal temperature (approximately 0.5°C to 1°C) (see below). Arterial blood gases and arterial blood glucose (Accu-Chek Easy Blood Glucose Monitor, Boehringer Mannheim) were measured just before aortic occlusion and 10 minutes after the onset of reperfusion. Heparin (100 U/kg) was administered intra-arterially 5 minutes before aortic occlusion.
The rats were divided into three groups. Group A animals (17 SD and 4 LE rats) underwent surgical procedures according to a previously described method.11,12 The animals were placed in the supine position, and two short skin incisions were made at the left groin and the ventral aspect of the proximal tail. The left femoral artery and the tail artery were dissected free. A polyethylene catheter (PE-50) was inserted into the tail artery for monitoring the DABP and for collecting blood specimens. The proximal descending thoracic aorta was transiently occluded by inflating the balloon of a 2F Fogarty catheter (model 120602F, Baxter Healthcare) with 0.1 mL of water after its insertion through the left femoral artery and its advancement 11 cm cephalad from the arteriotomy site. Our own preliminary studies, including dissections of the aorta and its major branches, indicated that in both SD and LE rats this length of catheter advancement resulted in positioning of the catheter tip just distal to the origin of the left subclavian artery. After aortic occlusion for 15, 20, 24, or 30 minutes, the balloon was deflated and the catheter was withdrawn. The efficiency of occlusion was confirmed by monitoring the DABP. The DABP could not be further lowered during the time of ischemia by greater balloon inflation.
Group B rats (10 LE rats) were placed in the supine position with the head and neck partially turned toward the right side. Left groin and ventral midline cervical skin incisions were made. The left femoral artery was cannulated with a PE-50 catheter and was used for monitoring the DABP and for collecting blood samples. An arteriotomy was made in the left common carotid artery, and a second PE-50 catheter was introduced and advanced cephalad into the left internal carotid artery. This catheter was used for monitoring the leftDICAP. A 2F Fogarty catheter was also inserted through the left common carotid arteriotomy and was advanced caudally into the descending thoracic aorta for approximately 8 cm. The catheter balloon was then partially inflated with 0.03 mL of water, and the catheter was gently withdrawn. When the balloon reached the origin of the common carotid artery, resistance to further catheter withdrawal was clearly felt by the operator. At this point the balloon was further inflated with water to a total of 0.10 mL. The above manipulations allowed the precise localization of the level of aortic occlusion in all animals, regardless of individual differences in weight or size. The duration of ischemia was 20 minutes based on data from a series of preliminary studies. At the end of the period of ischemia, the balloon was deflated and the catheter was withdrawn.
Animals in group C (10 LE rats) underwent a surgical procedure similar to animals in group B but with the following differences (Fig 1⇓). The tail artery catheter was used for monitoring the DABP and for collecting blood samples. The left femoral artery was partially incised transversely immediately after full inflation of the catheter balloon, equilibrating the arterial pressure to the atmospheric pressure. Blood that was extravasated through the femoral arteriotomy during the period of aortic occlusion was collected in a 10-mL heparinized syringe. Part of the recovered blood was administered to the animals during the later stages of the period of aortic occlusion through the left internal carotid artery catheter to maintain the mean DICAP at approximately 50 mm Hg. The remaining blood was administered to the animals within a period of 2 minutes after deflation of the aortic balloon. The period of aortic occlusion in group C was reduced to 12 minutes based on results of preliminary studies. Finally, two additional LE rats underwent sham aortic arch occlusion. A Fogarty catheter was inserted in the descending aorta for 20 minutes through the left carotid artery, but the balloon was never inflated and no blood was withdrawn. Serial assessments of motor function in the hind limbs of all animals were performed at 1, 6, 12, and 24 hours and thereafter daily for 4 days. Credé’s maneuver was used for evacuation of the urinary bladder at least twice a day.
Evaluation of Neurobehavioral Outcome
Motor function deficits in the hind limbs were evaluated according to the following system, which was modified from the system reported by Marsala and Yaksh.12 An MDS was given to each rat at each assessment according to the following criteria: 0=normal; 1=walks normally but legs are weak (cannot pull the legs if one holds them); 2=assumes normal body posture on a flat surface and is able to walk, but there is either ataxia or spasticity; 3=able to walk on the knuckles or able to walk on the feet without proper stepping; 4=drags legs but there is movement at the knees; and 5=drags legs without significant movement in the lower limbs, and either spasticity or flaccidity is present.
Animals with MDS ≥3 were considered paraplegic in the study, whereas animals with MDS <3 were considered nonparaplegic. Spasticity was defined as the continuous or intermittent tonic positioning of the hind limbs in extension, particularly with the feet in plantar flexion. The degree of spasticity ranged from mild (usually present intermittently or after provocation by lifting the tail or by stressing the animal) to conspicuous (continuous, present even at rest, and occasionally accompanied by tonic upward curvature of the tail). Flaccidity was defined as absence or great reduction of muscle tone as felt during passive movements at the ankle joints.
Animals were killed with an intraperitoneal pentobarbital injection (150 mg/kg). They were transcardially perfused with 0.9% NaCl solution for 1 minute followed by 400 mL 10% buffered formalin. The cadavers were kept at 4°C for 4 hours, and then the whole spinal cords were harvested and postfixed in the same fixative for 2 to 7 days before they were embedded in paraffin. Transverse sections were obtained through the middle of the 4th lumbar, 11th thoracic, 3rd thoracic, and 5th cervical spinal cord segments. Sections were stained with hematoxylin and eosin as well as with Nissl and were examined under the light microscope.
The physiological parameters were analyzed by one-way ANOVA followed by post hoc Tukey’s test. The neurological scores were analyzed by Kruskal-Wallis one-way ANOVA, corrected for ties. The paraplegia and mortality rates were analyzed by Fisher’s exact test. A value of P<.05 was considered significant. Data are expressed as mean±SEM.
The preocclusion arterial blood gases were within normal range for all animals (pH 7.4±0.1, Pao2 >10.5 kPa [80 mm Hg], Paco2 5±0.1 kPa [37±7 mm Hg]). The differences between the three groups were not statistically significant. Arterial blood gases taken 10 minutes after the balloon deflation were characteristic of severe metabolic acidosis with varying degrees of compensatory respiratory alkalosis (Table 1⇓). The difference in the preocclusion blood glucose levels between the three groups was not statistically significant (5.58±0.22, 5.42±0.26, and 5.46±0.28 mmol/L for groups A, B, and C, respectively). Balloon inflation with 0.10 mL of water caused a rapid decrease of the DABP in all rats (Fig 2⇓). The maximal DABP during the period of aortic occlusion was significantly lower in group C compared with groups A and B (Table 1⇓). The leftDICAP rose after balloon inflation. In group B animals, the DICAP remained elevated throughout the period of ischemia, whereas in group C animals it gradually dropped and was maintained at approximately 50 mm Hg by return of part of the drained blood in the circulation (Fig 2⇓). In group C animals, 9.1±0.5 mL of blood was removed during the period of ischemia. After deflation of the balloon, the DABP recovered to baseline levels within 5 to 10 minutes in groups A and B. In group C, the aortic pressure recovered almost immediately after the end of the administration of the previously drained blood (Fig 2⇓). The rectal temperature dropped during the period of aortic occlusion and recovered gradually during the reperfusion in all animals. The difference in the rectal temperatures between the three groups was not statistically significant at the end of the period of ischemia (Table 1⇓).
The neurobehavioral outcome for group A at 24 hours after reperfusion is shown in Table 2⇓. The outcome in group A did not change during the rest of the observation period. The 96-hour paraplegia rate (number of surviving paraplegic animals of the total number of operated animals) for group A was 5% (1 of 21). This rate was found to be significantly lower than the paraplegia rates in groups B (40% or 4 of 10, P<.03) and C (90% or 9 of 10, P=.00000) (Fig 3⇓). The 96-hour mortality rate (number of dead animals of the total number of operated animals) for group C was 10% (1 of 10). This rate was lower than the rate in groups B (40% or 4 of 10, P=.3) and A (19% or 4 of 21) (Fig 3⇓). At 96 hours after reperfusion, the mean MDS of surviving rats in group A (1.18±0.23) was significantly lower than the MDS of surviving rats in groups B (3.67±0.71, P=.0069) and C (4.56±0.24, P=.0000). The difference in the mean MDS between groups B or C was not statistically significant at any time. The mean MDS of surviving rats decreased during the first 48 hours after reperfusion in groups B (from 4.6±0.27 at 1 hour to 3.67±0.71 at 48 hours) and C (from 5±0 at 1 hour to 4.56±0.24 at 48 hours) but remained stable thereafter. The majority of the symptomatic animals had flaccid paraplegia after recovery from the anesthesia but developed spasticity within 1 to 6 hours. Only 1 animal (group B) showed deterioration of its hind limb motor function, which occurred on the second postoperative day. Animals with severe motor deficits (MDS 5) at 96 hours exhibited either persisting pronounced spasticity of the hind limb extensor muscles (2 and 5 animals in groups B and C, respectively) or complete flaccidity (1 animal in each of groups B and C). In addition, animals with no movement at the hind limbs at 96 hours (MDS 5) often had associated diminished or even absent responsiveness to pinching of the soles as well as urinary system disturbances in the form of hematuria, urinary retention, or incontinence. There were no complications from the permanent interruption of the blood flow in the femoral and tail arteries. In group A, all 4 deaths occurred within the first 24 hours after reperfusion. At autopsy, dilated urinary bladder was found in 3 cases, gangrenous small bowel in 2, and heart dilatation in 1. In group B, 3 deaths occurred during the first 24 hours and 1 death occurred between 24 and 48 hours after reperfusion. At autopsy, dilated urinary bladder was found in 3 cases, gangrene of small bowel in 2, and dilated heart in 2. In group C, 1 death occurred within the first 24 hours after reperfusion, with no gross abnormalities at autopsy.
A variety of histopathological changes were found in the spinal cord sections, which appeared to reflect the various degrees of neurological deficits and not the different operative techniques used. In all groups, the ischemic changes increased in the rostrocaudal direction. Sections from the lower thoracic and lumbar segments of spinal cords obtained from nonparaplegic animals (MDS <3) showed either no abnormalities or the occasional presence of few isolated eosinophilic (red) neurons in the base of the dorsal horns and the intermediate zone of the gray matter. Sections from animals with MDS equal to 3 or 4 (1 animal in group A, 1 in group B, and 3 in group C) showed more numerous eosinophilic neurons with a distribution within the gray matter similar to the one described above (Fig 4a⇓, 4a⇓′, 4d). The large alpha motor neurons appeared normal (Fig 4a⇓). There was mild vacuolization of the gray matter (Fig 4a⇓′). Pale areas representing infarcts were seen in the gray matter of animals with MDS equal to 5 (3 animals in group B and 6 animals in group C)(Fig 4b⇓, 4c⇓). The infarcted areas were characterized by destruction of the normal tissue architecture, often with formation of coalescent cavities in the tissue, as well as by the conspicuous presence of large numbers of infiltrating neutrophils and mononuclear phagocytes (Fig 4c⇓, 4c⇓′). The infarcts were well demarcated from the adjacent normal tissue that often contained surviving neurons (Fig 4c⇓′). The infarcts had variable size and distribution within the gray matter. In animals with severe spastic paraplegia the infarcts were seen mainly at the base of the dorsal horns (Rexed’s laminae 3 to 6) (Fig 4b⇓, 4d⇓) with variable extensions into the intermediate zone of the gray matter (Rexed’s lamina 7). In animals with severe flaccid paraplegia the infarcts were much larger in size, and they extended almost throughout the gray matter (Rexed’s laminae 2 to 10) (Fig 4c⇓, 4d⇓). Most of the large alpha motor neurons were dead. Nevertheless, in most sections, a few of these neurons, usually located close to the border with the white matter, appeared to have survived (Fig 4c⇓, 4c⇓′). Vacuoles were seen in the white matter of paraplegic animals, and these were more numerous in the anterior funiculus (Fig 4c⇓). The gray matter infarcts did not extend into the white matter even in cases with severe injury.
The Rat as Model for SCII
It is generally agreed that the availability of clinically relevant and reliable animal models greatly promotes scientific efforts to further understand disease and to develop more efficient treatments. A variety of factors, including the drive for improved cost-effectiveness, has led to the increasing use of the rat as model in the research laboratory. In addition, the rat features important advantages compared with other species as a model for the study of SCII, in particular. The nervous system of this animal has been extensively studied; perhaps more importantly, a great similarity of the spinal cord vascular anatomy between rats and humans has been found by several investigators.6–9 Reduction of the spinal cord perfusion is one of the most important etiopathogenetic factors in the production of SCII after transient interruption of the blood flow in the thoracic aorta.2,13 For a given tissue perfusion pressure, the vascular architecture influences regional differences in tissue perfusion. The similar spinal cord vascular anatomy between rats and humans may allow the close experimental reproduction of the spatial differences in spinal cord perfusion and the associated neuropathology that occurs after occlusion of the descending thoracic aorta in patients.
Development of Rat SCII Models
Coston et al11 first described the use of a 2F Fogarty catheter inserted through the femoral artery to transiently occlude the descending thoracic aorta in the rat. Ischemia for 15 to 16 minutes resulted in permanent hind limb motor deficits in 75% of the operated animals. Marsala and Yaksh12 used the same method to characterize the neurological and histopathological outcome after occlusion of the rat aorta for periods from 0 to 30 minutes. Twenty minutes of ischemia was reported to produce paraplegia in 22 of 24 operated animals with significant histopathological changes in the spinal cord at 8 hours after reperfusion. In the present study, use of the same method failed to produce rats with significant hind limb motor function deficits. Our experience appears to be in agreement with a subsequent report by Taira and Marsala.14 Simple balloon occlusion of the descending thoracic aorta for 10 to 30 minutes, according to the previously described method,12 failed to produce hind limb neurological motor deficits in 8 rats. Likewise, in an invasive rat model of SCII, simple occlusion of the descending thoracic aorta for up to 20 minutes produced hind limb neurological deficits in only one of five rats.15 It therefore appears that simple transient occlusion of the proximal descending thoracic aorta is not a satisfactory method to reliably produce SCII and paraplegia in outbred rats. This is probably due to the presence of significant collateral circulation to the spinal cord during the period of aortic occlusion. Taira and Marsala14 have shown that one way to decrease the collateral flow to the rat spinal cord is by reducing the aortic blood pressure above the occlusion. As an alternative, both the left subclavian and the right internal thoracic arteries10 or both subclavian arteries16 have been occluded in addition to the occlusion of the descending thoracic aorta in invasive models of SCII. In our study, inflation of the balloon in the aortic arch after insertion of the Fogarty catheter through the left common carotid artery (groups B and C) occluded not only the aortic lumen but at the same time the origins of the left carotid and subclavian arteries. The feasibility of the multiple occlusion was confirmed with autopsy and was attributed to the invariable close proximity of the orifices of the left carotid and the left subclavian arteries.
The equilibration of the left femoral artery pressure to the atmospheric pressure during the period of ischemia resulted in increased paraplegia and decreased mortality rates in group C compared with group B. The reliable achievement of lower levels of DABP, which is the major determinant of spinal cord perfusion pressure during a period of aortic occlusion,2,17 may be the main factor behind the reproducible production of paraplegia in group C. In addition, the lower DABP also allowed the use of shorter periods of aortic occlusion in group C, which may have contributed to the lower mortality. On the other hand, the additional procedures required to induce hypovolemia in group C increase the complexity of the surgical technique and could potentially introduce artifacts in the evaluation of the SCII. A similar effect could also be produced by other factors, such as the need to induce spinal cord ischemia while the animals are anesthetized.
Twelve minutes of aortic occlusion resulted in severe hind limb motor deficits in all group C animals at normothermia. This time period is comparable to periods of aortic occlusion previously reported as sufficient to cause motor deficits in rats,16 dogs,18 or humans.13
Distal Internal Carotid Arterial Pressure
The DICAP is an end pressure that reflects the pressure in the large arteries at the base of the brain (or more accurately, at the left side of the circle of Willis). Since the longitudinal spinal arteries originate from the vertebral arteries at the level of the medulla oblongata, the DICAP is a close approximation of the pressure in the origin of these arteries. In addition, the DICAP provides insight into the proximal aortic pressure, which cannot be easily measured in combination with the aortic arch occlusion method used in group B and C animals. The proximal aortic pressure is always higher than or equal to the DICAP. Monitoring and control of the DICAP are necessary in view of reported evidence that the proximal aortic pressure influences outcome after aortic occlusion in rats,14 dogs,19 and humans.17
Evolution of Neurobehavioral Outcome
In the present study, symptomatic animals in all groups showed an overall tendency for recovery from the initial neurological deficits during the 48 hours after the onset of reperfusion. Only one animal showed delayed deterioration of hind limb motor function between 24 to 36 hours after reperfusion. Improvement of neurological deficits after SCII has been previously described in the rat10 as well as in humans.20 On the other hand, deterioration of the neurological deficits or even paraplegia delayed in onset is also known to occur 1 to 21 days after aortic reconstructive surgery in humans.21 Deterioration of the neurological condition at approximately 24 hours of reperfusion has been reported in the rabbit model of SCII when periods of ischemia that are just longer than the ischemic tolerance threshold are used22 or when neuroprotective agents are administered before the insult.23 It is possible that similar experimental conditions could result in an increased incidence of delayed deterioration in the neurological condition after SCII in the proposed rat model.
Our study supports the previously demonstrated correlation between histopathological changes in the gray matter at the lower lumbar spinal cord segments and hind limb motor deficits after transient occlusion of the aorta.12,24 The selective distribution of dead neurons in the central area of the lumbar gray matter in animals with moderate motor deficits as well as the somewhat similar distribution of infarcts in animals with severe spastic paraplegia gives us insight into the differential vulnerability of various areas in the spinal cord gray matter to ischemia. Areas that are centrally located in the cord, such as the intermediate zone (Rexed’s lamina 7), the adjacent area of the dorsal horns (Rexed’s laminae 3 to 6), and the area surrounding the central canal (Rexed’s lamina 10) may be the most susceptible to an ischemic insult secondary to transient aortic occlusion. Areas of the gray matter more peripherally located, such as the area of the large alpha motor neurons (Rexed’s lamina 9), the ventromedial anterior horn (Rexed’s lamina 8), and the tips of the dorsal horns (Rexed’s laminae 1 and 2) appear to sustain less severe injury. Our findings are in accord with findings from previous experimental studies of SCII after aortic occlusion in various animals.12,25,26
In summary, this study demonstrated that SCII may be reliably reproduced in the rat with a minimal access surgical technique. In addition, this new model was shown to be relevant to a significant clinical problem in terms of anatomic, hemodynamic, neurobehavioral, and histopathological parameters.
Selected Abbreviations and Acronyms
|DABP||=||distal arterial blood pressure|
|DICAP||=||distal internal carotid arterial pressure|
|MDS||=||motor deficit score|
|SCII||=||spinal cord ischemic injury|
This study was supported by the Shoenberg Foundation, National Institutes of Health grants NS 25542 and NS 32000, and the American Paralysis Association
- Received May 30, 1997.
- Revision received September 10, 1997.
- Accepted September 11, 1997.
- Copyright © 1997 by American Heart Association
Satran R. Spinal cord infarction. Stroke. 1988;19:529–532.
Allen BT. Spinal cord ischemia complicating aortic surgery. In: Callow AD, Ernst CB, eds. Vascular Surgery: Theory and Practice. 1st ed. Stamford, Conn: Appleton & Lange; 1995:1403–1425.
Zivin JA, DeGirolami U. Spinal cord infarction: a highly reproducible stroke model. Stroke. 1980;11:200–202.
Fazio V. Vascular pathology of the spinal cord. In: Minkler J, ed. Pathology of the Nervous System. New York, NY: McGraw-Hill Publishing Co; 1971:1548–1567.
Domisse GF. The blood supply of the spinal cord: a critical vascular zone in spinal surgery. J Bone Joint Surg Br. 1974;56B:225–235.
Tveten L. Spinal cord vascularity, IV: the spinal cord arteries in the rat. Acta Radiol Diagn (Stock). 1976;17:385–398.
Scremin OU. Cerebral vascular system. In: Paxinos GT, ed. The Rat Nervous System. 2nd ed. San Diego, Calif: Academic Press; 1995:3–35.
Taira Y, Marsala M. Effect of proximal arterial perfusion pressure on function, spinal cord blood flow, and histopathological changes after increasing intervals of aortic occlusion in the rat. Stroke. 1996;27:1850–1858.
Woloszyn TT, Marini CP, Coons MS, Tachmes L, Nathan IM, Acinapura AJ, Cunningham JN. Partial exsanguination effectively controls proximal hypertension and protects the spinal cord during thoracic aortic crossclamping. Surg Forum. 1989;40:193–195.
Gelfan S, Tarlov IM. Altered neuron population in L7 segment of dogs with experimental hind-limb rigidity. Am J Physiol. 1963;205:606–616.