Delayed Paraplegia After Spinal Cord Ischemic Injury Requires Caspase-3 Activation in Mice
Background and Purpose—Delayed paraplegia remains a devastating complication after ischemic spinal cord injury associated with aortic surgery and trauma. Although apoptosis has been implicated in the pathogenesis of delayed neurodegeneration, mechanisms responsible for the delayed paraplegia remain incompletely understood. The aim of this study was to elucidate the role of apoptosis in delayed motor neuron degeneration after spinal cord ischemia.
Methods—Mice were subjected to spinal cord ischemia induced by occlusion of the aortic arch and left subclavian artery for 5 or 9 minutes. Motor function in the hind limb was evaluated up to 72 hours after spinal cord ischemia. Histological studies were performed to detect caspase-3 activation, glial activation, and motor neuron survival in the serial spinal cord sections. To investigate the impact of caspase-3 activation on spinal cord ischemia, outcome of the spinal cord ischemia was examined in mice deficient for caspase-3.
Results—In wild-type mice, 9 minutes of spinal cord ischemia caused immediate paraplegia, whereas 5 minutes of ischemia caused delayed paraplegia. Delayed paraplegia after 5 minutes of spinal cord ischemia was associated with histological evidence of caspase-3 activation, reactive astrogliosis, microglial activation, and motor neuron loss starting at approximately 24 to 48 hours after spinal cord ischemia. Caspase-3 deficiency prevented delayed paraplegia and motor neuron loss after 5 minutes of spinal cord ischemia, but not immediate paraplegia after 9 minutes of ischemia.
Conclusions—The present results suggest that caspase-3 activation is required for delayed paraplegia and motor neuron degeneration after spinal cord ischemia.
Delayed paraplegia is a devastating complication of spinal cord ischemia (SCI), which can occur after thoracic and abdominal aortic surgery for a variety of aortic pathologies, including aneurysm and trauma.1 Rates of immediate and delayed-onset neurological deficits after major thoracic aortic repairs range between 4% and 11%.1 Of all incidences of paraplegia (immediate and delayed combined) associated with aortic surgery, the reported incidence of delayed paraplegia varied from 12% to 73%.1 Although introduction of several adjunct procedures, including cerebrospinal fluid drainage, reduced the incidence of immediate paraplegia after aortic surgery, the incidence of delayed paraplegia has not changed.2
Although immediate paraplegia is thought to be caused by an irreversible ischemic neuronal injury in the spinal cord, the mechanism responsible for delayed paraplegia is incompletely understood.3 Several potential mechanisms responsible for the development of delayed paraplegia have been proposed, including delayed apoptotic neuronal death executed by caspase-3 activation.4–6 Nonetheless, the role of motor neuron apoptosis and caspase-3 activation in the pathogenesis of delayed paraplegia remains controversial; some studies show the presence of apoptosis in the spinal cord of animals exhibiting delayed paraplegia, whereas others do not.7,8 Because the majority of these studies examined the role of apoptosis using immunohistochemical detection of caspase-3 and/or DNA fragmentation, no causal relationship between caspase-3 activation and delayed motor neuron death has been established to date. Furthermore, elucidation of the role of apoptosis in delayed paraplegia has been hindered by the lack of reproducible animal models in which genetic modification can be exploited to determine the molecular mechanisms.
To define the molecular mechanisms responsible for the delayed paraplegia after SCI, we have recently developed a mouse model of delayed paraplegia by modifying a previously reported mouse model of SCI.9 This model is unique in which mice that are subjected to SCI initially recover from surgery and anesthesia and exhibit ability to walk for approximately 24 hours. Subsequently, however, all mice develop delayed paraplegia starting at approximately 30 to 36 hours after surgery. Using this robust model, we sought to determine the role of apoptotic neuronal death in immediate and delayed paraplegia after SCI. We report that caspase-3 activation is required for delayed paraplegia but not for immediate paraplegia in mice.
Materials and Methods
Mouse Model of SCI
After approval by the Massachusetts General Hospital Subcommittee on Research Animal Care, male wild-type mice (WT, C57BL/6J, 8 to 10 weeks old; Jackson Laboratory, Bar Harbor, ME) and male mice deficient for caspase-3 backcrossed onto C57BL/6 background >10 generations (caspase-3−/−, 8 to 10 weeks old)10 were anesthetized with isoflurane and subjected to SCI according to the method described by Lang-Lazdunski and colleagues with modifications.9 See http://stroke.ahajournals.org for the details of surgical procedures, measurements of physiological parameters, quantal bioassay, and histological studies.
Assessment of Motor Neuron Function
Motor function was quantified serially at pre-SCI, 8, 24, 48, and 72 hours after SCI by the Basso Mouse Scale (BMS).11 The maximum deficit is indicated by a score of 0. Although BMS score <6 (0 to 5) indicates paraplegia, BMS score ≥6 (6 to 9) indicates ability to walk.
Parametric data were presented as mean±SD. Analysis of variance followed by Bonferroni and Tukey-Kramer tests or Student t test was used to compare parametric data. The quantal bioassay was based on logistic analysis.12 The difference between the probability of producing delayed and immediate paraplegia in 50% of mice (P50d and P50i, respectively) was graphically demonstrated by computer construction of an ischemic duration probability of paraplegia curve for each group.13 Changes in BMS were analyzed by 2-way repeated-measures analysis of variance followed by Tukey test. The number of viable neurons between experimental groups at individual time points was compared with Mann-Whitney U test. Probability values <0.05 were considered significant.
SCI Causes Immediate or Delayed Paraplegia
All mice subjected to 9 minutes SCI showed immediate paraplegia that persisted up to 72 hours after SCI (P<0.01 versus sham and pre-SCI; Figure 1). All mice subjected to 5 minutes SCI showed modest and transient motor weakness (BMS=6 to 9) at 30 minutes after reperfusion followed by gradual recovery over the next 24 to 30 hours (Figure 1). Up to approximately 8 hours after 5 minutes SCI, motor deficit manifested as ataxia and partial weakness in place-stepping reflex with preserved ability to walk (P<0.05 versus 9 minutes SCI at 8 and 24 hours after SCI; Figure 1). The motor function of lower extremities of mice subjected to 5 minutes SCI gradually worsened starting at approximately 30 hours after reperfusion and all mice exhibited complete paraplegia by 48 hours after reperfusion (Figure 1). Long-term outcomes after SCI (6 weeks) were examined in subgroups of mice that were subjected to 9 or 5 minutes SCI. Although all mice that were subjected to 9 minutes SCI died within 7 days, 20% of mice survived up to 6 weeks after 5 minutes of SCI (Supplemental Figure I). Once they became paraplegic, no mice showed any recovery of hind limb motor function until death or up to 6 weeks after SCI of either duration. Sham-operated mice did not show any neurological deficits throughout the experiments (Figure 1).
Quantal Bioassay for the Relationship Between the Duration of SCI and Neurological Function
The incidence of paraplegia (BMS=0 to 5) increased as the duration of SCI increased (Supplemental Table I; Figure 2). In the current model, <3 minutes of SCI appeared to be well tolerated that did not cause motor deficit. Although many mice exhibited ability to walk (BMS=6 to 9) up to 24 hours after SCI after shorter durations of aortic clamping (from 3.5 to 7.5 minutes), the majority of the mice developed complete paraplegia by 48 hours of reperfusion (delayed paraplegia). All mice subjected to >8 minutes of SCI showed motor deficit in the hind limb immediately after recovering from anesthesia without any recovery up to 72 hours after SCI (immediate paraplegia). Quantal bioassay analysis revealed that the P50i and P50d were 6.6±0.1 minutes and 3.5±0.1 minutes, respectively (Figure 2, P<0.01).
Time Course of Neurodegeneration After SCI
In the spinal cord of mice subjected to 9 minutes SCI, neurons in the ventral horn started to degenerate at 8 hours after SCI. Extensive neuronal loss appeared to have completed by 24 hours of reperfusion without further changes up to 72 hours after 9 minutes SCI (Figure 3A–B). In contrast, 5 minutes SCI did not affect the number and appearance of neurons in the ventral horn at 8 and 24 hours after reperfusion (Figure 3A–B). However, the ventral horns of mice subjected to 5 minutes SCI exhibited extensive cellular loss with marked cavitation at 48 and 72 hours after SCI. Although the structure of the ventral horn appeared to be better preserved at 72 hours after 9 minutes of SCI than after 5 minutes of SCI, extensive neuronal loss after SCI of both durations was confirmed by the absence of NeuN-positive neurons at 72 hours after 9 or 5 minutes of SCI (Figure 3C).
SCI Activates Astrocytes and Microglia
To examine the mechanisms responsible for the delayed extensive cavitation observed in the spinal cord of mice subjected to 5 minutes SCI, activation of astrocytes and microglia was assessed by immunohistochemical staining of glial fibrillary acidic protein and ionized calcium binding adaptor molecule 1, respectively. Although only few reactive astrocytes and activated microglia were observed at 8 and 24 hours after 5 minutes SCI (Figure 4A), the number of reactive astrocytes and activated microglia markedly increased at 48 hours after 5 minutes SCI (Figure 4A). In contrast, in the spinal cord of mice subjected to 9 minutes SCI, no reactive astrocyte or activated microglia were observed up to 48 hours after reperfusion (Figure 4A). These results suggest that delayed inflammatory reaction associated with glial activation may have contributed to the extensive spinal cord damage and delayed paraplegia after 5 minutes SCI.
Caspase-3 Activation Precedes the Onset of Delayed Paraplegia
A small number of cleaved caspase-3-positive neurons was observed in the ventral horn of the spinal cord of mice at 8 hours after 9 minutes SCI (Figure 4B). However, only a few cleaved caspase-3-positive neurons were found at 24 and 48 hours after 9 minutes SCI, presumably because of the extensive necrosis and disappearance of the spinal ventral neurons. In contrast, the number of ventral horn neurons with positive cleaved caspase-3 immunoreactivity markedly increased at 24 and 48 hours after 5 minutes SCI (Figure 4). These results suggest caspase-3 activation precedes the onset of extensive neuroinflammation, neurodegeneration, and delayed paraplegia after 5 minutes SCI.
Caspase-3 Deficiency Prevents Delayed Paraplegia in Mice
To determine the role of caspase-3 on the immediate and delayed onset paraplegia after SCI, we examined outcomes of SCI in caspase-3−/− mice. There was no difference in physiological parameters, including regional spinal cord blood flow during SCI between WT and caspase-3−/− mice (Supplemental Tables II, III, and IV). Similar to WT mice, caspase-3−/− mice that were subjected to 9 minutes SCI exhibit flaccid paraplegia immediately after recovery from anesthesia (Figure 5). Histopathologic analysis revealed extensive neuronal loss in the spinal cord of caspase-3−/− mice harvested at 72 hours after 9 minutes SCI (data not shown). In contrast to WT mice that developed delayed paraplegia after 5 minutes SCI, caspase-3−/− mice that were subjected to 5 minutes SCI exhibited near normal motor function of lower extremity up to 72 hours after reperfusion (P<0.01 versus WT 48 and 72 hours after 5 minutes SCI; Figure 5). Histopathologic analysis of the spinal cord of caspase-3−/− mice subjected to 5 minutes SCI revealed normal appearances of the spinal ventral gray matter at 72 hours after reperfusion (Figure 6A). There was a statistically significant difference (P<0.01) in the number of Nissl-positive spinal cord neurons at 72 hours after reperfusion between WT and caspase-3−/− mice subjected to 5 minutes SCI (Figure 6B).
In the present study, we found that 9 and 5 minutes of SCI produced immediate- and delayed-onset paraplegia, respectively, in mice. Once they became paraplegic, no mice showed any recovery of motor function in this model up to 6 weeks after SCI of either duration. Although degeneration of the spinal cord neurons and motor deficit became apparent 8 hours after 9 minutes SCI, neurodegeneration and paralysis did not manifest until 36 to 48 hours after 5 minutes SCI in WT mice. The delayed neurodegeneration and paraplegia after 5 minutes SCI were associated with reactive astrogliosis and microglial activation as well as marked cavitation in the ventral horn at 48 hours after reperfusion. Caspase-3 activation after 5 minutes SCI precedes the onset of glial activation and extensive neuronal loss. Of note, although 9 minutes SCI produced immediate paraplegia in caspase-3−/− mice, caspase-3 deficiency prevented delayed motor neuron loss and paraplegia after 5 minutes SCI. These observations suggest that caspase-3 activation is required for delayed neurodegeneration and paraplegia after spinal cord ischemic injury.
A variety of animal models including the dog,14 rabbit,4 baboon,15 pig,16 and rat17 have been used to study SCI. The rabbit model of SCI has been particularly popular because it can produce delayed and immediate paraplegia by temporarily occluding the infrarenal aorta. However, the arterial blood supply of the spinal cord in rabbit is almost purely segmental and different from that in humans.18,19 In rats and mice,9 1 anterior and 2 posterior spinal arteries supply the spinal cord, which is similar to humans. Moreover, availability of the variety of genetically altered mouse strains is a clear advantage of mouse models of human diseases compared with rats and rabbit.
We modified a previously reported mouse model of SCI by Lang-Lazdunski and colleagues9 by improving the stability of the model using endotracheal intubation and mechanical ventilation during anesthesia and surgery. The robustness of our model enabled us to determine the relationship between the duration of SCI and neurological outcomes (Figure 2). The duration-dependent acceleration of the onset of paraplegia after SCI observed in this study is consistent with what was reported in rabbits.20 Whether or not duration of ischemia affects outcomes of SCI in humans is incompletely understood. An individual difference in the extent of collateral blood supply and/or surgical techniques is likely to have a significant impact on the outcome of SCI. Nonetheless, our results suggest that severity of ischemia determines the timing of onset of neurodegeneration and paraplegia after SCI in mice. Furthermore, the high long-term mortality rate of paraplegic mice observed in the current study is consistent with a clinical report that showed a decreased survival rate in patients who developed delayed paraplegia after thoracoabdominal aortic operations.21
Delayed paraplegia and neurodegeneration after 5 minutes SCI were associated with marked cavitation in the ventral horn at 48 and 72 hours after reperfusion. The extensive cellular loss was associated with reactive astrogliosis and microglial activation at 48 hours after 5 minutes SCI. Although the role of activated astrocytes and microglia after SCI is incompletely understood, it is possible that necrosis and/or apoptosis activates proinflammatory and phagocytotic activity of glial cells promoting tissue damage.22 The current observation that caspase-3 activation preceded the glial activation supports this hypothesis. In contrast to 5 minutes SCI, no cavitation or glial activation was observed after 9 minutes SCI. Although the reason for the differing impact of 5 and 9 minutes SCI may be multifactorial, it is conceivable that 9 minutes of ischemic insult is too severe for any cells to survive in the ischemic spinal cord regions.
The role of caspase-3 activation and apoptosis in the pathogenesis of delayed motor neuron death has been implicated primarily based on immunohistochemical detection of cleaved caspase-3 in the spinal cord of rabbits that exhibited delayed paraplegia after SCI.4–6 However, the role of caspase-3 in delayed motor degeneration has been questioned because of the failure of a conventional caspase inhibitor to prevent delayed paraplegia7 and lack of caspase-3 activation and apoptosis in the spinal cord of rabbits.8 These conflicting results are at least in part due to the limited potency and/or toxicity of chemical caspase inhibitors and lack of reproducible and standardizable animal models of SCI. In the current study, using caspase-3−/− mice, we demonstrated that caspase-3 is required for the SCI-induced development of delayed paraplegia, but not immediate paraplegia, in mice. Time-dependent changes of the SCI-induced caspase-3 activation in the spinal cord support the critical role of caspase-3 in the pathogenesis of delayed paraplegia (see Figure 4). Although the precise mechanism whereby caspase-3 deficiency prevents SCI-induced delayed paraplegia remains to be determined, our results suggest that inhibition of apoptotic cell death and glial activation in caspase-3−/− mice contributed to the preserved spinal cord neurons after 5 minutes of SCI. To elucidate the role of caspases in the pathogenesis of delayed paraplegia, further studies using genetically altered mouse models and new-generation caspase inhibitors are warranted.
The current study demonstrates that SCI in mice can cause immediate or delayed-onset motor neuron degeneration and paraplegia depending on the duration of ischemia. Our results also provide definitive evidence that caspase-3 activation is required for the development of delayed paraplegia after 5 minutes of SCI in mice. These observations have important clinical implications suggesting the preventive effects of caspase-3 inhibition in SCI induced by surgery or trauma. Furthermore, the unique mouse model of delayed paraplegia described here provides an important experimental platform to further examine the molecular mechanisms of delayed motor neuron degeneration.
Sources of Funding
This work was supported by grant-in-aid for scientific research from the Ministry of Education of Japan (B) 22390299 to M.K. and National Institutes of Health grant R01HL101930 to F.I.
The online-only Data Supplement is available at http://stroke.ahajournals.org/cgi/content/full/STROKEAHA.110.600429/DC1.
Costantino Iadecola, MD, was the Consulting Editor for this paper.
- Received August 19, 2010.
- Revision received February 8, 2011.
- Accepted February 23, 2011.
- © 2011 American Heart Association, Inc.
- Abe K,
- Aoki M,
- Kawagoe J,
- Yoshida T,
- Hattori A,
- Kogure K,
- et al
- Hayashi T,
- Sakurai M,
- Abe K,
- Sadahiro M,
- Tabayashi K,
- Itoyama Y,
- et al
- Lang-Lazdunski L,
- Matsushita K,
- Hirt L,
- Waeber C,
- Vonsattel JP,
- Moskowitz MA,
- Dietrich WD
- Woo M,
- Hakem R,
- Soengas MS,
- Duncan GS,
- Shahinian A,
- Kagi D,
- et al
- Zivin JA,
- Waud DR
- Zivin JA,
- DeGirolami U
- Svensson LG,
- Von Ritter CM,
- Groeneveld HT,
- Rickards ES,
- Hunter SJ,
- Robinson MF,
- et al
- Taira Y,
- Marsala M