Experimental Neurology 172, 199 –210 (2001) doi:10.1006/exnr.2001.7765, available online at http://www.idealibrary.com on
The Immunophilin Ligand FK506 Attenuates the Axonal Damage Associated with Rapid Rewarming Following Posttraumatic Hypothermia Eiichi Suehiro,* ,† Richard H. Singleton,* James R. Stone,* and John T. Povlishock* *Department of Anatomy, Medical College of Virginia, Campus of Virginia Commonwealth University, Box 980709, Richmond, Virginia 23298-0709; and †Department of Neurosurgery, Yamaguchi University School of Medicine, 1-1-1 Minamikogushi, Ube, Yamaguchi 755-8505, Japan Received December 26, 2000; accepted June 14, 2001; published online September 13, 2001
INTRODUCTION Our laboratory has shown that traumatically induced axonal injury (TAI) is significantly reduced by posttraumatic hypothermia followed by slow rewarming. Further, TAI can be exacerbated by rapid rewarming, and the damaging consequences of rapid rewarming can be reversed by cyclosporin A, which is believed to protect via blunting mitochondrial permeability transition (MPT). In this communication, we continue investigating the damaging consequences of rapid posthypothermic rewarming and the protective role of immunophilin ligands using another member of the immunophilin family, FK506, which does not affect MPT but rather inhibits calcineurin. Rats were subjected to impactacceleration brain injury followed by the induction of hypothermia with subsequent rapid or slow posthypothermic rewarming. During rewarming, animals received either FK506 or its vehicle. Three hours postinjury, animals were prepared for the visualization of TAI via antibodies targeting impaired axoplasmic transport (APP) and/or overt neurofilament alteration (RMO-14). Rapid rewarming exacerbated TAI, which was attenuated by FK506. This protection was statistically significant for the APPimmunoreactive fibers but not for the RMO-14positive fibers. Combined labeling, using one chromagen to visualize both axonal changes, suggested that these two immunoreactive profiles revealed two distinct pathologies not occurring along the same axon. Collectively, these studies confirmed previous observations identifying the adverse consequences of rapid rewarming while also showing the complexity of the pathobiology of TAI. Additionally, the demonstration that FK506 is protective suggests that calcineurin may be a major target for neuroprotection. © 2001 Academic Press Key Words: traumatic axonal injury; hypothermia; rewarming; FK506; ␤-amyloid precursor protein (APP); RMO-14.
Traumatic brain injury (TBI) has long been associated with concomitant axonal damage which has been linked to some of the morbidity and mortality associated with this condition (27, 28). To date, considerable insight has been gained into the pathogenesis of traumatic axonal injury (TAI), with the recognition that the traumatic episode does not immediately tear the axon. Rather it results in delayed or secondary axotomy (28). This delayed axotomy has been linked to early focal axolemmal and intraaxonal cytoskeletal changes which progress, over time, to axonal disconnection, resulting, in many cases, in axonal swelling adjacent to the site of disconnection (24, 25, 29). Recently, the use of hypothermia in the treatment of TBI has gained considerable attention based upon the suggestion that it may stabilize the axolemma and/or the underlying cytoskeleton, resulting in the attenuation of TAI. In the laboratory setting, we have demonstrated that the use of both pre- and early postinjury hypothermia provides axonal protection (9, 10). This has been shown by quantifying the total number of swollen/damaged axons per unit area in hypothermic-treated versus nonhypothermic-treated animals and by the analysis of cytoskeletal or other intraaxonal changes linked to the progressive pathobiology associated with TAI (2). In addition to establishing the potential usefulness of early posttraumatic hypothermia, we also demonstrated, in a convincing fashion, that the rate of posthypothermic rewarming is an important independent variable (37). Specifically, we demonstrated that slow or gradual rewarming results in maximal axonal protection, while rapid rewarming, in contrast, significantly exacerbates the magnitude of axonal damage (37). Further, drawing upon data from other fields examining the adverse consequences of rapid posthypothermic rewarming (11), we examined the possibility that this exacerbation of TAI was related to the induc-
0014-4886/01 $35.00 Copyright © 2001 by Academic Press All rights of reproduction in any form reserved.
SUEHIRO ET AL.
tion of mitochondrial permeability transition within damaged axons, thereby contributing to their demise. We demonstrated that the use of cyclosporin A with rapid posthypothermic rewarming was protective, a finding consistent with literature from other organ systems (37). Further, in relation to TAI alone, we demonstrated that cyclosporin A also exerts profound protection, purportedly through its role in blunting mitochondria permeability transition (22, 23); however, in these studies we could not exclude the possibility that cyclosporin A via its modulation of caleineurin also exerted protection. Recently our overall understanding of the pathobiology of TAI and its potential treatment was called into question by the observation that therapies targeting intraaxonal change independent of mitochondrial permeability transition inhibition were also efficacious in blunting the progression of traumatically induced axonal injury. Specifically, we recently demonstrated that the use of FK506, which has no effect on mitochondrial permeability transition, but inhibits the protein phosphatase calcineurin (14), also provides profound axonal protection in normothermic animals (34). In contrast, recent studies in normothermic animals sustaining traumatic contusional injury showed that while cyclosporin A reduced contusional volume, FK506 did not demonstrate comparable protection (32). Companion studies also revealed that this protection was derived from cyclosporin A’s action on the mitochondrial front (38). In view of these somewhat discrepant findings, we sought to evaluate if FK506 could exert a neuroprotective effect in paradigms utilizing hypothermia followed by rapid rewarming. The premise of these studies was that their successful conduct would provide potentially further insight into the pathogenesis of traumatically induced axonal injury, while also shedding new insight into its potential treatment via the use of posttraumatic hypothermia. In addition to our primary goals focusing on the neuroprotective effects of FK506 and hypothermia, we also evaluated the overall utility of these approaches based upon a comprehensive analysis of total axonal damage. As we have recently recognized that TAI may result in disconnected swollen as well as nonswollen axons (31), the current study evaluated both types of axonal injury to provide a full accounting of the axonal populations involved and the degree of neuroprotection provided. METHODS
General Preparation Fifteen male Sprague–Dawley rats weighing 376 to 397 g were used. After anesthesia was induced with 4% isoflurane in 70% nitrous oxide and 30% oxygen, the animals were rapidly intubated and mechanically ventilated with isoflurane at a concentration maintained at 1–2%. The rate of mechanical ventilation was ad-
justed to maintain normocapnia. The femoral artery was cannulated with a PE50 catheter for continuous monitoring of arterial blood pressure and blood gases. Blood gas samples (0.15 cc) were taken prior to the injury, and follow-up blood gas studies (0.15 cc) were performed at 60-min intervals for the duration of the individual experiments. At the same time, the femoral vein was also cannulated for injection of the chosen drugs, which were given before rewarming. This catheter was filled with heparinized saline to maintain venous access. In all animals, both the temporalis muscle and the rectal temperatures were monitored. All animals were prepared and maintained in accordance with our institution’s policies on animal use and care. Experimental Traumatic Brain Injury All animals were subjected to impact-acceleration TBI in a manner consistent with that previously described (4, 16). Briefly, a midline sagittal scalp incision was made over the parietal bones and the skull exposed with blunt dissection. The exposed area was cleaned and dried and a 10 ⫻ 3-mm stainless steel disc helmet was secured with dental acrylic between bregma and lambda. The animal was then placed prone on a foam bed under a 2.5-m Plexiglas tube, with the disc centered immediately under the lower end of the tube. Next a 450-g brass weight was allowed to fall through the tube from a height of 2 m to impact the disk. Based upon criteria established by Marmarou and colleagues, this injury was considered severe (4, 16). After injury the rat was immediately removed from the injury device and ventilated with 100% O 2 for a period of 1 min, at which point the animal was returned to isoflurane anesthesia. This brief period of 100% oxygen was utilized because the injury procedure required that the animal be removed from the ventilator for the induction of injury. Thus, this brief period of postinjury 100% oxygenation was employed to blunt any potential hypoxic episode. Following injury and this brief period of increased oxygen delivery, the helmet was removed, and the skull was examined for sign of fracture which, if found, disqualified the animal from further evaluation. Of the 20 animals used, 3 animals died and 2 sustained skull fractures. Those animals were not included in the study. Hypothermic Procedure Rewarming and FK506 Delivery Both the temporalis muscle and the rectal temperatures of the animals were maintained at 37°C prior to and during the induction of trauma. Hypothermia, which was induced by means of ice packs applied to the body immediately after TBI, was maintained at 32°C for 1 h as verified by both the temporalis muscle and the rectal temperatures. The temporalis muscle temperature was employed based upon the fact that these
POSTHYPOTHERMIC REWARMING AND FK506 IN AXONAL INJURY
readings closely parallel brain temperature (7). Using these general protocols, the animals were assigned randomly to one of three groups. In Group 1 (n ⫽ 5), animals were given 0.48 ml of vehicle (normal saline) intravenously 50 min following the induction of hypothermia. Following vehicle delivery, 60 min after the onset of hypothermia the animals were gradually rewarmed to normothermic levels over a 90-min period. This rewarming was accomplished according to a protocol detailed previously (9, 10, 37) in which the brain temperature was allowed to rise 1°C over a 16-min interval until the temperature reached normothermic levels. In Group 2 (n ⫽ 5), following vehicle delivery in the same fashion described above, the animals were rapidly rewarmed to normothermic levels within a 20min period. In Group 3 (n ⫽ 5), FK506 (2 mg/kg) was injected intravenously with its vehicle (1.67 mg FK506/ml vehicle; total volume 0.48 ml) 10 min prior to rewarming. This dosage was based upon our previous experience with this drug in which we demonstrated that at this dosage the FK506 crosses the blood– brain barrier, achieving intraparenchymal levels consistent with the provision of neuroprotection (34). After injection, the animals were rapidly rewarmed to normothermic levels within a 20-min period. Following the TBI, all rats were allowed to survive for 3 h. Immunocytochemical Studies At 3 h postinjury, all animals were injected intraperitoneally with an overdose of pentobarbital and perfused with 4% paraformaldehyde and 0.1% glutaraldehyde in 0.1 M Millonig’s phosphate buffer. Postperfusion, the brains were removed and transferred to fixative for storage overnight. The next day, brains were placed in a sagittal brain-blocking device, with the middle 2 mm parasagittally blocked to include the pyramids, basal pons, and interpeduncular region of the midbrain. This blocking strategy was based on the fact that, in this model system, maximal axonal damage is seen in the pontomedullary junction within the descending corticospinal tracts (CSpT) and ascending medial lemnisci (ML) (30). After harvesting, this 2-mm-wide sagittal block was serially sectioned with a Vibratome Series 1000 (Polysciences, Inc., Warrington, PA) at a thickness of 40 m. All sections were collected in separate plastic wells in a serial fashion. For every three sections, the first and second sections were processed for visualization with antibodies targeted to the C-terminus of the ␤-amyloid precursor protein (APP) and RMO-14, respectively. Antibodies to the C-terminus of the APP, a marker of impaired axonal transport and axonal damage (1, 5), are more sensitive than those targeting the N-terminus and demonstrate significantly reduced background immunoreactivity (36). The RMO-14 antibodies were chosen based upon our previous experience that these antibodies also identi-
fied sites of axonal damage associated with intraaxonal cytoskeletal change involving neurofilament compaction and sidearm modification (3, 13, 30). Specifically, this antibody has been characterized as recognizing those epitopes on the neurofilament rod domains of the NF-M subunit visualized only upon alteration/loss of their phosphorylated sidearms (6). In the evolution of events progressing to axotomy, it is thought that these immunoreactive foci detected early sites of injury that then progressed to the swollen profiles seen with APP. Finally, using the remaining third group of serial sections, the validity of this premise was assessed via the simultaneous use of both the RMO-14 and the APP antibodies visualized through a common chromagen. The premise here was if both RMO-14 and APP recognized the same injured axonal segment, the resulting immunoreactivity would be quantitatively the same as that seen with the use of either antibody alone. Conversely, if this was not the case and these antibodies identified different populations of injured axons, the resulting immunoreactivity would be additive. In all studies, appropriate controls were used, including primary antibody deletion and/or preabsorption studies. Single-Labeling Protocol Sections from the first and second groups were rinsed 3 ⫻ 10 min in phosphate-buffered saline (PBS) and endogenous peroxidase activity was blocked with 0.5% H 2O 2 in PBS for 30 min. The sections were placed into a sodium-citrate buffer (pH 6.0) and transferred to a temperature-controlled microwave (Ted Pella, Redding, CA) in order to achieve optimal immunoreactivity with the retention of excellent structural detail (35). Following microwave processing, sections were allowed to cool in the same buffer used for microwaving for 20 min, rinsed 3 ⫻ 10 min in PBS, and preincubated for 40 min with 0.2% Triton X (Sigma Chemical Co., St. Louis, MO) in 10% normal goat serum (NGS), for sections to be reacted with APP, or normal horse serum (NHS), for sections to be reacted with RMO-14 in PBS. The tissue was then incubated overnight either in rabbit C-terminus APP antibody at a dilution of 1:5000 (Zymed Laboratories) in 1% NGS in PBS (first group) or in mouse RMO-14 antibody at a dilution of 1:600 (kindly provided by Dr. John Q. Trojanowski, University of Pennsylvania Department of Pathology) in 1% NHS in PBS (second group). After 3 ⫻ 10-min washes in PBS containing 1% NGS or NHS, sections were incubated in biotinylated goat anti-rabbit immunoglobulin (diluted 1:200 in 1% NGS/PBS) (Vector, Burlingame, CA) or in 1:200 dilution of biotinylated antimouse/rat adsorbed immunoglobulin derived from horse (Vector, Burlingame, CA) for 60 min followed by 3 ⫻ 10-min rinses in PBS. After incubation in an avidin– biotin–peroxidase complex (ABC Standard Elite Kit; Vector) at a dilution of 1:200 for 60 min and rins-
SUEHIRO ET AL.
ing in PBS and 0.1 M phosphate buffer 3 ⫻ 10 min and 2 ⫻ 10 min, respectively, sections were processed for visualization of the immunohistochemical complex using 0.05% diaminobenzidine (DAB) (Sigma Chemical Co., St. Louis, MO), 0.01% hydrogen peroxide, and 0.3% imidazole in 0.1 M phosphate buffer. Next the sections were mounted on gelatin-coated glass slides and serially dehydrated, and coverslips were applied. Combined Single-Labeling Protocol Sections from the third group were subjected to the combined single-labeling strategy, which included the use of a single chromagen for the immunohistochemical visualization of both RMO-14 and C-APP antibodies. Following the above-described incubations for the RMO-14 primary antibody and the biotinylated antimouse/rat adsorbed secondary antibody, sections were rinsed three times for 10 min in PBS containing 1% NHS and incubated for 40 min with PBS containing 10% NGS. Next, the sections were incubated overnight in rabbit C-terminus APP antibody at a dilution of 1:5000 in 1% NGS in PBS, washed in 1% NGS/PBS 3 ⫻ 10 min, and incubated in 1:200 dilution of biotinylated anti-rabbit immunoglobulin. After washing in PBS, sections were incubated in 1:200 dilution of ABC Standard Elite Kit for 60 min and then reacted with DAB/ H 2O 2 with imidazole enhancement for visualization of secondary antibody. The sections were mounted on gelatin-coated glass slides, serially dehydrated, and coverslipped. Quantitative Analysis After the completion of the immunocytochemical procedure, the slides were transferred to a microscope interfaced with a computer-assisted image analysis system (Vision Gauge) to count the APP, RMO-14, or combined immunoreactive axonal profiles within specific sites detailed below. Five semiserial brain-stem sections from each of the three immunocytochemical groups (APP, RMO-14, and combined) were taken from consistent regions of brain stem wherein the CSpT and ML were best delineated. The sections were examined with a Nikon Eclipse 800 light microscope fitted with a Sony Catseye digital camera. Sections from each group were maintained in strict serial order. Based upon our previous experience in this model (2, 8, 27), the CSpT and ML at the pontomedullary junction were captured and enlarged using a 20⫻ objective lens with a 2.5⫻ intermediate lens (total 50⫻). All of the images were then coded and randomized by another investigator to allow for blinded analysis. The image was viewed on a monitor and the sampling area was delineated by a 100,000m 2 rectangle. The numbers of damaged APP, RMO14, or combined immunoreactive axonal profiles larger than 5 m in diameter within this area were then
counted and expressed as the density of APP-positive axons, RMO-14-positive axons, or combined-positive axons per square millimeter. For statistical analysis, the numbers of damaged axons in the experimental groups were tested by analysis of variance (ANOVA) followed by Dunn’s procedure test for multiple comparisons. Differences were considered statistically significant at P ⬍ 0.05. Last, the axonal numbers obtained via RMO-14 and APP single labeling were summed and divided by the numbers obtained via combined labeling to generate a ratio. The premise here is that if the ratio approximates 1, it would indicate that the antibodies targeted antigenic sites on different axons. Alternatively, if this ratio exceeded 1, it would suggest labeling of antigenic sites colocalized on the same axon. RESULTS
General Physiological Observations As assessed by the temporalis muscle and rectal temperature probes, the animals subjected to hypothermia required on average 12.1 ⫾ 0.2 min (mean ⫾ SEM) to reach the target temperature of 32°C. After 1 h of hypothermic intervention, Group 1 was rewarmed slowly over a 90-min interval. In this rewarming phase, care was taken to ensure graduated rewarming, with a typical temperature rise of 1°C per 21.7 ⫾ 0.1 min. Groups 2 and 3 were rewarmed rapidly within 14.0 ⫾ 0.6 and 13.6 ⫾ 0.4 min, respectively (Fig. 1). All blood gas levels were subjected to statistical analysis using ANOVA. Although the PaCO 2 levels rose following hypothermia with rewarming, these values were within the normal physiological limits (Table 1). Further, the PaO 2 and pH remained within normal limits throughout the duration of the experiment (Table 1). Qualitative Immunocytochemical Findings Light microscopic examination of animals subjected to TAI showed focal immunocytochemical reaction products within injured axons in CSpT and ML consistent with our previous studies (30, 36). The greatest number of immunoreactive axons was localized to brain-stem white matter tracts at the pontomedullary junction. The anatomical distribution and morphological characteristics of the immunopositive, damaged axons were similar in all experimental groups. APPimmunoreactive axonal profiles appeared as swollen bulbs, many of which appeared disconnected. These profiles were diffusely scattered throughout both CSpT and ML (Figs. 2A and 2a). In contrast, RMO-14 immunoreactivity was confined to elongate, nonswollen axonal segments, with a distribution similar to that of APP (Figs. 2B and 2b). The combined labeling strategy
POSTHYPOTHERMIC REWARMING AND FK506 IN AXONAL INJURY
FIG. 1. The changes in the mean temporalis muscle temperature throughout the surgical procedure are shown. The data points represent 5-min intervals. In all experimental groups, it required an average of 12.1 ⫾ 0.2 min to reduce the temporalis muscle temperature to 32°C. During the 1-h period of hypothermia, the temperatures were maintained; thereafter, the animals were rewarmed to normothermic levels over a 90-min period in Group 1 or within a 20-min period in Groups 2 and 3.
did not allow the discrimination of specific immunoreactive foci; however, the use of a common chromagen consistently revealed swollen immunoreactive profiles reminiscent of those seen with APP as well as elongate immunoreactive axonal profiles reminiscent of those seen with RMO-14 (Figs. 2C and 2c). Interestingly, these immunoreactive swollen and elongate profiles did not commonly overlap or share proximity, suggesting that they represented different types of damaged
axons. The axonal number seen via this combined strategy also appeared greater than that seen with the use of either APP or RMO-14 alone. Although the finding of immunopositive axonal profiles identified using APP, RMO-14, and the combined strategy was consistent in all animals studied, qualitative analysis alone did suggest potential differences in axonal density in each experimental subgroup analyzed. Specifically, in those animals subjected to rapid
TABLE 1 Physiological Variables in Rats Subjected to TBI Time course Parameter MABP (mm Hg)
PaO 2 (mm Hg)
PaCO 2 (mm Hg)
1 2 3 1 2 3 1 2 3 1 2 3
104.7 ⫾ 12.4 105.1 ⫾ 7.4 96.7 ⫾ 11.8 7.42 ⫾ 0.03 7.42 ⫾ 0.02 7.41 ⫾ 0.03 151.1 ⫾ 18.9 126.2 ⫾ 6.6* ,† 147.3 ⫾ 13.9 37.8 ⫾ 1.9 39.3 ⫾ 3.8 39.2 ⫾ 2.6
107.3 ⫾ 15.2 107.2 ⫾ 11.1 97.3 ⫾ 25.3 7.40 ⫾ 0.02 7.40 ⫾ 0.02 7.42 ⫾ 0.02 154.4 ⫾ 49.8 144.4 ⫾ 47.9 152.2 ⫾ 36.6 38.0 ⫾ 2.6 39.1 ⫾ 2.5 38.6 ⫾ 2.8
89.3 ⫾ 12.8 95.1 ⫾ 15.9 100.5 ⫾ 8.8 7.41 ⫾ 0.02 7.37 ⫾ 0.02* 7.37 ⫾ 0.03* 153.7 ⫾ 26.3 147.3 ⫾ 48.3 126.7 ⫾ 18.1 36.6 ⫾ 2.0 41.3 ⫾ 2.5* 39.3 ⫾ 3.6
87.9 ⫾ 6.7 88.1 ⫾ 8.5 86.8 ⫾ 10.5 7.40 ⫾ 0.02 7.40 ⫾ 0.02 7.39 ⫾ 0.02 127.7 ⫾ 19.7 136.2 ⫾ 20.6 136.5 ⫾ 10.7 37.8 ⫾ 1.4 38.2 ⫾ 1.9 37.4 ⫾ 2.1
Note. Blood gas values analyzed at 60-min intervals are shown. Note that these values are within the normal physiological range. All values are expressed as means ⫾ SEM. * P ⬍ 0.05 compared to Group 1. † P ⬍ 0.05 compared to Group 3.
SUEHIRO ET AL.
FIG. 2. These photomicrographs taken from rat brains 3 h after the impact-acceleration brain injury show the damaged/APP-immunoreactive axonal profiles (A, a), the damaged/RMO-14-immunoreactive axonal profiles (B, b), and the damaged/combined immunoreactive axonal profiles (C, c). (A, B, C) In the corticospinal tract within the pontomedullary junction. (a, b, c) In the medial lemniscus. Note that at this time point the APP-immunoreactive axons appear as bulbs (arrowheads), while the RMO-14-positive axons are more compact and linear (arrows) (original magnification 50⫻).
POSTHYPOTHERMIC REWARMING AND FK506 IN AXONAL INJURY
FIG. 3. Comparison of the mean number of APP- (A, C) and RMO-14 (B, D) immunoreactive axonal profiles with the corticospinal and medial lemniscal tracts in the slow rewarming (1), rapid rewarming (2), and rapid rewarming/FK506 (3) groups. (A, B) The mean axonal number in the corticospinal tract. (C, D) The mean axonal number in the medial lemniscus. Error bars represent standard error of the mean.
rewarming, qualitative assessment suggested numerous APP-immunoreactive profiles in the CSpT. In contrast, those animals undergoing gradual rewarming appeared to manifest reduced numbers in the CSpT. Further, animals receiving FK506 following rapid rewarming also appeared to show reduced numbers of APP-positive axons in the CSpT compared to vehicletreated animals. In contrast, no apparent differences could be recognized in the ML of any of the experimental/treatment groups. Quantitative Findings Single labeling. In the CSpT, the mean number of damaged/APP-immunoreactive axonal profiles/mm 2 significantly decreased from 860.8 ⫾ 52.0 (mean ⫾
SEM) in the rapid rewarming group to 458.0 ⫾ 48.7 in the slow rewarming group (P ⬍ 0.01). The use of rapid rewarming clearly exacerbated TAI (Fig. 3A). Of further significance was the finding that in the group subjected to rapid rewarming but treated with FK506, the mean number of damaged/APP-immunoreactive axonal profiles/mm 2 decreased to 398.0 ⫾ 41.3 in comparison to the vehicle-treated, rapidly rewarmed group (P ⬍ 0.01). Further, no significant difference was detected between the slow rewarming group and the rapid rewarming group treated with FK506 (Fig. 3A). A decrease in RMO-14-immunoreactive axons was also detected in CSpT. The mean number of damaged/ RMO-14-immunoreactive axonal profiles/mm 2 in the slow rewarming group was 405.6 ⫾ 72.7, a statistically
SUEHIRO ET AL.
significant reduction from the mean density in the rapid rewarming group, 838.4 ⫾ 140.6, P ⬍ 0.05. The mean axonal number/mm 2 in the FK506-treated group, 632.8 ⫾ 127.7, was not statistically significant in comparison to the rapid rewarming group (Fig. 3B). In contrast, in the ML no significant differences were detected between any experimental group (Figs. 3C and 3D). The mean number of damaged/APP-immunoreactive axonal profiles/mm 2 varied from 122.0 ⫾ 24.3 in the rapid rewarming group to 67.6 ⫾ 7.7 and 87.2 ⫾ 27.4 in the slow rewarming and FK506-treated groups, respectively. Further, the mean density of RMO-14immunoreactive profiles/mm 2 varied from 168.0 ⫾ 52.5 to 300.0 ⫾ 76.3 and 250.0 ⫾ 142.3 in the slow rewarming, rapid rewarming, and FK506-treated groups, respectively. Combined labeling. In the combined-labeling strategy, quantitative assessment of the CSpT from the slow rewarming group revealed a total number of immunoreactive profiles which approximated the sum of the number of axons seen with RMO-14 and APP alone. Specifically, as noted above, RMO-14 alone revealed a mean number of 405.6 ⫾ 72.7 axons/mm 2. APP alone revealed a mean number of 458.0 ⫾ 48.7 axons/mm 2, and the combined analysis showed a mean number of 735.2 ⫾ 84.6 axons/mm 2 (Fig. 4A). In this combined analysis, individual swollen or linear immunoreactive axons were counted individually as one immunoreactive profile. Those immunoreactive segments showing swelling contiguous with a linear segment were counted as one. As swellings contiguous with linear immunoreactive segments were not typically observed within the same axon, this was consistent with the quantitative data which suggested two distinct populations of reactive change. In the rapidly rewarmed groups, this combined-labeling approach, when used in the CSpT, provided data comparable to that described above. Namely, rapid rewarming exacerbated axonal injury while FK506 blunted the increase in traumatic axonal injury seen with rapid rewarming. Specifically, the mean densities of combined immunoreactive axonal profiles/mm 2 were 1583.2 ⫾ 180.5 in the rapid rewarming group, and 735.2 ⫾ 84.6 and 813.2 ⫾ 99.5 in the slow rewarming and FK506-treated groups, respectively. Importantly, as also noted previously, the mean numbers of APPand RMO-14-immunoreactive axonal profiles in each experimental group, when summed, approximated the numbers seen via the combined approach (Figs. 4A, 4C, and 4E). Collectively these data suggest that APP- and RMO-14-immunoreactive profiles revealed separate populations of reactive axons in the CSpT. Further, the ratios obtained by dividing the sum of these individual counts by the combined group count approximated 1 (Fig. 4), again suggesting that these antibodies identified epitopes on two different axon populations.
The use of combined labeling in the ML, in contrast, did not reveal the same picture seen in the CSpT. Here the counts obtained via combined labeling did not approximate the sums of the immunoreactive profiles seen through the use of RMO-14 and APP immunoreactivity alone (Figs. 4B, 4D, and 4F). In contrast to the CSpT, the numbers of axons displaying swellings contiguous with a linear immunoreactive segment were more conspicuous in the ML. Specifically, the mean densities of combined immunoreactive profiles/mm 2 were 163.6 ⫾ 54.5, 326.4 ⫾ 100.2, and 304.8 ⫾ 196.3 in the slow rewarming, rapid rewarming, and FK506treated groups, respectively. Interestingly, the number of combined immunoreactive profiles paralleled the number of RMO-14-immunoreactive profiles visualized through the single-labeling strategy in the slow and rapid rewarming groups (Fig. 4B and 4D), and when the individual RMO-14 and APP counts were divided by these combined counts, the ratios obtained (1.4 –1.7) suggested that within the ML RMO-14 and APP antibodies colocalized to some of the same injured axonal segments (Figs. 4B, 4D, and 4F). DISCUSSION
The results of the current communication are considered intriguing from several perspectives. First and foremost, they confirm and supplement observations from a previous communication from our lab showing that, in the corticospinal system, rapid posthypothermic rewarming significantly exacerbates axonal damage compared with slow posthypothermic rewarming (37). Importantly, in the current communication, this exacerbation of axonal damage was identified using two specific markers of axonal injury, APP and RMO14, which denote two different features of traumatic axonal injury, impairment of axoplasmic transport (5, 31, 33, 35, 36) and intraaxonal neurofilament compaction (30, 31), respectively. As we have recently posited that these two forms of traumatic axonal change may not share the same pathogenesis (31), we believe that the involvement of these two markers reflects the overall widespread adverse consequences of rapid posthypothermic rewarming in terms of the ensuing axonal response. Not only do these studies confirm the damaging consequences of rapid posthypothermic rewarming, but also they suggest the potential usefulness of various therapeutic approaches to blunt these damaging consequences. Particularly noteworthy within the CSpT is the demonstration that FK506, a member of the immunophilin ligand family that inhibits calcineurin (14), significantly protects against the damaging consequences associated with rapid rewarming in the CSpT as evidenced by the reduction of APP-immunoreactive fibers therein. Importantly, although protection of APP-immunoreactive fibers was seen, it is of
POSTHYPOTHERMIC REWARMING AND FK506 IN AXONAL INJURY
FIG. 4. The total immunoreactive axonal numbers seen with the use of RMO-14 and APP antibodies visualized either individually or in combination are plotted. These plots were performed in both the corticospinal (A, C, E) and the medial lemniscal (B, D, F) tracts in the slow rewarming (1), the rapid rewarming (2), and the rapid rewarming/FK506 (3) groups. Additionally, ratios (r) were generated by summing the axonal numbers obtained via RMO-14 and APP single labeling and dividing these by the numbers of immunoreactive axons obtained via combined labeling.
note that similar protection was not provided in terms of RMO-14 immunoreactivity. The reason for this potential lack of neuroprotection is, at present, unclear. Previous studies of traumatic axonal injury in para-
digms using hypothermia followed with rapid rewarming, coupled with the use of cyclosporin A to attenuate mitochondrial permeability transition, did not directly assess comparable ROM-14 immunoreactivity (37).
SUEHIRO ET AL.
However, in studies examining the protective effects of cyclosporin A, in normothermic as well as hypothermic animals subjected to gradual rewarming, we have shown that cyclosporin A is capable of reducing the number of ROM-14- and APP-immunoreactive axons, a finding consistent with neuroprotection. Thus, in the traumatic situation alone, not complicated by rapid rewarming, cyclosporin A is capable of protecting or blunting the generation of increased AAP-immunoreactive axons while also blunting a similar rise in ROM14-immunoreactive fibers (22, 23). It is of interest that previous studies focusing on RMO-14 immunoreactivity within injured axons have shown that the RMO-14 immunoreactivity correlates with the presence of calpain-mediated spectrin proteolysis and mitochondrial damage (2, 30, 31), all of which speak to catastrophic intraaxonal damage. Perhaps the cyclosporin A proves successful in these paradigms because of its ability to blunt the mitochondrial damage found at these sites, thereby attenuating the related RMO-14-linked change and calpain-mediated spectrin proteolysis. It will be of obvious interest to determine if the RMO-14targeted actions of cyclosporin A persist with hypothermia followed by rapid rewarming. The fact that FK506 was protective in the current paradigm within the corticospinal system was somewhat unexpected based upon our previous experience with cyclosporin A. Based upon past studies, we posited that any protection provided following posthypothermic rewarming was based solely upon its ability to protect mitochondria from undergoing mitochondrial permeability transition (22, 23). The results of the current communication force us to reevaluate this issue, as FK506 has no known action on mitochondrial permeability transition (14). Rather, it exerts its influence through the inhibition of calcineurin, which may have multiple effects upon the intraaxonal cytoskeletal front, including the inhibition of calcineurin-mediated, dephosphorylation-dependent processes (14, 20). Some investigations have reported that changes in neurofilament phosphorylation status can significantly alter the axonal cytoskeleton. Specifically, changes in neurofilament side-arm phosphorylation status alters the amount of side-arm extension, thus changing interfilament spacing and, indirectly, total axonal caliber (3, 13, 17, 19, 21). Recently, we evaluated the neuroprotective effects of cyclosporin A versus FK506 within the context of traumatic brain injury not complicated by the use of posttraumatic hypothermia and/or any rewarming paradigm (34). Specifically, in traumatic brain injury followed by the administration of cyclosporin A and/or FK506, we found significant axonal protection, which was even more dramatic within the dosing range used for FK506. Collectively, these findings together with the findings of the current communication emphasize the potential role of calcineurin in the pathobiology of traumatic axonal injury as well as
the involvement of calcineurin in the exacerbation of traumatic axonal injury caused by rapid rewarming. One unanticipated finding of the current communication was the fact that the above-described exacerbation of axonal damage by rapid rewarming and its subsequent blunting through the use of FK506 could not be confirmed within the ML. The reason for this at the moment is unclear, but two possibilities may explain this discrepancy. First, it is known that the fibers of the ML are different from those found in the CSpT, as the ML contains large-caliber, heavily myelinated axons of cytoskeletal composition different from that of those found in the CSpT (18). Recent communications from our lab have highlighted the differences in pathogenesis of axonal injury within the ML versus the CSpT (31). Conceivably, such differences in pathogenesis may explain these unanticipated results. Second, and perhaps more likely, is the fact that the number of damaged fibers within the ML is relatively small in comparison to the numbers seen in the corticospinal system, and further, the standard error is high. Thus, this small fiber number, coupled with the high standard error, may explain our failure to find statistically significant changes in this fiber system. Obviously, this will mandate further investigation, perhaps revisiting this issue in a higher order animal such as a micropig, wherein proportionally more damaged axons per unit area will be found within the ML. One caveat that should be added to our consideration of the potential protective effects of FK506 is the fact that we have not ruled out other potential systemic and/or local factors that could also modulate the progression of axonal damage. For example, it is known that FK506 use can result in hyperkalemia and/or hypomagnemia (8, 15), both of which can influence CNS injury and repair. Although these ions were not assessed in the current communication, we would point out that all studies reporting such ionic alteration involved more chronic FK506 administration in comparison to the acute, one-time injection strategy employed in the current communication. Thus, we do not believe that this is a major confound. On a different issue, FK506 has also been shown to inhibit APP transcription, which can be activated in various pathological states, including traumatic brain injury (12). While we cannot rule out a transcription-related mode of protection in terms of the intraaxonal APP accumulation assessed in the current communication, we suggest that the time frames considered in the current study would seem too short to invoke a major transcriptionrelated effect. In addition to the potential implications of this study in terms of posttraumatic hypothermic rewarming and potential neuroprotective effects of FK506, we believe the current communication has other important implications for our understanding of the pathobiology of traumatic axonal injury. Specifically, in the present
POSTHYPOTHERMIC REWARMING AND FK506 IN AXONAL INJURY
communication the use of combined labeling which utilized the same chromagen to identify sites of neurofilament alteration (RMO-14) and impaired axoplasmic transport (APP) indicates that the numbers of axons seen with this combined approach did not equal the number of axons seen through the use of either RMO-14 or APP alone, which we initially assumed to be the case. This assumption was based upon our belief that these immunoreactive profiles (RMO-14 and APP) delineated different sequences of pathologic change occurring within the same axon at different points in time. Specifically, in our original studies we believed that the neurofilament compaction, identified via the RMO-14 antibody, accompanied other local cytoskeletal changes that led to an upstream impairment of axoplasmic transport (24, 25, 30). This was believed to result in local swelling which was associated with APP accumulation. In contrast, more recent studies from our laboratory (31) demonstrated that few RMO-14immunoreactive axons were spatially or temporally linked to local APP-positive swellings, thereby demonstrating that two different axonal responses were ongoing. The lack of swelling-associated APP immunoreactivity in the RMO-14-positive axons was posited to be linked to calpain activation, with the conversion of anterograde to retrograde transport and the subsequent failure of local axonal swelling (31). This potential mechanism, however, requires confirmation. The findings of the current communication appear to confirm the existence of such differential axonal responses to injury. As noted, the total number of axons seen using the combined approach frequently approximated the sum of both the RMO-14- and the APPreactive populations in the CSpT and the ML. Further, even with single labeling alone the numbers of RMO14- and APP-positive axons did not typically correspond, collectively suggesting that these two individual markers labeled two disparate populations of axons with two different forms of initiating pathogenesis and fate. Why variation existed between the axonal counts obtained via combined labeling and the total counts seen with both single-labeling protocols is not entirely clear; however, this most likely reflects the fact, as we reported recently (31), that in some cases RMO-14 and APP can colocalize to the same axonal segment thereby skewing the total counts obtained. Collectively then, these combined immunocytochemical studies clearly confirm our recent observations (31) regarding the diverse pathogenesis of traumatically induced axonal change, suggesting that caution must be exercised when focusing on one marker of axonal change. Further, the fact that the use of the combined labeling provides a comprehensive overview of the total number of axons involved in a given pathogenesis suggests that this may be a useful approach in addressing the overall magnitude of traumatically induced reactive change.
In sum, the current communication provides unique insight into multiple features of traumatic axonal injury. It emphasizes the adverse consequences of rapid posthypothermic rewarming, again suggesting the efficiency of various immunophilin ligand family members and their potential protective roles via calcineurin inhibition. Finally, this study confirms and extends recent communications from our lab, speaking to the complexity of the pathobiology of traumatic axonal injury and strongly suggesting that the APP-immunoreactive swellings and RMO-14-immunopositive compacted axonal segments do not target identical damaged axonal segments undergoing the same biological fate. ACKNOWLEDGMENTS This work is supported by Grant NS 20193. We thank Susan Walker, Lynn Davis, and Thomas Coburn for their excellent technical support.
Blumbergs, P. C., G. Scott, J. Manavis, H. Wainwright, D. A. Simpson, and Mclean. 1995. Topography of axonal injury as defined by amyloid precursor protein and the sector scoring method in mild and severe closed head injury. J. Neurotrauma 12: 565–572. Bu¨ki, A., H. Koizumi, and J. T. Povlishock. 1999. Moderate posttraumatic hypothermia decreases early calpain-mediated proteolysis and concomitant cytoskeletal compromise in traumatic axonal injury. Exp. Neurol. 159: 319 –328. Carden, M. J., W. W. Schlaepfer, and V. M. Lee. 1985. The structure, biochemical properties, and immunogenicity of neurofilament peripheral regions are determined by phosphorylation state. J. Biol. Chem. 260: 9805–9817. Foda, M. A., and A. Marmarou. 1994. A new model of diffuse brain injury in rats. II: Morphological characterization. J. Neurosurg. 80: 301–313. Gentleman, S. M., M. J. Nash, C. J. Sweeting, D. I. Graham, and G. W. Roberts. 1993. ␤-Amyloid precursor protein (␤APP) as a marker for axonal injury after head injury. Neurosci. Lett. 160: 139 –144. Hall, G. F., and M.-Y. Lee. 1995. Neurofilament sidearm proteolysis is a prominent early effect of axotomy in lamprey giant central neurons. J. Comp. Neurol. 353:38 – 49. Jiang, J. Y., B. G. Lyeth, G. L. Clifton, L. W. Jenkins, R. J. Hamm, and R. L. Hayes. 1991. Relationship between body and brain temperature in traumatically brain-injured rodents. J. Neurosurg. 74: 492– 496. Kaplan, B., Z. Wang, M. M. Abecassis, J. P. Fryer, F. P. Stuart, and D. B. Kaufman. 1996. Frequency of hyperkalemia in recipients of simultaneous pancreas and kidney transplant bladder drainage. Transplantation 62: 1174 –1175. Koizumi, H., and J. T. Povlishock. 1998. Posttraumatic hypothermia in the treatment of axonal damage in an animal model of traumatic axonal injury. J. Neurosurg. 89: 303–309. Koizumi, H., and J. T. Povlishock. 2001. Early posttraumatic hypothermia provides persisting axonal protection and improves behavioral outcome. Brain Res., in press. Leducq, N., M. C. Delmas-Beauvieux, I. Bourdel-Marchasson, S. Dufour, J. L. Gallis, P. Canioni, and P. Diolez. 1998. Mito-
SUEHIRO ET AL. chondrial permeability transition during hypothermic to normothermic reperfusion in rat liver demonstrated by the protective effect of cyclosporin A. Biochem. J. 336: 501–506. Lee, R. K. K., S. Knapp, and R. J. Wurtman. 1999. Prostaglandin E2 stimulates amyloid precursor protein gene expression: Inhibition by immunosuppressants. J. Neurosci. 19: 940 –947. Lee, V. M.-Y., M. J. Carden, W. W. Schlaepfer, and J. Q. Trojanowski. 1987. Monoclonal antibodies distinguish several differentially phosphorylated states of the two largest rat neurofilament subunits (NF-H and NF-M) and demonstrate their existence in the normal nervous system of adult rats. J. Neurosci. 7: 3474 –3488. Liu, J., J. D. Farmer Jr., W. S. Lane, J. Friedman, I. Weissman, and S. L. Schreiber. 1991. Calcineurin is a common target of cyclophilin-cyclosporin A and FKBP-FK506 complexes. Cell 66: 807– 815. Lote, C. J., A. Thewles, J. A. Wood, and T. Zafar. 2000. The hypomagnesaemic action of FK506: Urinary excretion of magnesium and calcium and the role of parathyroid hormone. Clin. Sci. 99: 285–292. Marmarou, A., M. A. Foda, W. van den Brink, J. Campbell, H. Kita, and K. Demetriadou. 1994. A new model of diffuse brain injury in rats. I: Pathophysiology and biomechanics. J. Neurosurg. 80: 291–300. Martin, R., R. Door, A. Ziegler, W. Warchol, J. Hahn, and D. Breitig. 1999. Neurofilament phosphorylation and axon diameter in the squid giant fibre system. Neuroscience 88: 327–336. Martz, D., J. Garner, and R. J. Lasek. 1989. Protein changes during anterograde-to-retrograde conversion of axonally transported vesicles. Brain Res. 476: 199 –203. Mata, M., N. Kupina, and D. J. Fink. 1992. Phosphorylationdependent neurofilament epitopes are reduced at the node of Ranvier. J. Neurocytol. 21: 199 –210. Morioka, M., J. Hamada, Y. Ushio, and E. Miyamoto. 1999. Potential role of calcineurin for brain ischemia and traumatic injury. Prog. Neurobiol. 58: 1–30. Nixon, R. A., P. A. Paskevich, R. K. Sihag, and C. Y. Thayer. 1994. Phosphorylation on carboxyl terminus domains of neurofilament proteins in retinal ganglion cell neurons in vivo: Influences on regional neurofilament accumulation, interneurofilament spacing, and axon caliber. J. Cell. Biol. 126: 1031–1046. Okonkwo, D. O., A. Buki, R. Siman, and J. T. Povlishock. 1999. Cyclosporin A limits calcium-induced axonal damage following traumatic brain injury. NeuroReport 10: 353–358. Okonkwo, D. O., and J. T. Povlishock. 1999. An intrathecal bolus of cyclosporin A before injury preserves mitochondrial integrity and attenuates axonal disruption in traumatic brain injury. J. Cereb. Blood Flow Metab. 19: 443– 451. Pettus, E. H., C. W. Christman, M. L. Giebel, and J. T. Povlishock. 1994. Traumatically induced altered membrane permeability: Its relationship to traumatically induced reactive axonal change. J. Neurotrauma 11: 507–522.
Pettus, E. H., and J. T. Povlishock. 1996. Characterization of a distinct set of intra-axonal ultrastructural changes associated with traumatically induced alteration in axolemmal permeability. Brain Res. 722: 1–11. Pijak, D. S., G. F. Hall, P. J. Tenicki, A. S. Boulos, D. I. Lurie, and M. E. Selzer. 1996. Neurofilament spacing, phosphorylation, and axon diameter in regenerating and uninjured lamprey axons. J. Comp. Neurol. 368: 569 –581. Povlishock, J. T., D. E. Erb, and J. Astruc. 1992. Axonal response to traumatic brain injury: Reactive axonal change, deafferentation, and neuroplasticity. J. Neurotrauma 9(Suppl): S189 –S200. Povlishock, J. T., and C. W. Christman. 1995. The pathobiology of traumatically induced axonal injury in animals and humans: A review of current thoughts. J. Neurotrauma 12: 555–562. Povlishock, J. T., and E. H. Pettus. 1996. Traumatically induced axonal damage: Evidence for enduring changes in axolemmal permeability with associated cytoskeletal change. Acta Neurochir. Suppl. 66: 81– 86. Povlishock, J. T., A. Marmarou, T. McIntosh, J. Q. Trojanowski, and J. Moroi. 1997. Impact acceleration injury in the rat: Evidence for focal axolemmal change and related neurofilament sidearm alteration. J. Neuropathol. Exp. Neurol. 56: 347–359. Povlishock, J. T., and J. R. Stone. 2001. Traumatic axonal injury. In Head Trauma: Basic, Preclinical and Clinical Directions. (L. P. Miller and R. L. Hayes, Eds.), Wiley, New York. Scheff, S. W., and P. G. Sullivan. 1999. Cyclosporin A significantly ameliorates cortical damage following experimental traumatic brain injury in rodents. J. Neurotrauma 16: 783–792. Sherriff, F. E., L. R. Bridges, and S. Sivaloganathan. 1994. Early detection of axonal injury after human head trauma using immunocytochemistry for beta-amyloid precursor protein. Acta Neuropathol. 87: 55– 62. Singleton, R. H., J. R. Stone, D. O. Okonkwo, A. J. Pellicane, and J. T. Povlishock. 2001. The immunophilin ligand FK506 attenuates axonal injury in an impact-acceleration model of traumatic brain injury. J. Neurotrauma 18(6): 607– 614. Stone, J. R., S. A. Walker, and J. T. Povlishock. 1999. The visualization of a new class of traumatically injured axons through the use of a modified method of microwave antigen retrieval. Acta Neuropathol. 97: 335–345. Stone, J. R., R. H. Singleton, and J. T. Povlishock. 2000. Antibodies to the C-terminus of the ␤-amyloid precursor protein (App): A site specific marker for the detection of traumatic axonal injury. Brain Res. 871: 288 –302. Suehiro, E., and J. T. Povlishock. 2001. Exacerbation of traumatically induced axonal injury by rapid posthypothermic rewarming and attenuation of axonal change by cyclosporin A. J. Neurosurg. 94: 493– 498. Sullivan, P. G., M. B. Thompson, and S. W. Scheff. 1999. Cyclosporin A attenuates acute mitochondrial dysfunction following traumatic brain injury. Exp. Neurol 160: 226 –234.