Cyclosporin-A Treatment Attenuates Delayed Cytoskeletal Alterations and Secondary Axotomy Following Mild Axonal Stretch Injury

: Following central nervous system trauma, diffuse axonal injury and secondary axotomy result from a cascade of cellular alterations including cytoskeletal and mitochondrial disruption. We have examined the link between intracellular changes following mild/ moderate axonal stretch injury and secondary axotomy in rat cortical neurons cultured to relative maturity (21 days in vitro ). Axon bundles were transiently stretched to a strain level between 103% and 106% using controlled pressurized ﬂuid. Double-immunohistochemical analysis of neuroﬁlaments, neuronal spectrin, a -internexin, cyto-chrome-c, and ubiquitin was conducted at 24, 48, 72, and 96-h postinjury. Stretch injury resulted in delayed cytoskeletal damage, maximal at 48-h postinjury. Accumulation of cytochrome-c and ubiquitin was also evident at 48 h following injury and colocalized to axonal regions of cytoskeletal disruption. Pretreatment of cultures with cyclosporin-A, an inhibitor of calcineurin and the mitochondrial membrane transitional pore, reduced the degree of cytoskeletal damage in stretch-injured axonal bundles. At 48-h postinjury, 20% of untreated cultures demonstrated secondary axotomy, whereas cyclosporin A-treated axon bundles remained intact. By 72-h postinjury, 50% of control preparations and 7% of cyclosporin A-treated axonal bundles had progressed to secondary axotomy, respectively. Statistical analyses demonstrated a signiﬁ-cant ( p < 0.05) reduction in secondary axotomy between treated and untreated cultures. In summary, these results suggest that cyclosporin-A reduces progressive cytoskeletal damage and secondary axotomy following transient axonal stretch injury in vitro . ' 2007 Wiley Periodicals, Inc. Develop Neurobiol67:1831–1842, 2007


INTRODUCTION
Widely distributed \diffuse axonal injury" is an important determinant of neurological outcome following traumatic brain injury (Povlishock et al., 1992;Büki et al., 1999;Chen et al., 1999;Adams et al., 2000). Diffuse axonal injury likely results from a progressive molecular and cellular cascade of pathological changes within the axon following initial shear stress at the time of injury, resulting in delayed degeneration and secondary axotomy Smith et al., 2003). Key cytoskeletal changes that lead to diffuse axonal injury include neurofilament compaction (Tomei et al., 1990;Pettus et al., 1994;Jafari et al., 1997Jafari et al., , 1998Povlishock et al., 1997;Okonkwo et al., 1998;King et al., 2001), microtubule loss Jafari et al., 1997;Fitzpatrick et al., 1998;Maxwell et al., 2003), and the formation of neurofilament ring-like structures (Dickson et al., 2000;Chung et al., 2005). Axon shear stress may also cause alterations in focal axolemmal permeability, leading to increased intracellular calcium and the harmful activation of calcium-dependant enzymatic cascades, involving calpain, caspase, and calcineurin, which contribute to cytoskeletal damage (Povlishock, 1993;Povlishock and Pettus, 1996;Wolf et al., 2001). However, it has also been demonstrated in in vitro models of moderate axonal stretch injury that cytoskeletal disruption can occur in the absence of focal axolemmal disruption (Smith et al., 1999).
The perturbation of the cytoskeleton in relevant experimental models has been linked to impaired axonal transport and, consequently, the formation of axonal swellings Okonkwo et al., 1998;Saatman et al., 1998Saatman et al., , 2000Saatman et al., , 2003Postmantur et al., 2000). The role of specific cytoskeletal changes in axon transport deficits, such as neurofilament compaction, is unclear (Stone et al., 2001). However, there is increasing evidence of structural and functional damage to mitochondria following diffuse axonal injury (Lifshitz and McIntosh, 2003;Lifshitz et al., 2004Lifshitz et al., , 2006, which may contribute to the pathophysiology of traumatic brain injury via either metabolic dysfunction or the release of proapoptotic factors such as cytochrome-c (Sullivan et al., 1999(Sullivan et al., , 2002. In this respect, inhibition of the formation of the mitochondrial permeability transition pore has been reported to prevent cytoskeletal alterations and axonal degeneration following in vivo impact acceleration brain injury (Büki et al., 1999;Suehiro and Povlishock, 2001).
The present study utilizes an in vitro model of transient stretch injury to the axons of cortical neurons in long-term primary culture. We have investigated the temporal relationship between delayed cytoskeletal damage, axolemmal disruption, mitochondrial abnormalities (cytochrome-c release), and secondary axotomy. Furthermore, we have determined whether cyclosporin-A, a mitochondrial permeability transition pore and calcineurin inhibitor, attenuates cytoskeletal damage and secondary axotomy.

Cortical Cell Culture
Neuron cultures were prepared as reported previously (Dickson et al., 2000). This culture technique produces nerve cells that are relatively mature and reflective of the adult CNS (Chuckowree and Vickers, 2003;Chung and West, 2004;King et al., 2006;Haas et al., 2007). Briefly, neocortical tissue was removed from embryonic day (E)18 (sperm-positive day ¼ E1) Hooded Wistar rat embryos and incubated in sterile 10 mM HEPES buffer (378C). This was followed by trypsin digestion (0.25%) for 20 min and gentle washes of the cell pellet using fresh HEPES buffer. The cell suspension was disassociated carefully using a 1-mL pipette and the cells plated onto glass coverslips (132 mm 2 ) (Marienfeld, Germany) precoated overnight with 0.01% poly-L-lysine (Sigma, St. Louis, MO), at a cell density of 5 3 10 4 cells/well. Cultures were maintained at 378C in humidified air containing 5% CO 2 . Neurons were initially plated into a culture medium consisting of Neurobasal TM medium, supplemented with 10% fetal bovine serum, 0.1% (f/c) B-27 supplement, 0.1 mM (f/c) L-glutamine, 25 lM glutamate, and 200 U/mL gentamicin (all from Invitrogen BRL, Life Technologies, Grand Island, NY). At 24-h postplating, the medium was replaced with fresh medium, which lacked fetal bovine serum and glutamate, and was changed every 3-4 days until cultures were 21 days old. The fresh medium contained factors that inhibited substantial glial cell proliferation (Chuckowree and Vickers, 2003;Chung et al., 2005). As initially reported (Dickson et al., 2000;Chuckowree and Vickers, 2003;Chung et al., 2005), this long-term culture technique results in the formation of large neuronal clusters across the coverslip interconnected by thick fasciculated axonal bundles. These axonal bundles form synapses, and contain a cytoskeletal architecture characteristic of mature axons (Chuckowree and Vickers, 2003). Immunolabeling of cortical cultures with antibodies to glial fibrillary acidic protein (GFAP, marker for astrocytes, DAKO), ferritin(marker of activated microglia, ICN Biomedicals) and CD11b/c (marker of microglial/macrophages, CalTag), as well as counterstaining with the nuclear stain, Nuclear Yellow (Molecular Probes, Eugene, OR), was conducted to determine the presence/activation of nonneuronal cells relative to axonal injury.

Experimental Axon Stretch Injury and Immunocytochemical Analysis
The experimental axonal stretch injury was conducted using a novel in vitro model of mild-to-moderate axon injury (Chung et al., 2005). At 21 days in vitro, single coverslips were placed into individual 35-mm culture dishes (Iwaki, Japan) with 2 mL of fresh culture medium and placed upon a Leica DMIR inverted microscope. Using a glass micropipette (bore thickness 5 lm), a single pulse of sterile fluid was applied at 20 psi over a period of 13-20 ms using the Picospritzer III (Parker Instrumentation) causing the axon bundle to stretch and twist briefly before returning to its original orientation. The pulse of fluid produced a strain level of 2-10% and resulted in a transient 1-6% increase in original axon length. Previous data on this injury model reported axonal deformation to be 0.5-5%, 3-27%, and 6-43% at 24-, 48-, and 72-h postinjury, respectively (Chung et al., 2005). It is important to note that this form of stretch injury does not result in primary axotomy.
At 24-, 48-, and 72-h postinjury, cells were fixed with 4% paraformaldehyde for 20 min prior to incubation overnight with primary antibodies in diluent (0.3% Triton X-100 in 0.01 M PBS), targeting a-internexin, neurofilament-M, neuronal spectrin, ubiquitin, and cytochrome-c ( Table 1). The cells were then washed and incubated with the respective goat anti-mouse Alexa Fluor 488 or goat anti-rabbit Alexa Fluor 594 secondary antibody (Molecular Probes), applied in diluent. Coverslips were mounted onto slides using Permafluor mounting medium (Immunotech, Marseilles, France). Immunolabeled preparations were visualized under an upright epifluorescence microscope (Leica DMLB2) with images taken using an Optronics Magnafire digital camera. Additionally, a series of digital images were captured at 1.2 lm z-axis intervals through the entire thickness of the injured axonal bundle using an Optiscan F900e krypton/argon confocal scanning system equipped to an Olympus BX50 epifluorescence microscope. Images collected in the z-plane were then reconstructed into a single stacked image using NIH ImageJ (version 1.37v) software. Secondary axotomy was identified as injured cells illustrating severed axons, and these were counted and presented as a percentage. Three axonal bundles were stretchinjured per coverslip, with two uninjured axon bundles on the same coverslip used for control purposes. Thirty-six axon bundles were injured per time-point, and to ensure reliability, three sets of injuries were conducted at each time-point in different cultures.

Axolemmal Permeability
The axolemmal permeability following stretch injury was assessed using Alexa 488 hydrazide (Molecular Probes), a membrane impermeant fluorescent dye with a molecular weight of 570 Da. Intracellular accumulation of the dye demonstrates the axolemma permeability of small molecules (Smith et al., 1999). Briefly, prior to injury, 200 lL of 580 lM Alexa 488 hydrazide dissolved in control saline solution (120 mM NaCl, 5.4 mM KCl, 0.8 mM MgCl 2 , 1.8 mM CaCl 2 , 15 mM glucose, and 25mM HEPES, pH 7.4) was added to culture medium. The axons were stretched in the presence of the dye solution. Prior to analysis on the epifluoresence microscope (Leica DMLB2), the dye solution was washed off using fresh control saline solution. High (HC PL FluoroTar 20X/0.50) and low (HC PL FluoroTar 40X/0.30) magnification images were collected using an Optronics Magnafire digital camera to determine the percentage of stretch injured axons with dye accumulation at each time-point. As a positive control, uninjured neuron cultures were permeabilized with a single rinse of 0.005% Saponin (Sigma) in control saline solution and incubated for 5 min with the dye solution. As a negative control, the cultures were incubated in the dye solution without injury or Saponin, according to the postinjury observation time intervals.

Determination of Cytochrome-c Release from Axonal Mitochondria Following Injury
MitoTracker TM green dye (Invitrogen) was added to cultures at predetermined time-points to label axonal mitochondria and determine whether cytochrome-c was localized to mitochondria or the cytosol following injury. Prior to injury and at 24-, 48-, and 72-h postinjury, the medium from the cultures was removed and replaced with fresh prewarmed Neurobasal medium containing 100 nM Mito-Tracker TM green. The cells were incubated in this medium for 20 min, the medium removed, and the cells washed in prewarmed medium without MitoTracker green. The cells were then fixed and immunolabeled for cytochrome-c. Cells retained MitoTracker green following fixation and permeabilization.

Cyclosporin-A Treatment
Cyclosporin-A, a known inhibitor of calcineurin and the mitochondrial permeability transition pore (Büki et al., 1999), was added to cultures prior to injury to investigate the potential role of postinjury mitochondrial disruption relative to the release of cytochrome-c and cytoskeletal alterations. One hour prior to injury, neuronal cultures were treated with 20 nM cyclosporin-A (Sigma) (Hansson et al., 2004;Mironov et al., 2005). This concentration of cyclosporin-A is within the range of the lowest dose concentration capable of producing a therapeutic effect in other experimental models (Brustovetsky and Dubinsky, 2000;Hansson et al., 2004;Mironov et al., 2005). Cyclosporin-A remained in the medium for up to 96 h until fixation with 4% paraformaldehyde. Following fixation, cyclosporin-A

Delayed Cytoskeletal Damage and Alteration to Axolemma Permeability
Axolemma permeability following injury was examined using a low-molecular-weight tracer (570 Da), Alexa 488 hydrazide dye. With no stretch injury or chemical permeabilization, axons did not take up the Alexa 488 hydrazide dye (i.e. the axolemma remained impermeant to small molecules) at 5 min, 30 min, and 1-hour postinjury. However, chemical permeabilization of unstretched axons with saponin resulted in a substantial uptake of the tracer (Fig. 1).
In comparison, at 12-h postinjury, only 9% of stretched axons exhibited any uptake of the fluorescent dye. At 24-h postinjury, 22% of stretch-injured axons demonstrated Alexa 488 dye uptake, whereas by 48-h postinjury, 65% of stretch-injured axons demonstrated dye uptake (Fig. 1). The increase in axolemma permeability with increasing postinjury intervals was associated with disruption to the axonal cytoskeleton. Intraaxonal labeling for neurofilament-M, a-internexin, and neuronal spectrin was relatively ordered and linear at 24-h postinjury (Fig. 2). However, 20% of stretchinjured axons demonstrated dense and compacted neurofilament immunolabeling at this time-point. By Figure 1 Axolemmal permeability in stretch-injured axons. A: Alexa 488 hydrazide dye was readily taken up in uninjured axons following chemical permeabilization of the axolemma with saponin. B: A small but detectable amount of dye was taken up into stretch injured axons at 24-h postinjury. C: At 48-h postinjury, dye was readily detected within stretch-injured axons. Table 1 shows the percentage of stretch injured axons with permeability alterations over a period of 48-h postinjury. Scale bar, 50 lm. 48-h postinjury, there was an increased disruption of the axonal cytoskeleton, particularly in relation to the loss of linear labeling for spectrin, neurofilament-M and a-internexin, and the development of localized axonal swellings, the latter also focally labeled for ubiquitin (Fig. 2). Forty eight hours postinjury was the peak time-point for axon bundles (67%) to demonstrate neurofilament compaction. Cytoskeletal alterations within individual axons were investigated by laser scanning confocal microscopy, confirming that neuronal alterations observed with fluorescence microscopy were due to intraaxonal cytoskeletal changes rather than just a perturbation in axon bundling (Fig. 3). By 48-h postinjury, 18% of stretchinjured axons had progressed to frank secondary axotomy, increasing to 50% by 72-h postinjury (Fig. 2).

Mitochondrial Disruption and Cytochrome-c Release
Focal accumulation of cytochrome-c labeling occurred at 48-and 72-h postinjury, showing colocali-zation to axonal segments demonstrating neurofilamentous compaction and axonal swelling (Fig. 4). There was a similar pattern of colocalization of cytochrome-c and MitoTracker green (a mitochondrial marker) labeling in uninjured axons as well as stretch-injured axons at 24-h postinjury [ Fig. 4(C)]. Separation of MitoTracker Green fluorescence and cytochrome-c immunolabeling occurred within stretchinjured axons by 48-and 72-h postinjury. This included depletion of cytochrome-c from MitoTrackerstained mitochondria, the latter also accumulating in focal regions of axonal damage [ Fig. 4(D)].
To determine whether disrupting the release of cytochrome-c through the mitochondrial permeability transition pore attenuates the cytoskeletal damage observed following stretch injury, cyclosporin-A was added to cultures prior to injury. At 24-h postinjury, 0 and 20% of stretch-injured axon bundles demonstrated neurofilament compaction in cyclosporin Atreated and control preparations, respectively. Relatively reduced cytoskeletal and morphological change in cyclosporin A-treated stretched axons was also evident at 48-h postinjury (Fig. 5). At this time-point,  20% of control preparations demonstrated secondary axotomy, whereas no cyclosporin A-treated cultures had progressed to frank degeneration. At 72-h postinjury, 50% of control preparations and 7% of cyclosporin A-treated bundles had progressed to secondary axotomy. Kaplan-Meier survival analyses demonstrated that there was a significant decrease in axot-omy (p < 0.05) in cyclosporin-A treated cultures in comparison to untreated cultures. In cyclosporin-A treated cultures, there was a reduction in the number of injured axons demonstrating axon swellings in comparison to untreated cultures. At 72-h postinjury, 57% of untreated cultures and 12% of treated injured axons displayed swellings, respectively. Furthermore, there was a reduction in the number of injured axons demonstrating release and accumulation of cytochrome-c in cyclosporin-A treated cultures in comparison to untreated cultures.

Glial Cells and Axonal Stretch Injury
The potential perturbation and role of nonneuronal cells relative to the stretch injury was examined using Nuclear Yellow staining as well as immunocytochemical labeling for microglia (CD11b/c), activated microglia (ferritin), and astrocytes (GFAP). Small numbers of nonneuronal cells were present within the culture both prior and after stretch injury (Fig. 6). Substantial glial cell proliferation was inhibited by the culture medium. Nonneuronal cells were largely associated with the neuronal cell clusters and did not come in contact with the axon bundles. Stretch injury resulted in damage to the axon bundle independent to any disruption to the nonneuronal cells or subsequent glial proliferation/activation (Fig. 6). Figure 6 Examination of nonneuronal cells relative to stretch injury. A, B: Few astrocytes (red, GFAP) were present under these culture conditions. GFAP-labeled astrocytes were not associated with axonal bundles (green, neurofilament-M) either in uninjured cultures (B) or following stretch injury (A, 48-h postinjury). C (24-h postinjury): There was no increase in activated microglia (red, ferritin) following axonal injury, and no association of these glial cells with stretch-injured axon bundles (green, neurofilament-M). D: Similarly, there were few cells labeled for a marker of macrophages/microglia (CD11b/c, green) in either uninjured (D) or stretch-injured cultures, nor any association between microglia and axonal bundles (red, neurofilament-M). E, F: Nuclear Yellow staining (E) at 48-h postinjury confirmed a lack of association of nonneuronal cells with stretchinjured axon bundles (F, phase microimage). Scale bar, 50 lm.

DISCUSSION
The model of axonal injury in the current study delivered dynamic uniaxial and biaxial deformation of an axon at a strain rate of *4 s À1 (Chung et al., 2005) in primary cortical neurons grown to relative maturity in vitro (Chuckowree and Vickers, 2003). This study has demonstrated that axon bundles that receive this transient stretch-injury undergo a series of stereotypical cellular changes leading to delayed axotomy. Cellular changes included compaction of neurofilaments, derangement of other cytoskeletal proteins, cytoplasmic accumulation of cytochrome-c, and axon swelling. Alterations in axolemmal permeability were also demonstrated, which were maximal 24-48 h following injury. Mitochondrial disruption may be a key pathological alteration leading to progressive axonal degeneration, as blocking of the mitochondrial permeability transition pore and/or calcineurin with cyclosporin-A attenuated many of these cellular changes, as well as secondary axotomy.
Other studies have indicated that disruption of cytoskeletal proteins such as neurofilaments are likely to be important for the focal axonal transport abnormalities leading to axon swelling and degeneration following axonal trauma (Frappier et al., 1992;Pettus and Povlishock, 1996;Povlishock and Pettus, 1996;Povlishock et al., 1997;Okonkwo et al., 1998;Saatman et al., 1998Saatman et al., , 2000Saatman et al., , 2003Smith et al., 1999). Cytoskeletal changes in stretch-injured neurons in the current in vitro model included neurofilament compaction, the loss of linear organization of filamentous proteins, and accumulation in axonal swellings. Neurofilament compaction within axons has been described in in vivo models of brain injury (Gallyas et al., 2002;Marmarou et al., 2005). It is unclear whether our data indicate that neurofilament compaction following axonal stretch injury necessarily predates secondary axotomy. However, it may not be that all axons showing such postinjury cytoskeletal alterations will inevitably progress to secondary degeneration (Stone et al., 2001). Indeed, secondary degeneration is more complex than previously thought, and is possibly initiated via multiple pathways, as highlighted by gene expression analyses (Morrison et al., 2000).
Shear strain injury to axons has been proposed to lead to transient, focal disruption of the axonal membrane, concomitant with an overall increase in intraaxonal calcium (Tomei et al., 1990;Povlishock, 1993;Maxwell et al., 1995;Povlishock and Pettus, 1996;Wolf et al., 2001;Stone et al., 2004). Following axonal stretch injury to NT2 cells, Smith et al. (1999) reported initial lack of membrane permeability to dyes. In further work with the same model (Wolf et al., 2001), it was demonstrated that stretch-induced strain on the axonal membrane lead to acute and abnormal Na þ influx through mechanosensitive channels, triggering a reverse of the Na þ /Ca 2þ exchanger, activation of voltage-gated Ca 2þ channels, and subsequent pathological influx of Ca 2þ . Geddes-Klein et al. (2006) demonstrated that biaxial rather than axial stretch of the axons of cultured cortical neurons was more effective at causing acute axolemmal permeability and Ca 2þ influx. Our study involved transient axonal stretch in cultured primary neurons maintained for longer periods of time, and we have shown in this model that there were no axolemmal permeability alterations immediately and up to 12 h following injury. Rather, permeability alterations were delayed and closely associated with the delayed disruption of the axon cytoskeleton following stretch injury. These results indicate that the relatively mild stretch injury we have utilized is not sufficient to cause immediate perturbation of the axolemmal membrane and/or trigger opening of membrane channels. In this respect, delayed disruption of the cytoskeleton may lead to deficits in the integrity of linked membrane proteins and channels (Denker and Barber, 2002). Our data indicate that there may be some value in the examination of axolemmal permeability at longer postinjury time-points in models involving axonal shear strain, reflecting also the delayed nature of diffuse axonal injury and secondary axotomy.
In this study, mild axonal stretch injury also resulted in the delayed accumulation of cytochrome-c and ubiquitin within highly localized regions of the injured axon, in association with regions of cytoskeletal disruption. Cytochrome-c release within traumatically injured axons has been proposed to have a role in stimulating apoptosis and neuronal degeneration (Büki et al., 2000). The delayed release and accumulation of cytochrome-c has also been reported in ischemic and traumatic brain injury models (Fujimura et al., 1998;Pike et al., 1998), as well as following excitotoxic injury (Luetjens et al., 2000). Our study demonstrated that cytochrome-c accumulation in the axon represents abnormal release into the cytosol rather than simply indicating the accumulation of mitochondria containing this protein. Release of cytochrome-c from mitochondria may have several detrimental effects, for example, disruption of normal mitochondrial respiration and/or activation of the caspase cascade resulting in apoptosis (Liu et al., 1996;Li et al., 1997). With respect to traumatic brain injury, Büki et al. (2000) have suggested that localized release of cytochrome-c in damaged axons results in the activation of caspases (e.g., caspase 3) that may act directly on cytoskeletal elements such as spectrin. However, calpain activity may also contribute to progressive intraaxonal cytoskeletal damage and secondary axotomy in an experimental optic nerve stretch injury model (Saatman et al., 2003). As previously stated, cytoskeletal damage and secondary apoptosis may be initiated via multiple pathways; however, it is clear that mitochondrial damage is likely to be an important determinant of evolving axonopathy (Morrison et al., 2000). Mitochondrial damage may result in disruption to mitochondrial respiration (Vink et al., 1990), release of reactive oxidative species (Lifshitz and McIntosh, 2003;Lifshitz et al., 2004), and disruption to calcium homeostasis (Xiong et al., 1997). Cyclosporin-A has been found to attenuate mitochondrial dysfunction in a cortical impact model of traumatic brain injury (Sullivan et al., 1999). Additionally, cyclosporin-A pretreatment before impact-acceleration brain injury of rats reduced subsequent immunohistochemically identifiable axonal damage . Subsequently, in a variety of in vivo experimental models involving rodents and sheep, postinjury treatment with cyclosporin-A has been shown to reduce lesion volume following cortical contusion (Sullivan et al., 2000), improve motor and sensorimotor function following lateral fluid percussion injury (Riess et al., 2001), reduce amyloid precursor protein-labeled axonal damage after impact-acceleration (Okonkwo et al., 2003), and expression of this protein subsequent to focal head impact ( Van den Heuvel, 2004). Although previous in vivo experiments report attenuation of cytoskeletal damage following constant injections with cyclosporin-A (Riess et al., 2001), it is difficult to translate the therapeutic time window and dosage into in vitro models, particularly as constant addition of cyclosporin-A, or changes in media, may have a further toxic effect on neurons maintained in vitro. Hence, a single pretreatment of cyclosporin-A was utilized in the current study to attenuate acute and possibly chronic axonal disruption.
Our studies show that a single pretreatment with cyclosporin-A reduced the progression of axonal injury into neuritic swellings and secondary axotomy in vitro. Furthermore, as neurons were isolated in culture and the injury did not promote a change in glial cells, the neuroprotective effect of cyclosporin-A is likely to be axon-specific, potentially reducing damage to mitochondrial membrane integrity and the release of cytochrome-c into the cytosol. Interestingly, cyclosporin-A did not have an effect on the development of specific cytoskeletal changes such as neurofilament compaction. Likewise, another immunophilin ligand, Tacrolimus (FK506), reduced axonal swelling, but not neurofilamentous compaction, following impact-acceleration injury in rats (Marmarou and Povlishock, 2006).These data indicate that neurofilament compaction, which likely follows cleavage of filament side-arms, occurs through the activation of intracellular processes unrelated to the site of action of the immunophilin ligands, and also strongly supports the proposal that neurofilament compaction may not be a critical pathological change leading to axonal swelling, axon transport deficits, and degeneration following traumatic brain injury (e.g. Stone et al., 2001;Marmarou et al., 2005). FK506 itself does not appear to inhibit the mitochondrial permeability transition pore, indicating the intriguing possibility that both FK506 and cyclosporin-A may be acting through inhibition of calcineurin activity to attenuate axonal pathology following injury.
In summary, although limitations of a tissue culture model are an important caveat for direct translation to human traumatic brain injury, the interpretation of these results, in conjunction with previous in vivo and in vitro studies, provides evidence that cyclosporin-A treatment and the targeting of specific postinjury cellular changes that develop within axons may be a useful approach for reducing diffuse axonal injury and secondary axotomy following brain trauma.