1. The Problem
Spinal cord injury results in devastating morbidity and immense life-time health care costs. The immediate changes that occur following various types of injury to the spinal cord are well-described
[1] and advances have been made in the acute care of these patients
[2]. However, the long-term prognosis of such injuries remains poor. This is largely due to the inability of the injured tissue to restore the pathways and connections that have been damaged. Axonal regeneration following injury to spinal cord white matter is particularly limited. While in some cases, injury may involve complete transection of the spinal cord, far more common are contusion injuries in which the portions of the cord rostral and caudal to the injury site remain in continuity
[3]. Nevertheless, axonal regeneration generally fails to occur across the injury site. Failure of regeneration is axiomatically assumed to underlie the etiology of the functional deficits associated with spinal cord injury. Consequently, establishing conditions that facilitate robust axonal regeneration and functional re-connectivity is generally assumed to be a prerequisite for successful treatment paradigms.
Despite the lack of axonal regeneration in spinal cord white matter, there is evidence that injured axons in the central nervous system (CNS) are capable of regeneration. Some of the earliest observations were made by Ramón y Cajal
[4], who described “certain axons, few in number…that initiate regeneration…developing a small cone of growth”. This initial regeneration proceeds for as many as nine days, but “these sprouts penetrate very rarely into the wound itself (and) this neoformative action decreases or is arrested from the tenth to the fourteenth day”. There is abundant evidence that central axons are capable of regeneration if presented with an alternative environment. Grafts of non-CNS tissues such as peripheral nerve
[5][6][7][8][9], striated muscle
[8][10], iris
[11][12], mitral valve
[11], skin, tendon, thyroid and salivary glands
[8] placed within the CNS have been shown to be invaded by central axons, although these results are often difficult to reproduce
[8][10][11][13][14].
In contrast to those in spinal cord white matter, axons generally regenerate following peripheral nerve injury. Peripheral axonal regeneration is generally preceded by significant reorganization of the tissue milieu ahead of the advancing front of axons. Degeneration and clearance of distal axonal segments occurs
[15], followed by reorganization of Schwann cells into longitudinal bands
[16]. However, peripheral axonal regeneration is delayed in mutant mice in which there is a significant reduction in macrophage infiltration into the distal nerve stump
[17][18][19] and delayed Wallerian degeneration
[18][19]. These observations suggest that clearance and reorganization of the cytoarchitecture of the distal stump may be required before regeneration can occur.
Likewise, alterations to the cytoarchitecture of spinal cord white matter occur following injury. Within the injury zone, small perivascular hemorrhages appear within minutes along the longitudinal axis of the fiber tract
[20]. Within 30 min, spaces appear between axons and their myelin sheaths
[21][22]. Within one hour, the myelin sheaths disintegrate leaving unmyelinated segments of axons and scattered myelin debris within the injury zone
[21]. Over the next few hours, axonal segments within the injury zone degenerate completely
[4][20][23] and oligodendrocytes disappear
[24]. Although macrophage invasion and the onset of debris clearance are rapid in the direct injury zone
[24][25][26], myelin debris and lipid-laden cells can still be observed as late as 60 days following injury
[24][26]. Within one week after injury, reactive astrogliosis occurs within and around the injury zone
[27][28]. The resulting glial scar is characterized by “a dense mesh-work of hypertrophied astrocytic processes”
[28] that form a complete border around the injury site between 1 week and 3 months post-injury
[27]. Within the boundary of the injury site, the hypertrophied astrocytic processes appear anisomorphic in orientation but transition toward an orientation that is parallel with the longitudinal axis of the tract within the adjacent white matter
[27].
Distal to the injury zone, a process similar to Wallerian degeneration occurs within CNS white matter
[23][26][29][30] but appears slower or less robust than that in peripheral nerves. Compared with peripheral nerves, there is less macrophage invasion of injured optic nerve and spinal cord tracts
[26][30][31]. The rates of Wallerian degeneration and pathway clearance are also comparatively reduced in the distal stump of central fiber tracts
[26][29][30]. In fact, myelin debris can still be observed for as long as four months or three years following injury
[23]. Thus, as appears to be the case in peripheral nerves of Wallerian degeneration-delayed mutant mice, slow clearance of debris and incomplete reorganization of normal white matter architecture distal to the injury site may pose a barrier to axonal regeneration.
Although the focus is on the evidence for a role of tissue geometry, there are other factors that are relevant to any strategy for promoting functional recovery following spinal cord injury. For example, the decline in the intrinsic growth potential of CNS neurons with maturity appears to play an important role in limiting regeneration
[32]. Treatments that target intracellular pathways involved in axonal transport, such as the mechanistic target of rapamycin (mTOR), phosphatase and tensin homolog (PTEN) and growth-associated protein 43 (GAP-43), offer potential therapeutic targets for enhancing axonal growth and functional recovery
[33][34][35][36][37][38][39][40]. Similar evidence has been obtained for the use of neurotrophic factors, which can also contribute to greater regeneration and/or collateral sprouting
[41][42][43][44]. Enhancing the intrinsic growth potential of axons, however, is unlikely to be sufficient if disruption of tissue organization prevents axonal navigation of the injury site.
2. Factors Associated with Glial Scars and Myelin Have Been Implicated in Growth Cone Collapse
The glial scar has been the focus of much research into the causes of regeneration failure
[45][46][47][48][49][50][51]. In microtransplantation studies designed to reduce glial scarring, robust axonal growth occurs within white matter in vivo
[27][52]. However, in cases in which more disruption of the host white matter and glial scarring occurred than intended, axonal growth halted at the boundary of the glial scar
[27][52].
Much of the focus on glial scars as a barrier to regeneration within the CNS has sought to identify specific putative inhibitory molecules associated with the glial scar. Chondroitin sulfate proteoglycans upregulated in association with glial scars in the peri-injury zone have been shown to cause growth cone collapse or turning in vitro
[28][53][54]. However, the extent to which the presence of these factors alone serves as a barrier to growth is difficult to assess. Within the glial scar, reactive astrocytes and putative inhibitors associated with them are not organized in parallel with the longitudinal axis of the fiber tract, unlike the astrocytes that characterize uninjured white matter
[23]. Chondroitin sulfate proteoglycans are present in white matter distal to the injury site
[55][56], which supports axonal growth, but there is relatively limited upregulation of these factors in distal white matter following injury
[23] and these factors appear to be limited to a geometry that “surround axonal profiles”
[55]. Microglia and macrophages at the glial scar site promote tissue repair and prevent infection, but also exhibit neurotoxic properties
[57]. The fibrotic scar that lies in the cascade of reactive astrocytes has been referred to as an additional barrier to neuronal growth, since the dense network of fibroblasts produce a multitude of inhibitory factors to axonal growth
[48][58].
The glial scar is now recognized to be more complicated than originally assumed in terms of its cellular composition and the expression of factors that either inhibit or enhance axonal growth
[50][59]. In fact, Anderson et al.
[60] provided evidence for a supportive, rather than inhibitory role of the scar in promoting regeneration. Furthermore, there is substantial evidence for positive roles for the glial scar, at least during acute phases of injury
[48][51]. Studies have demonstrated that ablation of the astrocytic portion of the glial scar leads to neurological deficits and incomplete recovery, indicating that the presence of the astrocytic scar is important for functional recovery
[61][62]. There is additional recent evidence that a glial scar exposed to an appropriate chemical environment may aid in functional recovery
[51][63]. Other glial scar components such as microglia and oligodendrocytes may also have beneficial roles
[64]. These findings suggest that the glial scar plays a much more complicated role in CNS injuries than previously assumed and that its removal may impede functional recovery.
Other putative inhibitors of axonal growth have been identified as normal components of myelin. As a result, the persistent presence of myelin debris distal to the injury site has been proposed to pose not only a physical barrier, but also to present molecular inhibitors of growth, thus depriving an advancing growth cone of a supportive growth environment. Several factors associated with myelin have been shown to induce growth cone collapse in tissue culture systems including Nogo, Myelin-Associated Glycoprotein (MAG), and Oligodendrocyte-Myelin Glycoprotein (OMgp)
[65][66][67][68][69][70][71][72][73]. Lipid components of myelin have also been shown to be inhibitory to axonal growth
[74][75]. Significant progress has been made in identifying the cellular mechanisms by which these factors lead to inhibition of neurite growth, as well as serving roles in regulating neuronal development and neuronal plasticity
[76].
Much of the evidence supporting the designation of myelin-associated factors as putative inhibitors of growth come from experiments in which myelin, or its components, have been reconstituted as a substrate for cultured cells. For example, purified CNS myelin absorbed onto poly-lysine-coated culture dishes has been shown to inhibit cell attachment, neurite outgrowth and 3T3 fibroblast spreading
[66][77][78], but not when myelin from MAG knockout mice is used
[71]. Cultured oligodendrocytes and culture dishes coated with the myelin-associated protein Nogo have been shown to be non-permissive substrates for neuronal attachment, neurite outgrowth and 3T3 fibroblast spreading
[65][66][72][78][79][80]. MAG-transfected Chinese Hamster Ovary cells were shown to inhibit neurite growth when used as substrates for postnatal cerebellar and adult sensory neurons
[77][81]. The extracellular domain of MAG in solution inhibits neurite growth in vitro
[73][82]. However, in these studies, myelin or myelin-associated factors are not presented to cultured cells or growing neurites in a manner mimicking their normal cytoarchitecture within the spinal cord. In healthy tissue, myelin and its associated factors are organized in parallel with the longitudinal axis of a white matter tract, a geometric organization that is obliterated in cases of homogeneous exposure to such factors, as is the case with reconstituted myelin absorbed onto culture dishes, cultured cells expressing myelin-associated factors, or myelin-associated factors in solution.
Deactivation of myelin-associated inhibitors has been shown to promote axonal regeneration in systems that more closely resemble their normal organization. Monoclonal antibodies raised against Nogo can enhance axonal regeneration through white matter lesions sites in vivo
[83][84][85][86][87][88][89][90] or in vitro
[66][91][92]. Immunization with myelin, Nogo, or MAG has promoted corticospinal regeneration
[93]. X-irradiation to block myelination has been shown to increase regeneration in crushed optic nerve
[89]. Increased axonal regeneration occurs through lesions in the corticospinal tracts of MAG-knockout mice
[69] and laser-inactivation of MAG increases regeneration of cultured retina and optic nerve
[94]. Increased regeneration in CNS white matter has, likewise, been observed in mice in which various isoforms of Nogo have been knocked out
[95][96], although these experiments have produced variable results
[97][98][99].
Other studies have targeted the Nogo Receptor, which binds all three known myelin-associated inhibitor proteins
[100][101][102]. Immunization with recombinant Nogo Receptor promotes regeneration in the spinal cord
[61]. Transfection of retinal ganglion cells with a dominant negative form of Nogo Receptor promoted optic nerve regeneration in vivo
[103]. Treatment with a competitive antagonist of the Nogo Receptor or the ectodomain of the Nogo Receptor in soluble form promotes regeneration of rubrospinal, raphespinal, and corticospinal fibers, often bypassing the lesion through adjacent gray matter
[102][104][105][106][107]. Overall, these studies support the concept that to achieve optimal axonal regeneration in CNS white matter, the effects of myelin-associated inhibitors must be blocked. However, there are data supporting the concept that myelin-associated inhibitors are not entirely antagonistic to axonal regeneration and functional recovery.
3. Intact White Matter Can Support Axonal Growth
Despite the presence of myelin-associated factors that induce growth cone collapse, there is clear evidence that CNS white matter supports axonal growth in cases where the cytoarchitecture remains intact. Systematic studies and serendipitous observations have been made of axonal growth over long distances in vivo within CNS white matter from transplanted embryonic neurons
[27][52][108][109][110][111][112][113][114][115][116] or postnatal neurons
[27][52][117].
The use of fresh-frozen tissue sections as substrates for neuronal culture has also provided evidence that white matter can support neurite growth. The technique, often referred to as cryoculture, has shown a variety of results, depending on the culture conditions
[118][119][120][121][122][123][124][125]. When successful outgrowth from neurons cultured on cryostat sections of white matter occurs, it is generally directed in parallel with the longitudinal axis of the fiber tract but is limited on transverse sections through the same tissue
[126][127]. Furthermore, on sections of the corpus callosum from myelin-deficient rats, the pattern of axonal growth is not parallel with the fiber tract and is morphologically indistinguishable from that on adjacent gray matter
[127]. These observations are most consistent with the hypothesis that it is the spatial organization of the tissue components that determines whether axonal growth is successful.
A direct test of this hypothesis was carried out in which the spinal cord was removed, crushed, and then immediately frozen to prevent glial scarring or other post-injury changes. Longitudinal cryostat sections through the crush site were then used as substrates on which primary neurons were cultured. Neurite growth occurred in parallel with the longitudinal axis of the tract on the uninjured areas of the white matter but did not extend across the crush site
[126]. In contrast, the gray matter within the same injury site supported growth. These results demonstrate that disruption of the white matter tissue geometry is sufficient to inhibit growth, presumably due to the disorganized distribution of myelin and its associated inhibitors, that otherwise occurs on segments of white matter where the geometry is intact.
To test the idea that similar disruption of the geometry of a peripheral nerve, whose myelin has inhibitory properties
[71][128], might also impede axonal growth, additional experiments were conducted in which the sciatic nerve was removed, crushed, and then immediately frozen to prevent subsequent injury-induced changes. Longitudinal cryostat sections of the crushed nerve were then used as substrates to assess axonal growth. As found for the growth on longitudinal sections of crushed spinal cord, axons extended in parallel with the longitudinal axis of the nerve on segments where the nerve geometry is intact. However, these same axons did not cross injury sites
[126]. Again, the assumption is that the disorganized myelin within the crush site served to inhibit growth through the region. The collective results support the conclusion that disruption of the tissue geometry, both in spinal cord white matter and in peripheral nerve tissues, is sufficient to create a largely non-permissive environment for axonal growth in the absence of scar formation. Beyond its potential role in contributing to axonal growth failure in injury, myelin-associated factors may normally serve to guide axonal growth in intact and developing tissue
[70][87][127][129][130][131].
In newborn rats, X-irradiated locally to prevent myelination or treated with anti-Nogo antibodies, corticospinal axons regenerated over a larger cross-sectional area, intermixing with ascending axons running in the Fasciculi Gracilis and Cuneatus
[70][87], suggesting that the absence of myelin eliminated guardrails that normally limit aberrant growth. Following spinal cord hemisection and combinatorial treatment including anti-Nogo antibodies and neurotrophin-3, regenerating axons appeared to cross the midline to bypass the lesion
[132]. Treatment with anti-Nogo antibodies appeared to induce sprouting by the intact corticospinal tract across the midline, following a cervical spinal cord lesion
[133] or an experimentally-induced ischemic stroke affecting the sensorimotor cortex
[134]. Sprouting of the rubrospinal tract has been observed following treatment with anti-Nogo antibodies
[135]. Similar collateral sprouting in the CNS has been observed in vivo following X-irradiation to suppress myelination
[136][137], treatment with anti-Nogo antibodies
[138][139] or targeted Nogo knockouts
[140]. Following sciatic nerve transection, focal and temporary treatment with MAG reduced hyperinnervation and improved the accuracy of target reinnervation
[141]. These studies suggest that myelin-associated factors may play an important role in guiding growth, constraining growth, and limiting collateral sprouting (which may or may not be beneficial).
Such a role for myelin-associated factors is supported by in vitro studies. Axons growing on patterned substrates consisting of alternating lanes of poly-lysine/laminin and myelin extend preferentially on poly-lysine/laminin but are oriented in the direction of the lanes, growing not on, but parallel to the myelin
[142]. Treatment with MAG decreases axonal branching from dissociated dorsal root ganglion neurons
[141]. Together, these in vivo and in vitro data support a role for putative inhibitors, normally found within white matter, in guiding or constraining axonal growth.
This entry is adapted from the peer-reviewed paper 10.3390/medicina58040542