Brain Advance Access originally published online on February 21, 2007
Brain 2007 130(4):940-953; doi:10.1093/brain/awl374
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Growth-modulating molecules are associated with invading Schwann cells and not astrocytes in human traumatic spinal cord injury
1Department of Neurology, Aachen University Hospital, Germany, 2Centre for Neuromuscular and Neurological Disorders, University of Western Australia, Perth, Australia, 3Department of Neurosurgery, Sart Tilman Hospital, 4Department of Neurology and Neuroanatomy and 5Center for Cellular and Molecular Neuroscience, University of Liège, Liège, Belgium
Correspondence to: Armin Buss, Pauwelsstrasse 30, 52074 Aachen, Germany E-mail: arminbuss{at}hotmail.com
 |
Summary |
|---|
 Top Summary IntroductionMaterial and methodsResultsEarly survival times (2-11...DiscussionReferences |
|---|
Despite considerable progress in recent years, the underlying mechanisms responsible for the failure of axonal regeneration after spinal cord injury (SCI) remain only partially understood. Experimental data have demonstrated that a major impediment to the outgrowth of severed axons is the scar tissue that finally dominates the lesion site and, in severe injuries, is comprised of connective tissue and fluid-filled cysts, surrounded by a dense astroglial scar. Reactive astrocytes and infiltrating cells, such as fibroblasts, produce a dense extracellular matrix (ECM) that represents a physical and molecular barrier to axon regeneration. In the human situation, correlative data on the molecular composition of the scar tissue that forms following traumatic SCI is scarce. A detailed investigation on the expression of putative growth-inhibitory and growth-promoting molecules was therefore performed in samples of post-mortem human spinal cord, taken from patients who died following severe traumatic SCI. The lesion-induced scar could be subdivided into a Schwann cell dominated domain which contained large neuromas and a surrounding dense ECM, and a well delineated astroglial scar that isolated the Schwann cell/ECM rich territories from the intact spinal parenchyma. The axon growth-modulating molecules collagen IV, laminin and fibronectin were all present in the post-traumatic scar tissue. These molecules were almost exclusively found in the Schwann cell-rich domain which had an apparent growth-promoting effect on PNS axons. In the astrocytic domain, these molecules were restricted to blood vessel walls without a co-localization with the few regenerating CNS neurites located in this region. Taken together, these results favour the notion that it is the astroglial compartment that plays a dominant role in preventing CNS axon regeneration. The failure to demonstrate any collagen IV, laminin or fibronectin upregulation associated with the astroglial scar suggests that other molecules may play a more significant role in preventing axon regeneration following human SCI.
Key Words: spinal cord injury; human; regeneration; glial scar
Abbreviations: CNS, central nervous system; ECM, extracellular matrix; NF, neurofilament; PNS, peripheral nervous system; SCI, spinal cord injury
Received June 30, 2006. Revised December 8, 2006. Accepted December 13, 2006.
|
Introduction |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsEarly survival times (2-11...DiscussionReferences |
|---|
In contrast to the peripheral nervous system (PNS), lesioned long distance projecting axons within the white matter of the adult mammalian central nervous system (CNS) do not undergo regeneration. The loss of function following traumatic spinal cord injury (SCI) is often permanent and results in serious limitations to the patients' quality of life (McDonald and Sadowsky, 2002
). Despite considerable progress in recent years, the underlying mechanisms responsible for the failure of axonal regeneration after SCI remain only partially understood.
Experimental studies have demonstrated that following SCI, the initial parenchymal damage is followed by a complex cascade of secondary events including breakdown of the blood–spinal cord barrier, infiltration of blood-derived inflammatory cells, oedema, excitotoxicity and ischaemia. This early phase of secondary parenchymal damage is followed by the removal of tissue debris. Finally, severe lesions become dominated by scar tissue composed of connective tissue and fluid-filled cysts, surrounded by a dense astroglial scar (Schwab and Bartholdi, 1996). The astroglial scar is largely composed of reactive astrocytes and a dense irregular network of their processes. In traumatic injuries leading to damage of nerve roots and the pial surface, Schwann cells and meningeal fibroblasts can also invade the lesion site (Brook et al., 1998; Fawcett and Asher, 1999; Grimpe and Silver, 2002). The reactive astrocytes and infiltrating fibroblasts produce a dense extracellular matrix (ECM) that represents a physical and molecular barrier to axonal regrowth (Schwab and Bartholdi, 1996; Fawcett and Asher, 1999; Condic and Lemons, 2002).
Collagen IV belongs to the network-forming collagens and is a major component of the basal lamina. In the developing nervous system this protein is extensively distributed and its expression pattern suggests guiding functions for growing PNS and CNS axons (Venstrom and Reichardt, 1993). In vitro studies have demonstrated a growth-promoting effect of collagen IV on some neuronal populations, including sympathetic neurons (Shiga and Oppenheim, 1991). In the unlesioned spinal cord parenchyma, collagen IV is restricted to blood vessel walls and the meninges, however in experimental SCI it is found in the evolving extracellular scar tissue. In contrast to the above direct growth-promoting effects of collagen IV, it has been reported that intervention strategies which delay the post-traumatic synthesis of collagen IV result in enhanced axonal regeneration (Stichel et al.,). However, since other groups 1999; Hermanns et al., 2001have not been able to reproduce this effect (Weidner et al., 1999; Iseda et al., 2003) and even some degree of spontaneous axonal regeneration has been observed within collagen IV rich territories following experimental SCI (Joosten et al., 2000) a clear understanding of the influence of collagen IV expression on plasticity and regeneration at the lesion site in vivo remains to be elucidated.
Laminin is the major non-collagenous glycoprotein of basement membranes (Timpl et al., 1979; Kleinman et al., 1982) and is involved in a number of cell–basement membrane interactions e.g. adhesion, migration and proliferation. During development, its expression pattern suggests a guiding role in axonal elongation (Venstrom and Reichardt, 1993) and numerous in vitro studies have demonstrated a growth-promoting effect on a range of neuronal populations (Condic and Lemons, 2002; Grimpe and Silver, 2002). Similar to collagen IV, the distribution of laminin in the normal spinal cord is restricted to blood vessel walls, and following experimental SCI it is up-regulated in the lesion site. Although laminin supports axonal regeneration in the lesioned PNS, in vivo experiments have, so far, failed to unequivocally demonstrate a similar role in CNS injury. In fact, it has even been suggested that high concentrations of laminin may be responsible for the entrapment of growth cones—thereby preventing any further axonal extension (Condic and Lemons, 2002).
Fibronectin (FN) is an axon growth-promoting ECM protein that is widely expressed in the developing central and peripheral nervous system and has been reported to guide growing axons (Venstrom and Reichardt, 1993). In the mature nervous system, its distribution is more restricted. In the CNS, FN is found in the vasculature, the ependyma and the meninges. In the PNS it is present in the endo-, peri- and epineurium. After PNS and CNS injuries, fibronectin becomes dramatically upregulated and, at least in the PNS, plays a role in the regeneration of the lesioned axons (Lefcort et al., 1992). In the lesioned CNS, its role is, so far, not clearly defined. However, implantation of fibronectin guidance channels into experimental spinal cord lesions has resulted in enhanced axonal regeneration (King et al., 2005).
In the human situation, data on the molecular composition of the scar tissue at the lesion site following traumatic SCI is scarce. A few studies have used histological stains to demonstrate the appearance of myelin or collagen at the lesion site. More recently, the use of immunohistochemistry has demonstrated the expression of chondroitin sulphate proteoglycan (CSPG) in the lesion site of a subset of cases, which correlated with the presence of infiltrating Schwann cells (Bruce et al., 2000). However, no data have been presented regarding the spatiotemporal pattern of collagen IV, laminin or fibronectin after SCI. We, therefore, performed an immunohistochemical investigation on the expression of several putative growth-inhibitory and growth-promoting molecules in samples of post-mortem human spinal cord, taken from patients who died at a range of survival times following severe traumatic SCI. In the present cases, the lesion induced scar could be subdivided into a Schwann cell dominated domain containing large neuromas and a dense surrounding ECM, and the well delineated astroglial scar with a dense network of hypertrophic processes. Collagen IV, laminin and fibronectin were all present in the post-traumatic scar tissue and could almost exclusively be found in the Schwann cell domain, both in the neuromas and the ECM. In the astrocytic domain, their presence was restricted to blood vessel walls. The presence of all three growth-influencing ECM molecules in the Schwann cell area of the lesion site suggests an important role for this cell population in the post-traumatic human lesion site and highlights these cells as potential targets for future intervention strategies aiming to enhance axon regeneration in the lesioned adult human spinal cord.
|
Material and methods |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsEarly survival times (2-11...DiscussionReferences |
|---|
Post-mortem, the spinal cords were removed from 4 control patients who had not suffered from any neurological disease (see Table 1) and from 15 patients who died at a range of time points after traumatic spinal cord injury (see Table 2). The study was approved by the Aachen University Ethics Committee. Patients with traumatic injury had been diagnosed as having ‘complete’ injuries and presented with paraplegia or tetraplegia. The spinal columns were removed at autopsy,
15–48 h after death. Following incision of the dura mater, the spinal cord was fixed in 10% buffered formalin for at least 2 weeks. Thereafter, blocks of the lesion site and tissue from regions rostral and caudal to the lesion (1 cm thickness) were embedded in paraffin wax.
|
|
Peroxidase immunohistochemistry
Transverse sections (5 µm thick) were collected onto poly-L-lysine-coated slides and allowed to dry. Sections were de-waxed in xylene, rehydrated and endogenous peroxidase activity was blocked by incubation in 0.1 M PBS containing 3% H2O2 for 30 min. Next, the slides were microwaved in 10 mM citrate buffer (pH 6) for 3 x 3 min. Sections for laminin staining were treated with 10 µg/ml proteinase K for 30 min at 37°C. Subsequently, non-specific binding was blocked by incubation in 0.1 M PBS containing 3% normal goat serum and 0.5% Triton X-100 for 30 min. Next, sections were incubated in the primary antibody, overnight at room temperature. The primary antibodies used were: polyclonal rabbit anti-collagen IV (ICN, diluted 1 : 200), polyclonal rabbit anti-laminin (Sigma, diluted 1 : 200), polyclonal rabbit anti-fibronectin (Dako, diluted 1 : 10 000), polyclonal rabbit anti-GFAP (Dako, diluted 1 : 2500), monoclonal mouse anti-low affinity nerve growth factor receptor (NGFr or p75; Dako, diluted 1 : 100), polyclonal rabbit anti-myelin basic protein (MBP, Chemicon, diluted 1 : 1000), polyclonal rabbit anti-neurofilament (NF, Sigma-Aldrich, diluted 1 : 1000). Following extensive rinsing steps in 0.1 M PBS, sections were incubated in biotinylated horse anti-mouse or anti-rabbit antibody (Vector Laboratories, diluted 1 : 500) for 1 h at room temperature. This was followed by the Vector ABC Elite system and a subsequent incubation in diaminobenzidine for visualization of the reaction product. For negative controls the primary antibody was omitted.
Immunofluorescence
For double immunofluorescence, sections were de-waxed in xylene and rehydrated. Microwave treatment in 10 mM citrate buffer (pH 6) for 3 x 3 min was followed by blockade of non-specific binding by incubation in 3% normal goat serum and 0.5% Triton X-100 in 0.1 M PBS for 30 min and subsequent incubation overnight at room temperature with the following primary antibodies: monoclonal mouse anti-NOGO-A antibody (Oertle et al., 2003, 1 µg/ml), monoclonal anti-NF antibody (Sigma, Clone52, diluted 1 : 100), monoclonal anti-P0 (gift from Dr J. Archelos) polyclonal rabbit anti-NF (Sigma-Aldrich, diluted 1 : 1000), polyclonal rabbit anti-collagen IV (ICN, diluted 1 : 200), polyclonal anti-calcitonin gene related protein (CGRP) (Sigma, C8198, diluted 1 : 1000) and polyclonal rabbit anti-MBP (Chemicon, diluted 1 : 1000). After the subsequent incubation with Alexa 594 (red-fluorescence)-conjugated goat anti-mouse and Alexa 488 (green fluorescence)-conjugated goat anti-rabbit secondary antibodies (Molecular Probes, diluted 1 : 500) for 3 h at room temperature, slides were coverslipped with Moviol. To check for unspecific cross reactivity co-incubation of the different combinations of non-corresponding primary and secondary antibodies as well as both secondary antibodies was performed.
For triple immunofluorescence, the tyramide signal amplification kit (TSA Cyanine 3 system, NEL704A, PerkinElmer Life Sciences) was used. Briefly, following the blockade of endogenous tissue peroxidase, sections were rinsed in 0.1 M PBS and incubated with the anti-collagen IV (diluted 1 : 10 000) or the anti-MBP antibody (diluted 1 : 50 000) overnight. Incubation with a biotinylated horse anti-rabbit antibody (1 : 500, BA2001, Vector) for 1 h and blocking with the provided blocking reagent for 30 min was followed by streptavidin–HRP (diluted 1 : 500) in blocking reagent for 30 min and cyanine 3-tyramide working solution (diluted 1 : 100) for 10 min. After rinsing, the slides were incubated with the monoclonal anti-NF primary antibody (Sigma, Clone52, diluted 1 : 100) and the polyclonal GFAP antibody (DAKO, diluted 1 : 2000) overnight, followed by the Alexa 350 (blue-fluorescence)-conjugated goat anti-mouse (Molecular Probes, diluted 1 : 100) and the Alexa 488 (green-fluorescence)-conjugated goat anti-rabbit secondary antibody (Molecular Probes, diluted 1 : 250) for 3 h at room temperature. Finally, the sections were cover-slipped with Moviol. In correspondence with the double immunofluorescence, all combinations of primary and secondary antibodies were checked for unspecific immunoreactivity.
|
Results |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsEarly survival times (2-11...DiscussionReferences |
|---|
The spinal cords of 19 individuals were examined using immunohistochemistry for collagen IV, laminin, fibronectin, NGFr, GFAP, MBP, NOGO-A, P0, CGRP and NF. The brains of all individuals were carefully examined and declared to be without pathological findings. The spinal cords of the four control cases were also without pathological findings. The pathological cases have been subdivided into two groups according to the postlesion survival times (i.e. early and late survival times) because distinct morphological stages in the formation of the scar were found.
Normal distribution of collagen IV, laminin, fibronectin, NGFr and GFAP in the spinal cord
In control cases, staining for both collagen IV and laminin was overlapping and was confined to the basal lamina and ECM of the meninges as well as to meningeal and intra-parenchymal blood vessels (Fig. 1A and C). Nerve roots revealed endo- and perineurial immunoreactivity as well as staining in blood vessel walls (Fig. 1B and D). Similar to collagen IV and laminin, fibronectin immunoreactivity in the spinal cord parenchyma could also be seen in the meninges and in both meningeal and parenchymal blood vessel walls (Fig. 1E). In nerve roots, staining was detected in the endo-, peri- and the epineurium as well as in the blood vessel walls (Fig. 1F). NGFr immunoreactivity was restricted to the nerve roots and was located in the peri- and epineurium as well as in some Schwann cells (Fig. 1G). GFAP immunohistochemistry demonstrated astrocytic cell bodies in between the network of fine processes detected in both grey and white matter of the spinal cord (Fig. 1H) and no immunoreactivity in the nerve roots. The normal distribution of NOGO-A and MBP in the normal human spinal cord was as previously described elsewhere (Buss et al., 2004, 2005). NOGO-A immunohistochemistry revealed neuronal staining in motoneurons, Clarke's nucleus neurons and subpopulations of interneurons. Furthermore, oligodendrocytic cell bodies as well as the peri- and abaxonal membranes of CNS myelin sheaths were immunopositive (Fig. 1I). MBP immunoreactivity was present in the compact myelin around both CNS and PNS axons (Fig. 1J).
|
Morphological appearance of the lesion site
The lesion sites of the severe human traumatic SCI cases could be subdivided into the lesion epicentre, an intermediate zone and the perilesional area. The lesion epicentre was initially characterized by the complete destruction of cytoarchitecture to the extent that it was difficult, if not impossible, to distinguish grey and white matter regions. Furthermore, haemorrhagic infiltration was visible in between the amorphous regions of tissue destruction (not shown). At 24 days after injury, the first indication of Schwann cell migration into the lesion core was seen and in cases with survival times of 4 months and longer the lesion epicentre was characterized by a dense ECM with embedded nerve root-like structures as well as individual myelinated nerve fibres (Fig. 2A). Beside meningeal cells and fibroblasts, Schwann cells could be found in the neuromas and in the ECM (see subsequently). In contrast, no astrocytes were detectable in this region.
|
The intermediate zone included the extremities of the lesion site and their interface with the adjacent damaged, but non-degenerating, CNS parenchyma. After an initially dramatic loss of the astroglial framework, the remaining astroglia became activated and produced a dense, irregular scar at later time points (see subsequently). At survival times of 4 months and later, sections from this area, therefore, demonstrated a clear demarcation between the Schwann cell area and that of the astroglial scar, with the astrocyte-dominated area becoming increasingly more evident in sections further away from the lesion epicentre. In most cases, cysts could also be seen in between these 2 areas (Fig. 2B and C).
In the perilesional area, the spinal cord was largely intact but was clearly distorted. Nonetheless, grey and white matter regions were clearly distinguishable and the astroglial framework was detectable at all time points. At survival times ranging from 2 days to 4 months, hypertrophic activated/reactive astrocytes could be seen in the white matter. Furthermore, the perilesional area could be characterized by an early appearance of newly formed blood vessels, starting 4 days after injury (see subsequently). At no time point after injury could migrating Schwann cells be found in the perilesional CNS parenchyma (Fig. 2D).
|
Early survival times (2–11 days post insult) |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsEarly survival times (2-11...DiscussionReferences |
|---|
Lesion epicentre
In these cases, the lesion epicentre was characterized by the complete destruction of cytoarchitecture. The loss of staining for collagen IV, laminin and fibronectin indicated the destruction of blood vessels, but the incidence of vessel staining slowly increased from 2 to 11 days after injury (Fig. 3A and B). At all subsequent survival times, the number of blood vessels remained below normal. Staining for collagen IV, laminin and fibronectin revealed no other immunoreactive structures. GFAP immunoreactivity clearly revealed a loss of astrocytes and their processes at the lesion site early after SCI. In heavily damaged areas, neither GFAP nor NGFr immunoreactivity could be detected (not shown).
|
Intermediate zone
In the less severely affected areas at the border of the lesion epicentre, the density of astrocytic cells and their network of processes was dramatically decreased compared with control cases (compare Fig. 3C with Fig. 1H). Ten and eleven days after injury, the first signs of an astrocytic reaction to the lesion could be detected. Nests of reactive GFAP-positive cell bodies with a dense, irregular network of processes were observed (Fig. 3D). Similar to observations at the lesion epicentre, no NGFr immunopositive structures could be detected at this early survival time (data not shown).
Perilesional area
In the perilesional area, 1–2 segments distant from the primary lesion site, immunohistochemistry for collagen IV, laminin and fibronectin revealed the early post-traumatic appearance of newly formed blood vessels in both grey and white matter. Their appearance was irregular and they were often arranged in clusters. These vascular structures were first seen 4 days after injury when they were mostly of a small size (e.g. 50 µm, Fig. 3E). Ten and eleven days after SCI, their size and density were heterogeneous, with enlarged luminae (up to 250 µm) and thickened vessel walls often being detectable (Fig. 3E insert). No NGFr immunoreactivity was found in the spinal cord parenchyma; instead staining was restricted to nerve roots and was located in the peri- and epineurium as well as in some Schwann cells (not shown). GFAP immunoreactivity revealed an intact astroglial framework. At all survival times, large activated astrocytes could be seen spread over the white matter (Fig. 3F).
Late survival times (24 days–30 years post insult)
Lesion epicentre
At 24 days after trauma, NGFr staining revealed the first signs of glial cell migration. There was an elevated density of NGFr immunoreactivity on small diameter cell bodies and processes within the damaged spinal parenchyma in the vicinity of spinal nerve roots. This was interpreted as being an indicator of Schwann cell migration. The NGFr positive cells infiltrated up to 900 µm into the spinal cord tissue with a decreasing density towards the more central region (Fig. 4A). The region of infiltrating Schwann cells was found to be immunoreactive for collagen IV, laminin and fibronectin. The staining for these molecules was present on Schwann cells and their processes as well as on blood vessel walls (Fig. 4B and C). Fibronectin staining was also scattered as a loose extracellular network which enveloped the rounded phagocytic macrophages (Fig. 4D).
|
Double immunofluorescence revealed that most of the initial migration of Schwann cells into the lesioned spinal cord parenchyma was not directly accompanied by regenerating nerve fibres. In the regions of the lesion epicentre that were further away from nerve roots and thus not infiltrated by Schwann cells, some irregular NF-positive structures were detectable most likely representing debris which was still to be phagocytosed by infiltrating macrophages (Fig. 5A). However, occasionally densely packed and orientated Schwann cell processes were associated with individual, similarly aligned axons (Fig. 5B). In this area, collagen IV, laminin and fibronectin staining was restricted to blood vessels (e.g. Fig. 5C), the number of which had increased compared with earlier survival times, and which displayed an irregular appearance with thickened, often doubled ring-like, vessel walls (e.g. see Fig. 3E).
|
At survival times of 4 months and longer after SCI, the lesion site could be clearly divided into two regions: (i) a Schwann cell-containing area (which could be further subdivided into areas rich in ECM and also in neuromas) and (ii) an astrocyte-dominated scar. The lesion epicentre revealed massive infiltration by NGFr-positive Schwann cells (Fig. 6A). This area had, by now, become partially filled with sheet-like lamellae of ECM which were collagen IV-, laminin- and fibronectin-positive. Staining for the different ECM proteins revealed an overlapping distribution. The collagen IV- and laminin-positive lamellae were more loosely packed and highly irregular (Fig. 6B and C). Fibronectin immunoreactivity of the ECM revealed densely packed fibrils that were intensely stained (Fig. 6D). Round and oval structures resembling regrowing processes of nerve root fibres could also be seen. These neuromas contained larger numbers of NGFr-positive Schwann cells (Fig. 6E) as well as myelinated nerve fibres. Furthermore, individual nerve fibres were encircled by rings of collagen IV and laminin immunoreactive basal lamina (Fig. 6F and G). Immunohistochemistry for fibronectin revealed a more diffuse staining pattern within the neuromas, partially filling the endoneurial space between the axons (Fig. 6H). Double and triple immunofluorescence of the neuromas revealed that such axons were encompassed by P0- and MBP-positive myelin sheaths. These in turn were surrounded by the collagen IV-, laminin- and fibronectin-positive rings of basal lamina (Fig. 7A–E). Immunohistochemistry for CGRP demonstrated that at least a subpopulation of these axons originated from regenerated dorsal roots (Fig. 7F). Staining for NOGO-A revealed no immunoreative structures, demonstrating the absence of mature oligodendrocytes in this area (data not shown).
|
Intermediate zone
At 24 days after SCI, the intermediate zone was devoid of infiltrating Schwann cells. The accompanying ECM molecules collagen IV, laminin and fibronectin were restricted to blood vessel walls (Fig. 4E). Instead, the area revealed the first signs of astrocytic scar formation. In these regions of grey and white matter, a densely packed mass or network of diffusely stained GFAP-positive processes could be seen (Fig. 4F).
At survival times of 4 months and longer after SCI, the territory of the densely packed GFAP-positive astroglial scar was clearly distinguishable and distinct from that of the Schwann cell dominated regions (Fig. 7G and H). In the astrocytic scar, triple immunofluorescence revealed irregularly shaped, thin NF-positive axons with no indication of enlarged end bulbs and therefore not resembling dystrophic axons as described earlier in experimental animals. These axons were found individually or as bundles, which were scattered throughout the astroglial scar. None of the axons in this area was CGRP immunopositive (not shown). Most of these nerve fibres were myelinated as indicated by MBP immunohistochemistry (Fig. 7I). However, in contrast to the Schwann cell area, no P0 immunoreactivity could be demonstrated within the astrocytic domain, indicating myelin sheaths of the CNS origin (Fig. 7J). Furthermore, there was no co-localization with collagen IV, laminin or fibronectin positive structures (Fig. 7K). Immunoreactivity for these 4 ECM-related molecules was restricted to blood vessels with a mostly irregular appearance, partly with enlarged luminae and a thickened vessel wall as described earlier. NGFr immunohistochemistry revealed no Schwann cells or processes within the astroglial component of the scar (not shown). Staining for NOGO-A demonstrated oligodendrocytic cell bodies spread heterogeneously throughout the astroglial scar, individually or as nests of immunoreactive cells. Some oligodendrocytes were found in close proximity to the regenerating nerve fibres, but NOGO-A immunoreactivity could not be demonstrated on the myelin sheaths (Fig. 7L).
|
It is important to note that the lesion epicentre with the invading PNS structures (neuromas and Schwann cell dominated ECM) was strictly separated from the astroglial scar surrounding this area. Neurofilament immunohistochemistry revealed no indication of axons traversing the interface between the astroglia dominated compartment of the scar and the Schwann cell dominated compartment. Overall, the data gave the impression that an area of the lesioned spinal cord had adopted a PNS phenotype through the invasion of fibroblasts, meningeal cells, Schwann cells and axons. This PNS dominated lesion site had thus become effectively sealed off or encapsulated by the surrounding astrocytic scar.
|
Discussion |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsEarly survival times (2-11...DiscussionReferences |
|---|
Over recent years, remarkable progress in the understanding of the cellular and molecular events following SCI has been made through the use of experimental animals. Some of these advances have resulted in the development of experimental intervention strategies, the success of which have led to the initiation of clinical trials. Despite these rapid advances in the experimental domain, relatively little is known about the correlative events that take place in traumatically injured human tissues. In an attempt to address this imbalance, the present investigation has focused on the spatiotemporal distribution of a range of axon-growth promoting and axon-growth inhibitory molecules at the lesion site of post-mortem material obtained from patients who died at a range of survival times following neurologically ‘complete’, maceration-type SCIs.
Studies in experimental animals have shown that the intact network of astrocytes and their processes normally prevents Schwann cells from entering the CNS (Blakemore, 1992). However, after spinal cord injury-induced disruption of the glia limitans and the (complete) loss of the astroglial framework at the lesion epicentre, Schwann cells are able to migrate into the lesion site (e.g. Brook et al., 1998). This scenario also appears to take place in the present cases of severe human traumatic SCI, in which there was complete destruction of the CNS parenchyma at the lesion epicentre. Subsequently, cell populations normally not present within the CNS, such as fibroblasts, meningeal cells and Schwann cells can invade the spinal cord. The migrating Schwann cells would provide an ideal substrate for axonal regeneration also emanating from the damaged dorsal nerve roots. The present immunohistochemical data showed the presence of occasional NGFr-positive Schwann cells within the normal (undamaged) spinal nerve roots. The lesioned spinal cord, however, contained many more NGFr-positive profiles. It is likely that the lesion-induced dedifferentiation of Schwann cells from the damaged spinal nerve roots resulted in their massive upregulation of NGFr, as already demonstrated by others (Johnson et al., 1988). This upregulation of NGFr by the dedifferentiated Schwann cells would make them more readily detectable by the present immunohistochemical procedure. The identification of NGFr-positive cellular migration into human spinal cord lesion sites is consistent with experimental data. However, NGFr immunoreactivity alone is not sufficient for the unequivocal identification of Schwann cells, since previous studies in post-mortem human tissues have also revealed oligodendrocyte precursor cells in control CNS and macrophages/microglia as well as mature oligodendrocytes in pathological CNS tissue to be NGFr-immunopositive (Dowling et al., 1999; Chang et al., 2000; Petratos et al., 2004). Nevertheless, apart from the possible contribution of some oligodendrocyte precursor cells, the distribution pattern and morphology of NGFr-positive cells in the present investigation suggests that most immunoreactive cells are indeed Schwann cells.
Schwann cell migration into human SCI lesion sites confirms previous observations in post-mortem human material (Wolman, 1965; Hughes, 1978; Kakulas, 1984; Wang et al., 1996; Bruce et al., 2000; Guest et al., 2005) as well as numerous experimental investigations (e.g. Beattie et al., 1997; Brook et al., 1998; Brook et al., 2000). In the present study, the first indication of Schwann cell proliferation and migration occurred by 3 weeks after injury, which is somewhat delayed in comparison to the previously published data using experimental models of SCI where the first signs of Schwann cell migration into experimental rat SCI were detected during the first week after injury (e.g. Beattie et al., 1997; Brook et al., 1998; Brook et al., 2000). However, Schwann cell migration appeared substantially earlier than the 3–4 month time points reported by others in human SCI (Kakulas, 1984; Bruce et al., 2000;). The previous investigations identified schwannosis in the majority, but not all cases of human SCI, whereas in the present investigation, massive amounts of Schwann cell dominated neuromas and ECM scarring were found in all cases with survival times of 8 months and longer. It is likely that these differences are due to selection criteria employed for investigation. Here, only cases undergoing severe SCI (clinically complete injuries) were chosen for investigation whereas others have chosen a more heterogeneous group. In general, it is important to note that the present results were generated in cases with acute and massive traumatic SCI, representing the maceration type injury as classified earlier (Bunge et al., 1997). Future studies are needed to verify if similar Schwann cell invasion of the lesion site can be found in less severe as well as other types of SCI, in particular following laceration and contusion type injuries.
The invading Schwann cells form an integral part of the fibroadhesive-glial scar evolving at the lesion site epicentre following traumatic SCI. The intermediate zone of such lesions, containing the boundary or interface of the lesion with the adjacent intact spinal cord, consists of various cell populations, including reactive astrocytes and newly formed vasculature. Such cells and structures are embedded in an ECM (containing multiple growth-modulating molecules) that is widely acknowledged to be a major impediment to the regeneration of lesioned CNS axons (Kakulas, 1984; Schwab and Bartholdi, 1996).
The presence of collagen IV (as a major constituent of basal lamina) has been demonstrated in the connective tissue scar of experimental animals after SCI. However, there is, at present, an apparent lack of consensus regarding functional consequences of this expression. On the one hand, it has been reported that basal lamina plays a substantial role in preventing CNS axon regeneration, presumably by acting as a framework to which growth-repulsive ECM molecules (in particular highly sulphated proteoglycans) bind and exert their effects (Stichel et al., 1999; Hermanns et al., 2001). However, some groups have been unable to confirm this notion while others have even observed numerous NF-positive axons growing within the collagen IV-rich ECM of spinal cord lesions (Weidner et al., 1999; Joosten et al., 2000; Iseda et al., 2003). This lack of consensus is of particular interest because it has been suggested that preventing collagen IV deposition during the early stages of scar formation might be a clinically applicable approach to promote axon regeneration following human SCI (Klapka et al., 2005). There has, to date, been no published indication of the time scale over which collagen IV appears within the lesion site following human SCI, primarily due to the relative scarcity of appropriate material for immunohistochemical investigation. The temporal and spatial expression pattern of collagen IV at the lesion site after human SCI tends to support the notion that collagen IV rich basal lamina is axon growth-promoting (at least for peripherally derived axons) rather than inhibitory. At the lesion site, any close proximity of collagen IV immunoreactive structures with nerve fibres was confined to the Schwann cell dominated areas. In the surrounding astroglial component of the scar, only the basal lamina of newly formed blood vessels was stained. These vascular structures, however, did not co-localize with any sort of growing axons. The first co-localization of collagen IV-positive structures with nerve fibres could be seen 8 months after injury. At the 24 day survival time, infiltrating Schwann cells in the spinal cord parenchyma around nerve roots demonstrated collagen IV immunoreactivity associated with their cell bodies and processes. At this early time point however, the Schwann cell network was not yet accompanied by axonal growth. Eight months after SCI, the Schwann cell dominated region at the lesion epicentre was filled with structures resembling nerve root fibres and a surrounding ECM that strictly delineated this area from the adjoining astroglial scar. The regenerating axons in the neuromas were not only myelinated by Schwann cells, as indicated by P0 and MBP immunohistochemistry, but they were also surrounded by an arrangement of ECM molecules not usually present in mature nerve roots, including collagen IV that directly encased the nerve fibres. This pattern resembles outgrowing neural crest structures during development in which collagen IV plays an important role for the guidance of neuronal elongation (Venstrom and Reichardt, 1993). After developmental axonal growth has ceased, the distribution of the molecule becomes progressively more restricted, to the extent that, in mature peripheral nerves, collagen IV can be detected immunohistochemically as thin bands in the endo- and perineurial basement membrane and that of blood vessels (Hill and Williams, 2002). However, in the lesion induced neuromas, the expression pattern of the molecule remains strongly elevated in endo- and perineurium, even 20–30 years after injury. This co-localization of collagen IV and outgrowing peripheral nerve fibres could also be seen in the ECM surrounding the neuromas.
There was no indication of any regenerating axons crossing from the astrocytic region into the Schwann cell region or vice versa (i.e. across the CNS–PNS interface of the lesion site) nor of any axons reaching this interface and turning or being deflected away. This observation, combined with the distribution of oligodendrocyte-derived myelin, suggests that the few axons that were located in the astrocytic scar tissue were derived from the CNS. The inability to detect any axons traversing the CNS–PNS interface of the lesion site warrants the caveat that transverse sections of the spinal cord (as used in the present investigation) may not be ideal for the identification of such behaviour and that it remains conceivable that CNS axons may have traversed this boundary and subsequently undergone retraction or die-back.
A number of observations suggest that the axons observed within the Schwann cell rich domain were derived from a peripheral source. First of all, as direct evidence, a subpopulation of nerve fibres in the neuromata were CGRP-positive demonstrating an origin from regenerating dorsal root ganglion cells. Furthermore, no axons were detected traversing the CNS–PNS interface, the nerve fibres were mostly arranged in root-like structures and their myelin sheaths were P0-positive. Destruction of the astroglial framework would permit the migration of fibroblasts, Schwann cells and axons from damaged spinal nerve roots. However, in the absence of definitive proof of the peripheral source of all nerve fibres within the lesion core, it cannot be excluded that some Schwann cells were associated with regrowing CNS axons. Others have recently provided indirect evidence that Schwann cells are capable of associating with and remyelinating spared CNS axons following mild to moderate human SCI. This behaviour of migrating Schwann cells has been termed ‘atypical’ schwannosis (Guest et al., 2005).
It seems clear that the data obtained from the present post-mortem human material are not able to support the notion that collagen IV expression and basal lamina deposition contribute to the failure of CNS axon regeneration. This was particularly evident by the inability to detect any dystrophic axons or even markedly deviated axons at the collagen IV-rich lesion interface. Taken together, in human SCI, collagen IV is most likely produced by Schwann cells and fibroblasts, where it plays a supporting role in the outgrowth of PNS axons from lesioned nerve roots.
The temporal and spatial expression pattern of collagen IV was closely matched by that of laminin, another important constituent of basal lamina. Laminin was also co-localized with the regenerating PNS axons in the neuromas and the surrounding ECM. This pattern is in line with many experimental in vitro studies demonstrating a growth-promoting effect of laminin on various neuronal populations (Condic and Lemons, 2002; Grimpe and Silver, 2002). Furthermore, it recapitulates the guiding function of laminin during the developmental and regenerative outgrowth of the PNS. In the astroglial scar, laminin staining (like collagen IV) was restricted to blood vessel walls. This lack of proximity between any laminin immunopositive structure and the few regenerating CNS axons in this area differs from data from experimental animals. In experimental traumatic SCI, regenerating CNS axonal growth cones were reported to stop at the laminin-rich fibroglial scar which was interpreted as an entrapment of these nerve fibres in laminin rich regions of the scar tissue (Condic and Lemons, 2002). Such a growth-interfering effect of laminin on CNS axons could not be seen in the present human material.
In a recent experimental study, astrocyte-bound fibronectin was purported to promote regeneration of donor adult dorsal root ganglion (DRG) axons in a degenerating CNS white matter tract (Tom et al., 2004). The ability of axons to grow in the normal and lesioned adult CNS reflects the concept that the balance between growth-promoting and growth-inhibitory molecules determines the extent of any axonal regeneration. The present data revealed no association of fibronectin with reactive astrocytes at or around the lesion site. Fibronectin does, however, appear to be produced by cells invading the lesion site, including Schwann cells, and was associated with areas rich in PNS axon regeneration. The ability of fibronectin to support axonal growth following experimental SCI was clearly demonstrated by King and colleagues who used mats of fibronection implanted into the lesion site to induce significant axon regeneration (King et al., 2005).
The present data demonstrates no association of collagen IV, laminin or fibronectin with the reactive astrocytes located in the region of traumatic human SCI. The inevitable delays that occur before such post-mortem tissues undergo fixation will result in suboptimal antigen preservation. It remains possible that low levels of antigen, below the level of detection by the current immunohistochemical approach, may still be present at pathophysiologically relevant concentrations. These molecules were, nonetheless, clearly detectable in the Schwann cell-dominated areas of the lesion site and its associated ECM.
In summary, maceration of the human spinal cord following severe traumatic injury leads to the destruction of the astroglial framework and extensive proliferation and migration of fibroblasts and Schwann cells, associated with the local deposition of ECM-related molecules collagen IV, laminin and fibronectin. Thus, the post-traumatic expression of collagens, laminin and fibronectin in human SCI is not extensively derived from reactive astrocytes and astroglial scarring, but rather from populations of cells migrating into the lesion site, such as fibroblasts, meningeal cells and Schwann cells. The surrounding astroglial scar effectively isolates these Schwann cell/ECM-rich territories from the adjacent intact spinal parenchyma. The subsequent axonal growth into the lesion site appears to be mainly derived from damaged spinal nerve roots apart from a few CNS neurites which are located in the astroglial scar. In the separate and distinct astroglial dominated compartment, these ECM related molecules appear to be restricted to blood vessel walls. The observation that the perilesional astroglia surrounded the few axons that remained in this area might favour the notion that it is the astroglial compartment that is playing a dominant role in preventing CNS regeneration (e.g. for review see Silver and Miller, 2004). The failure to demonstrate any collagen IV, laminin or fibronectin upregulation associated with the astroglial scar could mean that other molecules may play a significant role in human SCI.
The neuromas in the Schwann cell dominated lesion epicentre have already been described following human SCI, not only after acute injuries but also in more chronic traumatic mechanisms such as compression due to a vertebral metastasis. These regenerating axons were reported to be derived from injured peripheral nerve roots (Hughes and Brownell, 1963; Bruce et al., 2000). Such neuromas probably generate no functional benefit and may even play a detrimental role, such as in post-traumatic pain or spasticity in patients with chronic SCI.
Finally, the present study highlights the importance of undertaking correlative investigations in post-mortem human material to clearly establish the relationship between data obtained in experimental animals with what actually takes place in the human spinal cord following traumatic injury. The rational development of any experimental intervention strategy that is intended to be translated to the clinical domain requires a detailed understanding of the spatial and temporal expression patterns of the key functional molecules in appropriate samples of damaged human tissues.
|
Acknowledgements |
|---|
The authors thank S. Lecouturier for excellent technical assistance.
|
References |
|---|
|
TopSummaryIntroductionMaterial and methodsResultsEarly survival times (2-11...DiscussionReferences |
|---|
Beattie MS, Bresnahan JC, Komon J, et al. (1997a) Endogenous repair after spinal cord contusion injuries in the rat. Exp Neurol 148:453–63.[CrossRef][Web of Science][Medline]
Blakemore WF. (1992) Transplanted cultured type-1 astrocytes can be used to reconstitute the glia limitans of the CNS: the structure which prevents Schwann cells from myelinating CNS axons. Neuropathol Appl Neurobiol 18:460–6.[Web of Science][Medline]
Brook GA, Houweling DA, Gieling RG, et al. (2000) Attempted endogenous tissue repair following experimental spinal cord injury in the rat: involvement of cell adhesion molecules L1 and NCAM? Eur J Neurosci 12:3224–38.[CrossRef][Web of Science][Medline]
Brook GA, Plate D, Franzen R, et al. (1998) Spontaneous longitudinally orientated axonal regeneration is associated with the Schwann cell framework within the lesion site following spinal cord compression injury of the rat. J Neurosci Res 53:51–65.[CrossRef][Web of Science][Medline]
Bruce JH, Norenberg MD, Kraydieh S, Puckett W, Marcillo A, Dietrich D. (2000) Schwannosis: role of gliosis and proteoglycan in human spinal cord injury. J Neurotrauma 17:781–8.[Web of Science][Medline]
Bunge RP, Puckett WR, Hiester ED. (1997) Observations on the pathology of several types of human spinal cord injury, with emphasis on the astrocyte response to penetrating injuries. Adv Neurol 72:305–15.[Medline]
Buss A, Brook GA, Kakulas B, et al. (2004) Gradual loss of myelin and formation of an astrocytic scar during Wallerian degeneration in the human spinal cord. Brain 127:34–44.
Buss A, Sellhaus B, Wolmsley A, Noth J, Schwab ME, Brook GA. (2005) Expression pattern of NOGO-A protein in the human nervous system. Acta Neuropathol (Berl) 110:113–9.[CrossRef][Medline]
Chang A, Nishiyama A, Peterson J, Prineas J, Trapp BD. (2000) NG2-positive oligodendrocyte progenitor cells in adult human brain and multiple sclerosis lesions. J Neurosci 20:6404–12.
Condic ML and Lemons ML. (2002) Extracellular matrix in spinal cord regeneration: getting beyond attraction and inhibition. Neuroreport 13:37–48.[CrossRef]
Dowling P, Ming X, Raval S, et al. (1999) Up-regulated p75NTR neurotrophin receptor on glial cells in MS plaques. Neurology 53:1676–82.
Fawcett JW and Asher RA. (1999) The glial scar and central nervous system repair. Brain Res Bull 49:377–91.[CrossRef][Web of Science][Medline]
Grimpe B and Silver J. (2002) The extracellular matrix in axon regeneration. Prog Brain Res 137:333–49.[Web of Science][Medline]
Guest JD, Hiester ED, Bunge RP. (2005) Demyelination and Schwann cell responses adjacent to injury epicenter cavities following chronic human spinal cord injury. Exp Neurol 192:384–93.[CrossRef][Web of Science][Medline]
Hermanns S, Klapka N, Muller HW. (2001) The collagenous lesion scar—an obstacle for axonal regeneration in brain and spinal cord injury. Restor Neurol Neurosci 19:139–48.[Web of Science][Medline]
Hill RE and Williams RE. (2002) A quantitative analysis of perineurial cell basement membrane collagen IV, laminin and fibronectin in diabetic and non-diabetic human sural nerve. J Anat 201:185–92.[CrossRef][Web of Science][Medline]
Hughes JT. (1978) Pathology of the spinal cord. Major Probl Pathol 6:1–257.[Medline]
Hughes JT and Brownell B. (1963) Aberrant nerve fibres within the spinal cord. J Neurol Neurosurg Psychiatry 26:528–34.
Iseda T, Nishio T, Kawaguchi S, Kawasaki T, Wakisaka S. (2003) Spontaneous regeneration of the corticospinal tract after transection in young rats: collagen type IV deposition and astrocytic scar in the lesion site are not the cause but the effect of failure of regeneration. J Comp Neurol 464:343–55.[CrossRef][Web of Science][Medline]
Johnson EM Jr, Taniuchi M, DiStefano PS. (1988) Expression and possible function of nerve growth factor receptors on Schwann cells. Trends Neurosci 11:299–304.[CrossRef][Web of Science][Medline]
Joosten EA, Dijkstra S, Brook GA, Veldman H, Bar PR. (2000) Collagen IV deposits do not prevent regrowing axons from penetrating the lesion site in spinal cord injury. J Neurosci Res 62:686–91.[CrossRef][Web of Science][Medline]
Kakulas BA. (1984) Pathology of spinal injuries. Cent Nerv Syst Trauma 1:117–29.[Medline]
King VR, Phillips JB, Hunt-Grubbe H, Brown R, Priestley JV. (2006) Characterization of non-neuronal elements within fibronectin mats implanted into the damaged adult rat spinal cord. Biomaterials 27:485–96.[CrossRef][Web of Science][Medline]
Klapka N, Hermanns S, Straten G, et al. (2005) Suppression of fibrous scarring in spinal cord injury of rat promotes long-distance regeneration of corticospinal tract axons, rescue of primary motoneurons in somatosensory cortex and significant functional recovery. Eur J Neurosci 22:3047–58.[CrossRef][Web of Science][Medline]
Kleinman HK, McGarvey ML, Liotta LA, Robey PG, Tryggvason K, Martin GR. (1982) Isolation and characterization of type IV procollagen, laminin, and heparan sulfate proteoglycan from the EHS sarcoma. Biochemistry 21:6188–93.[CrossRef][Medline]
Lefcort F, Venstrom K, McDonald JA, Reichardt LF. (1992) Regulation of expression of fibronectin and its receptor, alpha 5 beta 1, during development and regeneration of peripheral nerve. Development 116:767–82.[Abstract]
McDonald JW and Sadowsky C. (2002) Spinal-cord injury. Lancet 359:417–25.[CrossRef][Web of Science][Medline]
Oertle T, van der Haar ME, Bandtlow CE, et al. (2003) Nogo-A inhibits neurite outgrowth and cell spreading with three discrete regions. J Neurosci 23:5393–406.
Petratos S, Gonzales MF, Azari MF, et al. (2004) Expression of the low-affinity neurotrophin receptor, p75(NTR), is upregulated by oligodendroglial progenitors adjacent to the subventricular zone in response to demyelination. Glia 48:64–75.[CrossRef][Web of Science][Medline]
Schwab ME and Bartholdi D. (1996) Degeneration and regeneration of axons in the lesioned spinal cord. Physiol Rev 76:319–70.
Shiga T and Oppenheim RW. (1991) Immunolocalization studies of putative guidance molecules used by axons and growth cones of intersegemental interneurons in the chick embryo spinal cord. J Comp Neurol 310:234–52.[CrossRef][Web of Science][Medline]
Silver J and Miller JH. (2004) Regeneration beyond the glial scar. Nat Rev Neurosci 5:146–56.[CrossRef][Web of Science][Medline]
Stichel CC, Hermanns S, Luhmann HJ, et al. (1999) Inhibition of collagen IV deposition promotes regeneration of injured CNS axons. Eur J Neurosci 11:632–46.[CrossRef][Web of Science][Medline]
Timpl R, Rohde H, Robey PG, Rennard SI, Foidart JM, Martin GR. (1979) Laminin—a glycoprotein from basement membranes. J Biol Chem 254:9933–7.
Tom VJ, Doller CM, Malouf AT, Silver J. (2004) Astrocyte-associated fibronectin is critical for axonal regeneration in adult white matter. J Neurosci 24:9282–90.
Venstrom KA and Reichardt LF. (1993) Extracellular matrix. 2: role of extracellular matrix molecules and their receptors in the nervous system. FASEB J 7:996–1003.[Abstract]
Wang ZH, Walter GF, Gerhard L. (1996) The expression of nerve growth factor receptor on Schwann cells and the effect of these cells on the regeneration of axons in traumatically injured human spinal cord. Acta Neuropathol (Berl) 91:180–4.[CrossRef][Medline]
Weidner N, Grill RJ, Tuszynski MH. (1999) Elimination of basal lamina and the collagen "scar" after spinal cord injury fails to augment corticospinal tract regeneration. Exp Neurol 160:40–50.[CrossRef][Web of Science][Medline]
Wolman L. (1965) The disturbance of circulation in traumatic paraplegia in acute and late stages: a pathological study. Paraplegia 59:213–26.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||