Download Growth-modulating molecules are associated with invading

doi:10.1093/brain/awl374
Brain (2007), 130, 940 ^953
Growth-modulating molecules are associated with
invading Schwann cells and not astrocytes in human
traumatic spinal cord injury
Armin Buss,1 Katrin Pech,1 Byron A. Kakulas,2 Didier Martin,3,5 Jean Schoenen,4,5 Johannes Noth1 and
Gary A. Brook1
1
Department of Neurology, Aachen University Hospital, Germany, 2Centre for Neuromuscular and Neurological
Disorders, University of Western Australia, Perth, Australia, 3Department of Neurosurgery, Sart Tilman Hospital,
4
Department of Neurology and Neuroanatomy and 5Center for Cellular and Molecular Neuroscience, University of Lie'ge,
Lie'ge, Belgium
Correspondence to: Armin Buss, Pauwelsstrasse 30, 52074 Aachen, Germany
E-mail: [email protected]
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.
Keywords: spinal cord injury; human; regeneration; glial scar
Abbreviations: CNS ¼ central nervous system; ECM ¼ extracellular matrix; NF ¼ neurofilament;
PNS ¼ peripheral nervous system; SCI ¼ spinal cord injury
Received June 30, 2006. Revised December 8, 2006. Accepted December 13, 2006. Advance Access publication February 21, 2007
Introduction
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
results in serious limitations
(McDonald and Sadowsky,
progress in recent years,
(SCI) is often permanent and
to the patients’ quality of life
2002). Despite considerable
the underlying mechanisms
ß The Author (2007). Published by Oxford University Press on behalf of the Guarantors of Brain. All rights reserved. For Permissions, please email: [email protected]
Growth-modulating molecules in human SCI
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., 1999; Hermanns et al.,
2001). However, since other groups have 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
Brain (2007), 130, 940 ^953
941
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.
942
Brain (2007), 130, 940 ^953
A. Buss et al.
Material and methods
Table 1 Patients who served as the control group
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.
Case number
Age
Cause of death
1
2
3
4
29 years
47 years
62 years
83 years
Breast cancer
Pneumonia
Breast cancer
Myocardial infarction
Case number
Age
Injury level
Injury^ death
interval
Peroxidase immunohistochemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
21 years
51 years
84 years
65 years
63 years
18 years
72 years
85 years
76 years
80 years
72 years
44 years
71 years
47 years
57 years
T12
C1
C3- 4
C5
C6
T6
T11-12
C3
T8 -9
C5- 6
T12
L1
C3- 4
T5
T3- 4
2 days
4 days
5 days
8 days
11 days
12 days
24 days
4 months
10 months
1 year
2 years
8 years
20 years
26 years
30 years
Transverse sections (5 mm thick) were collected onto poly-L-lysinecoated 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 3 min.
Sections for laminin staining were treated with 10 mg/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 antineurofilament (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 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 mg/ml), monoclonal anti-NF
antibody (Sigma, Clone52, diluted 1 : 100), monoclonal anti-P0
(gift from Dr J. Archelos) polyclonal rabbit anti-NF (SigmaAldrich, 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.
Table 2 Patients who died after traumatic injury to the
spinal cord
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
The spinal cords of 19 individuals were examined using
immunohistochemistry
for
collagen
IV,
laminin,
Growth-modulating molecules in human SCI
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 intraparenchymal 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
Brain (2007), 130, 940 ^953
943
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)
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
944
Brain (2007), 130, 940 ^953
A. Buss et al.
Fig. 1 Distribution of collagen IV, laminin, fibronectin, NGFr, NOGO-A, MBP and GFAP in the normal human spinal cord and nerve roots.
Transverse sections of control human spinal cord and nerve roots. (A) Collagen IV immunohistochemistry reveals blood vessels (arrows)
as the only immunoreactive parenchymal structures in the spinal cord. (B) In a dorsal nerve root, collagen staining can be seen in blood
vessel walls (arrow) as well as in the endo- and perineurium. (C) In the spinal cord, laminin staining displays parenchymal and meningeal
blood vessels (arrows) as well as immunoreactivity in the meningeal ECM (asterisk). (D) In a dorsal nerve root, laminin is present in the
endo- and perineurium and in the wall of blood vessels (arrow). (E) In the spinal cord parenchyma, fibronectin is present in blood vessel
walls (arrow). (F) In a dorsal nerve root, fibronectin staining can be seen in blood vessel walls (arrow) as well as in the endo- and
perineurium. (G) In control dorsal nerve roots, NGFr immunoreactivity is restricted to some cell bodies (long arrow) and the peri- and
epineurium (short arrows). (H) In the spinal cord white matter, GFAP staining reveals the honeycomb-like framework of astroglial
processes. (I) NOGO-A immunohistochemistry demonstrates neuronal (short arrow) and oligodendrocytic (long arrows) cell bodies.
DAB immunohistochemistry is not suitable for the detection of NOGO-A in myelin rings. (J) MBP staining is present in compact myelin
of spinal cord white matter and nerve roots. (A^J magnification 320)
Growth-modulating molecules in human SCI
Brain (2007), 130, 940 ^953
945
Fig. 2 The typical morphological appearance of the lesion site in long term human traumatic SCI. Schematic diagrams showing the typical
transverse appearance of the lesion site in the present cases of severe human traumatic SCI at long survival times. (A) The lesion epicentre
is characterized by numerous regenerated root-like structures (short arrows) of different sizes embedded in a dense ECM. Furthermore,
individual spinal nerve roots (long arrows) and the entry zone of a nerve root into the spinal cord can be seen (asterisk). (B) In the intermediate zone, the lesion can often be subdivided into a centrally located cystic region surrounded by an astrocytic scar on one side and an
area with embedded root-like structures (arrows) on the other side. (C) When no cysts are present in the intermediate zone, the lesion is
largely divided into an astrocytic scar and the region with root-like structures including Schwann cells (arrows). (D) In the perilesional area,
the morphology of the intact spinal cord parenchyma appears distorted with a clearly distinguishable but deformed grey/white matter
interface. At the 4 months survival time activated astrocytes can still be seen in the white matter. No root-like structures or non-CNS cell
populations are detectable in the spinal cord. These schematics will be used to provide a broad idea of where, within sections of other
survival times, particular images have been taken.
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 mm, Fig. 3E). Ten and eleven days
after SCI, their size and density were heterogeneous, with
enlarged luminae (up to 250 mm) 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 mm 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
946
Brain (2007), 130, 940 ^953
Fig. 3 The cellular and molecular composition of the scar in
human SCI after early survival times. Transverse sections of the
human spinal cord of patients with survival times of 2^11 days after
SCI. The schematic diagrams in the upper right corner indicate the
region from where the actual picture was taken (black rectangle).
Asterisks represent areas of haemorrhagic infiltration. (A) Two
days after injury, collagen IV staining at the lesion epicentre shows
rare blood vessels with small diameters (up to 50 m). (B) Eleven
days after trauma, small collagen IV-immunopositive blood vessels
are still scarce at the site of injury. (C) Eight days after trauma,
GFAP staining in the intermediate zone shows areas with only
residual astrocytic processes (arrows) without clearly identifiable
cell bodies. (D) GFAP staining in the intermediate zone 11 days
after injury demonstrates nests of activated astrocytes (arrows)
with an irregular network of surrounding processes. (E) In a section from the perilesional area, collagen IV staining reveals many
blood vessels which, at 4 days after injury, are mostly small and
often arranged in cordons. The insert demonstrates an enlarged
blood vessel with a thickened vessel wall 10 days after trauma.
(F) Ten days after trauma, GFAP staining of the perilesional area
reveals large astrocytes (arrows) in between their network of
processes. (A^F magnification 320, inset 320)
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).
A. Buss et al.
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 fibronectinpositive
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).
Growth-modulating molecules in human SCI
Brain (2007), 130, 940 ^953
947
Fig. 4 The cellular and molecular composition of the scar in human SCI after long survival times. Transverse sections of the human spinal
cord of a patient who died 24 days after SCI. The schematic diagrams in the upper right corner indicate the region from where the actual
picture was taken (black rectangle). Asterisks represent areas of haemorrhagic infiltration. (A) NGFr immunohistochemistry in a section
from the lesion site demonstrates tangentially sectioned Schwann cell processes invading the spinal cord parenchyma (short arrows). Long
arrows indicate the meningeal surface with the parenchyma on the right. (B) Laminin staining shows immunopositive cells and processes
infiltrating the spinal cord at the dorsal root entry zone (arrow). The inset demonstrates an immunoreactive bipolar cell at high magnification. (C) In an adjacent section, collagen IV immunohistochemistry reveals a very similar staining pattern with cells and processes invading the spinal cord regions close to the nerve root entry zone. The inset demonstrates immunoreactive blood vessels at high magnification.
(D) In the corresponding area on a subsequent section, fibronectin positive fibrils can be seen in a very disordered pattern in the extracellular space between rounded macrophages. (E) In the intermediate zone, distant from the region of infiltrating Schwann cells, collagen IV
immunoreactivity is restricted to blood vessel walls. (F) In the corresponding area on a subsequent section, GFAP staining reveals a diffuse,
irregular network of astrocytic processes without identifiable cell bodies. (A^F magnification 320, inserts 400)
Fig. 5 The cellular and molecular composition of the scar in human SCI after long survival times II. Double immunofluorescence with
antibodies against laminin (red) and NF (green) on transverse sections of the human spinal cord of a patient who died 24 days after SCI.
The schematic diagrams in the upper right corner indicate the region from where the actual picture was taken (black rectangle). (A) In the
lesion epicentre, close to a dorsal nerve root, laminin immunoreactive processes can be seen (long arrows). NF immunoreactivity demonstrates rounded, probably degenerating axonal profiles (short arrows), none of which appear to be associated with the laminin positive
processes. (B) In the same section, an area with densely packed and orientated laminin-positive Schwann cell processes reveals scarce,
small diameter NF positive fibres (arrow). (C) In more central areas of the lesion core, laminin immunoreactivity is restricted to blood
vessels with a sometimes irregular morphology with a thickened vessel wall (arrow). Other scattered NF-positive profiles appear to show
no association with the laminin-positive structures and (as in Fig. 4A) may reflect degenerated but still immunoreactive fragments of axons
which are yet to be cleared by invading macrophages. (A^C magnification 400)
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
948
Brain (2007), 130, 940 ^953
A. Buss et al.
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
Fig. 6 The cellular and molecular composition of the scar in
human SCI after long survival times III. Transverse sections of the
human spinal cord of patients with a survival time of 20 years after
SCI. The schematic diagrams in the upper right corner indicate the
region from where the actual picture was taken (black rectangle).
(A) In the ECM in and around the neuromas, NGFr positive
Schwann cells are present in varying numbers. In the present area
a nest of these cell processes can be seen. (B ^C) Collagen IV (B)
and laminin (C) staining demonstrates sheet-like bundles of fibrils
in the Schwann cell dominated ECM. (D) The distribution of fibronectin positive fibres in this area is more diffuse with an irregular
and dense pattern of lamellae. (E) In the neuromas themselves,
NGFr immunoreactivity shows cell bodies which are irregularly
distributed. (F^G) Collagen IV (F) and laminin (G) staining reveals
these molecules in the endo- and perineurium and blood vessel
walls of the regenerated root-like structures. (H) Fibronectin
immunohistochemistry is present in the endo- and perineurium as
well as blood vessel walls. (A^H magnification 320)
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
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
Growth-modulating molecules in human SCI
Brain (2007), 130, 940 ^953
949
Fig. 7 The cellular and molecular composition of the scar in human SCI after long survival times IV. Transverse sections of the human
spinal cord of patients with a survival time of 20 years after SCI. The schematic diagrams in the upper right corner indicate the region from
where the actual picture was taken (black rectangle). (A) Triple immunofluorescence of NF (green), MBP (red) and P0 (blue) at the lesion
epicentre reveals multiple myelinated nerve fibres in the neuromas with PNS-like myelin sheaths. (B) Staining for MBP (green) and collagen
IV (red) demonstrates that the myelinated nerve fibres in the neuromas are surrounded by a collagen IV-positive basal lamina. (C) Double
immunofluorescence of NF (red) and laminin (green) shows that the basal lamina surrounding axons also contains laminin. (D) Staining for
NF (red) and fibronectin (green) demonstrates that beside some single axons (short arrows) the latter glycoprotein tightly surrounds
bundles of nerve fibres in the neuromas (long arrows). (E) At higher magnification, the fibronectin-positive ECM (green) around single NFpositive axons (red) can be seen. (F) Staining for CGRP (red) and P0 (green) reveals a subpopulation of CGRP-positive myelinated axons in
the neuromata. (G) Triple immunofluorescence of NF (blue), collagen IV (red) and GFAP (green) shows the strict separation of the Schwann
cell dominated lesion area (containing neuromas and collagen IV rich ECM) and the astroglial scar (with scarce nerve fibres and collagen IV
restricted to blood vessels). (H) Triple immunofluorescence of NF (blue), MBP (red) and GFAP (green) reveals the strict separation of the
Schwann cell containing neuromas from the astroglial scar. (I) At higher magnification, single MBP-positive myelinated nerve fibres can be
seen in the GFAP-positive astroglial scar. (J) Triple immunofluorescence of P0 (blue), MBP (red) and GFAP (green) reveals the PNS-like
myelin sheaths in the Schwann cell area with a complete overlap of P0 and MBP. In contrast, in the astrocytic scar, only MBP immunoreactivity can be seen and no P0 staining. (K) Triple immunofluorescence of NF (blue), collagen IV (red) and GFAP (green) demonstrates
individual regenerating nerve fibres in the astrocytic scar with no co-localization with collagen IV which is restricted to blood vessel walls.
(L) Staining for NF (green) and NOGO-A (red) demonstrates some thin nerve fibres in the astrocytic scar area in close proximity to
NOGO-A positive oligodendrocytes (arrows). However, the axons are not surrounded by NOGO-A immunoreactive myelin sheaths.
(A, C, E, F, I, K^L magnification 400; B, D, G, H, J magnification 200)
950
Brain (2007), 130, 940 ^953
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
A. Buss et al.
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 growthrepulsive 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 growthpromoting (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
Growth-modulating molecules in human SCI
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 endoand 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
Brain (2007), 130, 940 ^953
951
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 growthpromoting 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 celldominated 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
952
Brain (2007), 130, 940 ^953
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
Beattie MS, Bresnahan JC, Komon J, et al. Endogenous repair after spinal
cord contusion injuries in the rat. Exp Neurol 1997a; 148: 453–63.
Blakemore WF. 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 1992; 18: 460–6.
A. Buss et al.
Brook GA, Houweling DA, Gieling RG, et al. Attempted endogenous tissue
repair following experimental spinal cord injury in the rat: involvement
of cell adhesion molecules L1 and NCAM? Eur J Neurosci 2000; 12:
3224–38.
Brook GA, Plate D, Franzen R, et al. 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 1998; 53: 51–65.
Bruce JH, Norenberg MD, Kraydieh S, Puckett W, Marcillo A, Dietrich D.
Schwannosis: role of gliosis and proteoglycan in human spinal cord
injury. J Neurotrauma 2000; 17: 781–8.
Bunge RP, Puckett WR, Hiester ED. Observations on the pathology of
several types of human spinal cord injury, with emphasis on the
astrocyte response to penetrating injuries. Adv Neurol 1997; 72: 305–15.
Buss A, Brook GA, Kakulas B, et al. Gradual loss of myelin and formation
of an astrocytic scar during Wallerian degeneration in the human spinal
cord. Brain 2004; 127: 34–44.
Buss A, Sellhaus B, Wolmsley A, Noth J, Schwab ME, Brook GA.
Expression pattern of NOGO-A protein in the human nervous system.
Acta Neuropathol (Berl) 2005; 110: 113–9.
Chang A, Nishiyama A, Peterson J, Prineas J, Trapp BD. NG2-positive
oligodendrocyte progenitor cells in adult human brain and multiple
sclerosis lesions. J Neurosci 2000; 20: 6404–12.
Condic ML, Lemons ML. Extracellular matrix in spinal cord regeneration:
getting beyond attraction and inhibition. Neuroreport 2002; 13: 37–48.
Dowling P, Ming X, Raval S, et al. Up-regulated p75NTR neurotrophin
receptor on glial cells in MS plaques. Neurology 1999; 53: 1676–82.
Fawcett JW, Asher RA. The glial scar and central nervous system repair.
Brain Res Bull 1999; 49: 377–91.
Grimpe B, Silver J. The extracellular matrix in axon regeneration. Prog
Brain Res 2002; 137: 333–49.
Guest JD, Hiester ED, Bunge RP. Demyelination and Schwann cell
responses adjacent to injury epicenter cavities following chronic human
spinal cord injury. Exp Neurol 2005; 192: 384–93.
Hermanns S, Klapka N, Muller HW. The collagenous lesion scar—an
obstacle for axonal regeneration in brain and spinal cord injury. Restor
Neurol Neurosci 2001; 19: 139–48.
Hill RE, Williams RE. A quantitative analysis of perineurial cell basement
membrane collagen IV, laminin and fibronectin in diabetic and
non-diabetic human sural nerve. J Anat 2002; 201: 185–92.
Hughes JT. Pathology of the spinal cord. Major Probl Pathol 1978; 6:
1–257.
Hughes JT, Brownell B. Aberrant nerve fibres within the spinal cord.
J Neurol Neurosurg Psychiatry 1963; 26: 528–34.
Iseda T, Nishio T, Kawaguchi S, Kawasaki T, Wakisaka S. 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 2003;
464: 343–55.
Johnson EM Jr, Taniuchi M, DiStefano PS. Expression and possible
function of nerve growth factor receptors on Schwann cells. Trends
Neurosci 1988; 11: 299–304.
Joosten EA, Dijkstra S, Brook GA, Veldman H, Bar PR. Collagen IV
deposits do not prevent regrowing axons from penetrating the lesion site
in spinal cord injury. J Neurosci Res 2000; 62: 686–91.
Kakulas BA. Pathology of spinal injuries. Cent Nerv Syst Trauma 1984; 1:
117–29.
King VR, Phillips JB, Hunt-Grubbe H, Brown R, Priestley JV.
Characterization of non-neuronal elements within fibronectin mats
implanted into the damaged adult rat spinal cord. Biomaterials 2006; 27:
485–96.
Klapka N, Hermanns S, Straten G, et al. 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 2005; 22: 3047–58.
Kleinman HK, McGarvey ML, Liotta LA, Robey PG, Tryggvason K,
Martin GR. Isolation and characterization of type IV procollagen,
Growth-modulating molecules in human SCI
laminin, and heparan sulfate proteoglycan from the EHS sarcoma.
Biochemistry 1982; 21: 6188–93.
Lefcort F, Venstrom K, McDonald JA, Reichardt LF. Regulation of expression
of fibronectin and its receptor, alpha 5 beta 1, during development and
regeneration of peripheral nerve. Development 1992; 116: 767–82.
McDonald JW, Sadowsky C. Spinal-cord injury. Lancet 2002; 359: 417–25.
Oertle T, van der Haar ME, Bandtlow CE, et al. Nogo-A inhibits neurite
outgrowth and cell spreading with three discrete regions. J Neurosci
2003; 23: 5393–406.
Petratos S, Gonzales MF, Azari MF, et al. Expression of the low-affinity
neurotrophin receptor, p75(NTR), is upregulated by oligodendroglial
progenitors adjacent to the subventricular zone in response to
demyelination. Glia 2004; 48: 64–75.
Schwab ME, Bartholdi D. Degeneration and regeneration of axons in the
lesioned spinal cord. Physiol Rev 1996; 76: 319–70.
Shiga T, Oppenheim RW. Immunolocalization studies of putative
guidance molecules used by axons and growth cones of intersegemental
interneurons in the chick embryo spinal cord. J Comp Neurol 1991;
310: 234–52.
Silver J, Miller JH. Regeneration beyond the glial scar. Nat Rev Neurosci
2004; 5: 146–56.
Brain (2007), 130, 940 ^953
953
Stichel CC, Hermanns S, Luhmann HJ, et al. Inhibition of collagen IV
deposition promotes regeneration of injured CNS axons. Eur J Neurosci
1999; 11: 632–46.
Timpl R, Rohde H, Robey PG, Rennard SI, Foidart JM, Martin GR.
Laminin—a glycoprotein from basement membranes. J Biol Chem 1979;
254: 9933–7.
Tom VJ, Doller CM, Malouf AT, Silver J. Astrocyte-associated fibronectin
is critical for axonal regeneration in adult white matter. J Neurosci 2004;
24: 9282–90.
Venstrom KA, Reichardt LF. Extracellular matrix. 2: role of extracellular
matrix molecules and their receptors in the nervous system. FASEB J
1993; 7: 996–1003.
Wang ZH, Walter GF, Gerhard L. 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) 1996; 91: 180–4.
Weidner N, Grill RJ, Tuszynski MH. Elimination of basal lamina and the
collagen ‘‘scar’’ after spinal cord injury fails to augment corticospinal
tract regeneration. Exp Neurol 1999; 160: 40–50.
Wolman L. The disturbance of circulation in traumatic paraplegia in acute
and late stages: a pathological study. Paraplegia 1965; 59: 213–26.