Samples of spinal cord from all 19 cases were examined using immunohistochemistry for MMP-1, -2, -9 and -12, TIMP-1, -2 and -3 as well as CD68 and GFAP. The brains of all cases were carefully examined and were declared to be without pathological findings. The spinal cords of the control cases were also morphologically intact. For an overview of the results, especially the number of immunoreactive cells at the lesion epicentre at the various survival times, see table . The interpretation of the semi-quantitative presentation of MMP and TIMP expression is restricted, due to the limited comparability of staining quality (largely due to differences in the time between death and the dissection/fixation of tissue). However, immunohistochemistry was performed using the identical staining procedure and cases with closely matching survival times (e.g. 4–5 days and 10–11 days after SCI) demonstrated very similar amounts of MMP and TIMP immunoreactivity.
Amount of immunopositive cells at the lesion epicentre at different survival times following human SCI
Normal distribution of MMPs and TIMPs in the human spinal cord
The spinal cord parenchyma including the meninges was immunonegative for MMP-2 and -12 and TIMP-1 (not shown). Immunohistochemistry for MMP-1 revealed staining in most motoneurons and Clarke's nucleus neurons as well as scattered interneurons, mostly in laminae IV to VI (Fig.). Staining for MMP-9 demonstrated rare intravascular monocytes in both meningeal and parenchymal blood vessels (Fig.). TIMP-2 immunoreactivity was present in most motoneurons and individual Clarke's nucleus neurons and interneurons from laminae I to VI (Fig.). TIMP-3 revealed a similar distribution but with only approximately 50% of motoneurons stained in cervical, thoracic and lumbar segments (Fig.). This distribution was detected in all 4 control cases, and no evidence for age-related changes in the distribution of MMP and TIMP isoforms could be found in any of the normal human CNS samples.
Figure 1 Distribution of MMPs and TIMPs in the normal human spinal cord. Images were taken from control human spinal cords in transverse sections. A: Immunohistochemistry for MMP-1 revealed cytoplasmic staining in motoneurons; the nucleus was devoid of immunoreactivity. (more ...)
Macrophage/microglial responses to human SCI
In the normal unlesioned spinal cord, CD68 immunoreactivity was scarce. Occasional bipolar perivascular cells (e.g. arrowheads, Fig.) and intravascular monocytes were immunopositive. Although sections from the lesion site demonstrated no preservation of cytoarchitecture at 2 days after SCI, there was a slight increase in the incidence of CD68-positive cells. These cells were round to oval-shaped and were spread over the lesion core and the peri-lesional area (Fig.). By 4 and 5 days after injury, the number of immunopositive cells at the lesion epicentre was further increased. Microglia/macrophages could be found in the area of necrosis, either as individual cells or as small clusters (Fig.). In peri-lesional gray matter areas, reactive CD68-positive microglia/macrophages were sometimes found in close apposition to neuronal cell bodies (Fig.). By 8, 10, 11 and 24 days post injury, the lesion core contained numerous microglia/macrophages, mostly with the morphology of foamy macrophages (Fig. and ) but by 4 months, fewer CD68 immunoreactive cells could be detected at the lesion site (Fig.). Sections from tissue blocks of all later survival times demonstrated a few remaining immunoreactive cells with a peri-vascular distribution, similar to that seen in the control cases (not shown).
Figure 2 Macrophages/microglia in the human spinal cord after traumatic injury. Images were taken from control spinal cords and cases with different survival times after SCI in transverse sections stained with the CD68 antibody. A: In a control spinal cord, CD68 (more ...)
Astrocytic reaction to human SCI
GFAP immunohistochemistry in the normal, unlesioned spinal cord revealed astrocytic cell bodies evenly dispersed throughout spinal cord gray and white matter, scattered between the network of astrocytic processes (Fig.). At early survival times, spanning 2 to 11 days after trauma, the areas of massive tissue destruction demonstrated a massive reduction of GFAP immunoreactivity. Hardly any astroglial cell bodies and an irregular, loose arrangement of processes were detectable (Fig.). In cases with survival times of 4 days and longer, perilesional, highly GFAP-positive, activated astrocytes could be seen. These cells were distributed homogeneously over white and gray matter in sections up to 1 to 2 segments proximal and distal to the lesion site (Fig.). By eleven days, nests of large, intensely GFAP-positive astrocytes surrounded by a dense irregular network of processes could be observed (Fig.). The formation of the glial scar progressed, and in cases with survival times of 4 months and longer, a dense astroglial, GFAP-positive, scar was visible in which the cell bodies were hardly detectable (Fig.). In the peri-lesional areas up to 1 segment away from the lesion site, activated astroglia could still be detected up to 1 year after trauma (Fig.).
Figure 3 Astrocytes in the human spinal cord after traumatic injury. Images were taken from control spinal cords and cases with different survival times after SCI in transverse sections stained with the GFAP antibody. A: GFAP immunohistochemistry in the unlesioned (more ...)
Expression of MMP-1 after human SCI
At all survival times, the neuronal distribution of MMP-1 in the peri-lesional area remained unchanged (Fig. and ). At 2 days, some round to oval cells were seen at the site of injury. They were detectable in the heavily destroyed tissue but were present at a lower density than the CD68- positive macrophages/microglia. Four days after trauma, immunoreactive cells, morphologically resembling macrophages, could be seen in areas of bleeding and massive tissue destruction (Fig. ) and by 8 days, the lesion core contained substantial numbers of immunoreactive cells, corresponding in size, distribution and morphology to the CD68 (and MMP-9, see below)-positive cells observed in near adjacent sections (Fig. and ). By 10–24 days, the incidence of MMP-1 positive cells was dramatically decreased and only single round cells were detectable (Fig.). However, sections taken from the peri-lesional area between 4 months and 1 year post injury revealed large, MMP-1 positive cell bodies which were subsequently demonstrated to be astrocytes by double immunofluorescence (Fig. and ).
Figure 4 MMP-1 in the human spinal cord after traumatic injury. Images were taken from cases with different survival times after SCI in transverse sections stained with the MMP-1 antibody. A: Two days after injury, motoneurons in the perilesional area retained (more ...)
Figure 7 MMP-1, -2, -9 and -12 and TIMP-3 in macrophages and astrocytes after traumatic human spinal cord injury. Images were taken from cases with different survival times after SCI in transverse sections.A: 8 days after injury. Double immunofluorescence for (more ...)
Expression of MMP-2 after human SCI
Immunohistochemistry for MMP-2 demonstrated the first positive cells in sections from the lesion epicentre at 2 days after injury. In areas of massive destruction and hemorrhagic infiltration, a few positive round cells were detected whereas areas of the lesion site without bleeding demonstrated a higher number of immunoreactive cells with a round to oval morphology (Fig.). By 8 days, the morphology and distribution of the cells within the lesion site corresponded to those of CD68-positive macrophages (Fig.). However, by 10 days the number of MMP-2 positive macrophages was clearly reduced (Fig.) and by 24 days, only single immunoreactive cells could be detected. At all later survival times, no specific staining was visible, similar to control cases (not shown).
Figure 5 MMP-2, -9 and -12 in the human spinal cord after traumatic injury. Images were taken from cases with different survival times after SCI in transverse sections. A: Immunohistochemistry for MMP-2 two days after injury. A few positive round cells (arrows) (more ...)
Expression of MMP-9 after human SCI
By 2 days after injury (the earliest time point investigated) staining for MMP-9 demonstrated immunoreactive cells at and around the lesion site. In areas of severe tissue destruction and intensive bleeding, only few cells with a round morphology were stained (Fig.). At 4–5 days after injury, in areas at the lesion epicentre but without signs of hemorrhagic infiltration, the amount of immunoreactive cells was higher compared to other heavily destroyed areas. The cells were mostly of a round morphology and were relatively evenly distributed except for clusters of positive cells in and around blood vessels (Fig.). Double immunofluorescence with an anti-CD68 antibody demonstrated a microglia/macrophage origin of most MMP-9 positive cells (Fig.). Morphologically, the MMP-9 positive/CD68-negative cells at these early survival times resembled infiltrating neutrophils. In sections further away from the lesion, the number of immunoreactive cells decreased rapidly; such that hardly any MMP-9 positive cell could be detected one segment distal or proximal of the site of injury (Fig.). This distribution remained more or less constant up to 11 day post injury. By 24 days, large numbers of MMP-9 positive round cells were visible at the lesion core and filled the area of tissue destruction (Fig.). Double immunofluorescence revealed these cells to be phagocytosing macrophages (Fig.). At survival times of 4 months and later, only single MMP-9 immunoreactive cells were detectable at and around the lesion site, showing a distribution similar to that of control cases (not shown). Apart from microglia/macrophages and some neutrophils, no other cell population demonstrated immunoreactivity for MMP-9 at all time points (not shown).
Expression of MMP-12 after human SCI
In contrast to the other MMPs investigated, immunohistochemistry for MMP-12 only displayed rare immunopositive, rounded cells at and around the lesion site up to 11 days post injury (Fig.). This distribution changed dramatically by 24 days, when many intensely immunoreactive microglia/macrophages could be detected at the lesion core (Fig. and ). The number of stained cells decreased rapidly in sections further away from the lesion such that by 1 segment proximal and distal no immunoreactive cells were visible. In all cases with longer survival times, no specific MMP-12 staining was seen at and around the lesion site (not shown).
Expression of TIMP-1, -2 and -3 after human SCI
Only rare TIMP 1, 2 or 3 immunoreactive cells could be demonstrated at the lesion site up to 24 days post injury (Fig. and ). These cells were rounded and were usually seen in the vicinity of blood vessels. By 4 months and longer, no specific TIMP-1 staining could be detected at or around the lesion site (not shown).
Figure 6 TIMP-1, -2 and -3 in the human spinal cord after traumatic injury. Images were taken from cases with different survival times after SCI in transverse sections. A: In a section from patients who died 2 days after SCI, immunohistochemistry for TIMP-1 demonstrated (more ...)
Immunohistochemistry for TIMP-2 demonstrated a temporal loss of neuronal staining. Immunoreactivity could be detected in most motoneurons and some interneurons for up to 4 days (Fig.). However, at survival times ranging from 8 to 24 days only single immunoreactive neurons were visible. This was particularly evident for TIMP-2 staining in motoneurons close to the lesion site (Fig.). From 4 months, the remaining neurons displayed a staining pattern comparable to the control cases with most distinguishable motoneurons being TIMP-2 positive and some positive interneurons in laminae I to VI (not shown). Apart from intermittent immunopositive macrophages and stained neurons, no other cell population demonstrated TIMP-2 immunoreactivity. In contrast to TIMP-2, the neuronal staining pattern of TIMP-3 remained unchanged over the range of survival times. About half of the motoneurons were stained and single interneurons could always be seen at and around the site of injury (Fig. and ). Activated astrocytes became transiently immunopositive for TIMP-3 between 8 months to 1 year after SCI (Fig. and ). Survival times of greater than 1 year only revealed neuronal TIMP-3 immunoreactivity (not shown).