Repulsive guidance molecule (RGM) antibodies. A, Domain of human RGM-A (GenBank accession numbers NP_064596 and NM_020211) detected by the RGM276-293 peptide-specific RGM antibody (gray-shaded letters). Numbers on the right indicate amino acids. B, Western blot analysis performed with RGM-A–transfected cell membranes. Unspecific staining with the secondary antibody was used for immunodetection (control). Incubation with anti-RGM–A antibody reveals the expected 43-kDa band (mouse RGM-A). Selected apparent molecular weights are indicated on the left. The arrow indicates the interface between the stacking and running gel.
Repulsive guidance molecule (RGM) immunolocalization after central nervous system (CNS) injury. The specificity of the antibodies against RGM was confirmed by the inhibition of RGM immunoreactivity after preincubation for 3 hours on ice with the cognate peptide RGM276-293 (A) but not with the control peptide (B). C, In control brains, RGM immunoreactivity was detected on white matter–like fibers (similar to that obtained by myelin basic protein staining [data not shown]), oligodendrocytes, and the perikarya of some neurons. D, The RGM-immunopositive cells were also detected in the choroid plexus. E, Ependyma and frequently Purkinje cells of the cerebellum expressed RGM. After CNS injury of either ischemic or traumatic etiology, cellular RGM immunoreactivity revealed similar expression patterns. F, After CNS injury, RGM-immunopositive cells accumulated in infarctional white matter, hemorrhagic areas, the infarction core, and peri-infarctional areas. Arrow indicates infarction core. Early after ischemic damage (up to 2.5 days), RGM immunoreactivity was predominantly observed with (G) neurons and leukocytes of granulocytic, monocytic, and lymphocytic origin in vessels and infiltrative areas in ischemic tissue (H-J). Furthermore, in a widespread pattern, small and middle-sized vessels revealed RGM immunoreactivity confined to endothelial cells (H) and some fibromyocytes. I, Paralleled by edema formation, up to 1 to 7 days, RGM-immunopositive cells were found extravasating outside the vascular walls into the focal ischemic lesional parenchyma. Arrows indicate perpendicular invading routes. J, In perivascular regions, RGM-immunopositive cells formed clusters in Virchow-Robin–like spaces from days 1 to 7, which subsided later. K, At later stages, arising during the end of the first week after infarction, we detected RGM-immunopositive cells with fibroblast-like Morphologic features (arrow). L, Excessive RGM-immunopositive formations, constituting neo-laminae, were localized to areas of ongoing scar formation. M, With tissue reorganization of the lesion, “foamy,” lipid-laden, phagocytic RGM-immunopositive macrophages/lysosomal ingestions (arrow) were also observed. N, Double-labeling experiments identified RGM-immunopositive cellular structures localized to myelin basic protein–immunopositive myelin fibers and myelin basic protein–immunopositive oligodendrocytes. O, Most CD45+ leukocytes expressing RGM (70%) consist of CD68+ microglia/macrophages. With maturation of the lesion, we observed RGM co-expression by cellular and extracellular components of the developing scar. P, Abundant extracellular RGM-immunopositive laminae scar components were verified to be fibronectin immunopositive. Q and R, Oligodendroglial RGM expression was confirmed by additional 3′-cyclic nucleotide 3′-phosphodiesterase double labeling. Arrows indicate RGM-immunopositive cells and diaminobenzidine; arrowheads, 3′-cyclic nucleotide 3′-phosphodiesterase and fast blue (Q and R).
Cumulative repulsive guidance molecule (RGM) expression in brains with focal cerebral ischemia (FCI). A, Regional analysis demonstrated significant accumulation of RGM-immunopositive cells in peri-infarctional border zones (P<.001), whereas in remote regions and in control brains, rare RGM-immunopositive cells were labeled. B, After FCI, quantitative assessment revealed an early, significant increase in RGM-immunopositive cell numbers during the first 24 hours followed by maximum RGM-immunopositive cell density from day 1.5 to day 2.5 after FCI. Subsequently, the numbers of lesional RGM-immunopositive cells, although decreasing, remained enhanced compared with controls for up to months after FCI. Control brains were used as the equivalent to 0 hours after FCI. The y-axis refers to total RGM-immunopositive cell counts (means) per high-power field. Error bars represent SEM.
Mean cumulative repulsive guidance molecule (RGM) expression in brains with traumatic brain injury (TBI). A, Regional analysis demonstrated significant lesional accumulation of RGM-immunopositive cells in injured areas (P<.001), whereas in remote white matter regions and in control brains, only a few cells were labeled. B, The temporal RGM expression pattern demonstrated a cellular up-regulation already 1 hour after injury (P<.001), persisting at elevated levels for months after injury. Control brains were used as the equivalent to 0 hours after trauma. The y-axis refers to total RGM-immunopositive cell counts (means) per high-power field. Error bars represent SEM.
Schwab JM, Monnier PP, Schluesener HJ, Conrad S, Beschorner R, Chen L, Meyermann R, Mueller BK. Central Nervous System Injury–Induced Repulsive Guidance Molecule Expression in the Adult Human Brain. Arch Neurol. 2005;62(10):1561-1568. doi:10.1001/archneur.62.10.1561
The repulsive guidance molecule (RGM) is involved in formation of the central nervous system during development by moderating the repulsion of growing axons. However, the role of RGM in adult central nervous system lesions remains to be clarified.
To identify and determine RGM expression in adult brains with focal cerebral ischemia or traumatic brain injury and in neuropathologically unaffected control brains.
Twenty-one brains of patients with focal cerebral ischemia, 25 brains after traumatic brain injury, and 4 control brains.
Main Outcome Measure
Expression of RGM as assessed by immunohistochemical analysis.
In normal brains, RGM expression was detected on the perikarya of some neurons, choroid plexus, smooth muscle and endothelial cells, oligodendrocytes, and myelinated white matter fibers. After focal cerebral ischemia and traumatic brain injury, RGM-immunopositive cells accumulated in lesional and perilesional areas. In hemorrhagic lesions, a massive accumulation of RGM-immunopositive cells was observed. During the first week after insult, RGM expression remained confined to neurons, smooth muscle and endothelial cells, and leukocytes infiltrating the lesion. Thereafter, with maturation of the lesion, we observed RGM expression by components of the developing scar tissue, such as fibroblastoid cells, reactive astrocytes, and a pronounced extracellular RGM deposition resembling neo-laminae.
This is the first study of RGM in the human central nervous system. Following central nervous system injury, RGM, a novel, potent axonal growth inhibitor, is present in axonal growth impediments: the mature myelin, choroid plexus, and components of the developing scar.
Axonal regeneration in the adult nervous system is a suggested goal to improve clinical outcomes in a range of neurologic conditions after traumatic brain injury (TBI) and focal cerebral ischemia (FCI). Inhibitors present in central nervous system (CNS) myelin and the glial scar are considered the main obstacles to axonal regeneration.1- 3 Several inhibitory molecules in the adult myelin and scar were identified, including among others the myelin-associated glycoprotein (MAG),4 neurite outgrowth inhibitor A (Nogo-A),5- 7 oligodendrocyte-myelin glycoprotein (Omgp)/Arretin,8,9 and chondroitin sulfate proteoglycans (CSPGs).1,8 These repulsive factors are thought to act primarily on the growth cone at the distal tip of the growing axon,9 and it is assumed that growth cone collapse is the mechanism underlying the action of repulsive molecules on growing axons.10 Thus, an understanding of the distribution of inhibitory proteins after CNS injury is crucial to any consideration of their pathological roles in limiting axonal regeneration.
The repulsive guidance molecule (RGM) was recently reported as the long sought-after 33-kDa tectum repellent first characterized by the Bonhoeffer group.11 This membrane-associated glycoprotein shares no sequence homology with any other known guidance cues. Recombinant RGM induces the collapse of temporal but not nasal growth cones and guides temporal retinal axons in vitro,12 demonstrating its repulsive and axon-specific guiding activity. Herein, we describe the in vivo RGM expression pattern in normal and injured human brains.
Antibodies were raised against chick RGM276-293 (amino acid sequence RMPEEVVNAVEDRDSQGL), revealing greater than 94% homology with the corresponding human sequence (Figure 1A). The RGM antibodies were purified using a column loaded with the cognate peptide as described elsewhere.12 Peptide-specific polyclonal antibodies directed against RGM were diluted (1:10) in Tris balanced salt solution (TBS) (containing 0.025M Tris and 0.15M sodium chloride) in 1% bovine serum albumin (BSA) and incubated overnight at room temperature.
The RGM complementary DNA sequence has been deposited in GenBank (accession No. AY128507) at the National Center for Biotechnology Information, National Institutes of Health (Bethesda, Md).12 The human sequence of RGM is encoded on chromosome 15 and is accessible under GenBank accession No. AL136826. High homology of the detected sequence between chick and human was confirmed by the fact that binding of RGM antibodies to human tissue epitopes was blocked by preincubating with the cognate chick-RGM peptide visualized by abolished RGM immunoreactivity (Figure 2A and B).
Membrane suspensions were prepared from HEK293 cells transfected using an RGM-A–expressing plasmid. Two days after transfection, cells were homogenized in homogenization buffer (10mM Tris hydrochloride, pH 7.4, 1.5mM calcium chloride, and 1mM spermidine) by pressing the cells through 27-gauge syringe needles. The homogenate was layered on the top of a step gradient of 50% and 5% sucrose and was centrifuged for 10 minutes at 28 000g. The membrane fragments were collected from the interphase layer situated between 5% and 50% sucrose and were washed with phosphate-buffered saline solution. Membrane suspensions were adjusted to an optical density of 0.1 (measured at 220 nm). Sodium dodecyl sulfate gel electrophoresis and immunoblotting were performed according to standard protocols (Figure 1B).
Twenty-one brains of patients with a clinical history and a neuropathologically confirmed diagnosis of FCI and 25 brains of patients with TBI were included in this study. Infarctional and traumatically injured brain tissue samples were sustained in an updated stroke and trauma brain bank reported previously.13,14 Tissue specimen procurement was performed according to the ethical guidelines of the University of Tuebingen. Patients with altered immune status because of immunosuppressive therapy, meningitis, or encephalitis were excluded from this study. The results were compared with those in tissue from corresponding areas of 4 normal brains previously described elsewhere (Table 1>, Table 2>, and Table 3>).13,14 In addition to patient data, hematoxylin-eosin, Luxol fast blue, and iron staining were used to evaluate the typical histologic features defined as the standard indication of infarct15 and trauma age.16
After formaldehyde fixation and paraffin embedding, rehydrated 2-μm sections were boiled (in a 600-W microwave oven) 7 times for 5 minutes in citrate buffer (2.1 g of sodium citrate per liter, pH 6.0). Endogenous peroxidase was inhibited using 2.5% hydrogen peroxide in methanol (15 minutes). Sections were incubated in 10% normal porcine serum diluted in 0.1% TBS-BSA (Biochrom, Berlin, Germany) to block nonspecific binding of immunoglobulins. Monospecific polyclonal antibodies directed against RGM were diluted (1:10) in 0.1% TBS-BSA and incubated overnight at 4°C. Specific binding of the antibodies was detected using a secondary biotinylated swine anti–rabbit IgG F(ab)2 antibody fragment (1:400) for 30 minutes (DAKO, Hamburg, Germany), followed by incubation in a peroxidase-conjugated streptavidin-biotin complex (DAKO). The enzyme was visualized using diaminobenzidine as a chromogen (Fluka, Neu-Ulm, Germany). Sections were counterstained with Mayer hemalum (1:2 in distilled water). Negative controls consisted of sections incubated in the absence of the primary antibody. The specificity of the polyclonal RGM antibody was confirmed by inhibition of staining using human ischemic brain tissue after preincubation for 3 hours on ice by excess of the cognate RGM peptide (Figure 2A and B).
In double-labeling experiments, we first labeled a cell-type or activation-specific antigen using the ABC procedure in combination with alkaline phosphatase conjugates. Specific antigens were labeled with antibodies against monoclonal glial fibrillary acidic protein (Boehringer, Mannheim, Germany) (1:100) to detect astrocytes, polyclonal myelin basic protein (DAKO, Glostrup, Denmark) (1:500) and monoclonal 3′-cyclic nucleotide 3-phosphodiesterase (Chemicon International Inc, Temecula, Calif) (1:100) to detect oligodendrocytes, and, with CD68 (PG-M1; DAKO) (1:500) for microglia/macrophage identification. Activated microglia/macrophages were detected by using antibodies directed against HLA-DR, HLA-DP, and HLA-DQ (major histocompatibility complex class II) (DAKO) (1:100). Lymphocytic subpopulations were classified by using monoclonal antibodies against CD4 (T-helper lymphocytes) (1:10), CD8 (T-cytotoxic and T-supressor lymphocytes) (1:500), and CD20 (pan–B-cell marker) (1:200) (all from DAKO). For further characterization of the reactive astrocytic state, we used monoclonal antibodies against nestin (Pharmingen, Heidelberg, Germany) (1:100) and vimentin (DAKO) (1:15). To detect extracellular basal lamina structures in vessels and during scar formation, we used mouse laminin (Chemicon) antibodies (1:500), and rabbit fibronectin (DAKO) antibodies (1:100) were used to detect matrix deposition. Neuronal structures were identified by monoclonal antibodies–detecting neurofilament (NF; DAKO, 1:100) and neuron-specific enolase (NSE; Chemicon International Inc, 1:10). Briefly, slices were deparaffinized, irradiated in a microwave oven for antigen retrieval, and incubated in nonspecific porcine serum as described previously herein. Visualization was achieved by adding biotinylated secondary antibodies (1:400) for 30 minutes and alkaline phosphatase–conjugated ABC complex diluted in a ratio of 1:400 in TBS-BSA for 30 minutes. Consecutively, we developed with fast blue BB salt chromogen substrate solution, yielding a blue reaction product. Between double-labeling experiments, slices were irradiated in a microwave oven for 5 minutes in citrate buffer. Then RGM was immunolabeled as described previously herein.
Data were calculated as mean ± SEM of labeled cells from border zones or remote areas of the same tissue section and were compared with normal control brains using the 2-tailed unpaired t test. Border zones were defined as perilesional areas adjacent to the developing necrotic core that demarcate the region of major damage. The RGM-immunopositive cells were counted in 10 high-power fields (×200 magnification using an eyepiece grid representing 0.25 mm2).
Twenty-one brains of patients with FCI and 25 brains after TBI were evaluated for RGM expression by immunohistochemical analysis and were compared with 4 control brains.
To ensure antibody specificity, Western blotting was performed using HEK293 cell membrane preparations transfected using RGM-A. As expected, a specific 43-kDa band appears in blots developed using the RGM antibody (Figure 1B).
In control brains without neuropathologic alterations, RGM immunoreactivity was detected on white matter fibers (Figure 2C), oligodendrocytes, and the perikarya of some neurons. The RGM-immunopositive cells were also detected in the choroid plexus (Figure 2D), ependyma, and, frequently, Purkinje cells of the cerebellum (Figure 2E). Only single RGM-immunopositive cells were detected in perivascular spaces. Furthermore, rare smooth muscle cells and a few endothelial cells, but no astrocytes, were labeled.
We analyzed whether number and distribution of RGM immunoreactivity expression is altered after cerebral infarctions and observed accumulation of RGM expression confined to the injured regions. Lesion-associated cellular RGM immunoreactivity was confined to neurons, a few reactive astrocytes, invading leukocytes, and with lesion aging RGM-immunopositive extracellular depositions such as neo-laminae being components of the forming glial scar. The RGM-immunopositive cells accumulated in infarctional white matter, hemorrhagic areas, the infarction core, and peri-infarctional areas (Figure 2F and Figure 3A). Using the t test, we detected significantly more RGM-immunopositive cells in peri-infarctional areas (24 ± 1.1) than in remote areas (2 ± 0.2) or control tissue (6 ± 0.8) (P<.001) (Figure 3A). The morphologically described peri-infarctional areas were part of the physiologically defined penumbra.15 In these areas, RGM-immunopositive cells accumulated already at day 1 (32 ± 2.3; P<.001), reached maximal numbers 1.5 to 2.5 days after infarction (34 ± 3.2), and remained elevated during several weeks and months of survival (11 ± 1.4) (Figure 3B). Early after ischemic damage (up to 2.5 days), RGM immunoreactivity was predominantly observed with neurons (Figure 2G) and leukocytes of granulocytic, monocytic, and lymphocytic origin in vessels and infiltrative areas in ischemic tissue (Figure 2H-J). Furthermore, in a widespread pattern, small and medium-sized vessels revealed RGM immunoreactivity confined to endothelial cells (Figure 2H) and some fibromyocytes and vascular smooth muscle cells.
Paralleled by edema formation, up to 1 to 7 days, RGM-immunopositive cells were found extravasating outside the vascular walls into the focal ischemic lesioned parenchyma (Figure 2I). In perivascular regions, RGM-immunopositive cells formed clusters in the Virchow-Robin spaces from day 1 to 7, which subsided later (Figure 2J). These perivascular cells, also referred to as adventitial or perithelial cells, are characteristically alert immune cells. With lesion aging, from day 3 forward we also observed lesional RGM expression by a few reactive astrocytes (<5% of counterstained nuclei) restricted to the demarcating lesion core. Arising during the end of the first week after infarction, we detected RGM-immunopositive cells with fibroblast-like morphologic features (Figure 2K). In serial sections, these cells were nestin immunopositive and glial fibrillary acidic protein negative, a characteristic fibroblastoid phenotype. Excessive RGM-immunopositive formations, constituting pseudostratified neo-laminae, were localized in areas of ongoing scar formation (Figure 2L). These RGM-immunopositive laminae increased in magnitude and regional extent across time, “sealing” the lesional brain parenchyma. With tissue reorganization of the lesion, “foamy,” lipid-laden, phagocytic, RGM-immunopositive microglia/macrophages were observed (Figure 2M). Increases in numbers of RGM-immunopositive leukocytes correlated with the appearance of infiltrating leukocytes and the activation of microglia/macrophages after injury.17 In contrast, the up-regulation of extracellular RGM immunoreactivity correlated with the time course and the appearance of the scar after injury.1
In patients who died after TBI, in accordance with FCI, the immunohistologic evaluation revealed early cellular membraneous, cytoplasmatic, and nuclear RGM expression by leukocytes, a few reactive astrocytes, and neurons, demonstrating strong staining of the perikarya, dendrites, and axons. In the necrotic core and the bordering perinecrotic parenchyma, we detected the accumulation of RGM-immunopositive cells in border zones (22 ± 0.7) compared with remote areas (1 ± 0.1) and controls (6 ± 0.8) (P<.001) (Figure 4A). After TBI, RGM-immunopositive cell numbers reached maximal levels during the first 24 hours (29 ± 0.9; P<.001) and decreased subsequently (Figure 4B). With increasing time after TBI, most remarkable changes corresponded to areas of ongoing scar formation. By the end of the first week, RGM-immunopositive cells showed condensation and alignment along the lesion edge demarcating the lesion site. In these areas, RGM-immunopositive gliosis was composed of fibroblastoid cells and well-defined extracellular RGM-immunopositive laminae visibly condensing adjacent to the border zone. In these regions, RGM-immunopositive cells were localized close to the formation of neighboring basal lamina visualized by laminin expression. As observed after ischemic injury, RGM immunoreactivity was also detected in endothelial and vascular smooth muscle cells.
The RGM-immunopositive cells were characterized by double-labeling experiments, with monoclonal antibodies identifying cell entities and activation status of RGM-immunopositive cells. In control brains, RGM immunoreactivity was detected on myelin basic protein–immunopositive myelin fibers and a few myelin basic protein–immunopositive oliogodendrocytes (Figure 2N). In lesional areas, during the first week, RGM was expressed by NF and NSE immunopositive neurons (30%-40%) and CD45+ leukocytes (40%). These RGM-immunopositive, CD45+ leukocytes were observed in the lesional CNS parenchyma and in perivascular Virchow-Robin spaces, indicating drainage routes of extravasating cells into the brain. Most of the CD45+ leukocytes expressing RGM (70%) consist of CD68+ microglia/macrophages (70%) (Figure 2O). Furthermore, most RGM-immunopositive microglia/macrophages coexpress major histocompatibility complex class II molecules (>50%) (data not shown).
In addition, RGM immunoreactivity was most pronounced in vessels and hemorrhagic areas strictly confined to the lesion area. Rare RGM and CD68 coexpression was observed mainly with leukocytes of vessel lumina (data not shown). Occasionally, parenchymal RGM-immunopositive cells coexpress pan–B-cell (CD20) or CD4 and CD8 T-lymphocytic antigens (<5%) (data not shown). With maturation of the lesion, we observed RGM in the cellular and extracellular components of the developing scar. In detail, we detected RGM-immunopositive cells with fibroblastoid cell−like morphologic features. These cells were nestin immunopositive and glial fibrillary acidic protein negative, a characteristic fibroblastoid phenotype.14 Abundant extracellular RGM-immunopositive laminae scar components were verified to be fibronectin immunopositive (Figure 2P). Furthermore, we observed gemistocytic, reactive RGM-immunopositive, glial fibrillary acidic protein–immunopositive astrocytes confined to the lesion margin (data not shown). These RGM-immunopositive reactive astrocytes coexpressed the astrocytic activation antigens vimentin and nestin (data not shown). Oligodendroglial RGM expression was confirmed by additional 3′-cyclic nucleotide 3′-phosphodiesterase double labeling (Figure 2Q and R).
After TBI or FCI, glial cells initiate a classic wound-healing response with the formation of a barrier between the injured and healthy tissues. The barrier is a cellular and molecular boundary essential for the protection of injured cells by confining the area of injury, thus limiting secondary injury.1,17,18 Together with the presence of myelin compounds, formation of this glial barrier has been characterized as a negative event that prevents regeneration of injured neurons.1- 3 In the past decade, substantial progress has been made concerning our understanding of the molecular nature of this boundary. A variety of individual proteins that inhibit axonal growth have been identified, including, among others, CSPGs,1,19,20 Nogo-A,5- 7 MAG,4 and, recently, OMgp/Arretin.21,22
In the present study, we investigated the spatial and temporal expression of the glycosylphosphatidylinositol (GPI)–linked protein RGM in normal adult brains and in brains with TBI or FCI. In the normal brain, RGM is expressed on the perikarya of some neurons, smooth muscle and endothelial cells, and the choroid plexus. The choroid plexus has been identified as a repellent for axons owing to the secretion of chemorepulsive-acting molecules.23 Notably, RGM is also expressed by oligodendrocytes and is co-localized to myelin basic protein–immunopositive myelinated white fibers. The distribution of this protein supports the hypothesis that it might interact at contact sites between axons and myelin and thus may contribute to the inhibitory activity associated with myelin. Further studies are required to determine the importance of RGM in myelin inhibitory activity. Antibodies to Nogo-A have been shown to stimulate axonal sprouting and plasticity.24- 26 Because RGM and Nogo-A seem to share myelin localization and axonal growth inhibitory activity, it may be hypothesized that RGM function in the normal adult mammalian CNS contributes to the regulation of axonal sprouting and thus plasticity.
Evidence10,22,27,28 is now emerging that the proteins that govern developmental axonal guidance may contribute to the failure of injured central neurons to regenerate. The expression of chemorepulsive semaphorins and ephrins has been shown to relate to the success or failure of injured axons to regenerate10,28 through damaged CNS. As a next step toward elucidation of the role of chemorepulsive molecules in axonal regeneration, we studied the effects of TBI and FCI on the expression of RGM. After FCI and TBI, RGM-immunopositive cells accumulated in lesional and perilesional areas. During the first week after insult, RGM expression remained confined to neurons, smooth muscle and endothelial cells, activated microglia/macrophages, and leukocytes infiltrating the lesion. Thereafter, with maturation of the lesion, RGM was observed in components of the scar tissue, such as reactive astrocytes and fibroblastoid cells. Pronounced accumulation of RGM immunopositive cells localized in areas of ongoing scar formation was also observed. Thus, findings from pathological studies support the hypothesis that RGM participates in the extreme limitation of axonal regeneration after TBI or FCI; RGM may be part of the inhibitory activity associated with scar tissue. To date, RGM is one of the first potent growth cone collapse–inducing proteins present in several axonal growth impediments: the mature myelin, choroid plexus, and components of the glial scar.
Because RGM is present on cell surfaces and in extracellular components of scar tissue, RGM may be released from numerous RGM-immunopositive cells that undergo apoptosis after CNS injury, or a secretion mode of RGM can be proposed. First, RGM might be made from a splice variant that is missing its lipid anchor (GPI motif). Second, intracellular proteases process the mature full-length RGM at its proteolytic cleavage site. Third, endogenous GPI motif–specific hydrolyzing enzymes (GPI-specific phospholipases) exist in the mammalian brain, and in the lesion milieu they might convert the membrane-bound RGM to a soluble protein.
The domain of human RGM-A detected by the RGM276-293 peptide-specific RGM antibody is depicted in Figure 1. The specificity of antibodies against RGM immunization peptide was confirmed by the inhibition of RGM immunoreactivity after preincubation for 3 hours on ice with the cognate peptide RGM276-293 but not with the control peptide. In mice, recent studies identified 3 RGM family members: RGM-A, RGM-B, and RGM-C. However, only RGM-A and RGM-B were expressed in the CNS; RGM-C was exclusively detected in striated muscle and in the myocardium.29,30 To date, functional aspects of RGM in the CNS have been solely confined to RGM-A.10,29,31 Investigation of RGM-B was not the focus of this work and requires a functional characterization of RGM-B for justification.
In vivo RGM immunopositive cells are present in structures that guide axons by repulsion during development. It has been suggested that RGM repels retinal axons in the optic tectum, allowing the formation of retinotectal maps. The presence of RGM in the glial barrier and its inhibitory activity in vitro make it entirely plausible that it would exert inhibitory effects on regenerating axons in the glial scar. This result is important for several reasons. First, it represents an important step toward a molecular understanding of the inhibition of axonal regeneration by the glial scar after TBI and FCI. Second, because neutralization of inhibitory proteins by antibodies25 or peptides32 promotes long-distance regeneration and results in partial functional regeneration after CNS insult, neutralization of RGM may offer a means to support regeneration after TBI or FCI. Finally, myelin-derived growth inhibitory proteins are not only important for regeneration but also in regulating axonal-glial interactions and plasticity. Thus, these novel findings open new fields to explore the role of RGM in neuroregeneration and in neuronal plasticity.
Correspondence: Jan M. Schwab, MD, PhD, Brigham and Women’s Hospital, Center for Experimental Therapeutics, Thorn Bldg 724, 75 Francis St, Harvard Medical School, Boston, Mass 02115 (Jschwab@zeus.bwh.harvard.edu).
Accepted for Publication: January 7, 2005.
Author Contributions:Study concept and design: Schwab and Mueller. Acquisition of data: Schwab, Monnier, Conrad, and Chen. Analysis and interpretation of data: Schwab, Schluesener, Beschorner, and Meyermann. Drafting of the manuscript: Schwab, Monnier, Schluesener, Conrad, and Mueller. Critical revision of the manuscript for important intellectual content: Schwab, Schluesener, Beschorner, Chen, and Meyermann. Statistical analysis: Schwab. Obtained funding: Schwab and Schluesener. Administrative, technical, and material support: Schwab, Monnier, Schluesener, Conrad, Beschorner, Chen, and Mueller. Study supervision: Schluesener and Meyermann. Drs Schwab and Monnier contributed equally to this work.
Funding/Support: This study was supported by the Wings for Life Spinal Cord Research Foundation, Salzburg, Austria (Dr Schwab); by an international “poste rouge” award/scholarship from the Centre National de la Recherche Scientifique of the French government, Paris (Dr Schwab); by grant 1164/1-1 from the German Research Council, Bonn (Dr Schwab); and by the Hertie Foundation, Frankfurt am Main, Germany (Dr Schluesener).
Acknowledgment: We thank Ingrid Nagel and Simone Mucha for their critical reading of and help with the manuscript and Michel Mittelbronn, MD, for providing neuropathologically unaltered brain specimens.