Mol Neurobiol DOI 10.1007/s12035-015-9277-0

Myelin Basic Protein Cleaves Cell Adhesion Molecule L1 and Improves Regeneration After Injury David Lutz 1 & Hardeep Kataria 1 & Ralf Kleene 1 & Gabriele Loers 1 & Harshita Chaudhary 1 & Daria Guseva 1,4 & Bin Wu 1 & Igor Jakovcevski 1 & Melitta Schachner 2,3

Received: 19 March 2015 / Accepted: 1 June 2015 # Springer Science+Business Media New York 2015

Abstract Myelin basic protein (MBP) is a serine protease that cleaves neural cell adhesion molecule L1 and generates a transmembrane L1 fragment which facilitates L1-dependent functions in vitro, such as neurite outgrowth, neuronal cell migration and survival, myelination by Schwann cells as well as Schwann cell proliferation, migration, and process formation. Ablation and blocking of MBP or disruption of its proteolytic activity by mutation of a proteolytically active serine residue abolish L1-dependent cellular responses. In utero injection of adeno-associated virus encoding proteolytically active MBP into MBP-deficient shiverer mice normalizes differentiation, myelination, and synaptogenesis in the developing postnatal spinal cord, in contrast to proteolytically inactive MBP. Application of active MBP to the injured wild-type spinal cord and femoral nerve augments levels of a transmembrane L1 fragment, promotes remyelination, and improves functional recovery after injury. Application of MBP antibody impairs recovery. Virus-mediated expression of active MBP in the lesion site after spinal cord injury results in improved functional recovery, whereas injection of virus encoding

* Melitta Schachner [email protected] 1

Zentrum für Molekulare Neurobiologie, Universitätsklinikum Hamburg-Eppendorf, Martinistr. 52, 20246 Hamburg, Germany

2

Melitta Schachner, Center for Neuroscience, Shantou University Medical College, 22 Xin Ling Road, Shantou, Guangdong 515041, People’s Republic of China

3

Keck Center for Collaborative Neuroscience and Department of Cell Biology and Neuroscience, Rutgers University, 604 Allison Road, Piscataway, NJ 08854, USA

4

Present address: Department of Cellular Neurobiology, Medical School Hannover, Hannover, Germany

proteolytically inactive MBP fails to do so. The present study provides evidence for a novel L1-mediated function of MBP in the developing spinal cord and in the injured adult mammalian nervous system that leads to enhanced recovery after acute trauma. Keywords Cell adhesion molecule L1 . Development . Nerve injury . Myelin basic protein . Regeneration . Spinal cord injury

Introduction Development of a functioning nervous system depends on a number of cellular processes, such as cell proliferation and migration as well as neuritogenesis, fasciculation of neurites, myelination, and synaptogenesis. The cell adhesion molecule L1 participates in these important events [1, 2]. Furthermore, L1 not only plays crucial roles in the development of the central and peripheral nervous systems but also modulates important functions in adulthood, such as synaptic plasticity in learning and memory [1, 3]. L1 is also involved in regeneration after injury fostering axonal regrowth and remyelination after injuries to the spinal cord and peripheral nerve in mammals and fish [4–9]. In humans, L1 is associated with neural disorders, such as the L1 syndrome, fetal alcohol syndrome, Hirschsprung’s disease, schizophrenia, and Alzheimer’s disease [1, 10–14]. The beneficial functions of L1 depend on its proteolytic cleavage, and evidence has been presented that several proteases can mediate this proteolytic cleavage [15–23]. In a recent study, we have introduced a novel protease, the serine protease myelin basic protein [MBP] that cleaves L1 [24], thus augmenting the functional spectrum of MBP which is generally known to play important roles in formation and maintenance

Mol Neurobiol

of myelin [25–27]. MBP-mediated proteolytic processing of L1 leads to the generation and nuclear import of a sumoylated 70 kDa transmembrane L1 fragment which promotes neurite outgrowth, neuronal cell migration and survival, and myelination of axons by Schwann cell in vitro [23, 24]. Alterations in the generation of this fragment after spinal cord injury or in a mouse model of Alzheimer’s disease [28] indicated that MBP-mediated cleavage of L1 is functionally important in vivo and improves recovery in acute and chronic nervous system trauma. In the present study, we show that the serine protease activity of MBP is essential during embryonic development of the spinal cord and promotes morphological as well as functional recovery after femoral nerve and spinal cord injury. We demonstrate that MBP mediates L1-dependent functions in vitro and in vivo, highlighting the importance of the MBP-generated L1 fragment under in vivo conditions.

mouse P0, proteolipid protein (PLP), or glyceraldehyde-3phosphate dehydrogenase (GAPDH) were from Santa Cruz Biotechnology, and antibodies against vesicular glutamate transporter (VGLUT) neurofilament-L, glial fibrillary acidic protein (GFAP), distal-less homeobox protein 2 (Dlx-2), and anti-ionized binding calcium adapter molecule 1 (Iba-1) were from Synaptic Systems, Covance, Dako, Sigma-Aldrich, and Wako Chemicals, respectively. Normal goat serum was from Dianova. Aprotinin was from Sigma-Aldrich. Generation of adeno-associated virus 1 (AAV1) encoding active and inactive MBP and production of viral particles have been described [23, 24]. MBP purified from bovine brain (#64200100) and goat pan-MBP antibody (sc-13914) were obtained from AbD Serotec and Santa Cruz Biotechnology, respectively, and used for application to injured femoral nerves. Western Blot Analysis Western blot analysis was performed as described [34, 35].

Materials and Methods Animals L1-deficient mice [29] and MBP-deficient shiverer mice [30, 31] were maintained as heterozygous breeding pairs. L1deficient mice were maintained on a mixed genetic background (129SVJ × C57BL/6 × Black Swiss), and MBPdeficient shiverer mice were maintained on an inbred C57BL/6J background. Wild-type littermates were used as controls in experiments where mutant mice were investigated. C57BL/6J mice were used as wild-type mice in all other experiments. Mice were maintained under standard laboratory conditions with food and water supply ad libitum and with an artificial 12 h light/dark cycle. All experiments were conducted in accordance with the German and European Community laws on protection of experimental animals, and all procedures used were approved by the responsible committee of The State of Hamburg (TVA 6/14, TVA 098/09). Experiments were carried out and the manuscript prepared following the ARRIVE guidelines for animal research. Antibodies and Reagents Monoclonal antibodies 555 and 557 against mouse L1 have been described [32]. The MBP antibody against the exon IIencoded domain of mouse MBP [33] was kindly provided by David Colman (Montreal Neurological Institute and Hospital of McGill University, Montreal, Canada). Monoclonal antibody G3G4 against BrdU was from Developmental Studies Hybridoma Bank. Secondary antibodies coupled to horseradish peroxidase (HRP) or conjugated to fluorescent dyes were purchased from Jackson ImmunoResearch. L1 was purified from mouse brains as described [34]. Antibodies recognizing

Neurite Outgrowth and Schwann Cell Proliferation and Migration Primary cultures of dorsal root ganglion (DRG) neurons and motoneurons were prepared as described [36, 37]. For neurite outgrowth assays, dissociated neurons (1–2×105 cells/ml) were seeded onto glass coverslips coated with 0.01 % polyL-lysine (PLL) (for DRG neurons) or poly-L-ornithine (for motoneurons) and treated with 10 μg MBP/ml culture medium. After 20–24 h, cells were fixed with 2.5 % glutaraldehyde and stained with 1 % toluidine blue, 1 % methylene blue in 1 % borax. Neurite outgrowth was quantified by measuring total neurite length of at least 100 neurons per condition using an AxioVision system 4.6 (Carl Zeiss). Schwann cells were prepared as described [38]. To determine process formation, cells were seeded at a density of 1× 105 cells/ml onto glass coverslips first coated with 0.01 % PLL and thereafter treated with 10 μg/ml of MBP, 10 ng/μl L1 antibody 557, 5 μg/ml L1, and 10 ng/μl MBP antibody. After 24 h in culture, cells were fixed and stained as described above and process formation was determined using the AxioVision system [38]. For the migration assay, confluent monolayers of cultured wild-type and shiverer Schwann cells were subjected to scratch injury [24] and treated for 16 h with 4× 1011 AAV1 viral particles encoding proteolytically active wild-type MBP or proteolytically inactive MBP mutant and with viral particles without MBP-encoding cDNA insert as vehicle control, fixed, mounted in Fluoromount solution containing DAPI, and gap closure was analyzed by measuring 10 gap widths of six coverslips. To assess proliferation, cells were seeded at a density of 2.5×105 cells/ml onto glass coverslips coated with PLL and cultured in the presence of 12 ng/ml neuregulin (ImmunoTools). Four hours after seeding, 20 μM

Mol Neurobiol

BrdU (Sigma-Aldrich) was added to the culture. After 48 h, cells were fixed with 4 % formaldehyde in 0.1 M phosphate buffer, pH 7.3. After incubation for 30 min with 2 N HCl at 37 °C, cells were washed, blocked with normal goat serum, and incubated overnight with BrdU antibody (1:100) at 4 °C and with secondary anti-mouse antibody (1:200) for 1 h at room temperature. The coverslips were then washed and mounted with RotiMount FluorCare (Carl Roth). To estimate the number of BrdU-labeled cells, 10 images per treatment were taken from different areas of each coverslip using an Axiophot 2 microscope (Zeiss) and a 20× objective. In the different samples, the chosen areas corresponded with each other. Each area was photographed using phase contrast and epifluorescence. The two digital images were then overlaid using ImageJ, and Image Tool 2.0 software (University of Texas, San Antonia, TX, USA) was used to count BrdU-positive Schwann cells and total number of Schwann cells. Schwann cells were distinguished from other cells (DRG neurons and fibroblasts) by their characteristic long spindle-shaped cell body and two processes in opposite directions. Approximately 1000 cells were counted for each experimental value. In Vitro Myelination in Co-Cultures of Dorsal Root Ganglion Neurons and Schwann Cells Co-cultures of dorsal root ganglion neurons and Schwann cells and assessment of in vitro myelination have been described [23]. Briefly, dissociated cells were isolated from dorsal root ganglia of 14-day-old mouse embryos and maintained on Matrigel-coated glass coverslips at 37 °C in a humidified 5 % CO2 atmosphere. Cells were treated with 10 μg/ml MBP, 1 μM aprotinin, or 4×1011 viral particles and fixed in 4 % paraformaldehyde/PBS at room temperature for 10 min followed by fixation and permeabilization with 100 % methanol at −20 °C for 10 min. Samples were incubated in phosphate-buffered saline (PBS), pH 7.3 containing 1 % bovine serum albumin (BSA) (BSA/PBS solution) for 30 min at room temperature and stained with goat P0 antibody and rabbit neurofilament-L antibody followed by secondary donkey anti-goat Cy2- and anti-rabbit Cy3-conjugated antibodies in BSA/PBS solution. Coverslips were mounted with Fluoromount solution containing DAPI, and confocal images were taken on the Olympus FV1000 confocal microscope. Myelination was analyzed by counting the number of internodes on myelinated axons at similar fiber density from 20 coverslip areas of six coverslips per condition from three independent experiments. In Utero Injection of Viral Particles Timed pregnant mice at embryonic day 16 were anesthetized by exposure to 4 % isoflurane which was supplied with an O2-

volumetric flow of 0.8 l/min. During surgery, the isoflurane concentration was lowered to 1.5–2 %. The abdominal cavity of the pregnant mice was opened, and the uterine horns bearing the embryos were exposed. For in utero injection, 4×1012 AAV1 encoding wild-type MBP or MBP mutated in the proteolytically active serine residue 176 [24] or virus without MBP-encoding insert (kindly provided by Ingke Braren, Vector Core Facility, University Hospital Hamburg-Eppendorf) and 0.01 % w/v Fastgreen/ PBS (Sigma Aldrich) were introduced via a thin capillary into the embryonic brachial plexus of the spinal cord. Embryos were re-placed into the abdominal cavity to allow development to continue until postnatal day 5 when mice were killed by decapitation and the thoracic part of the trunks was isolated and prepared for immunohistological analysis. Femoral Nerve Injury and Analysis of Femoral Nerve Regeneration and Motor Function Femoral nerve injury in 2- to 3-month-old wild-type and L1deficient mice was performed as described [36]. Locomotor function and cellular parameters of regeneration were analyzed also as described [36], with the exception that the polyethylene tubes (3 mm length, 0.58 mm inner diameter; Becton Dickinson) were filled either with scaffold peptide that forms a gel matrix support (0.5 % PuraMatrix Peptide Hydrogel; BD Biosciences) alone or with scaffold peptide supplemented with MBP (100 μg/ml) or MBP antibody (10 μg/ml). Semi-thin and ultrathin sections of fixed nerves were stained with toluidine blue, and the mean orthogonal diameters of the axon (inside the myelin sheath) and of the nerve fiber (including the myelin sheath) were measured. Analysis of degree of myelination was performed by counting the number of myelin sheaths and the myelinated axonal area from 50 ultrathin sections collected at tree adjacent points within the lesion areas from injured nerves of three different animals per group. Calculation of the g-ratio (axon to fiber diameter) has been described [38]. For Western blot analysis, tissue from spinal cord and femoral nerve was collected. An approximately 1-cm-long segment was taken from the lumbar part of the spinal cord (centered on L2-4 containing the motoneuron cell bodies that project their axons into the femoral nerve). From the femoral nerve, an approximately 1-cm-long segment proximal to the lesion site and an approximately 1.5-cm-long segment distal to the lesion site including approximately 1.2 cm of both the sensory and motor branches were taken. The 3-mm-long nerve fragment within the tube containing the lesion site could not be used for Western blot analysis because it did not contain enough protein. Tissue was homogenized in RIPA buffer (25 mM Tris–HCl, 150 mM NaCl, 1 % NP-40, 1 % sodium deoxycholate, 0.1 % SDS, pH 7.6) and processed for Western blot analysis.

Mol Neurobiol

Spinal Cord Injury and Analysis of Motor Function For surgery, female mice were anesthetized by intraperitoneal injections of ketamin and xylazin (100 mg Ketanest® and 5 mg Rompun®, per kg body weight). Laminectomy was performed at the T7–T9 level with mouse laminectomy forceps (Fine Science Tools). A mouse spinal cord compression device was used to elicit compression injury [39]. Compression force (degree of closure of the forceps) and duration were controlled by an electromagnetic device: The spinal cord was maximally compressed (100 %) [39] for 1 s by a timecontrolled current flow through the electromagnetic device. Viral particles (4×1011) were injected into the dorsal lemniscal area of spinal cord, bilaterally at the proximal and distal sites of the lesion (1–2 mm) with a thin capillary to minimize spinal cord damage. Muscles and skin were then closed using 6–0 nylon stitches (Ethicon). After surgery, mice were kept in a heated room (37 °C) for several hours to prevent hypothermia and thereafter singly housed in a temperature-controlled (22 °C) room with water and standard food provided ad libitum. During the postoperative time period, the bladders of the female animals were manually voided twice daily. The recovery of ground locomotion was evaluated using the Basso Mouse Scale (BMS) [40]. In addition, we used a numerically more precise assessment of locomotion: singleframe motion analysis to determine the foot-stepping angle and rump-height index using the beam walking test and the foot-stepping angle as defined by a line parallel to the dorsal surface of the hind paw and the horizontal line [41]. The footstepping angle was measured with respect to the posterior aspect at the beginning of the stance phase which is well defined in intact mice with an angle being approximately 20°. After spinal cord injury and severe loss of locomotor abilities, the mice drag behind their hind limbs with dorsal paw surfaces facing the beam surface. The foot-stepping angle is increased to >150°. The rump-height index was estimated from the recordings used for measurements of the foot-stepping angle. The rump-height parameter is defined as height of the rump, i.e., the vertical distance from the dorsal aspect of the animal’s tail base to the beam, normalized to the thickness of the beam measured along the same vertical line. Assessment was performed before and at 1, 2, 4, 8, and 12 weeks after the injury. Values for the left and right extremities were averaged. Tissue Preparation, Immunohistochemistry, and In Situ Hybridization

perfusion, spinal cords were left in situ for 2 h at room temperature, after which they were dissected out and postfixed overnight at 4 °C in the same fixative. Tissue was then immersed in a 15 % w/v sucrose solution in 0.1 M cacodylate buffer, pH 7.4, for 2 days at 4 °C, embedded in Tissue Tek (Sakura Finetek), frozen by a 2-min immersion into 2-methylbutane (isopentane) precooled to −80 °C and stored in liquid nitrogen until sectioned. Serial transverse 25-μm-thick sections were cut in a cryostat (Leica CM3050, Leica Instruments) and collected on SuperFrost Plus glass slides (Carl Roth). For immunohistochemistry, antigen retrieval was performed by incubating the sections in 10 mM sodium citrate solution (pH 9.0) for 30 min in a water bath at 80 °C, followed by blocking of non-specific binding sites with PBS containing 0.2 % Triton X-100 (Sigma), 0.02 % sodium azide (Merck), and 5 % normal donkey serum (Jackson ImmunoResearch) for 1 h at room temperature. The sections were incubated overnight at 4 °C in a humid chamber with antibodies against mouse MBP, GFAP, Iba-1, neurofilament-L, PLP, VGLUT, or Dlx-2. All primary antibodies were used at optimal dilutions in PBS. Following washing in PBS (3×15 min at room temperature), the sections were incubated for 2 h at room temperature with the appropriate secondary antibodies conjugated to fluorescent dyes (1:200 in PBS). After a subsequent wash in PBS, cell nuclei were stained for 10 min at room temperature with bis-benzimide solution (Hoechst dye 33258, 5 μg/ml in PBS, Sigma). Finally, the sections were washed again, mounted with Fluoromount G (Southern Biotechnology Associates), and stored at 4 °C in the dark. Microphotographs were taken on a Leica confocal laser scanning microscope (Leica) at the digital resolution 1024×1024. Photographs were further processed by using Adobe Photoshop CS5 software (Adobe Systems) to adjust brightness and contrast. As a measure of glial scar area, the area immunostained by GFAP was determined. Numbers of activated microglia were determined by counting Iba-1-positive amoeboid cells. Numbers of axons or myelinated fibers were determined by counting neurofilament-L- or PLP-positive profiles. To evaluate the number of synaptic contacts of glutamatergic neurons and to assess neuronal differentiation, VGLUT immunostaining area intensity and numbers of Dlx-2-positive cells were determined. All quantifications were performed on a Keyence fluorescent microscope using the Hybrid cell count software (Keyence). In situ hybridization was performed as described [24]. ELISA

Tissue fixation and sectioning was performed as previously described [38]. Twelve weeks after injury mice were anesthetized with a 16 % w/v solution of sodium pentobarbital (Narcoren, 5 μl/g body weight) and then transcardially perfused with fixative (4 % w/v paraformaldehyde and 0.1 % w/v CaCl 2 in 0.1 M cacodylate buffer, pH 7.4). Following

Bovine MBP, goat serum, and crude myelin fractions isolated from wild-type or shiverer mice as described [24] were coated (all 5 μg protein/well) overnight in 384-well microtiter plates with high binding surface (Corning) at 4 °C. After washing with Tris-buffered saline, pH 7.3 (TBS), blocking with 1 %

Mol Neurobiol

BSA (essentially fatty acid free; PAA Laboratories) in TBS for 1 h at room temperature, and washing with TBST (TBS with 0.05 % Tween 20), serum collected from injured and treated mice 3 month after injury and diluted 1:20 in TBS was added at different concentrations and incubated for 1 h at room temperature. After washing with TBST, anti-mouse HRP-coupled secondary antibodies (Dianova) and o-phenylenediamine dihydrochloride (Pierce) as HRP substrate were used for detection. The reaction was stopped by addition of 2.5 M sulfuric acid. Absorbance was measured at 490 nm.

triggering L1 antibody 557 or mouse L1 led to a pronounced increase in process lengths of wild-type Schwann cells, while a less pronounced enhancement by L1 antibody 557 or L1 was observed in the presence of MBP antibody or when MBPdeficient Schwann cells were used for stimulation (Fig. 1e). Application of MBP enhanced process lengths of wild-type and MBP-deficient Schwann cells to a similar degree (Fig. 1e). These results show that L1-stimulated Schwann cell process formation depends on the activity of MBP generated in the cultures and that ablation of this MBP activity affects L1-induced cellular functions.

Results

MBP Enhances Schwann Cell Migration and Myelination of Axons of Dorsal Root Ganglion Neurons by Schwann Cells

MBP Promotes L1-Mediated Neurite Outgrowth of Dorsal Root Ganglion Neurons and Motoneurons, Enhances Schwann Cell Process Formation, and Reduces Proliferation of Schwann Cells Because application of MBP to cultured cerebellar neurons promotes L1-dependent neurite outgrowth and neuroprotection against oxidative stress and excitotoxicity [24], we investigated whether MBP also triggers other well-established L1dependent functions, such as neurite outgrowth of motoneurons and dorsal root ganglion neurons as well as Schwann cell proliferation and process formation [35, 36, 38, 42–45]. The lengths of neurites of wild-type dorsal root ganglion neurons and motoneurons and the length of processes of wild-type Schwann cells were increased in the presence of MBP when compared to the lengths measured in the absence of MBP (Fig. 1a–c). No effect of MBP on neurite outgrowth of dorsal root ganglion neurons and motoneurons or on Schwann cell process formation was observed when cells from L1-deficient mice were analyzed (Fig. 1a–c). The number of proliferating BrdU-labeled wild-type Schwann cells was decreased in the presence of MBP when compared to the number determined in the absence of MBP, while MBP had no effect on the number of proliferating BrdUlabeled Schwann cells from L1-deficient mice (Fig. 1d). These results indicate that MBP promotes neurite outgrowth of dorsal root ganglion neurons and motoneurons, enhances Schwann cell process formation, and reduces Schwann cell proliferation in an L1-dependent manner. By ablation of MBP or blocking of endogenous MBP functions with MBP antibody, we have demonstrated that L1induced neurite outgrowth from cerebellar neurons and L1triggered neuroprotection of cerebellar neurons against oxidative and glutamate stress depend on MBP [24]. To investigate whether L1-induced promotion of Schwann cell process formation also depends on MBP, we analyzed this response by application of MBP antibody to wild-type Schwann cells or MBP to MBP-deficient Schwann cells from shiverer mice. Stimulation of L1 signaling by addition of function-

Since Schwann cell process formation and migration as well as myelination of dorsal root axons by Schwann cells depends on the generation of the 70 kDa L1 fragment [23] and since MBP mediates the generation of this L1 fragment [24], we tested whether MBP affects myelination of dorsal root ganglion neurons by Schwann cells. To this aim, co-cultures of Schwann cells and dorsal root ganglion neurons were treated either with MBP in the absence or presence of the serine protease inhibitor aprotinin which prevents the generation of the 70 kDa fragment [24, 28] or with MBP antibody to inhibit MBP functions and to block the generation of the L1 fragment by endogenous MBP in the culture. Staining with antibody against the peripheral nervous system myelin marker protein P0 was used for determination of numbers of P0-positive myelin internodes. Application of MBP enhanced the number of internodes on myelinated axons relative to the control treatment, where MBP was not added, while this MBP-induced enhancement of myelination was not observed in the concomitant presence of aprotinin (Fig. 2a). Treatment with MBP antibody reduced the number of P0-positive internodes by more than 90 % relative to the control treatment in the absence of this antibody (Fig. 2a). These results indicate that MBP-mediated proteolysis of L1 plays an important role in myelination. To investigate whether MBP and its serine protease activity is required to promote myelination of axons from dorsal root ganglion neurons by Schwann cells, we applied AAV1 coding for the wild-type 21.5 kDa MBP isoform (AAV1-wtMBP) or the proteolytically inactive S/A176 mutant (AAV1-mut-MBP) [24] to the co-cultures of Schwann cells and dorsal root ganglion neurons from shiverer mice and measured myelination by counting the numbers of P0-positive internodes. After transduction with AAV1-wtMBP, the numbers of internodes increased when compared with control conditions, where viral particles without MBPencoding insert were applied, while similar numbers were observed for cells transduced with control viral particles and

Mol Neurobiol Fig. 1 MBP-triggered cellular responses depend on L1. Dorsal root ganglion neurons (DRG) (a), motoneurons (MN) (b), and Schwann cells (SC) from wildtype (WT) and L1-deficient (L1-/ y) mice were treated without (−) or with MBP (MBP). The total lengths of neurites (a, b) and of Schwann cell processes (c) were determined by counting at least 100 cells per sample and experiment. d Schwann cells from wild-type and L1-deficient mice were incubated with BrdU in the absence or presence of MBP. The number of BrdU labeled Schwann cells was determined. e Schwann cells (SC) from wild-type (WT) and MBPdeficient shiverer (shi) mice were incubated with L1 antibody 557 or brain L1 in the absence or presence of a MBP antibody (αMBP) or were treated with MBP. The total lengths of Schwann cell processes were determined counting at least 100 cells per sample and experiment. a–e Mean values±SEM from three independent experiments with duplicates relative to those determined without additives (set to 100 %) are shown (*p<0.05, **p<0.01, ***p<0.001; one-way ANOVA with Tukey’s multiple comparison test)

for cells transduced with viral particles encoding the proteolytically inactive MBP mutant (Fig. 2b). This result shows that the serine protease activity of MBP is required for MBPtriggered myelination of dorsal root ganglion neuron axons by Schwann cells. Next, we analyzed whether Schwann cell migration also depends on MBP-mediated cleavage of L1 in a scratch assay using monolayers of MBP-deficient Schwann cells from shiverer mice transduced with AAV1-wt-MBP, AAV1-mutMBP, or AAV1 without MBP-encoding insert. Closure of the gap, which had resulted from scratching, was enhanced after transduction of MBP-deficient Schwann cells with

AAV1-wt-MBP relative to gap closure by cells transduced with control virus (Fig. 2c). The extent of gap closure by MBP-deficient Schwann cells transduced with wild-type MBP was even more pronounced than that observed for wild-type Schwann cells (Fig. 2c). Upon transduction of MBP-deficient Schwann cells with AAV1-mut-MBP, the values for gap closure by Schwann cells expressing the proteolytically inactive MBP mutant were similar to those of MBP-deficient Schwann cells transduced with control virus (Fig. 2c). This result suggests that Schwann cell migration depends on the serine protease activity of MBP and, thus, on the MBP-mediated generation of the 70 kDa L1 fragment.

Mol Neurobiol

Fig. 2 MBP-mediated generation of the 70 kDa L1 fragment enhanced myelination. a Co-cultures of dorsal root ganglion neurons and Schwann cells from wild-type mice were incubated without (control) and with MBP in the absence (MBP) or presence of aprotinin (MBP+apro) or with MBP antibody (α-MBP). b Co-cultures of dorsal root ganglion neurons and Schwann cells from MBP-deficient shiverer mice were transduced with viral particles without MBP-encoding insert (AAV control) or with AAV1 coding for wild-type MBP (AAV1-wt-MBP) or MBP mutant (AAV1-mut-MBP). a, b Cells were immunostained with an antibody recognizing the myelin marker P0 (green) and nuclei were st ain ed wi th DA PI (bl ue). R epre sentati ve ima ges of the immunostainings are shown. Numbers of internodes were assessed, and mean values±SEM relative to those observed after control treatment (set to 100 %) from two independent experiments are shown. Differences

between groups were analyzed by one-way ANOVA with Tukey’s multiple comparison test (***p<0.005). c Schwann cell monolayers from MBP-deficient shiverer (shi) mice were mock-treated (shi/mock) or transduced with AAV1 coding for wild-type MBP (shi/AAV1-wtMBP) or MBP mutant (shi/AAV1-mut-MBP) followed by scratch injury. Mock-treated Schwann cell monolayers from wild-type (WT/ mock) mice were used as control. Gap borders are highlighted with continuous white lines. Schwann cell migration was analyzed by assessing the extent of gap closure 16 h after scratching as the ratio of gap widths determined after incubation and scratch injury and gap widths determined immediately after scratch injury. Mean values±SEM relative to those observed for mock-treated shiverer Schwann cells (set to 100 %) from two independent experiments are shown (***p<0.005; one-way ANOVA with Tukey’s multiple comparison test). Scale bars: 100 μm

Levels of the MBP-Generated 70 kDa L1 Fragment Are Increased After Femoral Nerve Injury

this fragment and full-length L1 were seen in the lumbar spinal cord of non-lesioned mice (Fig. 3a, b). Full-length L1 and high levels of the 70 kDa fragment were found in the proximal and distal stumps of the injured femoral nerves, with higher levels of the 70 kDa fragment and lower levels of full-length L1 in the distal than in the proximal stump (Fig. 3a). In the lumbar spinal cord, levels of full-length L1 were increased up to approximately 1.5-fold upon femoral nerve injury relative to the level observed without injury, while the level of the 70 kDa fragment increased up to approximately 13.5-fold (Fig. 3b). These observations implicate the MBP-generated 70 kDa L1 fragment in early cellular events occurring within the first week after acute femoral nerve injury.

Since L1 promotes regeneration after injury [6, 7] and MBP cleaves L1 to generate the 70 kDa L1 fragment [24], we were interested whether the MBP-mediated generation of this fragment plays a role in events occurring upon acute injury in vivo. Thus, we performed femoral nerve injury in adult wild-type mice and determined L1 protein levels in femoral nerve segments proximal or distal to the lesion site and in lumbar spinal cord segments 1 week after femoral nerve injury. In femoral nerve segments of non-lesioned mice, neither the 70 kDa fragment nor full-length L1 were detectable, while

Mol Neurobiol Fig. 3 Increased levels of the 70 kDa L1 fragment are beneficial for functional recovery after femoral nerve injury. a, b Femoral nerve segments proximal or distal to the lesion site and lumbar spinal cord segments were isolated from mice 1 week after femoral nerve injury (injured), and corresponding segments were taken from non-injured (intact) mice. The samples were subjected to Western blot analysis with L1 antibody 555. GAPDH antibody was used to control for loading. Full-length L1 (L1-200) and 70 kDa fragment (L1-70) are indicated by arrows. c–f Footbase angle (c), heels-tail angle (d), protraction length ratio (e), and overall recovery index (f) before injury (day 0) and at different time points after injury and application of MBP, MBP antibody (αMBP), or vehicle solution (PBS) were determined. Mean values± SEM and statistical differences between the groups treated with MBP or MBP antibody relative to the PBS control group are shown (*p<0.05, **p<0.01, ***p<0.001, one-way ANOVA with Tukey’s multiple comparison test; n=8 mice per treatment)

MBP Promotes Functional Recovery After Femoral Nerve Injury in Adult Mice In a recent study, we showed that MBP enhances neurite outgrowth from cerebellar neurons and protects cerebellar neurons from oxidative and excitotoxic stress [24]. Here, we demonstrated that MBP also promotes neuritogenesis of dorsal root ganglion neurons and motoneurons and affects Schwann cell migration, proliferation, and process formation. Thus, we investigated whether these beneficial effects of MBP on L1dependent cellular responses would also improve regeneration and functional outcome after peripheral nerve injury in an L1dependent manner. Functional recovery after femoral nerve injury of wild-type and L1-deficient mice was assessed by monitoring the function of the quadriceps muscle after a single application of MBP, MBP antibody, or vehicle solution to the lesion site immediately following injury. Femoral nerve injury impairs extensor function of the quadriceps muscle and leads to abnormal external rotation of the ankle and high heel

position at defined gait cycle phases [46]. Changes in gait can be quantitatively evaluated by measuring the foot-base angle and heels-tail angle during beam walking, and we used these parameters to evaluate the effect of MBP and MBP antibody on locomotor recovery. Another parameter for assessment of functional recovery is the protraction length ratio, which measures the ability of a mouse to extend hind limbs while grasping an object with its hind paws. When compared with the pre-operative values, a similar increase in foot-base angle and protraction length ratio and a similar decrease of heels-tail angle were observed in all mice 1 week after injury (Fig. 3c–e), showing the same degree of functional impairment at this stage. One, 2, 4, 8, and 12 weeks after injury, mice treated with MBP showed better functional improvement than control mice treated with vehicle solution, while the mice treated with MBP antibody showed poorer locomotor recovery than the control mice (Fig. 3c–e). The recovery index is a measure of the individual degree of recovery and reflects the degree of postoperative

Mol Neurobiol

normalization of the quadriceps extensor function during ground locomotion. The overall recovery index, combining the individual recovery indices for foot-base angle and heels-tail angle, also showed that application of MBP improved recovery, while application of MBP antibody impaired recovery between 2 and 12 weeks after injury (Fig. 3f). These results indicate that MBP application enhances functional recovery after nerve injury. The application of MBP antibody reduces functional recovery, indicating that endogenous MBP plays an important role in nerve regeneration. To analyze whether this beneficial effect of MBP on recovery is L1-dependent, MBP or vehicle solution was applied to L1-deficient mice. Two, 4, 8, and 12 weeks after injury, injured L1-deficient mice treated with MBP showed no better functional recovery than mice treated with vehicle solution alone as indicated by similar values for foot-base and heelstail angle as well as protraction length ratio (Fig. 4a–c). This result confirms that L1 mediates the beneficial effects of MBP on functional recovery after femoral nerve injury. Application of MBP or MBP Antibody Affects Generation of the 70 kDa L1 Fragment and Remyelination After Femoral Nerve Injury To evaluate whether the functional outcome after application of MBP or MBP antibody correlates with the degree of remyelination, we morphologically analyzed ultrathin sections of regenerated nerves for myelin thickness by counting the number of the myelin sheaths, by measuring the myelinated axonal area and by calculating g-ratios (axonal diameter divided by total fiber diameter). Treatment with MBP led to increased numbers of the myelin sheaths and myelinated cross-sectional axon areas, when compared to the values observed after application of MBP antibody or control treatment (Fig. 5a–c). The frequency distributions of the g-ratios for axons of MBP-treated injured nerves shifted toward lower (normal non-injured) values when compared with values obtained for axons of injured nerves treated with MBP antibody or control solution (Fig. 5d). These results indicate that MBP application enhances the thickness of myelin sheaths around injured axons compared to the control, while the MBP antibody has neither an effect on the thickness of myelin sheaths nor on axon diameters. Since we observed that levels of the 70 kDa L1 fragment were increased after femoral nerve injury (Fig. 3a), we analyzed whether application of MBP or MBP antibody affected MBP-mediated generation of the 70 kDa fragment in the spinal cords and proximal stumps of the femoral nerves. One day after femoral nerve injury, the levels of 70 kDa L1 fragment in the proximal nerve stumps were reduced by approximately 30 % in the presence of MBP antibody and enhanced approximately 2-fold in the presence of MBP when compared to the levels seen after vehicle application (Fig. 5e). Full-length L1

Fig. 4 Improvement of functional recovery by MBP depends on L1. Foot-base angle (a), heels-tail angle (b), and protraction length ratio (c) before injury (day 0) and at different time points after injury and application of MBP or vehicle solution (PBS) were determined for L1deficient mice. Mean values±SEM are shown, and statistical analysis revealed no significant (ns) differences between the MBP-treated and the PBS control groups (one-way ANOVA with Tukey’s multiple comparison test; n=8 mice per treatment)

levels in the proximal femoral nerve stumps were increased approximately 2- and 10-fold after MBP antibody application or MBP application, respectively (Fig. 5e). In the lumbar spinal cord, neither the L1 fragment levels nor levels of fulllength L1 were altered by application of MBP or MBP antibody 1 day after injury L1 relative to vehicle application (Fig. 5e). These results indicate not only that MBP mediates proteolytic processing of L1 and generation of the 70 kDa fragment but also that MBP is involved in upregulation of L1 expression. Since the applied MBP and MBP antibody were not derived from mouse but from bovine or goat, respectively, we

Mol Neurobiol

Fig. 5 Application of MBP to injured femoral nerves leads to increased remyelination of axons and enhanced 70 kDa L1 fragment levels. a–c Representative images from transmission electron microscopy of injured nerves (a) and determination of numbers of myelin sheaths and myelinated axonal areas (b, c) after treatment with MBP, MBP antibody (α-MBP), or vehicle control solution are shown. d Normalized overall frequency distributions of g-ratios (axon/fiber diameter) in the motor nerve branch of mice 4 months after injury and application of MBP, MBP antibody, or vehicle control solution are shown (n=8 mice per treatment). e Femoral nerve segments proximal or distal to the lesion site and lumbar spinal cord were taken from mice 1 day after femoral nerve injury (injured) and application of MBP and MBP antibody. Corresponding segments and spinal cords from non-injured (intact) mice (n=8) were used as controls.

The samples were subjected to Western blot analysis with L1 antibody 555. GAPDH antibody was used to control for loading. Blots of samples from one animal per group are shown, and full-length L1 (L1-200) and 70 kDa fragment (L1-70) are indicated by arrows. f Serum was prepared from noninjured (intact) or injured mice treated with vehicle control (inj/PBS), MBP (inj/MBP), or MBP antibody (inj/α-MBP) 3 months after injury. For ELIS A, sera were applied to substrate-coated bovine MBP, crude myelin from wild-type (WT myelin), and MBP-deficient (shi myelin) shiverer mice, and goat serum and binding of antibodies were determined using anti-mouse secondary antibody. Mean values±SEM and statistical differences between the groups are shown (***p<0.001, one-way ANOVA with Tukey’s multiple comparison test; n=8 mice per group). Scale bars in a: 5 μm

wanted to rule out that the effects on function recovery were due to stimulation of an immune response against the applied

proteins. Therefore, we determined the titer of endogenous antibodies directed to MBP, to total myelin proteins, and to

Mol Neurobiol

goat antibodies in the sera of non-injured and injured mice using ELISA. In comparison to sera from non-injured mice, sera of vehicle-treated injured mice showed a slight but nonsignificant increase in immunoreactivity against substratecoated bovine MBP (Fig. 5f). Sera from injured mice treated

with MBP or MBP antibody showed a similar titer of MBPdirected antibodies as sera of vehicle-treated injured mice (Fig. 5f). In comparison to the immunoreactivity observed with sera from non-injured mice, an increase in immunoreactivity was observed when sera of vehicle-treated and MBP-

Fig. 6 Expression of proteolytically active MBP promotes functional recovery after spinal cord injury. a AAV1 coding for wild-type MBP (AAV1-wt-MBP) or for a proteolytically inactive MBP mutant (AAV1mut-MBP) and viral particles without MBP-encoding insert (AAV control) was injected bilaterally in the Th7–Th9 region of the dorsal lemniscus rostral and caudal to the lesion site after spinal cord injury as indicated in the scheme. b–d Mean values±SEM of Basso Mouse Scale (BMS) scores (b), rump-height ratio (c), and foot-stepping angle (d) are shown. Statistical differences relative the control group are indicated (*p<0.05, **p<005, ***p<0.0005, one-way ANOVA with Tukey’s multiple comparison test for repeated measurements, n=8). e Spinal

cord segments from the lesion site (L) and proximal (P) or distal (D) to the lesion site were taken from mice 12 weeks after spinal cord injury and injection of AAV1 coding for wild-type MBP (AAV1-WT-MBP) or coding for proteolytically inactive MBP mutant (AAV1-mut-MBP) and of viral particles without MBP-encoding insert (AAV control). The samples were subjected to Western blot analysis with L1 antibody 555, exon II-specific MBP antibody, and GAPDH antibody. Blots of samples from one animal per group are shown. Full-length L1 (L1-200) and 70 kDa fragment (L1-70) are indicated by arrows, and the 21.5 kDa MBP isoform is indicated by an arrowhead

Mol Neurobiol

treated mice were applied to substrate-coated crude myelin from wild-type mice and MBP-deficient shiverer mice (Fig. 5f). Sera from injured mice treated with MBP antibody displayed a similar immunoreactivity to antibodies in substrate-coated goat serum as sera from non-injured mice and injured mice treated with vehicle solution or MBP (Fig. 5f). These results indicate that femoral nerve transection leads to an increase in the titer of antibodies against myelin proteins, but not against MBP, and that neither the in vivo application of bovine MBP nor of goat MBP antibody triggers the production of antibodies against bovine MBP or goat MBP antibody under the conditions of this study.

Acute Posttraumatic Administration of AAV1 Encoding Proteolytically Active MBP Improves Locomotor Recovery After Spinal Cord Injury To investigate whether MBP also improves functional recovery after an acute central nervous system injury and whether this effect of MBP depends on its serine protease activity, we applied either AAV1 coding for the proteolytically active wild-type 21.5 kDa MBP isoform or AAV1 coding for the proteolytically inactive mutated isoform [24] to the lesion site after a severe spinal cord compression injury in wild-type mice (Fig. 6a). As control, viral particles without MBPencoding insert were used. Locomotor disabilities and recovery were assessed by calculating BMS scores [40]. One week after injury, severe locomotor disability and no functional recovery was observed for all mice (Fig. 6b). Between 1 and 6 weeks, the walking ability of mice which received AAV1 expressing wild-type MBP improved more than that of mice treated with AAV1 encoding the mutant MBP when compared to mice treated with viral particles without MBPencoding insert (Fig. 6b). Furthermore, 6 weeks after injury, we observed that the rump-height index—a measure of the ability to support body weight during ground locomotion [41]—was higher for injured mice treated with AAV1 expressing wild-type MBP when compared with injured mice which received viral particles expressing the proteolytically inactive MBP mutant or not expressing MBP (Fig. 6c). Additionally, we determined the foot-stepping angle as a parameter for plantar stepping ability [41] and observed improved recovery of stepping for mice treated with AAV1 expressing proteolytically active MBP in contrast to mice treated with AAV1 expressing proteolytically inactive MBP or treated with control virus (Fig. 6d). The combined results show that expression of functionally active MBP improves the locomotor abilities after spinal cord injury and that expression of the proteolytically inactive MBP does not benefit locomotor functions after injury, indicating that the serine protease activity of MBP is required for the beneficial effect of MBP on regeneration after injury.

To evaluate where in the spinal cord virus-derived MBP exerts its beneficial effects and whether the beneficial effects of MBP are correlated with generation of the 70 kDa L1 fragment, we determined the levels of L1 and MBP in spinal cord segments from the lesion site and from the proximal and distal part of the injured spinal cords 12 weeks after injury. Equal amounts of endogenous MBP isoforms were found in the segments from mock-treated spinal cords and in the segments after transduction with AAV1 coding for wild-type and mutant MBP (Fig. 6e). Also, similar levels of the AAV1-derived 21.5 kDa wild-type and mutant MBP were observed (Fig. 6e). However, elevated levels of the 70 kDa L1 fragment were only seen in the lesion site and the proximal and distal spinal cord segments upon AAV1-derived expression of wild-type MBP (Fig. 6e), indicating that the serine protease activity of MBP is required for the generation of the 70 kDa L1 fragment. We then determined whether AAV1-derived MBP-messenger RNA (mRNA) could diffuse or be transported along regrown axons. In situ hybridization using an exon II-specific MBP probe showed very low endogenous expression of exon II-containing MBP forms around the lesion site of injured mock-treated mice (Fig. 7a). Upon injection of AAV1 coding for wild-type 21.5 kDa MBP, a significant evenly distributed AAV1-derived expression of MBP-mRNA was seen distally and proximately as well as within the lesion site, whereas an accumulation of MBP-mRNA near the injection sites was observed upon injection of AAV1 coding for the mutated 21.5 kDa MBP (Fig. 7a). These results suggested that the AAV1-derived expression of wild-type MBP promotes axonal regrowth through the lesion site and that AAV1-derived MBPmRNA could diffuse or be transported along regrown axons leading to an overall even distribution of AVV1-derived MBP-mRNA. Since improved functional recovery correlates with enhanced axonal regrowth as well as reduced scar formation and reduced activation of microglia, we determined the number of neurofilament profiles representing axon fibers as a measure of axon regrowth distal to the lesion site, the number Fig. 7 Virus-mediated expression of proteolytically active MBP after„ spinal cord injury leads to enhancement of axonal regrowth and remyelination and reduces scar formation and microglia activation. a In situ hybridization with RNA probe for the 21.5 kDa MBP isoform (green) in injured spinal cords after treatment with AAV1 coding for proteolytically active wild-type MBP (AAV1-wt-MBP), mutant MBP (AAV1-mut-MBP), or with virus without MBP-encoding insert (AAV control). Asterisks indicate the lesion site; closed white arrowheads indicate the injection area, and rectangles highlight the magnified area. b Assessment of glial scar area, number of axon fibers, and number of activated microglia in injured spinal cords after treatment with AAV1-wtMBP or AAV1-mut-MBP (asterisks indicate resting ramified microglia; arrowheads indicate amoeboid activated microglia). Mean values±SEM relative to those observed for mock-treated spinal cords are shown. Statistical differences relative the control group are indicated (***p<0.0001, one-way ANOVA with Tukey’s multiple comparison test, n=8 mice per group). Scale bars: 100 μm

Mol Neurobiol

Mol Neurobiol

of activated microglia and estimated the area of scar tissue at the lesion site. The scar area is proportional to the scar volume and allows determination of the extent of lesion-induced scar formation. The scar area and the number of axon fibers in injured mice treated with control viral particles was similar to that seen in mice which received AAV1 coding for the mutated 21.5 kDa MBP, whereas a pronounced reduction in scar area and an increased number of axon fibers was observed upon application of AAV1 encoding wild-type 21.5 kDa MBP (Fig. 7b). In comparison to control treatment, application of AAV1 encoding mutated MBP enhanced and application of AAV1 coding for wild-type MBP reduced the number of activated microglia (Fig. 7b). These results indicate that proteolytically active MBP contributes to attenuation of tissue scarring, enhancement of axonal regrowth, and prevention of microglia activation. The Serine Protease Activity of MBP Promotes Differentiation, Synaptogenesis, and Myelination During Spinal Cord Development Since proteolytically active MBP is important for regeneration of the spinal cord after injury and regeneration may be considered, at least to some extent, to recapitulate some aspects of development, we investigated whether MBP and its proteolytic activity play a role in development of the spinal cord. AAV1 encoding proteolytically active or inactive MBP was injected in utero into the area of plexus brachialis of shiverer embryos at embryonic day 16, and spinal cords were analyzed at postnatal day 5 by immunohistological staining to determine differentiation of neuronal precursor cells, myelination of nerve fibers, and formation of synaptic contacts at the injection site (Fig. 8a–e). As control, spinal cords of wild-type and shiverer mice were analyzed after infection of viral particles not coding for MBP (Fig. 8a–e). For measurement of myelination, we determined the number of fibers positive for the myelin marker PLP and found that the number of PLP-positive fibers was higher in wild-type than in shiverer mice, particularly in the ventral part of the spinal cord (Fig. 8b, e). Virus-mediated expression of proteolytically active MBP, but not of inactive MBP, in shiverer mice led to an increase in the number of PLP-positive fibers, reaching values observed in wild-type mice (Fig. 8b, e). Of note, expression of virus-derived mutant MBP was restricted to cells in the future grey matter, whereas virus-derived wild-type MBP was also found in cells and fibers in the prospective white matter (Fig. 8e). To determine the effect of MBP on differentiation, we performed immunostaining with an antibody against Dlx-2, which controls differentiation of progenitor cells in neurons or oligodendrocytes and represents a marker for neurogenesis versus oligodendrogenesis [47]. Numbers of Dlx-2-positive cells were higher in wild-type than in shiverer mice (Fig. 8c, e). After application of AAV1

encoding proteolytically active MBP to shiverer mice, the number of Dlx-2-positive cells was as high as in wild-type mice, whereas application of AAV1 encoding proteolytically inactive MBP did not enhance the number of Dlx-2-positive cells (Fig. 8c, e). As an indicator for the number of synaptic contacts of glutamatergic excitatory neurons and thus for synaptogenesis, we counted the intensity of the VGLUT in the ventral part of the spinal cords. In wild-type mice, VGLUT intensity was higher than in shiverer mice and similar to that observed in shiverer mice after application of AAV1 encoding wild-type MBP (Fig. 8d, e). When AAV1 encoding mutant MBP was applied to shiverer mice, VGLUT intensity was not enhanced and similar to that in shiverer mice after transduction with viral particles without MBP-encoding insert (Fig. 8d, e). These results indicate an involvement of the proteolytically active MBP in neuronal differentiation, myelination and synaptogenesis in the developing spinal cord, suggesting that MBP and its protease activity is required for proper development of the spinal cord.

Discussion MBP is known as one of the major players in formation and maintenance of myelin sheaths and as a major structural component of compact myelin in the central and peripheral nervous systems. In previous studies [24, 28] and in the present study, we provide evidence that MBP acts as a serine protease that cleaves L1 in vitro and in vivo. The MBP-mediated proteolytic cleavage of L1 yields a transmembrane 70 kDa L1 fragment which regulates cellular functions, such as neurite outgrowth, neuroprotection against oxidative stress, and excitotoxicity as well as Schwann cell proliferation, migration, and process formation [present study; 23, 24, 28]. These MBP-mediated and L1-dependent cellular events are not only important for the development of the nervous system but also play an important role in regeneration after injury of the adult peripheral and central nervous systems. Here, we observed an improvement in functional recovery which correlates with increased 70 kDa L1 fragment levels after application of MBP to the injured femoral nerve. Conversely, we also observed an impaired recovery and reduced 70 kDa L1 fragment levels when MBP antibody was applied to the injured nerve. Moreover, an improvement in functional recovery and increased levels of the 70 kDa L1 fragment were also found after application of viral particles encoding proteolytically active MBP to the lesioned spinal cord, whereas application of viral particles encoding proteolytically inactive MBP had no effect on functional recovery. These in vivo observations indicate that proteolytic cleavage of L1 by MBP and the MBP-mediated generation of the 70 kDa L1 fragment play pivotal roles for regeneration after injury in the adult. Application of

Mol Neurobiol

Fig. 8 Serine protease activity of MBP is required for myelination, cell differentiation, and synaptogenesis during spinal cord development. a In utero injection of AAV1 coding for proteolytically active wild-type MBP (AAV1-wt-MBP) or inactive mutant MBP (AAV1-mut-MBP) in the area of plexus brachialis of shiverer (shi) embryos at embryonic day 16 (E16). Thoracic trunks were isolated at postnatal day 5 (P5) and transversally sectioned for immunohistological staining. Wild-type (WT) and shiverer (shi) littermate embryos were used as controls. b–d Numbers of PLPpositive fibers (b), numbers of Dlx-2-positive cells (c), and VGLUT-

intensity (d) in thoracic spinal cord section were determined in three independent experiments by analyzing 10 sections of three mice per condition. Mean values±SEM relative to those observed for mocktreated shiverer embryos are shown. Statistical differences are indicated (***p<0.0005, one-way ANOVA with Tukey’s multiple comparison test). e Representative images from immunohistological staining of thoracic spinal cord sections are shown. Asterisks indicate the ventral side of the spinal cord. Scale bars: 100 μm

proteolytically active MBP leads to reduced glial scar volumes, enhanced numbers of axon fibers, and attenuated microglial activation. These morphological parameters correlate with enhanced regeneration and are similar to those underlying regeneration after application of recombinant L1 or L1 expressing cells [4–9]. In femoral nerves, MBP

improves remyelination correlating with functional recovery. These observations are in agreement with findings on improved remyelination after femoral nerve injury in mice lacking functional T- and B-lymphocytes [48], in mice treated with mimetics for polysialic acid [38] or in mice overexpressing L1 [45].

Mol Neurobiol

We also provide evidence that proteolytically active MBP is required for neuronal differentiation, myelination, and synaptogenesis in spinal cord development. It is, therefore, conceivable that the MBP-generated 70 kDa L1 fragment regulates cell interactions during development. It is also possible that alterations in the MBP-mediated generation of the 70 kDa L1 fragment play a role in neurodegeneration and aging of the nervous system. Altered levels of the 70 kDa L1 fragment in a mouse model of Alzheimer’s disease [28] suggest that the MBP-generated L1 fragment is implicated in the pathogenesis of this neurodegenerative disease. Using AAV encoding L1 in the Alzheimer’s disease mouse model, we observed that AAV-mediated L1 overexpression in neurons and glia reduces several histopathological parameters of this disease, such as amyloid-β plaque load [49], suggesting that L1 has a neuroprotective role. The importance of MBP in myelination during development and in degenerative diseases associated with demyelination, such as multiple sclerosis [50], is not only underscored by the lack of CNS myelin in MBP-deficient shiverer mice but also by the finding that MBP and a higher serine protease activity of MBP [50, 51] is associated with the pathogenesis of multiple sclerosis [52, 53] which is characterized by demyelination. Based on these findings and the results of our in vitro and in vivo experiments, we propose that proteolytic cleavage of L1 by MBP and the MBP-mediated generation of the 70 kDa L1 fragment are important for nervous system development and for regeneration after injury in an adult mammal. We infer that impaired or enhanced MBP-mediated generation of the L1 fragment could be associated with L1-linked diseases, such as the L1 syndrome, fetal alcohol syndrome, Hirschsprung’s disease, and schizophrenia as well as with degenerative diseases, other than multiple sclerosis and Alzheimer’s disease. Moreover, enhanced expression of L1 and dysregulated cleavage of L1 by MBP may also contribute to tumorigenesis and metastasis. Acknowledgments We are very grateful to David Colman for the gift of antibody against the exon II-encoded domain of mouse MBP. We appreciate the excellent technical assistance from Ute Bork, Emanuela Szpotowicz, Dagmar Drexler, Barbara Holstermann, Fritz Kutschera and Torsten Renz, Eva Kronberg and Ulrike Wolters for breeding and maintenance of mice, and Gabriele Rune for the possibility to use the electron microscope. Bin Wu was supported by the National Natural Science Foundation of China (81000520). Melitta Schachner is supported by the New Jersey Commission for Spinal Cord Research and the Li Ka-Shing Foundation at Shantou University Medical College. Conflict of Interest The authors declare that they have no competing interests.

2.

3. 4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

References 18. 1.

Maness PF, Schachner M (2007) Neural recognition molecules of the immunoglobulin superfamily: signaling transducers of axon guidance and neuronal migration. Nat Neurosci 10(1):19–26

Schmid RS, Maness PF (2008) L1 and NCAM adhesion molecules as signaling coreceptors in neuronal migration and process outgrowth. Curr Opin Neurobiol 18(3):245–250 Loers G, Schachner M (2007) Recognition molecules and neural repair. J Neurochem 101(4):865–882 Roonprapunt C, Huang W, Grill R, Friedlander D, Grumet M, Chen S, Schachner M, Young W (2003) Soluble cell adhesion molecule L1-Fc promotes locomotor recovery in rats after spinal cord injury. J Neurotrauma 20(9):871–882 Barbin G, Aigrot MS, Charles P, Foucher A, Grumet M, Schachner M, Zalc B, Lubetzki C (2004) Axonal cell-adhesion molecule L1 in CNS myelination. Neuron Glia Biol 1(1):65–72 Chen J, Wu J, Apostolova I, Skup M, Irintchev A, Kügler S, Schachner M (2007) Adeno-associated virus-mediated L1 expression promotes functional recovery after spinal cord injury. Brain 130(Pt 4):954–969 Lavdas AA, Chen J, Papastefanaki F, Chen S, Schachner M, Matsas R, Thomaidou D (2010) Schwann cells engineered to express the cell adhesion molecule L1 accelerate myelination and motor recovery after spinal cord injury. Exp Neurol 221(1):206–216 Cui YF, Xu JC, Hargus G, Jakovcevski I, Schachner M, Bernreuther C (2011) Embryonic stem cell-derived L1 overexpressing neural aggregates enhance recovery after spinal cord injury in mice. PLoS One 6(3), e17126 Xu JC, Bernreuther C, Cui YF, Jakovcevski I, Hargus G, Xiao MF, Schachner M (2011) Transplanted L1 expressing radial glia and astrocytes enhance recovery after spinal cord injury. J Neurotrauma 28(9):1921–1937 Poltorak M, Khoja I, Hemperly JJ, Williams JR, el-Mallakh R, Freed WJ (1995) Disturbances in cell recognition molecules (NCAM and L1 antigen) in the CSF of patients with schizophrenia. Exp Neurol 131(2):266–272 Kurumaji A, Nomoto H, Okano T, Toru M (2001) An association study between polymorphism of L1CAM gene and schizophrenia in a Japanese sample. Am J Med Genet 105(1):99–104 Strekalova H, Buhmann C, Kleene R, Eggers C, Saffell J, Hemperly J, Weiller C, Müller-Thomsen T et al (2006) Elevated levels of neural recognition molecule L1 in the cerebrospinal fluid of patients with Alzheimer disease and other dementia syndromes. Neurobiol Aging 27(1):1–9 Wakabayashi Y, Uchida S, Funato H, Matsubara T, Watanuki T, Otsuki K, Fujimoto M, Nishida A et al (2008) State-dependent changes in the expression levels of NCAM-140 and L1 in the peripheral blood cells of bipolar disorders, but not in the major depressive disorders. Prog Neuropsychopharmacol Biol Psychiatry 32(5):1199–1205 Schäfer MK, Altevogt P (2010) L1CAM malfunction in the nervous system and human carcinomas. Cell Mol Life Sci 67(14): 2425–2437 Sadoul K, Sadoul R, Faissner A, Schachner M (1988) Biochemical characterization of different molecular forms of the neural cell adhesion molecule L1. J Neurochem 50(2):510–521 Nayeem N, Silletti S, Yang X, Lemmon VP, Reisfeld RA, Stallcup WB, Montgomery AM (1999) A potential role for the plasmin(ogen) system in the posttranslational cleavage of the neural cell adhesion molecule L1. J Cell Sci 112(Pt24):4739–4749 Silletti S, Mei F, Sheppard D, Montgomery AM (2000) Plasmin-sensitive dibasic sequences in the third fibronectinlike domain of L1-cell adhesion molecule (CAM) facilitate homomultimerization and concomitant integrin recruitment. J Cell Biol 149(7):1485–1502 Mechtersheimer S, Gutwein P, Agmon-Levin N, Stoeck A, Oleszewski M, Riedle S, Postina R, Fahrenholz F et al (2001) Ectodomain shedding of L1 adhesion molecule promotes cell migration by autocrine binding to integrins. J Cell Biol 155:661–673

Mol Neurobiol 19.

20.

21.

22.

23.

24.

25.

26.

27. 28.

29.

30.

31.

32.

33.

34.

35.

36.

Kalus I, Schnegelsberg B, Seidah NG, Kleene R, Schachner M (2003) The proprotein convertase PC5A and a metalloprotease are involved in the proteolytic processing of the neural adhesion molecule L1. J Biol Chem 278(12):10381–10388 Matsumoto-Miyai K, Ninomiya A, Yamasaki H, Tamura H, Nakamura Y, Shiosaka S (2003) NMDA-dependent proteolysis of presynaptic adhesion molecule L1 in the hippocampus by neuropsin. J Neurosci 23(21):7727–7736 Maretzky T, Schulte M, Ludwig A, Rose-John S, Blobel C, Hartmann D, Altevogt P, Saftig P et al (2005) L1 is sequentially processed by two differently activated metalloproteases and presenilin/gamma-secretase and regulates neural cell adhesion, cell migration, and neurite outgrowth. Mol Cell Biol 25(20):9040–9053 Riedle S, Kiefel H, Gast D, Bondong S, Wolterink S, Gutwein P, Altevogt P (2009) Nuclear translocation and signalling of L1-CAM in human carcinoma cells requires ADAM10 and presenilin/ gamma-secretase activity. Biochem J 420(3):391–402 Lutz D, Wolters-Eisfeld G, Schachner M, Kleene R (2014) Cathepsin E generates a sumoylated intracellular fragment of the cell adhesion molecule L1 to promote neuronal and Schwann cell migration as well as myelination. J Neurochem 128(5):713–724 Lutz D, Loers G, Kleene R, Oezen I, Kataria H, Katagihallimath N, Braren I, Harauz G et al (2014) Myelin basic protein cleaves cell adhesion molecule L1 and promotes neuritogenesis and cell survival. J Biol Chem 289(19):13503–13518 Readhead C, Takasashi N, Shine HD, Saavedra R, Sidman R, Hood L (1990) Role of myelin basic protein in the formation of central nervous system myelin. Ann N Y Acad Sci 605:280–285 Readhead C, Hood L (1990) The dysmyelinating mouse mutations shiverer (shi) and myelin deficient (shimld). Behav Genet 20(2): 213–234 Boggs JM (2006) Myelin basic protein: a multifunctional protein. Cell Mol Life Sci 63(17):1945–1961 Lutz D, Wolters-Eisfeld G, Joshi G, Djogo N, Jakovcevski I, Schachner M, Kleene R (2012) Generation and nuclear translocation of a sumoylated transmembrane fragment of the cell adhesion molecule L1. J Biol Chem 287(21):17161–17175 Dahme M, Bartsch U, Martini R, Anliker B, Schachner M, Mantei N (1997) Disruption of the mouse L1 gene leads to malformations of the nervous system. Nat Genet 17(3):346–349 Mikoshiba K, Takamatsu K, Tsukada Y (1983) Peripheral nervous system of shiverer mutant mice: developmental change of myelin components and immunohistochemical demonstration of the absence of MBP and presence of P2 protein. Brain Res 283(1):71–79 Wolf MK, Billings-Gagliardi S (1984) CNS hypomyelinated mutant mice (jimpy, shiverer, quaking): in vitro evidence for primary oligodendrocyte defects. Adv Exp Med Biol 181:115–133 Appel F, Holm J, Conscience JF, von Bohlen und Halbach F, Faissner A, James P, Schachner M (1995) Identification of the border between fibronectin type III homologous repeats 2 and 3 of the neural cell adhesion molecule L1 as a neurite outgrowth promoting and signal transducing domain. J Neurobiol 28(3):297–312 Pedraza L, Fidler L, Staugaitis SM, Colman DR (1997) The active transport of myelin basic protein into the nucleus suggests a regulatory role in myelination. Neuron 18(4):579–589 Kleene R, Yang H, Kutsche M, Schachner M (2001) The neural recognition molecule L1 is a sialic acid-binding lectin for CD24, which induces promotion and inhibition of neurite outgrowth. J Biol Chem 276(24):21656–21663 Makhina T, Loers G, Schulze C, Ueberle B, Schachner M, Kleene R (2009) Extracellular GAPDH binds to L1 and enhances neurite outgrowth. Mol Cell Neurosci 41(2):206–218 Simova O, Irintchev A, Mehanna A, Liu J, Dihné M, Bächle D, Sewald N, Loers G et al (2006) Carbohydrate mimics promote functional recovery after peripheral nerve repair. Ann Neurol 60(4):430–437

37.

Lieberoth A, Splittstoesser F, Katagihallimath N, Jakovcevski I, Loers G, Ranscht B, Karagogeos D, Schachner M et al (2009) Lewis(x) and alpha2,3-sialyl glycans and their receptors TAG-1, Contactin, and L1 mediate CD24-dependent neurite outgrowth. J Neurosci 29(20):6677–6690 38. Mehanna A, Mishra B, Kurschat N, Schulze C, Bian S, Loers G, Irintchev A, Schachner M (2009) Polysialic acid glycomimetics promote myelination and functional recovery after peripheral nerve injury in mice. Brain 132(Pt 6): 1449–1462 39. Curtis R, Green D, Lindsay RM, Wilkin GP (1993) Up-regulation of GAP-43 and growth of axons in rat spinal cord after compression injury. J Neurocytol 22(1):51–64 40. Basso DM, Fisher LC, Anderson AJ, Jakeman LB, McTigue DM, Popovich PG (2006) Basso Mouse Scale for locomotion detects differences in recovery after spinal cord injury in five common mouse strains. J Neurotrauma 23(5):635–659 41. Apostolova I, Irintchev A, Schachner M (2006) Tenascin-R restricts posttraumatic remodeling of motoneuron innervation and functional recovery after spinal cord injury in adult mice. J Neurosci 26(30): 7849–7859 42. Chen S, Mantei N, Dong L, Schachner M (1999) Prevention of neuronal cell death by neural adhesion molecules L1 and CHL1. J Neurobiol 38(3):428–439 43. Loers G, Chen S, Grumet M, Schachner M (2005) Signal transduction pathways implicated in neural recognition molecule L1 triggered neuroprotection and neuritogenesis. J Neurochem 92(6): 1463–1476 44. Guseva D, Angelov DN, Irintchev A, Schachner M (2009) Ablation of adhesion molecule L1 in mice favours Schwann cell proliferation and functional recovery after peripheral nerve injury. Brain 132(Pt 8):2180–2195 45. Guseva D, Zerwas M, Xiao MF, Jakovcevski I, Irintchev A, Schachner M (2011) Adhesion molecule L1 overexpressed under the control of the neuronal Thy-1 promoter improves myelination after peripheral nerve injury in adult mice. Exp Neurol 229(2):339–352 46. Irintchev A, Simova O, Eberhardt KA, Morellini F, Schachner M (2005) Impacts of lesion severity and tyrosine kinase receptor B deficiency on functional outcome of femoral nerve injury assessed by a novel single-frame motion analysis in mice. Eur J Neurosci 22(4):802–808 47. Petryniak MA, Potter GB, Rowitch DH, Rubenstein JL (2007) Dlx1 and Dlx2 control neuronal versus oligodendroglial cell fate acquisition in the developing forebrain. Neuron 55(3): 417–433 48. Mehanna A, Szpotowicz E, Schachner M, Jakovcevski I (2014) Improved regeneration after femoral nerve injury in mice lacking functional T- and B-lymphocytes. Exp Neurol 261:147–155 49. Djogo N, Jakovcevski I, Müller C, Lee HJ, Xu JC, Jakovcevski M, Kügler S, Loers G et al (2013) Adhesion molecule L1 binds to amyloid beta and reduces Alzheimer’s disease pathology in mice. Neurobiol Dis 56:104–115 50. D’Souza CA, Wood DD, She YM, Moscarello MA (2005) Autocatalytic cleavage of myelin basic protein: an alternative to molecular mimicry. Biochemistry 44(38):12905–12913 51. Liao MC, Ahmed M, Smith SO, Van Nostrand WE (2009) Degradation of amyloid beta protein by purified myelin basic protein. J Biol Chem 284(42):28917–28925 52. Musse AA, Harauz G (2007) Molecular Bnegativity^ may underlie multiple sclerosis: role of the myelin basic protein family in the pathogenesis of MS. Int Rev Neurobiol 79:149–172 53. Derfuss T, Meinl E (2012) Identifying autoantigens in demyelinating diseases: valuable clues to diagnosis and treatment? Curr Opin Neurol 25(3):231–238