- Email: [email protected]
In Vivo Immunological Suppression of Spinal Cord Myelin Development
Brain Research Bulletin, Vol. 44, No. 6, pp. 727–734, 1997 Copyright © 1997 Elsevier Science Inc. Printed in the USA. All rights reserved 0361-9230/97 $17.00 1 .00
PII S0361-9230(97)00374-2
In Vivo Immunological Suppression of Spinal Cord Myelin Development H. S. KEIRSTEAD,1 D. M. PATAKY, J. McGRAW AND J. D. STEEVES CORD (Collaboration On Repair Discoveries), Departments of Zoology, Anatomy, and Surgery, University of British Columbia, Vancouver, B.C., CANADA V6T 1Z4 [Received 5 June 1997; Revised 11 September 1997; Accepted 15 September 1997] ABSTRACT: The onset of myelination in the embryonic chick spinal cord begins on embryonic day (E) 12 or E13 of the 21 day in ovo developmental period. This event coincides with a loss of functional axonal regeneration following complete transection of the thoracic spinal cord. In this study, we have characterised an immunological method for delaying the developmental onset of myelination in vivo until later stages of development (developmental myelin-suppression). A single injection of heterologous or homologous serum complement proteins plus myelinspecific, complement-binding antibodies into the spinal cord prior to E13 delayed the onset of myelination until E17. The state of spinal cord myelin was assessed with immunohistochemical, histological and ultrastructural techniques. Northern blot analysis indicated that myelin basic protein mRNA was not downregulated in myelin-suppressed spinal cords, which suggests that oligodendrocytes survived developmental myelin-suppression. Glial fibrillary acidic protein immunostaining of normal and treated tissue indicated that myelin-suppression did not alter the resident astrocyte population of the spinal cord or elicit astrogliosis. Immunostaining with microtubule-associated protein-2 and thionine staining of normal and myelin-suppressed tissue further indicated that the neuronal architecture was unaffected by the immunological protocol. © 1997 Elsevier Science Inc.
therapies to promote functional axonal regeneration following spinal cord injury. Experimental methods for in vivo myelin-suppression include: 1) X-ray irradiation [2]; 2) exposure to drugs such as ethidium bromide [19], lysolecithin [3], tellurium [21], cuprizone [33] or diptheria toxin [14]; 3) viruses such as Theiler’s virus [9] or the JMH strain of mouse hepatitis virus [25]; 4) nerve compression [5,9]; and 5) exposure to sera from animals afflicted with experimental allergic encephalomyelitis [31,39]. Although effective in suppressing myelin, none of the above methods appear to be specific for myelin alone. However, myelin-suppression initiated by complement-binding antibodies directed against CNS myelin is highly specific in vitro [11,12,14,18,26,27], although in vivo studies in the optic nerve [17,38,43] and spinal cord [35] report equivocal results with regards to cellular specificity. In the present study, we have examined the ability of two complement-binding, oligodendrocyte-specific antibodies to suppress the developmental onset of myelination in the chick embryo. The onset of myelination in the developing chick spinal cord begins around embryonic day (E) 13 of the 21 day developmental period [1,22,29,34]. The first appearance of oligodendrocytes in the chick precedes the initial formation of myelin by 2–3 embryonic days [1] and is characterised by the expression of galactocerebroside (GalC), an abundant sphingolipid produced by oligodendrocytes [40]. GalC is highly conserved across species [40]. The oligodendrocyte-specific antigen recognised by the 04 antibody is also expressed prior to the onset of myelination [47]. We have found that a single injection of serum complement proteins plus (monoclonal or polyclonal) GalC or 04 antibodies into the thoracic spinal cord of an E9 –E12 chick embryo (i.e. around the time of oligodendrocyte differentiation) resulted in a delay in the onset of spinal cord myelination until E17 (developmental myelin-suppression). Myelin basic protein (MBP) mRNA was not downregulated in myelin-suppressed spinal cords, which suggests that oligodendrocytes survived developmental myelin-suppression. Neither astrocytic or neuronal populations of the spinal cord were detectably disturbed by this intervention.
KEY WORDS: Demyelination, Remyelination, Galactocerebroside, 04 antibody, Oligodendrocyte.
INTRODUCTION Central nervous system (CNS) myelin is a highly differentiated membrane structure produced by oligodendrocytes. We have previously reported that CNS myelin is inhibitory to neuroanatomical repair and recovery of physiological activity following complete transection of late embryonic or hatchling chick spinal cord [28,29]. A few CNS myelin proteins have been identified as potential inhibitors of axonal regeneration, including NI-35, NI250 [6,44] and myelin-associated glycoprotein [36,37]. Implantation of IN-1 antibody-secreting hybridoma cells into the parietal cortex of adult rats promotes limited axonal regeneration by severed corticospinal neurons [4,44]. Given that CNS myelin has an inhibitory effect on the regenerative ability of previously severed axons, further characterisation of in vivo methods for transiently suppressing CNS myelin should provide insight into potential 1
MATERIALS AND METHODS Fertilized White Leghorn eggs were incubated at 37°C in an automatic rotating incubator (Peacock Equipment Limited, Aldergrove, B.C.) and staged using the protocol outlined by Hamburger
To whom requests for reprints should be addressed: Presently at MRC Cambridge Centre for Brain Repair, Robinson Way, Cambridge, U.K. CB2 2PY.
727
728 and Hamilton [20]. After surgery, eggs were sealed with paraffin wax and a sterile coverslip then returned to the incubator. Developmental Myelin-Suppression Thoracic spinal cord injections were performed at E9 –E12 in chick embryos using a glass micropipette (tip diameter 5 30 – 40 mm; A-M Systems, Everett, Washington #6045) connected to a Picospritzer II pressure injection system (General Valve Corp., Fairfield, New Jersey). Developmental myelin-suppression was evoked by injecting either; 1) an IgG mouse galactocerebroside (GalC) antibody (a gift from B. Ranscht; 2.67 mg/ml hybridoma supernatant) at a dilution of 1:25 with 20% homologous serum (as a source of complement) in 0.1M phosphate buffered saline (PBS), pH 7.4 (providing an effective concentration of 63.0 ng of GalC hybridoma supernatant injected per gram body weight) or, 2) an IgG rabbit GalC antibody (Chemicon International Inc., Temecula, California #AB142) at a dilution of 1:25 with 20% heterologous (guinea pig) complement (Gibco BRL, Burlington, Ontario #19195-015) in 0.1M PBS pH 7.4 or, 3) an IgM polyclonal 04 antibody [47] at a dilution of 1:25 with 20% guinea pig complement in 0.1M PBS pH 7.4. Each intraspinal injection was performed with a new glass micropipette backloaded with 2–3 ml of solution immediately prior to surgery. Each animal received a total volume of 2–3 ml, over 1– 4 penetrations, injected directly into the mid-to-high thoracic spinal cord. It is conceivable that cells other than oligodendrocytes may non-specifically bind the GalC or 04 antibodies via Fc receptors on the cell surface and indirectly affect myelination. To assess the possibility of non-specific binding, control embryos were similarly injected with 20% homologous serum complement plus an antihuman antibody that does not cross-react with chicken. We chose a monoclonal IgG antibody to glial fibrillary acidic protein (GFAP; Amersham), a major constituent of astrocytes within the CNS. Other immunological control embryos received injections of either: 1) the monoclonal or polyclonal GalC antibody only, 2) homologous serum complement proteins only, 3) guinea pig complement only, 4) vehicle only (0.1M PBS, pH 7.4), or 5) the monoclonal GalC antibody plus homologous complement, following heat-inactivation of the complement by exposure to 50°C for 30 min. Animals were perfused intracardially at the appropriate developmental stage (see results) with 0.1M PBS containing 2500 USP units of heparin per 50 mls PBS, pH 7.4 (at 37°C) followed by perfusion with 4% paraformaldehyde in 0.1M phosphate buffer, pH 7.4 (at 4 –10°C). The dissected spinal cords were then immersed in the same fixative for 24 hr at 4°C and subsequently stored in PBS, pH 7.4 (at 4 –10°C) for further processing. Immunohistochemistry Antigens were localised using indirect immunofluorescence. GalC immunohistochemistry was performed on cryostat sectioned tissue from unoperated control embryos that were perfused as outlined above. All other antigens were localised on paraffinembedded tissue sections. The rabbit antihuman MBP antibody (Accurate Chemical Scientific Corp., #AXL746), the mouse antipig GFAP (Sigma Immunochemicals, #G-3893) and the mouse anticalf microtubule associated protein-2 (MAP-2; Amersham, #RPN 1194) were all used at a dilution of 1:100. The mouse anti-chicken myelin-associated glycoprotein (MAG; Boehringer Mannheim #1450 972) and the rabbit antibovine 29,39-cyclic nucleotide 39-phosphodiesterase (CNP; a gift from Peter Braun) were used at a dilution of 1:500. The rabbit antibovine GalC antibody (Chemicon AB 142) was used at a dilution of 1:10. The secondary antibody was a goat anti-rabbit FITC conjugated immunoglobulin
KEIRSTEAD ET AL. (Caltag Laboratories, #L42001) diluted 1:100 or a goat anti-mouse FITC conjugated immunoglobulin (Dimension Labs, #30112663) diluted 1:100. Standard immunocytochemical controls (e.g., omission of primary and/or secondary antibodies) were processed alongside tissue sections from experimental and control animals. Photomicrographs were taken on a Zeiss Axiophot using epifluorescent illumination with the appropriate filters. Histological Staining Animals designated for toluidine blue staining were perfused intracardially at the appropriate developmental stage (see results) with 0.1M PBS containing 2500 USP units of heparin per 50 mls PBS, pH 7.4 (at 37°C) followed by perfusion with 2.5% glutaraldehyde in 0.1M phosphate buffer, pH 7.4 (at 4 –10°C). The dissected tissue was then dehydrated and resin embedded according to standard protocols. The hardened blocks were cut into 1 mm transverse sections on an ultra-microtome and placed on glass slides for toluidine blue staining. 0.1% toluidine blue stain was filtered and dropped onto sections for two minutes. The slides were then rinsed with hot tap water followed by distilled water, then placed on a hot plate to dry. Photomicrographs were taken on a Zeiss Axiophot microscope. Electron Microscopy Animals designated for toluidine blue staining were perfused intracardially at the appropriate developmental stage (see results) with 4% glutaraldehyde in 0.1M phosphate buffer (PB) pH 7.4. The lesion-containing length of spinal cord was cut into 1 mm transverse blocks and processed so as to preserve the cranio-caudal sequence and orientation. The tissue blocks were rinsed in 0.1M phosphate buffer pH 7.4 for 30 minutes, post-fixed in 2% OsO4, dehydrated in ascending alcohols, and embedded in TAAB resin. Thin sections (1 mm) were cut from each block, stained with alkaline Toluidine blue and examined by light microscopy. For electron microscopic analysis blocks were trimmed then cut at 100nm, mounted on copper grids, uranyl acetate and lead citrate stained, and viewed under a Hitachi EM 600 electron microscope at 75 kV. Northern Blotting Tissue was harvested from chicks at various stages of development beginning on E10 through post-hatching day (P) 12. Spinal cord myelin-suppression was begun on E11, as described above. Animals were sacrificed 2 to 10 days later. All northern blots were performed in triplicate. Total RNA from the thoracic and caudal two thirds of the cervical spinal cord was isolated by a modified version [48] of the method of Chomczynski and Sacchi (1987) [8]. Briefly, the tissue was homogenized in GAP (Solution A: 4M guanidinium isothiocyanate, 25mM sodium citrate, 0.5% sarkosyl and freshly added 0.1M 2-mercaptoethanol; plus H2O-saturated phenol and 2M NaOAc 1:1:0.1). 1/10 volume of chloroform: isoamyl alcohol (1:0.1) was added, the mixture left on ice for 15–30 min. The organic and aqueous phases were separated by centrifugation at 12,000xg at 4°C for 20 min. The aqueous phase was recovered and the RNA precipitated with an equal volume of isopropanol, centrifuged as above and resuspended in TE buffer. Following densitometric quantification of the amount of RNA present, the RNA was denatured and electrophoretically size-separated on 1.5% agarose/formaldehyde denaturing gels. An RNA ladder (Gibco BRL) was run with the initial samples to confirm the length of the message detected by the cRNA probe. After electrophoresis, the samples were capillary blotted overnight to Nytran membrane, then baked and stored until hybridisation at 220°C.
SUPPRESSION OF SPINAL CORD MYELIN DEVELOPMENT Northern hybridisations were performed following standard protocols [42]. Prehybridisation and hybridisation were performed at 65°C in 5x PIPES buffer with 50% formamide, 5x Denhardt’s, and 100mg/ml each of denatured salmon sperm DNA, herring sperm DNA and yeast tRNA. The cDNA for chicken MBP (a gift from Dr. Peter Jeffrey, GenBank accession #CHKMBP) was directionally subcloned into pBluescript (Stratagene), linearised and used as a template for the in vitro transcription (Promega) of high specific activity (. 109 cpm/mg) 32P-labelled cRNA probes. Blots were then rinsed at increasing stringency to 0.1xSSC at 65°C, then exposed to pre-flashed Kodak XAR X-ray film. To demonstrate the amount of RNA in each lane, blots were stripped in 80% formamide at 80°C for 1 hour, then re-hybridised as above using either a chicken actin (a gift of Dr. P. Gunning, clone #LK154, subcloned into pBluescript) or rat cyclophilin (a gift of Dr. James Douglas, subcloned into pBluescript) cRNAs as above. Behavioural Assessments Locomotor behaviour of hatchlings previously myelin-suppressed was assessed on P2, P4 and P6, prior to sacrifice for immunohistochemical analysis. The ability to walk, run, effect postural adjustments and righting responses was visually assessed for each animal and scored as ‘‘comparable to unoperated controls’’ or ‘‘not comparable to unoperated controls’’. RESULTS Myelin Development in the Chick GalC expression is an early indicator of oligodendrocyte differentiation from progenitor cells [40]. GalC immunoreactivity (results not shown) was not detected at any level of the chick spinal cord on E8 (n 5 3) or E9 (n 5 2), however, GalC was first detected in 4 of 6 animals sacrificed on E10. GalC immunoreactivity in these animals was punctate, and located in the lateral regions of the white matter tracts. In four animals analysed on E11, GalC immunoreactivity was similarly organised. GalC immunoreactivity was present throughout the spinal cord white matter on E12 (n 5 3), E13 (n 5 4), E14 (n 5 3) and E15 (n 5 5). GalC immunostaining of P1, P3, P7 and P10 spinal cords (n 5 6 for each day) was qualitatively similar to each other, suggesting that oligodendrocyte differentiation and myelination are complete by hatching. CNP immunoreactivity was not detected at any level of the spinal cord on E10 (n 5 3) or E11 (n 5 3), but was detected in the spinal cord on E12.5 (n 5 6). CNP immunoreactivity was present throughout the spinal cord white matter on E13 (n 5 4), E14 (n 5 8), and E15 (n 5 8). CNP immunostaining of P1, P3, P5 and P10 spinal cords (n 5 6 for each day) was similar at all stages. MBP or MAG immunoreactivity was not detected at any level of the spinal cord on E9 (n 5 6 for MBP), E10 (n 5 6 for MBP, n 5 4 for MAG) or E11 (n 5 12 for MBP, n 5 4 for MAG). On E12 MBP immunoreactivity was highly variable (n 5 12). In the majority of cases (n 5 8 of 12), MBP was absent on E12 (Fig. 1A). MAG was consistently absent on E12 (n 5 4 for MAG). MBP and MAG immunoreactivity were reliably detected within the ventrolateral funiculi of the cervical spinal cord on E13 and proceeded in a ventro-dorsal and rostral-caudal direction during subsequent development (n 5 14 for MBP, Fig. 1B; n 5 6 for MAG). By E14, all levels of the spinal cord displayed MBP, MAG and CNP immunoreactivity. The ventral funiculi were markedly more immunoreactive than other white matter regions (n 5 18 for MBP, n 5 12 for MAG). On E15 a dense network of immunoreactivity was observed within all spinal cord white matter tracts (n 5 28 for MBP, n 5 8 for MAG). Toluidine Blue staining (results not
729 shown) and electron microscopic analysis (Fig. 1C) of E15 tissue in transverse section revealed compact myelin profiles in all regions of white matter (n 5 4). A dramatic increase in MBP and MAG immunoreactivity was observed in animals sacrificed on E17 and E18, suggesting an increase in myelination at this stage of development (n 5 18 for MBP at each stage, n 5 6 for MAG at each stage). The MBP and MAG immunostaining on P1, P3, P5, P7 and P10 was qualitatively similar, once again suggesting that myelination is complete by the time of hatching (n 5 6 for each day, for each of MBP and MAG; Fig. 1D). Developmental Myelin-Suppression in the Embryonic Chick The developmental onset of myelination in the embryonic chick spinal cord was delayed by pressure injection of homologous or heterologous serum complement proteins plus either monoclonal GalC antibodies, polyclonal GalC antibodies or polyclonal 04 antibodies directly into the thoracic spinal cord, 1– 4 days prior to the normal developmental onset of myelination (E13). Monoclonal GalC antibodies, polyclonal GalC antibodies and polyclonal 04 antibodies were selected for their abilities to fix complement and specifically recognise myelin or oligodendrocyte cell-surface antigens. Thoracic spinal cord injections of monoclonal GalC antibodies, plus either chick serum or guinea pig complement, delayed the developmental onset of spinal cord myelination until E17 (Fig. 2). Immunohistochemical assessments of myelin-suppressed spinal cords (previously injected on E11) sacrificed on either E12 (n 5 6), E13 (n 5 6), E14 (n 5 6), E15 (n 5 12) or E16 (n 5 6) revealed a complete and comparable (i.e., no inter-animal variability) lack of MBP immunoreactivity throughout the spinal cord, excluding the most rostral 1– 4 cervical segments. Thus, the effect of the immunological myelin-suppression protocol appears to be very rapid. The extent and degree of developmental myelin-suppression on E15 was similar when the complement proteins and monoclonal GalC antibodies were injected on either E9, E10, E11 or E12 (n 5 3 for each day). Likewise, thoracic spinal cord injections of polyclonal GalC antibodies plus guinea pig complement on E11 also delayed the developmental onset of spinal cord myelination until E17 (Fig. 2). Immunohistochemical assessments of myelin-suppressed spinal cords on E12 (n 5 4), E13 (n 5 4), E14 (n 5 4), E15 (n 5 5) and E16 (n 5 4) showed a complete and comparable (i.e. no interanimal variability) lack of MBP immunoreactivity throughout the spinal cord, except for the most rostral 1– 4 cervical segments. Immunohistochemical assessments of myelin-suppressed spinal cords on E13 (n 5 3), E14 (n 5 3) and E15 (n 5 3) also showed a complete lack of MAG immunoreactivity throughout the spinal cord excluding the most rostral 1– 4 cervical segments. Immunohistochemical assessments for CNP immunoreactivity in myelinsuppressed spinal cords on E13 (n 5 3), E14 (n 5 3) and E15 (n 5 3) did not reveal any CNP immunoreactivity within the cord, except for the most rostral 1– 4 cervical segments. All the immunohistochemical markers used in this study indicated the immunological protocol delayed the development of spinal cord myelin. Five spinal cords injected on E11 with polyclonal GalC antibodies plus guinea pig complement were sacrificed on E15, resin embedded and transversely microsectioned for electron microscopic analysis. Electron microscopic analysis revealed a lack of myelin (Fig. 2B). In all instances only a very few myelinated axons were visible throughout the spinal cord white matter. Unoperated E15 control tissue (n 5 5) processed in a similar manner contained an abundance of tightly compacted myelin seen as concentric rings surrounding axon profiles (Fig. 1C).
730
KEIRSTEAD ET AL.
FIG. 1. Developmental pattern of myelination in the chick thoracic spinal cord. (A) MBP immunofluorescence staining of unoperated control E12 spinal cord in parasagittal section showing no MBP immunoreactivity. (B) MBP immunofluorescence staining of unoperated control E13 spinal cord in parasagittal section showing the developmental onset of MBP immunoreactivity. (C) Electron micrograph of a transverse section of unoperated E15 spinal cord showing cross-sectional profile of myelinated axons. (D) MBP immunofluorescence staining of unoperated control P5 spinal cord in parasagittal section showing fully developed pattern of spinal cord myelination. (Bars 5 100mm for A, 50mm for B,D; X15000 for C.)
Finally, thoracic intraspinal injections of polyclonal 04 antibodies plus guinea pig complement on E11, also delayed the developmental onset of spinal cord myelination until E17. Immunohistochemical assessments of myelin-suppressed spinal cords on E13 (n 5 3), E14 (n 5 3), E15 (n 5 3) and E16 (n 5 3) showed a complete and comparable (i.e. no interanimal variability) lack of MBP immunoreactivity throughout the spinal cord excluding the most rostral 1– 4 cervical segments. On or about E17 a period of robust myelination occurred in both unoperated control spinal cords (see above) and experimental spinal cords previously subjected to developmental myelin-suppression (results not shown). MBP immunohistochemical analysis of E17 and E18 spinal cords (n 5 6 for each day) previously injected with guinea pig complement plus monoclonal GalC antibodies on E11, revealed immunoreactivity throughout all levels of the spinal cord. Similarly, MBP immunohistochemical analysis of E17 and E18 spinal cords (n 5 4 for each day) previously injected with polyclonal GalC antibodies plus guinea pig complement on E11, revealed some immunoreactivity throughout all levels of the spinal cord. All myelin-suppressed spinal cords analysed on E17 or E18 showed obvious MBP immunoreactivity, although the MBP immunoreactivity appeared to be less intense than that in unoperated control E17 spinal cord tissue (results not shown). MBP immunohistochemical analysis of P2, P4 and P6 chick spinal cords (n 5 4 for each day) previously injected with guinea
pig complement plus monoclonal GalC antibodies on E11, showed levels of MBP immunofluorescence comparable to unoperated control tissue (n 5 25; results not shown). MBP immunohistochemical analysis of P2, P4 and P6 spinal cords (n 5 4 for each day), previously injected with polyclonal GalC antibodies plus guinea pig complement on E11, also showed levels of MBP immunoreactivity similar to the unoperated control tissue. After hatching, animals that had previously been subjected to developmental myelin-suppression exhibited sensory motor behaviours that were comparable to unoperated control hatchlings of the same age. It is important to note in this regard that all myelinsuppressed chicks hatched unassisted on E21. Locomotor behaviour of myelin-suppressed hatchlings was assessed on P2, P4 and P6 (n 5 8 for each day), prior to sacrifice for immunohistochemical analysis (see above). Visual assessments of postural adjustments, walking, running and righting responses suggested that neuronal control of locomotion was not altered in hatchlings that had been subjected to developmental myelin-suppression. As a control for the possible influence of non-specific binding of the GalC antibody, 5 control embryos received E11 intraspinal injections of guinea pig complement plus an antibody to GFAP that does not cross-react with chicken. MBP immunohistochemistry on E15 revealed no suppression of myelin development when compared with unoperated control E15 animals. Other immunological control embryos received injections of the monoclonal
SUPPRESSION OF SPINAL CORD MYELIN DEVELOPMENT
731
FIG. 2. Developmental myelin-suppression in the thoracic spinal cord of the embryonic chick. (A) MBP immunofluorescence staining of an E15 spinal cord in parasagittal section from a myelin-suppressed animal that received a single injection of monoclonal GalC antibody plus heterologous complement on E11; note the absence of myelin. (B) Electron micrograph of a transverse section of E15 spinal cord from a myelin-suppressed animal that received a single injection of polyclonal GalC antibody plus heterologous complement on E11; note the absence of cross-sectional profiles of myelinated axons. (C) MBP immunofluorescence staining of an E15 spinal cord in parasagittal section from an animal that received a single injection of GalC antibody only on E11; note that myelin is unperturbed. (D) MBP immunofluorescence staining of an E15 spinal cord in parasagittal section from an animal that received a single injection of heterologous complement proteins only on E11; note that myelin is unperturbed. (Bars 5 50 mm for A,C,D and 25mm for B.)
GalC antibody only (Fig. 2C; n 5 6), guinea pig complement only (Fig. 2D; n 5 8), PBS vehicle only (n 5 4) or monoclonal GalC antibodies plus heat-inactivated serum (heating serum prior to use denatures and inactivates the complement proteins; n 5 8). In no instance was myelin development suppressed or detectably altered from that observed in untreated normal chicks. Direct pressure injection of control and experimental solutions into the thoracic spinal cord did not result in significant damage to the spinal cord tissue. The injected solution did not displace spinal cord tissue or result in a region of necrosis at the injection site at any of the developmental ages examined (see above). A thin needle tract in the thoracic region was occasionally visible in E12–E14 spinal cords previously injected on E9 –E12. After E14, the injection site was undetectable. The developing and mature state of the spinal cord astrocyte population did not appear to be disturbed by this immunological protocol, as assessed by GFAP immunohistochemistry (Fig. 3). Immunohistochemical assessments of myelin-suppressed and unoperated control E14 and E15 spinal cords (n 5 4 for each group) showed that the distribution of astrocytes was similar at each developmental stage (Fig. 3A and 3B). P6 spinal cord tissue from myelin-suppressed chicks (n 5 6) also showed a similar distribu-
tion of astrocytes, when compared to untreated control P6 spinal cord tissue (n 5 8; results not shown). Additionally, individual astrocytes in myelin-suppressed spinal cords did not appear to express higher levels of GFAP than individual astrocytes from unoperated controls at any of the developmental stages examined. The neuronal population of the spinal cord did not appear to be disturbed as a result of developmental myelin-suppression (Fig. 3). Neuronal development in the embryonic spinal cord was assessed with microtubule-associated protein 2 (MAP-2) immunohistochemistry to identify the dendritic morphology, as well as thionin staining to assess overall neuronal and axonal morphology. Myelin-suppressed E15 spinal cords analysed with MAP-2 antibodies (Fig. 3D; n 5 4) or thionin staining (n 5 3; results not shown) were indistinguishable from unoperated control E15 spinal cords also analysed with MAP-2 antibodies (Fig. 3C; n 5 4) or thionin staining (n 5 3; results not shown). Northern blot analysis of the developing chick spinal cord indicated that chick MBP mRNA expression developed in a pattern similar to MBP immunoreactivity (Fig. 4A). MBP mRNA was first detected on E12 and continued to increase in abundance until approximately E17; thereafter, MBP mRNA levels remained relatively constant. Developmental myelin-suppression did not ap-
732
KEIRSTEAD ET AL.
FIG. 3. Developmental myelin-suppression does not detectably disturb the astrocytic or neuronal populations of the spinal cord. (A) GFAP immunofluorescence staining of an unoperated control E15 thoracic spinal cord in parasagittal section showing the normal pattern of GFAP immunoreactivity. (B) GFAP immunofluorescence staining of an E15 thoracic spinal cord in parasagittal section from a myelin-suppressed animal that received a single injection of GalC antibody plus heterologous complement on E11; note that the pattern of GFAP immunoreactivity is indistinguishably different from A. (C) MAP-2 peroxidase staining of an unoperated control E15 thoracic spinal cord in parasagittal section showing the normal dendritic architecture of spinal cord neurons. (D) MAP-2 peroxidase staining of an E15 thoracic spinal cord in parasagittal section from a myelin-suppressed animal that received a single injection of GalC antibody plus heterologous complement on E11; note that the dendritic architecture of spinal cord neurons is similar to C. (Bars 5 50mm.)
parently alter the levels of expression of MBP mRNA (Fig. 4B). Chick MBP mRNA levels were similar in unoperated E15 control thoracic spinal cords (n 5 3) and experimental cords that were myelin-suppressed (starting on E11) and examined on either E12, E13, E14 or E15 (n 5 3 for each developmental stage). DISCUSSION Disruption of compact CNS myelin or inhibition of some specific myelin proteins facilitates neurite extension and axonal regeneration [4,6,28,29,36,37,44]. We have previously shown that an imposed developmental delay of the onset of CNS myelination extends the developmental period during which CNS axons can regenerate [28]. In the studies reported here, we have further developed and characterised immunological procedures which result in developmental myelin-suppression. In order to induce developmental myelin-suppression, we injected complement proteins together with antibodies that bind to oligodendrocyte antigens expressed prior to myelination (i.e., around the time of oligodendrocyte differentiation). A single intraspinal injection of serum complement plus myelin/oligodendrocyte-specific antibodies delayed the developmental onset of my-
elination until E17 (Fig. 2). Myelin-suppression was confirmed by MBP, MAG and CNP immunohistochemistry, toluidine blue staining, and electron microscopy. Monoclonal GalC antibodies, polyclonal GalC antibodies and polyclonal 04 antibodies were each effective in eliciting developmental myelin-suppression (Fig. 2). After hatching, myelin antigen immunoreactivity within the spinal cords of myelin-suppressed animals appeared equivalent to unoperated age-matched control animals. This suggests that developmental myelin-suppression did not permanently alter the state of myelin in the spinal cord. It is conceivable that other cell types may have non-specifically bound the antibodies, activated the complement cascade, and indirectly altered the state of myelination. We examined the possibility of other cell types influencing the state of myelination in this manner by injecting an astrocyte-specific, complement-binding antibody (to glial fibrillary acidic protein that does not cross-react with chicken) plus homologous complement into the thoracic spinal cord at E10; this procedure did not alter the state of myelination at any developmental stage (results not shown). Similarly, no effect on CNS myelination was observed if serum complement proteins only, GalC or 04 antibody only, PBS vehicle only, or
SUPPRESSION OF SPINAL CORD MYELIN DEVELOPMENT
FIG. 4. Representative northern blots demonstrating the onset of MBP mRNA expression around E12, increasing to stable levels around E17 (upper panel). Compared to aged matched control animals on E15, no change is seen in MBP mRNA expression four days following an E11 intraspinal injection of GalC antibody plus complement (lower panel). Comparable northern blots of MBP mRNA expression on E13 and E14 did not indicate a transient down-regulation of MBP expression (results not shown). Sizes indicated are in kilobases; cyclophilin and actin hybridizations performed on these blots are included to indicate the amounts of RNA loaded in each lane.
GalC antibody plus heat-inactivated serum were injected into the spinal cord (Fig. 2C,D). These immunological control injections indicated that both serum complement proteins and myelin/oligodendrocyte-specific antibodies were necessary to delay the developmental onset of myelination. Additionally, these findings discount the possibility that serum complement entry, due to a breach of the blood— brain barrier alone, elicited developmental myelinsuppression. In vitro studies have indicated that myelin/oligodendrocytespecific antibodies target oligodendrocyte membranes and bind complement [13]. Activation of the first component of the complement protein cascade (C1q) requires the presentation of multiple Fc sites on cell-surface bound antibodies [32]. Thus, to activate the complement cascade, two or more complement-binding monomeric IgG antibodies must bind in proximity to each other on antigenic membrane sites. The binding of one complement-binding pentameric IgM antibody to the membrane is sufficient to activate complement. Consequently, the minimum antigen density required for IgM-mediated cellular attack (04-mediated developmental myelin-suppression) may be relatively less than that required for IgG-mediated cellular attack (GalC-mediated developmental myelin-suppression). IgG antibodies to other oligodendrocyte-specific proteins such as PLP are reportedly less effective in evoking demyelination when applied with complement [32]. The apparent ineffectiveness of PLP antibodies may be due to the relative scarcity of PLP antigen on the oligodendrocyte membrane surface or the diminished ability of these PLP antibodies to activate complement effectively.
733 The delayed onset of myelination following developmental myelin-suppression indicates that the oligodendrocyte lineage cell population is not permanently destroyed by this immunological protocol. The maintained expression of MBP mRNA (Fig. 4) throughout the entire period of myelin-suppression supports this suggestion. In vitro studies indicate that GalC antisera cause an influx of extracellular calcium when applied to oligodendrocytes in culture, resulting in disassembly of microtubules and a concomitant retraction of the oligodendrocyte processes; removal of GalC antisera results in re-extension of oligodendrocyte processes [14]. Although these observations and the results reported here suggest that oligodendrocytes survive developmental myelin-suppression, we cannot exclude the possibility that oligodendrocyte progenitor cells rapidly replace the myelinating population of oligodendrocytes. Recent studies in the adult rat indicate that, although oligodendrocytes survive immunological demyelination, they do not contribute to remyelination [30]. These studies implicate the oligodendrocyte progenitor as the origin of remyelination-competent cells. Finally, we have observed that this immunological protocol is also effective in delaying the development of myelin within the murine spinal cord (Keirstead and Peterson, unpublished observations). The mouse spinal cord normally myelinates during the first two weeks of neonatal development [17]. The injection of serum complement proteins plus polyclonal GalC antibodies into the cord on postnatal day (P) 2 delayed the developmental onset of myelination. Implantation of anti-GalC hybridoma cells in neonatal rats has been shown to inhibit CNS myelin development [41]. These studies illustrate that the antibody-mediated cellular attack paradigm is not species restricted. We have previously demonstrated that developmental myelinsuppression (elicited with monoclonal GalC antibodies plus serum complement) facilitates complete neuroanatomical and functional regeneration of transected E15 chick spinal cord [28]. Untreated control (i.e., normally-myelinated) E15 chicks effect no repair whatsoever following a similar transection of the spinal cord [16,23,24,28,45,46]. We have also utilised a modified immunological protocol to disrupt myelin within the mature avian spinal cord to evoke axonal regeneration [29]. More recent results indicate that immunological demyelination can be maintained for as long as serum complement proteins and myelin/oligodendrocyte antibodies are infused into the CNS. These findings suggest that immunological demyelination may be a useful component in the therapeutic intervention to promote repair of the injured adult spinal cord. ACKNOWLEDGEMENTS
We thank Udo Bartsch and Melitta Schachner for the gift of the O4 antibody, Karin L. Mathias and Ania Wisniewska for technical assistance. This study was supported by grants to J.D.S. from the Medical Research Council of Canada and the Canadian Neuroscience Network. H.S.K. was supported by scholarships from the NCE and the Natural Sciences and Engineering Research Council of Canada.
REFERENCES 1. Bensted, J. P. M.; Dobbing, J.; Morgan, R. S., Reid, R. T. W.; Payling Wright, G. Neuroglial development and myelination in the spinal cord of the chick embryo. J. Embryol. exp. Morph. 5:428 – 437; 1957. 2. Blakemore, W. F. Remyelination of CNS axons by Schwann cells transplanted from the sciatic nerve. Nature 266:68 – 69; 1997. 3. Blakemore, W. F. Observations on remyelination in the rabbit spinal cord following demyelination induced by lysolecithin. Neuropathol. Appl. Neurobiol. 4:47–59; 1978. 4. Bregman, B. S.; Kunkel-Bagden, E.; Schnell, L.; Dai, H. N.; Gao, D.,
734
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
22.
23.
24. 25. 26. 27.
KEIRSTEAD ET AL. Schwab, M. E. Recovery from spinal cord injury mediated by antibodies to neurite growth inhibitors. Nature 378:498 –501; 1995. Bunge, M. P.; Bunge, R. P.; Ris, H. Ultrastructural study of remyelination in an experimental lesion in adult cat spinal cord. J. Biophys. Biochem. Cytol. 10:67–94; 1961. Caroni, P.; Schwab, M. E. Antibody against myelin-associated inhibitor of neurite growth neutralizes nonpermissive substrate properties of CNS white matter. Neuron 1:85–96; 1988. Carroll, W. M.; Jennings, A. R.; Mastiglia, F. L. Immunocytochemical study of the glial cell response in antibody-mediated optic nerve demyelination. Neurosci. Lett. Suppl. 19:S49; 1985. Chomczynski, P.; Saachi, N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156 –159; 1987. Clifford-Jones, R. E.; Landon, D. N.; McDonald, W. E. Remyelination during optic nerve compression. J. Neurol. Sci. 46:239 –243; 1980. DalCanto, M. C.; Lipton, H. L. Schwann cell remyelination and recurrent demyelination in the central nervous system of mice injected with attenuated Theiler’s virus. Am. J. Pathol. 98:101–122; 1980. Dorfman, S. H.; Fry, J. M.; Silberberg, D. H. Antiserum induced myelin inhibition in vitro without complement. Brain Res. 177:105– 114; 1979. Dubios-Dalcq, M.; Niedieck, B.; Buyse, M. Action of anti-cerebroside sera on myelinated tissue cultures. Pathol. Eur. 5:331–347; 1970. Dyer, C. A.; Benjamins, J. A. Antibody to galactocerebroside alters organization of oligodendroglial membrane sheets in culture. J. Neurosci. 8:4307– 4318; 1988. Dyer, C. A.; Benjamins, J. A. Glycolipids and transmembrane signaling: antibodies to galactocerebroside cause an influx of calcium in oligodendrocytes. J. Cell Biol. 111:625– 633; 1990. Eames, R. A.; Jacobson, S. G.; McDonald, W. I. Pathologic changes in the optic chiasm of the cat following local injection of diphtheria toxin. J. Neurol. Sci. 32:381–393; 1977. Eidelberg, E. Consequences on spinal cord lesions upon motor function, with special attention to locomotor activity [Review]. Prog. Neurobiol. 17:185–202; 1981. Foran, D. R.; Peterson, A. C. Myelin acquisition in the central nervous system of the mouse revealed by an MBP-Lac Z trangene. J. Neurosci. 12:4890 – 4897; 1992. Fry, J. M.; Weissbarth, S.; Lehrer, G. M.; Burnstein, M. B. Cerebroside antibody inhibits sulfatide synthesis and myelination and demyelination in cord tissue cultures. Science 183:540 –542; 1974. Graca, D. L.; Blakemore, W. F. Delayed remyelination in rat spinal cord following ethidium bromide injection. Neuropath. and Appl. Neurol. 12:593– 605; 1986. Hamburger, V.; Hamilton, H. L. A series of normal stages in the development of the chick embryo. J. Morphol. 165:49 –92; 1951. Harry, G. J.; Goofrum, J. F.; Bouldin, T. W.; Wagner-Recio, M.; Toews, A. D.; Morell, P. Tellurium-induced neuropathy: metabolic alterations associated with demyelination and remyelination in rat sciatic nerve. J. Neurochem. 52:938 –945; 1989. Hartman, B. K.; Agrawal, H. C.; Kalmbach, S.; Shearer, W. T. A comparative study of the immunohistochemical localization of basic protein to myelin and oligodendrocytes in rat and chicken brain. J. Comp. Neurol. 188:273–290; 1979. Hasan, S. J.; Nelson, B. H.; Valenzuela, J. I.; Keirstead, H. S.; Schull, S. E.; Ethell, D. W.; Steeves, J. D. Functional repair of transected spinal cord in embryonic chick. Restor. Neurol. Neurosci. 2:137–154; 1991. Hasan, S. J.; Keirstead, H. S.; Muir, G. D.; Steeves, J. D. Axonal regeneration contributes to repair of injured brainstem-spinal neurons in embryonic chick. J. Neurosci. 13:492–507; 1993. Herndon, R. M.; Price, D. L.; Weiner, L. P. Regeneration of oligodendrocytes during recovery from demyelinating disease. Science 195: 693– 694; 1977. Hirayama, M.; Silberberg, D. H.; Lisak, R. P.; Pleasure, D. Long-term culture of oligodendrocytes isolated from rat corpus callosum by percoll density gradient. J. Neuropathol. Exp. Neurol. 42:16 –28; 1983. Hruby, S.; Alroul, E. C. Jr.; Seil, F. J. Synthetic galactocerebroside evokes myelination-inhibiting antibodies. Science 195:173–175; 1977.
28. Keirstead, H. S.; Hasan, S. J.; Muir, G. D.; Steeves, J. D. Suppression of the onset of myelination extends the permissive period for the functional repair of embryonic spinal cord. Proc. Natl. Acad. Sci. 89:11664 –11668; 1992. 29. Keirstead, H. S.; Dyer, J. K.; Sholomenko, G. N.; McGraw, J.; Delaney, K. R.; Steeves, J. D. Axonal regeneration and physiological activity following transection and immunological disruption of myelin within the hatchling chick spinal cord. J. Neurosci. 15(10):6963– 6974; 1995. 30. Keirstead, H. S.; Blakemore, W. F. (1997). Identification of postmitotic oligodendrocytes incapable of remyelination in the demyelinated adult spinal cord. J. Neuropathol. Exp. Neurol. Exp. Neurol. 56:1191–1201; 1997. 31. Lampert, P. W. Demyelination and remyelination in experimental allergic encephalomyelitis. Further electron microscopic observations. J. Neuropathol. Exp. Neurol. 24:371–585; 1968. 32. Law, S. K. A.; Reid, K. B. M. (1985) Complement edited by David Male and David Rickwood IRL Press, Oxford Washington, D.C. 33. Ludwin, S. K. Central nervous system demyelination and remyelination in the mouse. An ultrastructural study of cuprizone toxicity. Lab. Invest. 39:597– 612; 1978. 34. Macklin, W. B.; Weill, C. L. Appearance of myelin proteins during development in the chick central nervous system. Dev. Neurosci. 7:170 –178; 1985. 35. Mastaglia, F. L.; Carrol, W. M.; Jennings, A. R. Spinal cord lesions induced by antigalactocerebroside serum. Clin. Exp. Neurol. 26:33– 44; 1989. 36. McKerracher, L.; David, S.; Jackson, D. L.; Kottis, V.; Dunn, R. J.; Braun, P. E. Identification of myelin-associated glycoprotein as a major myelin-derived inhibitor of neurite growth. Neuron 13:805– 811; 1994. 37. Mukhopadhyay, G.; Doherty, P.; Walsh, F. S.; Crocker, P. R.; Filbin, M. A novel role for myelin-associated glycoprotein as an inhibitor of axonal regeneration. Neuron 13:757–767; 1994. 38. Ozawa, K.; Saida, T.; Saida, K.; Nishitani, H.; Kameyama, M. In vivo CNS demyelination mediated by anti-galactocerebroside antibody. Acta Neuropathol. 77:621– 628; 1989. 39. Raine, C. S.; Bornstein, M. B. Experimental allergic encephalomyelitis: a light and electron microscope study of remyelination and ‘‘sclerosis’’ in vitro. J. Neuropathol. Exp. Neurol. 29:552–574; 1970. 40. Ranscht, B.; Clapshaw, P. A.; Price, J.; Noble, M.; Seifert, W. Development of oligodendrocytes and Schwann cells studies with a monoclonal antibody against galactocerebroside. Proc. Natl. Acad. Sci. USA 79:2709 –2713; 1982. 41. Rosenbluth, J.; Liu, Z.; Guo, D.; Schiff, R. Inhibition of CNS myelin development in vivo by implantation of anti-GalC hybridoma cells. J. Neurocytol. 23:699 –707; 1994. 42. Sambrook, J.; Fritsch, E. F.; Maniatis, T. (1989) Molecular Cloning, 2nd edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 43. Sergott, R. C.; Brown, M. J.; Silberberg, D. H.; Lisak, R. P. Antigalactocerebroside serum demyelinates optic nerve in vivo. J. Neurol. Sci. 64:297–303; 1984. 44. Schnell, L.; Schwab, M. E. Axonal regeneration in the rat spinal cord produced by an antibody against myelin-associated neurite growth inhibitors. Nature 343:269 –272; 1990. 45. Shimizu, I.; Oppenheim, R. W.; O’Brien, M.; Schneiderman, A. Anatomical and functional recovery following spinal cord transection in the chick embryo. J. Neurobiol. 21:918 –937; 1990. 46. Sholomenko, G. N.; Steeves, J. D. Effects of selective spinal cord lesions on hindlimb locomotion in birds. Exp. Neurol. 95:403– 418; 1987. 47. Sommer, I.; Schachner, M. Cells that are 04 antigen-positive and 01 antigen-negative differentiate into 01 antigen-positive oligodendrocytes. Neuroscience Letters 29:183–188; 1982. 48. Xie, W.; Rothblum, L. I. Rapid, small-scale RNA isolation from tissue culture cells. Biotechniques 11:323–327; 1991.
PII S0361-9230(97)00374-2
In Vivo Immunological Suppression of Spinal Cord Myelin Development H. S. KEIRSTEAD,1 D. M. PATAKY, J. McGRAW AND J. D. STEEVES CORD (Collaboration On Repair Discoveries), Departments of Zoology, Anatomy, and Surgery, University of British Columbia, Vancouver, B.C., CANADA V6T 1Z4 [Received 5 June 1997; Revised 11 September 1997; Accepted 15 September 1997] ABSTRACT: The onset of myelination in the embryonic chick spinal cord begins on embryonic day (E) 12 or E13 of the 21 day in ovo developmental period. This event coincides with a loss of functional axonal regeneration following complete transection of the thoracic spinal cord. In this study, we have characterised an immunological method for delaying the developmental onset of myelination in vivo until later stages of development (developmental myelin-suppression). A single injection of heterologous or homologous serum complement proteins plus myelinspecific, complement-binding antibodies into the spinal cord prior to E13 delayed the onset of myelination until E17. The state of spinal cord myelin was assessed with immunohistochemical, histological and ultrastructural techniques. Northern blot analysis indicated that myelin basic protein mRNA was not downregulated in myelin-suppressed spinal cords, which suggests that oligodendrocytes survived developmental myelin-suppression. Glial fibrillary acidic protein immunostaining of normal and treated tissue indicated that myelin-suppression did not alter the resident astrocyte population of the spinal cord or elicit astrogliosis. Immunostaining with microtubule-associated protein-2 and thionine staining of normal and myelin-suppressed tissue further indicated that the neuronal architecture was unaffected by the immunological protocol. © 1997 Elsevier Science Inc.
therapies to promote functional axonal regeneration following spinal cord injury. Experimental methods for in vivo myelin-suppression include: 1) X-ray irradiation [2]; 2) exposure to drugs such as ethidium bromide [19], lysolecithin [3], tellurium [21], cuprizone [33] or diptheria toxin [14]; 3) viruses such as Theiler’s virus [9] or the JMH strain of mouse hepatitis virus [25]; 4) nerve compression [5,9]; and 5) exposure to sera from animals afflicted with experimental allergic encephalomyelitis [31,39]. Although effective in suppressing myelin, none of the above methods appear to be specific for myelin alone. However, myelin-suppression initiated by complement-binding antibodies directed against CNS myelin is highly specific in vitro [11,12,14,18,26,27], although in vivo studies in the optic nerve [17,38,43] and spinal cord [35] report equivocal results with regards to cellular specificity. In the present study, we have examined the ability of two complement-binding, oligodendrocyte-specific antibodies to suppress the developmental onset of myelination in the chick embryo. The onset of myelination in the developing chick spinal cord begins around embryonic day (E) 13 of the 21 day developmental period [1,22,29,34]. The first appearance of oligodendrocytes in the chick precedes the initial formation of myelin by 2–3 embryonic days [1] and is characterised by the expression of galactocerebroside (GalC), an abundant sphingolipid produced by oligodendrocytes [40]. GalC is highly conserved across species [40]. The oligodendrocyte-specific antigen recognised by the 04 antibody is also expressed prior to the onset of myelination [47]. We have found that a single injection of serum complement proteins plus (monoclonal or polyclonal) GalC or 04 antibodies into the thoracic spinal cord of an E9 –E12 chick embryo (i.e. around the time of oligodendrocyte differentiation) resulted in a delay in the onset of spinal cord myelination until E17 (developmental myelin-suppression). Myelin basic protein (MBP) mRNA was not downregulated in myelin-suppressed spinal cords, which suggests that oligodendrocytes survived developmental myelin-suppression. Neither astrocytic or neuronal populations of the spinal cord were detectably disturbed by this intervention.
KEY WORDS: Demyelination, Remyelination, Galactocerebroside, 04 antibody, Oligodendrocyte.
INTRODUCTION Central nervous system (CNS) myelin is a highly differentiated membrane structure produced by oligodendrocytes. We have previously reported that CNS myelin is inhibitory to neuroanatomical repair and recovery of physiological activity following complete transection of late embryonic or hatchling chick spinal cord [28,29]. A few CNS myelin proteins have been identified as potential inhibitors of axonal regeneration, including NI-35, NI250 [6,44] and myelin-associated glycoprotein [36,37]. Implantation of IN-1 antibody-secreting hybridoma cells into the parietal cortex of adult rats promotes limited axonal regeneration by severed corticospinal neurons [4,44]. Given that CNS myelin has an inhibitory effect on the regenerative ability of previously severed axons, further characterisation of in vivo methods for transiently suppressing CNS myelin should provide insight into potential 1
MATERIALS AND METHODS Fertilized White Leghorn eggs were incubated at 37°C in an automatic rotating incubator (Peacock Equipment Limited, Aldergrove, B.C.) and staged using the protocol outlined by Hamburger
To whom requests for reprints should be addressed: Presently at MRC Cambridge Centre for Brain Repair, Robinson Way, Cambridge, U.K. CB2 2PY.
727
728 and Hamilton [20]. After surgery, eggs were sealed with paraffin wax and a sterile coverslip then returned to the incubator. Developmental Myelin-Suppression Thoracic spinal cord injections were performed at E9 –E12 in chick embryos using a glass micropipette (tip diameter 5 30 – 40 mm; A-M Systems, Everett, Washington #6045) connected to a Picospritzer II pressure injection system (General Valve Corp., Fairfield, New Jersey). Developmental myelin-suppression was evoked by injecting either; 1) an IgG mouse galactocerebroside (GalC) antibody (a gift from B. Ranscht; 2.67 mg/ml hybridoma supernatant) at a dilution of 1:25 with 20% homologous serum (as a source of complement) in 0.1M phosphate buffered saline (PBS), pH 7.4 (providing an effective concentration of 63.0 ng of GalC hybridoma supernatant injected per gram body weight) or, 2) an IgG rabbit GalC antibody (Chemicon International Inc., Temecula, California #AB142) at a dilution of 1:25 with 20% heterologous (guinea pig) complement (Gibco BRL, Burlington, Ontario #19195-015) in 0.1M PBS pH 7.4 or, 3) an IgM polyclonal 04 antibody [47] at a dilution of 1:25 with 20% guinea pig complement in 0.1M PBS pH 7.4. Each intraspinal injection was performed with a new glass micropipette backloaded with 2–3 ml of solution immediately prior to surgery. Each animal received a total volume of 2–3 ml, over 1– 4 penetrations, injected directly into the mid-to-high thoracic spinal cord. It is conceivable that cells other than oligodendrocytes may non-specifically bind the GalC or 04 antibodies via Fc receptors on the cell surface and indirectly affect myelination. To assess the possibility of non-specific binding, control embryos were similarly injected with 20% homologous serum complement plus an antihuman antibody that does not cross-react with chicken. We chose a monoclonal IgG antibody to glial fibrillary acidic protein (GFAP; Amersham), a major constituent of astrocytes within the CNS. Other immunological control embryos received injections of either: 1) the monoclonal or polyclonal GalC antibody only, 2) homologous serum complement proteins only, 3) guinea pig complement only, 4) vehicle only (0.1M PBS, pH 7.4), or 5) the monoclonal GalC antibody plus homologous complement, following heat-inactivation of the complement by exposure to 50°C for 30 min. Animals were perfused intracardially at the appropriate developmental stage (see results) with 0.1M PBS containing 2500 USP units of heparin per 50 mls PBS, pH 7.4 (at 37°C) followed by perfusion with 4% paraformaldehyde in 0.1M phosphate buffer, pH 7.4 (at 4 –10°C). The dissected spinal cords were then immersed in the same fixative for 24 hr at 4°C and subsequently stored in PBS, pH 7.4 (at 4 –10°C) for further processing. Immunohistochemistry Antigens were localised using indirect immunofluorescence. GalC immunohistochemistry was performed on cryostat sectioned tissue from unoperated control embryos that were perfused as outlined above. All other antigens were localised on paraffinembedded tissue sections. The rabbit antihuman MBP antibody (Accurate Chemical Scientific Corp., #AXL746), the mouse antipig GFAP (Sigma Immunochemicals, #G-3893) and the mouse anticalf microtubule associated protein-2 (MAP-2; Amersham, #RPN 1194) were all used at a dilution of 1:100. The mouse anti-chicken myelin-associated glycoprotein (MAG; Boehringer Mannheim #1450 972) and the rabbit antibovine 29,39-cyclic nucleotide 39-phosphodiesterase (CNP; a gift from Peter Braun) were used at a dilution of 1:500. The rabbit antibovine GalC antibody (Chemicon AB 142) was used at a dilution of 1:10. The secondary antibody was a goat anti-rabbit FITC conjugated immunoglobulin
KEIRSTEAD ET AL. (Caltag Laboratories, #L42001) diluted 1:100 or a goat anti-mouse FITC conjugated immunoglobulin (Dimension Labs, #30112663) diluted 1:100. Standard immunocytochemical controls (e.g., omission of primary and/or secondary antibodies) were processed alongside tissue sections from experimental and control animals. Photomicrographs were taken on a Zeiss Axiophot using epifluorescent illumination with the appropriate filters. Histological Staining Animals designated for toluidine blue staining were perfused intracardially at the appropriate developmental stage (see results) with 0.1M PBS containing 2500 USP units of heparin per 50 mls PBS, pH 7.4 (at 37°C) followed by perfusion with 2.5% glutaraldehyde in 0.1M phosphate buffer, pH 7.4 (at 4 –10°C). The dissected tissue was then dehydrated and resin embedded according to standard protocols. The hardened blocks were cut into 1 mm transverse sections on an ultra-microtome and placed on glass slides for toluidine blue staining. 0.1% toluidine blue stain was filtered and dropped onto sections for two minutes. The slides were then rinsed with hot tap water followed by distilled water, then placed on a hot plate to dry. Photomicrographs were taken on a Zeiss Axiophot microscope. Electron Microscopy Animals designated for toluidine blue staining were perfused intracardially at the appropriate developmental stage (see results) with 4% glutaraldehyde in 0.1M phosphate buffer (PB) pH 7.4. The lesion-containing length of spinal cord was cut into 1 mm transverse blocks and processed so as to preserve the cranio-caudal sequence and orientation. The tissue blocks were rinsed in 0.1M phosphate buffer pH 7.4 for 30 minutes, post-fixed in 2% OsO4, dehydrated in ascending alcohols, and embedded in TAAB resin. Thin sections (1 mm) were cut from each block, stained with alkaline Toluidine blue and examined by light microscopy. For electron microscopic analysis blocks were trimmed then cut at 100nm, mounted on copper grids, uranyl acetate and lead citrate stained, and viewed under a Hitachi EM 600 electron microscope at 75 kV. Northern Blotting Tissue was harvested from chicks at various stages of development beginning on E10 through post-hatching day (P) 12. Spinal cord myelin-suppression was begun on E11, as described above. Animals were sacrificed 2 to 10 days later. All northern blots were performed in triplicate. Total RNA from the thoracic and caudal two thirds of the cervical spinal cord was isolated by a modified version [48] of the method of Chomczynski and Sacchi (1987) [8]. Briefly, the tissue was homogenized in GAP (Solution A: 4M guanidinium isothiocyanate, 25mM sodium citrate, 0.5% sarkosyl and freshly added 0.1M 2-mercaptoethanol; plus H2O-saturated phenol and 2M NaOAc 1:1:0.1). 1/10 volume of chloroform: isoamyl alcohol (1:0.1) was added, the mixture left on ice for 15–30 min. The organic and aqueous phases were separated by centrifugation at 12,000xg at 4°C for 20 min. The aqueous phase was recovered and the RNA precipitated with an equal volume of isopropanol, centrifuged as above and resuspended in TE buffer. Following densitometric quantification of the amount of RNA present, the RNA was denatured and electrophoretically size-separated on 1.5% agarose/formaldehyde denaturing gels. An RNA ladder (Gibco BRL) was run with the initial samples to confirm the length of the message detected by the cRNA probe. After electrophoresis, the samples were capillary blotted overnight to Nytran membrane, then baked and stored until hybridisation at 220°C.
SUPPRESSION OF SPINAL CORD MYELIN DEVELOPMENT Northern hybridisations were performed following standard protocols [42]. Prehybridisation and hybridisation were performed at 65°C in 5x PIPES buffer with 50% formamide, 5x Denhardt’s, and 100mg/ml each of denatured salmon sperm DNA, herring sperm DNA and yeast tRNA. The cDNA for chicken MBP (a gift from Dr. Peter Jeffrey, GenBank accession #CHKMBP) was directionally subcloned into pBluescript (Stratagene), linearised and used as a template for the in vitro transcription (Promega) of high specific activity (. 109 cpm/mg) 32P-labelled cRNA probes. Blots were then rinsed at increasing stringency to 0.1xSSC at 65°C, then exposed to pre-flashed Kodak XAR X-ray film. To demonstrate the amount of RNA in each lane, blots were stripped in 80% formamide at 80°C for 1 hour, then re-hybridised as above using either a chicken actin (a gift of Dr. P. Gunning, clone #LK154, subcloned into pBluescript) or rat cyclophilin (a gift of Dr. James Douglas, subcloned into pBluescript) cRNAs as above. Behavioural Assessments Locomotor behaviour of hatchlings previously myelin-suppressed was assessed on P2, P4 and P6, prior to sacrifice for immunohistochemical analysis. The ability to walk, run, effect postural adjustments and righting responses was visually assessed for each animal and scored as ‘‘comparable to unoperated controls’’ or ‘‘not comparable to unoperated controls’’. RESULTS Myelin Development in the Chick GalC expression is an early indicator of oligodendrocyte differentiation from progenitor cells [40]. GalC immunoreactivity (results not shown) was not detected at any level of the chick spinal cord on E8 (n 5 3) or E9 (n 5 2), however, GalC was first detected in 4 of 6 animals sacrificed on E10. GalC immunoreactivity in these animals was punctate, and located in the lateral regions of the white matter tracts. In four animals analysed on E11, GalC immunoreactivity was similarly organised. GalC immunoreactivity was present throughout the spinal cord white matter on E12 (n 5 3), E13 (n 5 4), E14 (n 5 3) and E15 (n 5 5). GalC immunostaining of P1, P3, P7 and P10 spinal cords (n 5 6 for each day) was qualitatively similar to each other, suggesting that oligodendrocyte differentiation and myelination are complete by hatching. CNP immunoreactivity was not detected at any level of the spinal cord on E10 (n 5 3) or E11 (n 5 3), but was detected in the spinal cord on E12.5 (n 5 6). CNP immunoreactivity was present throughout the spinal cord white matter on E13 (n 5 4), E14 (n 5 8), and E15 (n 5 8). CNP immunostaining of P1, P3, P5 and P10 spinal cords (n 5 6 for each day) was similar at all stages. MBP or MAG immunoreactivity was not detected at any level of the spinal cord on E9 (n 5 6 for MBP), E10 (n 5 6 for MBP, n 5 4 for MAG) or E11 (n 5 12 for MBP, n 5 4 for MAG). On E12 MBP immunoreactivity was highly variable (n 5 12). In the majority of cases (n 5 8 of 12), MBP was absent on E12 (Fig. 1A). MAG was consistently absent on E12 (n 5 4 for MAG). MBP and MAG immunoreactivity were reliably detected within the ventrolateral funiculi of the cervical spinal cord on E13 and proceeded in a ventro-dorsal and rostral-caudal direction during subsequent development (n 5 14 for MBP, Fig. 1B; n 5 6 for MAG). By E14, all levels of the spinal cord displayed MBP, MAG and CNP immunoreactivity. The ventral funiculi were markedly more immunoreactive than other white matter regions (n 5 18 for MBP, n 5 12 for MAG). On E15 a dense network of immunoreactivity was observed within all spinal cord white matter tracts (n 5 28 for MBP, n 5 8 for MAG). Toluidine Blue staining (results not
729 shown) and electron microscopic analysis (Fig. 1C) of E15 tissue in transverse section revealed compact myelin profiles in all regions of white matter (n 5 4). A dramatic increase in MBP and MAG immunoreactivity was observed in animals sacrificed on E17 and E18, suggesting an increase in myelination at this stage of development (n 5 18 for MBP at each stage, n 5 6 for MAG at each stage). The MBP and MAG immunostaining on P1, P3, P5, P7 and P10 was qualitatively similar, once again suggesting that myelination is complete by the time of hatching (n 5 6 for each day, for each of MBP and MAG; Fig. 1D). Developmental Myelin-Suppression in the Embryonic Chick The developmental onset of myelination in the embryonic chick spinal cord was delayed by pressure injection of homologous or heterologous serum complement proteins plus either monoclonal GalC antibodies, polyclonal GalC antibodies or polyclonal 04 antibodies directly into the thoracic spinal cord, 1– 4 days prior to the normal developmental onset of myelination (E13). Monoclonal GalC antibodies, polyclonal GalC antibodies and polyclonal 04 antibodies were selected for their abilities to fix complement and specifically recognise myelin or oligodendrocyte cell-surface antigens. Thoracic spinal cord injections of monoclonal GalC antibodies, plus either chick serum or guinea pig complement, delayed the developmental onset of spinal cord myelination until E17 (Fig. 2). Immunohistochemical assessments of myelin-suppressed spinal cords (previously injected on E11) sacrificed on either E12 (n 5 6), E13 (n 5 6), E14 (n 5 6), E15 (n 5 12) or E16 (n 5 6) revealed a complete and comparable (i.e., no inter-animal variability) lack of MBP immunoreactivity throughout the spinal cord, excluding the most rostral 1– 4 cervical segments. Thus, the effect of the immunological myelin-suppression protocol appears to be very rapid. The extent and degree of developmental myelin-suppression on E15 was similar when the complement proteins and monoclonal GalC antibodies were injected on either E9, E10, E11 or E12 (n 5 3 for each day). Likewise, thoracic spinal cord injections of polyclonal GalC antibodies plus guinea pig complement on E11 also delayed the developmental onset of spinal cord myelination until E17 (Fig. 2). Immunohistochemical assessments of myelin-suppressed spinal cords on E12 (n 5 4), E13 (n 5 4), E14 (n 5 4), E15 (n 5 5) and E16 (n 5 4) showed a complete and comparable (i.e. no interanimal variability) lack of MBP immunoreactivity throughout the spinal cord, except for the most rostral 1– 4 cervical segments. Immunohistochemical assessments of myelin-suppressed spinal cords on E13 (n 5 3), E14 (n 5 3) and E15 (n 5 3) also showed a complete lack of MAG immunoreactivity throughout the spinal cord excluding the most rostral 1– 4 cervical segments. Immunohistochemical assessments for CNP immunoreactivity in myelinsuppressed spinal cords on E13 (n 5 3), E14 (n 5 3) and E15 (n 5 3) did not reveal any CNP immunoreactivity within the cord, except for the most rostral 1– 4 cervical segments. All the immunohistochemical markers used in this study indicated the immunological protocol delayed the development of spinal cord myelin. Five spinal cords injected on E11 with polyclonal GalC antibodies plus guinea pig complement were sacrificed on E15, resin embedded and transversely microsectioned for electron microscopic analysis. Electron microscopic analysis revealed a lack of myelin (Fig. 2B). In all instances only a very few myelinated axons were visible throughout the spinal cord white matter. Unoperated E15 control tissue (n 5 5) processed in a similar manner contained an abundance of tightly compacted myelin seen as concentric rings surrounding axon profiles (Fig. 1C).
730
KEIRSTEAD ET AL.
FIG. 1. Developmental pattern of myelination in the chick thoracic spinal cord. (A) MBP immunofluorescence staining of unoperated control E12 spinal cord in parasagittal section showing no MBP immunoreactivity. (B) MBP immunofluorescence staining of unoperated control E13 spinal cord in parasagittal section showing the developmental onset of MBP immunoreactivity. (C) Electron micrograph of a transverse section of unoperated E15 spinal cord showing cross-sectional profile of myelinated axons. (D) MBP immunofluorescence staining of unoperated control P5 spinal cord in parasagittal section showing fully developed pattern of spinal cord myelination. (Bars 5 100mm for A, 50mm for B,D; X15000 for C.)
Finally, thoracic intraspinal injections of polyclonal 04 antibodies plus guinea pig complement on E11, also delayed the developmental onset of spinal cord myelination until E17. Immunohistochemical assessments of myelin-suppressed spinal cords on E13 (n 5 3), E14 (n 5 3), E15 (n 5 3) and E16 (n 5 3) showed a complete and comparable (i.e. no interanimal variability) lack of MBP immunoreactivity throughout the spinal cord excluding the most rostral 1– 4 cervical segments. On or about E17 a period of robust myelination occurred in both unoperated control spinal cords (see above) and experimental spinal cords previously subjected to developmental myelin-suppression (results not shown). MBP immunohistochemical analysis of E17 and E18 spinal cords (n 5 6 for each day) previously injected with guinea pig complement plus monoclonal GalC antibodies on E11, revealed immunoreactivity throughout all levels of the spinal cord. Similarly, MBP immunohistochemical analysis of E17 and E18 spinal cords (n 5 4 for each day) previously injected with polyclonal GalC antibodies plus guinea pig complement on E11, revealed some immunoreactivity throughout all levels of the spinal cord. All myelin-suppressed spinal cords analysed on E17 or E18 showed obvious MBP immunoreactivity, although the MBP immunoreactivity appeared to be less intense than that in unoperated control E17 spinal cord tissue (results not shown). MBP immunohistochemical analysis of P2, P4 and P6 chick spinal cords (n 5 4 for each day) previously injected with guinea
pig complement plus monoclonal GalC antibodies on E11, showed levels of MBP immunofluorescence comparable to unoperated control tissue (n 5 25; results not shown). MBP immunohistochemical analysis of P2, P4 and P6 spinal cords (n 5 4 for each day), previously injected with polyclonal GalC antibodies plus guinea pig complement on E11, also showed levels of MBP immunoreactivity similar to the unoperated control tissue. After hatching, animals that had previously been subjected to developmental myelin-suppression exhibited sensory motor behaviours that were comparable to unoperated control hatchlings of the same age. It is important to note in this regard that all myelinsuppressed chicks hatched unassisted on E21. Locomotor behaviour of myelin-suppressed hatchlings was assessed on P2, P4 and P6 (n 5 8 for each day), prior to sacrifice for immunohistochemical analysis (see above). Visual assessments of postural adjustments, walking, running and righting responses suggested that neuronal control of locomotion was not altered in hatchlings that had been subjected to developmental myelin-suppression. As a control for the possible influence of non-specific binding of the GalC antibody, 5 control embryos received E11 intraspinal injections of guinea pig complement plus an antibody to GFAP that does not cross-react with chicken. MBP immunohistochemistry on E15 revealed no suppression of myelin development when compared with unoperated control E15 animals. Other immunological control embryos received injections of the monoclonal
SUPPRESSION OF SPINAL CORD MYELIN DEVELOPMENT
731
FIG. 2. Developmental myelin-suppression in the thoracic spinal cord of the embryonic chick. (A) MBP immunofluorescence staining of an E15 spinal cord in parasagittal section from a myelin-suppressed animal that received a single injection of monoclonal GalC antibody plus heterologous complement on E11; note the absence of myelin. (B) Electron micrograph of a transverse section of E15 spinal cord from a myelin-suppressed animal that received a single injection of polyclonal GalC antibody plus heterologous complement on E11; note the absence of cross-sectional profiles of myelinated axons. (C) MBP immunofluorescence staining of an E15 spinal cord in parasagittal section from an animal that received a single injection of GalC antibody only on E11; note that myelin is unperturbed. (D) MBP immunofluorescence staining of an E15 spinal cord in parasagittal section from an animal that received a single injection of heterologous complement proteins only on E11; note that myelin is unperturbed. (Bars 5 50 mm for A,C,D and 25mm for B.)
GalC antibody only (Fig. 2C; n 5 6), guinea pig complement only (Fig. 2D; n 5 8), PBS vehicle only (n 5 4) or monoclonal GalC antibodies plus heat-inactivated serum (heating serum prior to use denatures and inactivates the complement proteins; n 5 8). In no instance was myelin development suppressed or detectably altered from that observed in untreated normal chicks. Direct pressure injection of control and experimental solutions into the thoracic spinal cord did not result in significant damage to the spinal cord tissue. The injected solution did not displace spinal cord tissue or result in a region of necrosis at the injection site at any of the developmental ages examined (see above). A thin needle tract in the thoracic region was occasionally visible in E12–E14 spinal cords previously injected on E9 –E12. After E14, the injection site was undetectable. The developing and mature state of the spinal cord astrocyte population did not appear to be disturbed by this immunological protocol, as assessed by GFAP immunohistochemistry (Fig. 3). Immunohistochemical assessments of myelin-suppressed and unoperated control E14 and E15 spinal cords (n 5 4 for each group) showed that the distribution of astrocytes was similar at each developmental stage (Fig. 3A and 3B). P6 spinal cord tissue from myelin-suppressed chicks (n 5 6) also showed a similar distribu-
tion of astrocytes, when compared to untreated control P6 spinal cord tissue (n 5 8; results not shown). Additionally, individual astrocytes in myelin-suppressed spinal cords did not appear to express higher levels of GFAP than individual astrocytes from unoperated controls at any of the developmental stages examined. The neuronal population of the spinal cord did not appear to be disturbed as a result of developmental myelin-suppression (Fig. 3). Neuronal development in the embryonic spinal cord was assessed with microtubule-associated protein 2 (MAP-2) immunohistochemistry to identify the dendritic morphology, as well as thionin staining to assess overall neuronal and axonal morphology. Myelin-suppressed E15 spinal cords analysed with MAP-2 antibodies (Fig. 3D; n 5 4) or thionin staining (n 5 3; results not shown) were indistinguishable from unoperated control E15 spinal cords also analysed with MAP-2 antibodies (Fig. 3C; n 5 4) or thionin staining (n 5 3; results not shown). Northern blot analysis of the developing chick spinal cord indicated that chick MBP mRNA expression developed in a pattern similar to MBP immunoreactivity (Fig. 4A). MBP mRNA was first detected on E12 and continued to increase in abundance until approximately E17; thereafter, MBP mRNA levels remained relatively constant. Developmental myelin-suppression did not ap-
732
KEIRSTEAD ET AL.
FIG. 3. Developmental myelin-suppression does not detectably disturb the astrocytic or neuronal populations of the spinal cord. (A) GFAP immunofluorescence staining of an unoperated control E15 thoracic spinal cord in parasagittal section showing the normal pattern of GFAP immunoreactivity. (B) GFAP immunofluorescence staining of an E15 thoracic spinal cord in parasagittal section from a myelin-suppressed animal that received a single injection of GalC antibody plus heterologous complement on E11; note that the pattern of GFAP immunoreactivity is indistinguishably different from A. (C) MAP-2 peroxidase staining of an unoperated control E15 thoracic spinal cord in parasagittal section showing the normal dendritic architecture of spinal cord neurons. (D) MAP-2 peroxidase staining of an E15 thoracic spinal cord in parasagittal section from a myelin-suppressed animal that received a single injection of GalC antibody plus heterologous complement on E11; note that the dendritic architecture of spinal cord neurons is similar to C. (Bars 5 50mm.)
parently alter the levels of expression of MBP mRNA (Fig. 4B). Chick MBP mRNA levels were similar in unoperated E15 control thoracic spinal cords (n 5 3) and experimental cords that were myelin-suppressed (starting on E11) and examined on either E12, E13, E14 or E15 (n 5 3 for each developmental stage). DISCUSSION Disruption of compact CNS myelin or inhibition of some specific myelin proteins facilitates neurite extension and axonal regeneration [4,6,28,29,36,37,44]. We have previously shown that an imposed developmental delay of the onset of CNS myelination extends the developmental period during which CNS axons can regenerate [28]. In the studies reported here, we have further developed and characterised immunological procedures which result in developmental myelin-suppression. In order to induce developmental myelin-suppression, we injected complement proteins together with antibodies that bind to oligodendrocyte antigens expressed prior to myelination (i.e., around the time of oligodendrocyte differentiation). A single intraspinal injection of serum complement plus myelin/oligodendrocyte-specific antibodies delayed the developmental onset of my-
elination until E17 (Fig. 2). Myelin-suppression was confirmed by MBP, MAG and CNP immunohistochemistry, toluidine blue staining, and electron microscopy. Monoclonal GalC antibodies, polyclonal GalC antibodies and polyclonal 04 antibodies were each effective in eliciting developmental myelin-suppression (Fig. 2). After hatching, myelin antigen immunoreactivity within the spinal cords of myelin-suppressed animals appeared equivalent to unoperated age-matched control animals. This suggests that developmental myelin-suppression did not permanently alter the state of myelin in the spinal cord. It is conceivable that other cell types may have non-specifically bound the antibodies, activated the complement cascade, and indirectly altered the state of myelination. We examined the possibility of other cell types influencing the state of myelination in this manner by injecting an astrocyte-specific, complement-binding antibody (to glial fibrillary acidic protein that does not cross-react with chicken) plus homologous complement into the thoracic spinal cord at E10; this procedure did not alter the state of myelination at any developmental stage (results not shown). Similarly, no effect on CNS myelination was observed if serum complement proteins only, GalC or 04 antibody only, PBS vehicle only, or
SUPPRESSION OF SPINAL CORD MYELIN DEVELOPMENT
FIG. 4. Representative northern blots demonstrating the onset of MBP mRNA expression around E12, increasing to stable levels around E17 (upper panel). Compared to aged matched control animals on E15, no change is seen in MBP mRNA expression four days following an E11 intraspinal injection of GalC antibody plus complement (lower panel). Comparable northern blots of MBP mRNA expression on E13 and E14 did not indicate a transient down-regulation of MBP expression (results not shown). Sizes indicated are in kilobases; cyclophilin and actin hybridizations performed on these blots are included to indicate the amounts of RNA loaded in each lane.
GalC antibody plus heat-inactivated serum were injected into the spinal cord (Fig. 2C,D). These immunological control injections indicated that both serum complement proteins and myelin/oligodendrocyte-specific antibodies were necessary to delay the developmental onset of myelination. Additionally, these findings discount the possibility that serum complement entry, due to a breach of the blood— brain barrier alone, elicited developmental myelinsuppression. In vitro studies have indicated that myelin/oligodendrocytespecific antibodies target oligodendrocyte membranes and bind complement [13]. Activation of the first component of the complement protein cascade (C1q) requires the presentation of multiple Fc sites on cell-surface bound antibodies [32]. Thus, to activate the complement cascade, two or more complement-binding monomeric IgG antibodies must bind in proximity to each other on antigenic membrane sites. The binding of one complement-binding pentameric IgM antibody to the membrane is sufficient to activate complement. Consequently, the minimum antigen density required for IgM-mediated cellular attack (04-mediated developmental myelin-suppression) may be relatively less than that required for IgG-mediated cellular attack (GalC-mediated developmental myelin-suppression). IgG antibodies to other oligodendrocyte-specific proteins such as PLP are reportedly less effective in evoking demyelination when applied with complement [32]. The apparent ineffectiveness of PLP antibodies may be due to the relative scarcity of PLP antigen on the oligodendrocyte membrane surface or the diminished ability of these PLP antibodies to activate complement effectively.
733 The delayed onset of myelination following developmental myelin-suppression indicates that the oligodendrocyte lineage cell population is not permanently destroyed by this immunological protocol. The maintained expression of MBP mRNA (Fig. 4) throughout the entire period of myelin-suppression supports this suggestion. In vitro studies indicate that GalC antisera cause an influx of extracellular calcium when applied to oligodendrocytes in culture, resulting in disassembly of microtubules and a concomitant retraction of the oligodendrocyte processes; removal of GalC antisera results in re-extension of oligodendrocyte processes [14]. Although these observations and the results reported here suggest that oligodendrocytes survive developmental myelin-suppression, we cannot exclude the possibility that oligodendrocyte progenitor cells rapidly replace the myelinating population of oligodendrocytes. Recent studies in the adult rat indicate that, although oligodendrocytes survive immunological demyelination, they do not contribute to remyelination [30]. These studies implicate the oligodendrocyte progenitor as the origin of remyelination-competent cells. Finally, we have observed that this immunological protocol is also effective in delaying the development of myelin within the murine spinal cord (Keirstead and Peterson, unpublished observations). The mouse spinal cord normally myelinates during the first two weeks of neonatal development [17]. The injection of serum complement proteins plus polyclonal GalC antibodies into the cord on postnatal day (P) 2 delayed the developmental onset of myelination. Implantation of anti-GalC hybridoma cells in neonatal rats has been shown to inhibit CNS myelin development [41]. These studies illustrate that the antibody-mediated cellular attack paradigm is not species restricted. We have previously demonstrated that developmental myelinsuppression (elicited with monoclonal GalC antibodies plus serum complement) facilitates complete neuroanatomical and functional regeneration of transected E15 chick spinal cord [28]. Untreated control (i.e., normally-myelinated) E15 chicks effect no repair whatsoever following a similar transection of the spinal cord [16,23,24,28,45,46]. We have also utilised a modified immunological protocol to disrupt myelin within the mature avian spinal cord to evoke axonal regeneration [29]. More recent results indicate that immunological demyelination can be maintained for as long as serum complement proteins and myelin/oligodendrocyte antibodies are infused into the CNS. These findings suggest that immunological demyelination may be a useful component in the therapeutic intervention to promote repair of the injured adult spinal cord. ACKNOWLEDGEMENTS
We thank Udo Bartsch and Melitta Schachner for the gift of the O4 antibody, Karin L. Mathias and Ania Wisniewska for technical assistance. This study was supported by grants to J.D.S. from the Medical Research Council of Canada and the Canadian Neuroscience Network. H.S.K. was supported by scholarships from the NCE and the Natural Sciences and Engineering Research Council of Canada.
REFERENCES 1. Bensted, J. P. M.; Dobbing, J.; Morgan, R. S., Reid, R. T. W.; Payling Wright, G. Neuroglial development and myelination in the spinal cord of the chick embryo. J. Embryol. exp. Morph. 5:428 – 437; 1957. 2. Blakemore, W. F. Remyelination of CNS axons by Schwann cells transplanted from the sciatic nerve. Nature 266:68 – 69; 1997. 3. Blakemore, W. F. Observations on remyelination in the rabbit spinal cord following demyelination induced by lysolecithin. Neuropathol. Appl. Neurobiol. 4:47–59; 1978. 4. Bregman, B. S.; Kunkel-Bagden, E.; Schnell, L.; Dai, H. N.; Gao, D.,
734
5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21.
22.
23.
24. 25. 26. 27.
KEIRSTEAD ET AL. Schwab, M. E. Recovery from spinal cord injury mediated by antibodies to neurite growth inhibitors. Nature 378:498 –501; 1995. Bunge, M. P.; Bunge, R. P.; Ris, H. Ultrastructural study of remyelination in an experimental lesion in adult cat spinal cord. J. Biophys. Biochem. Cytol. 10:67–94; 1961. Caroni, P.; Schwab, M. E. Antibody against myelin-associated inhibitor of neurite growth neutralizes nonpermissive substrate properties of CNS white matter. Neuron 1:85–96; 1988. Carroll, W. M.; Jennings, A. R.; Mastiglia, F. L. Immunocytochemical study of the glial cell response in antibody-mediated optic nerve demyelination. Neurosci. Lett. Suppl. 19:S49; 1985. Chomczynski, P.; Saachi, N. Single-step method of RNA isolation by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal. Biochem. 162:156 –159; 1987. Clifford-Jones, R. E.; Landon, D. N.; McDonald, W. E. Remyelination during optic nerve compression. J. Neurol. Sci. 46:239 –243; 1980. DalCanto, M. C.; Lipton, H. L. Schwann cell remyelination and recurrent demyelination in the central nervous system of mice injected with attenuated Theiler’s virus. Am. J. Pathol. 98:101–122; 1980. Dorfman, S. H.; Fry, J. M.; Silberberg, D. H. Antiserum induced myelin inhibition in vitro without complement. Brain Res. 177:105– 114; 1979. Dubios-Dalcq, M.; Niedieck, B.; Buyse, M. Action of anti-cerebroside sera on myelinated tissue cultures. Pathol. Eur. 5:331–347; 1970. Dyer, C. A.; Benjamins, J. A. Antibody to galactocerebroside alters organization of oligodendroglial membrane sheets in culture. J. Neurosci. 8:4307– 4318; 1988. Dyer, C. A.; Benjamins, J. A. Glycolipids and transmembrane signaling: antibodies to galactocerebroside cause an influx of calcium in oligodendrocytes. J. Cell Biol. 111:625– 633; 1990. Eames, R. A.; Jacobson, S. G.; McDonald, W. I. Pathologic changes in the optic chiasm of the cat following local injection of diphtheria toxin. J. Neurol. Sci. 32:381–393; 1977. Eidelberg, E. Consequences on spinal cord lesions upon motor function, with special attention to locomotor activity [Review]. Prog. Neurobiol. 17:185–202; 1981. Foran, D. R.; Peterson, A. C. Myelin acquisition in the central nervous system of the mouse revealed by an MBP-Lac Z trangene. J. Neurosci. 12:4890 – 4897; 1992. Fry, J. M.; Weissbarth, S.; Lehrer, G. M.; Burnstein, M. B. Cerebroside antibody inhibits sulfatide synthesis and myelination and demyelination in cord tissue cultures. Science 183:540 –542; 1974. Graca, D. L.; Blakemore, W. F. Delayed remyelination in rat spinal cord following ethidium bromide injection. Neuropath. and Appl. Neurol. 12:593– 605; 1986. Hamburger, V.; Hamilton, H. L. A series of normal stages in the development of the chick embryo. J. Morphol. 165:49 –92; 1951. Harry, G. J.; Goofrum, J. F.; Bouldin, T. W.; Wagner-Recio, M.; Toews, A. D.; Morell, P. Tellurium-induced neuropathy: metabolic alterations associated with demyelination and remyelination in rat sciatic nerve. J. Neurochem. 52:938 –945; 1989. Hartman, B. K.; Agrawal, H. C.; Kalmbach, S.; Shearer, W. T. A comparative study of the immunohistochemical localization of basic protein to myelin and oligodendrocytes in rat and chicken brain. J. Comp. Neurol. 188:273–290; 1979. Hasan, S. J.; Nelson, B. H.; Valenzuela, J. I.; Keirstead, H. S.; Schull, S. E.; Ethell, D. W.; Steeves, J. D. Functional repair of transected spinal cord in embryonic chick. Restor. Neurol. Neurosci. 2:137–154; 1991. Hasan, S. J.; Keirstead, H. S.; Muir, G. D.; Steeves, J. D. Axonal regeneration contributes to repair of injured brainstem-spinal neurons in embryonic chick. J. Neurosci. 13:492–507; 1993. Herndon, R. M.; Price, D. L.; Weiner, L. P. Regeneration of oligodendrocytes during recovery from demyelinating disease. Science 195: 693– 694; 1977. Hirayama, M.; Silberberg, D. H.; Lisak, R. P.; Pleasure, D. Long-term culture of oligodendrocytes isolated from rat corpus callosum by percoll density gradient. J. Neuropathol. Exp. Neurol. 42:16 –28; 1983. Hruby, S.; Alroul, E. C. Jr.; Seil, F. J. Synthetic galactocerebroside evokes myelination-inhibiting antibodies. Science 195:173–175; 1977.
28. Keirstead, H. S.; Hasan, S. J.; Muir, G. D.; Steeves, J. D. Suppression of the onset of myelination extends the permissive period for the functional repair of embryonic spinal cord. Proc. Natl. Acad. Sci. 89:11664 –11668; 1992. 29. Keirstead, H. S.; Dyer, J. K.; Sholomenko, G. N.; McGraw, J.; Delaney, K. R.; Steeves, J. D. Axonal regeneration and physiological activity following transection and immunological disruption of myelin within the hatchling chick spinal cord. J. Neurosci. 15(10):6963– 6974; 1995. 30. Keirstead, H. S.; Blakemore, W. F. (1997). Identification of postmitotic oligodendrocytes incapable of remyelination in the demyelinated adult spinal cord. J. Neuropathol. Exp. Neurol. Exp. Neurol. 56:1191–1201; 1997. 31. Lampert, P. W. Demyelination and remyelination in experimental allergic encephalomyelitis. Further electron microscopic observations. J. Neuropathol. Exp. Neurol. 24:371–585; 1968. 32. Law, S. K. A.; Reid, K. B. M. (1985) Complement edited by David Male and David Rickwood IRL Press, Oxford Washington, D.C. 33. Ludwin, S. K. Central nervous system demyelination and remyelination in the mouse. An ultrastructural study of cuprizone toxicity. Lab. Invest. 39:597– 612; 1978. 34. Macklin, W. B.; Weill, C. L. Appearance of myelin proteins during development in the chick central nervous system. Dev. Neurosci. 7:170 –178; 1985. 35. Mastaglia, F. L.; Carrol, W. M.; Jennings, A. R. Spinal cord lesions induced by antigalactocerebroside serum. Clin. Exp. Neurol. 26:33– 44; 1989. 36. McKerracher, L.; David, S.; Jackson, D. L.; Kottis, V.; Dunn, R. J.; Braun, P. E. Identification of myelin-associated glycoprotein as a major myelin-derived inhibitor of neurite growth. Neuron 13:805– 811; 1994. 37. Mukhopadhyay, G.; Doherty, P.; Walsh, F. S.; Crocker, P. R.; Filbin, M. A novel role for myelin-associated glycoprotein as an inhibitor of axonal regeneration. Neuron 13:757–767; 1994. 38. Ozawa, K.; Saida, T.; Saida, K.; Nishitani, H.; Kameyama, M. In vivo CNS demyelination mediated by anti-galactocerebroside antibody. Acta Neuropathol. 77:621– 628; 1989. 39. Raine, C. S.; Bornstein, M. B. Experimental allergic encephalomyelitis: a light and electron microscope study of remyelination and ‘‘sclerosis’’ in vitro. J. Neuropathol. Exp. Neurol. 29:552–574; 1970. 40. Ranscht, B.; Clapshaw, P. A.; Price, J.; Noble, M.; Seifert, W. Development of oligodendrocytes and Schwann cells studies with a monoclonal antibody against galactocerebroside. Proc. Natl. Acad. Sci. USA 79:2709 –2713; 1982. 41. Rosenbluth, J.; Liu, Z.; Guo, D.; Schiff, R. Inhibition of CNS myelin development in vivo by implantation of anti-GalC hybridoma cells. J. Neurocytol. 23:699 –707; 1994. 42. Sambrook, J.; Fritsch, E. F.; Maniatis, T. (1989) Molecular Cloning, 2nd edition. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. 43. Sergott, R. C.; Brown, M. J.; Silberberg, D. H.; Lisak, R. P. Antigalactocerebroside serum demyelinates optic nerve in vivo. J. Neurol. Sci. 64:297–303; 1984. 44. Schnell, L.; Schwab, M. E. Axonal regeneration in the rat spinal cord produced by an antibody against myelin-associated neurite growth inhibitors. Nature 343:269 –272; 1990. 45. Shimizu, I.; Oppenheim, R. W.; O’Brien, M.; Schneiderman, A. Anatomical and functional recovery following spinal cord transection in the chick embryo. J. Neurobiol. 21:918 –937; 1990. 46. Sholomenko, G. N.; Steeves, J. D. Effects of selective spinal cord lesions on hindlimb locomotion in birds. Exp. Neurol. 95:403– 418; 1987. 47. Sommer, I.; Schachner, M. Cells that are 04 antigen-positive and 01 antigen-negative differentiate into 01 antigen-positive oligodendrocytes. Neuroscience Letters 29:183–188; 1982. 48. Xie, W.; Rothblum, L. I. Rapid, small-scale RNA isolation from tissue culture cells. Biotechniques 11:323–327; 1991.