UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Neurite inhibiting factors from the developing and mature spinal cord Ethell, Douglas Wayne 1993

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1993_fall_phd_ethell_douglas.pdf [ 8.75MB ]
JSON: 831-1.0086433.json
JSON-LD: 831-1.0086433-ld.json
RDF/XML (Pretty): 831-1.0086433-rdf.xml
RDF/JSON: 831-1.0086433-rdf.json
Turtle: 831-1.0086433-turtle.txt
N-Triples: 831-1.0086433-rdf-ntriples.txt
Original Record: 831-1.0086433-source.json
Full Text

Full Text

NEURITE INHIBITING FACTORS FROM THE DEVELOPINGAND MATURE SPINAL CORDbyDOUGLAS WAYNE ETHELLB.Sc., The University of British Columbia, 1987.M.Sc., The University of British Columbia, 1991.A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THEREQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHYinTHE FACULTY OF GRADUATE STUDIES(Zoology)We accept this thesis as conformingto the required standardTHE UNIVERSITY OF BRITISH COLUMBIAJULY, 1993© Douglas Wayne Ethell, 1993.In presenting this thesis in partial fulfilment of the requirements for an advanceddegree at the University of British Columbia, I agree that the Library shall make itfreely available for reference and study. I further agree that permission for extensivecopying of this thesis for scholarly purposes may be granted by the head of mydepartment or by his or her representatives. It is understood that copying orpublication of this thesis for financial gain shall not be allowed without my writtenpermission.Department of  Zo0 ( The University of British ColumbiaVancouver, CanadaDate^;(); 19-3(Signature)DE-6 (2/88)ABSTRACTThe developing spinal cord supports the outgrowth of ascending anddescending pathways as well as propriospinal connections between spinal cordsegments (Okado and Oppenheim, 1985). During this early embryonic periodaxonal injury is readily repaired, in part due to the similar environments requiredby developing and regenerating axons (Shimizu et al., 1990). Soon aftercompleting the development of long ascending and descending tracts theenvironment of the spinal cord changes such that subsequent damage to axonswithin longitudinal spinal pathways does not result in regeneration. Severalrecent studies on the embryonic chick (Shimizu et al, 1990;Hasan et al.,1991,1993) have established embryonic day (E) 13 as the approximate timewhen the transition from a "permissive" to "restrictive" environment for spinalcord repair occurs. What developmental changes bring about this physiologicalloss of regenerative ability remains unclear.Several reports have documented that neurite inhibiting proteins,associated with myelin, are one group of factors preventing axonal repair in theadult CNS (Caroni and Schwab, 1988;Schwab and Caroni, 1988). Thisdissertation describes an examination of several biochemical changesaccompanying the loss of regenerative ability during spinal cord development.Proteins isolated from an individual developing thoracic spinal cord wereseparated using high resolution 2D gel electrophoresis and compared withproteins expressed at other developmental ages. Initially, ten protein spots wereidentified as showing consistent changes in expression over a developmentalperiod encompasssing the transition. Two of these protein spots were ofsufficient abundance to be further isolated for internal amino acid sequencing.One novel protein has been identified (DSP 7) and is currently the subject ofcloning experiments. Plasma membranes purified from early embryonic chickand rat spinal cord were shown to provide a supportive substrate for thedifferentiation and outgrowth of neuroblastoma x glioma hybrid NG108-15 cells,in vitro. However, plasma membranes isolated from spinal cords later indevelopment displayed an increasing neurite inhibiting activity. Components ofplasma membranes were further purified and tested as substrates for neuriteoutgrowth, in vitro. No inhibitory effects were observed when purified lipidsalone were used as substrates for neurite outgrowth. Although a specific newiteinhibiting protein was not isolated, a crude lipid isolation protocol co-purifiedproteins with strong neurite inhibitory effects.Major components of the extracellular matrix, heparan sulfate andchondroitin sulfate proteoglycans, were also examined for changes over thedevelopmental period encompassing the transition from permissive to restrictiveperiods for spinal cord repair. The ratio of neurite promoting heparan sulfateproteoglycans to neurite inhibiting chondroitin sulfate proteoglycans decreasesas spinal cord development proceeds.TABLE OF CONTENTSTitle Page^ iAbstract iiTable of Contents ivList of Tables viList of Figures^ viiList of Abbreviations xiAcknowledgements )diChapter 1 General Introduction^ 1Chapter 2 Changes in Protein Expression Associated with the^15Developmental Transition from Permissive to Restrictive Statesof Spinal Cord Repair in Embryonic ChickIntroduction^ 16Methods 18Results 26Discussion 39Chapter 3 Developing Chick and Rat Spinal Cords Change from^41Permissive to Restrictive Substrates for Neurite Outgrowth, InVitro.Introduction^ 42Methods 44Results 50Discussion 68Chapter 4 Spinal Cord Lipids Tested for Neurite Inhibition, In Vitro.^71Introduction^ 72Methods 73Results 79Discussion 93Chapter 5 Fractionation of Plasma Membrane Proteins for Neurite^95Inhibiting Assay, In Vitro.Introduction^ 96Methods 98Results 106Discussion 122ivChapter 6 Sulfated Proteoglycan Synthesis During a Developmentally^124Critical Period for Spinal Cord Repair, in Embryonic Chick.Introduction^ 125Methods 127Results 131Discussion 141Chapter 7 General Discussion^ 143References^ 149VLIST OF TABLESTable L Calculated molecular weights and isoelectric points for DSP 1- 2710.Table 2. Statistics of NG108-15 neurite outgrowth on rat spinal cord^65plasma membrane extracts from different developmental ages.Table 3. Neurite inhibitiong activity of plasma membrane protein^107fractions.Table 4. Radio-labelling of chick embryos with inorganic Na235SO4 for^128the evaluation of proteoglycan synthesis in the thoracic spinalcord.Table 5. Specific neurite promoting activity of proteoglycan fractions on 138lain inviLIST OF FIGURESFigure 1.^Representative 2D gel of E14 thoracic spinal cord proteins,^29with DSP 1-10, tubulin and actin labelled.Figure 2.^Preparative 2D gel showing DSP 7.^ 31Figure 3.^Representative areas of 2D gels from E8, E12, E14, E18 and^33P2, showing DSP 1,3,5 and 7.Figure 4.^Representative areas of 2D gels from E8, E12, E14, and E18, 36showing DSP 2,4,6,8,9 and 10Figure 5.^DSP 1 partial amino acid sequence^ 38Figure 6.^Photomicrographs of NG108-15 cells showing how they^49were categorized as possessing short medium or longneurites.Figure 7.^Photomicrographs of NG108-15 cells grown on plasma^52membranes from embryonic chick spinal cord E10.5 andE18.Figure 8.^Photomicrographs of NG108-15 cells grown on plasma^56membranes from embryonic chick spinal cord E12, E14 andE 1 8.viiFigure 9.^Graph of cell adhesion character for embryonic chick spinal^58cord plasma membranes E12, E14 and E18.Figure 10. Photomicrograph showing the inhibitory effects of E18 chick 60spinal cord plasma membranes for NG108-15 cells.Figure 11. Photomicrographs of NG108-15 cells growing on fetal,^63neonatal and adult rat spinal cord plasma membranes.Figure 12. Histogram of the percentage of cells possessing short,^67medium or long neurites when cultured on fetal, neonatal oradult rat spinal cord plasma membranes.Figure 13. Thin layer chromatogram of bovine lipids separated and^76illuminated with uv light.Figure 14. NG108-15 cells grown on lipids purified from the spinal^82cords of adult cow, and El() and hatchling chicks.Figure 15. Pheochromocytoma (PC12) cells grown on lipids purified^84from the spinal cords of adult cow, and El0 and hatchlingchicks.Figure 16. Superior cervical ganglion (SCG) cells grown on lipids^86purified from the spinal cords of adult cow, and El0 andhatchling chicks.viiiFigure 17. NG108-15 cells grown on fractioned bovine spinal cord^88lipids.Figure 18^PC12 cells grown on fractioned bovine spinal cord lipids.^90Figure 19^SCG cells gown on fractioned bovine spinal cord lipids.^92Figure 20. SDS-PAGE gels of bovine plasma membranes, stained with 100coomassie blue, indicating how fractions were divided intosub-fractions for in vitro inhibitory assay.Figure 21. An HPLC absorbance trace from hatchling chick spinal cord 105plasma membrane proteins separated using ion exchangechromatography.Figure 22. NG108-15 cells grown on BSA and fraction 26,^109reconstituted into liposomes.Figure 23. NG108-15 cells on adult rat plasma membrane showing^111boundaryFigure 24. NG108-15 cells growing at the boundary of an area pre-^115coated with purified bovine spinal cord lipids.Figure 25. Cells growing at the boundaries of areas pre-coated with^117dilutions of bovine spinal cord lipid, and BSA control.ixFigure 26. Cells growing at the boundaries of areas pre-coated with^119dilutions of bovine spinal cord lipid, andphosphatidylcholine/cholesterol control.Figure 27. SDS-PAGE of crudely purified lipid fraction, showing co-^121purifying proteins, stained with coomassie blue.Figure 28. A plot of 35S incorporation into embryonic chick spinal cord 133proteoglycans.Figure 29. Plot of the ratio for heparan sulfate to chondroitin sulfate^135synthesis during embryonic chick spinal cord development.Figure 30. Graph of neurite promoting activities of proteoglycans^140isolated from E9 and E17 thoracic spinal cords.LIST OF ABBREVIATIONSAgNO3AMPBDNFBSA°CCACAMscAMPCAPScDNACHAPScmcm2CNBrCNSCO2CSPGdBcAMPDaDEAEd1120DMEMDSPDTTECMEDTAEGFe.g.Et0HFCSFig.GAGHAcHC1HEPESHPLChrHSPGH3PO4IDi.e.1ECIEFSilver nitrateAdenosine monophoshatebrain derived neurotrophic factorBovine serum albumindegrees CentigradeCaliforniacell adheshion moleculescyclic Adenosine monophosphate3-cyclohexylamino-1-propanesulfonic acidcomplementary DNA3-[(3-Cholamidopropyl)dimethyl-ammonio1-1-propane-sulfonatecentimetrecentimetres squareCyanogen Bromidecentral nervous systemCarbon dioxidechondroitin sulfate proteoglycandiButpyl-cyclic AdenosinemonophosphateDaltondiethylaminoethanedistilled waterDulbecco's minimal essential mediadevelopmental spinal cord proteindithiothreitolembryonic dayextracellular matrixethylenediaminetetraacetic acidepidermal growth factorfor exampleethanolfetal calf serumfigureforce of gravityglycosaminoglycanAcetic acidhydrochloride4-(2-hydroxyethyl)-1-piperazine-ethanesulfonic acidhigh pressure liquid chromatographyhourheparan sulfate proteoglycanPhosphoric acidinternal diameterId estion exchange chromatographyisoelectric focusingincorporatedPotassium hydroxidelitremolarmilliAmperemethanolmilligrammillilitremillimetremillimolarmotoneuronrelative molecular weightmicrogrammicrolitremolecular weightSodium chlorideSodium hydroxideneuroblastoma x glioma hybrid cellnerve growth factornonionic detergent-40NeurotrophinNew Yorkpost-hatching/natalpolyacrylamide gel electrophoresisphoshate buffered salineProtein Databases Quest Inc.proteoglycanacidityisolelectric pointplasma membranephenylmethylsulfonylfluorideperipheral nervous systempolysialic acidpolyvinyldifluoridesecondsisotopic sulfur-35Sodium Dodecyl SulfatesulfateN,N,N',N1-tetramethylethylene-diaminethin layer chromatographytrade marktyrosine protein ldnaseUniversity of British Columbiavoltvolumevolume by weightvolume by volumetime/byInc.KOHmAMe0HmgmlnuninMMN1-tgMWNaC1NaOHNG108NGFNP-40NTN.Y.PAGEPBSPDIQuestPGpHpIPMPMSFPNSPSAPVDF355SDSSO4TEMEDTLCTMIrkUBCvoiviwxiACKNOWLEDGEMENTSFirst and foremost I would like to thank my parents for having me, whichotherwise would have made this research considerably more difficult. Early onthey fostered my curiosity by continuing to provide nice toys, even with theknowledge that I would promptly dismantle them, to find out how they worked.I also extend heartfelt thanks to my M.Sc./Ph.D. supervisor, John Steeves for hisfriendship, encouragement and an opportunity to pursue the most interestingcareer imaginable. Working with John over the past six years has provided mewith a very solid scientific foundation on which to build, and an insider'sperspective of neuroscience. Many thanks to Peter Hochachka, Chris McIntoshand Martin Hollenberg for providing enthusiastic and learned supervision, whilechallenging me to pursue excellence. My sincere appreciation also to the peoplewho have helped me along the way: Dave (Davidicus Paticus) & Soiban Patakyfor the many conferences and beers, Dee Webster and Dan Wilson for tips onstress management, Paul Stolortz, Joe Babity, Paul Bone, James Pernu (Jam usDomesticus), Michael Harris, Barbara Taylor, Eric Gagne, Dave Ginzinger andZ-95 FM for providing the tunes that kept me going on those many late nights.Great appreciation is also extended to the faculty, staff and graduate students inthe department of Zoology for providing me a much broader appreciation ofbiology than I could have obtained anywhere else.xi iCHAPTER 1GENERAL INTRODUCTIONNeurons with axonal projections to the peripheral nervous system (PNS)have long been recognized as having a remarkable capacity for regeneration(Waller, 1852;Ranvier, 1871;Ranvier, 1873,Ramon y Cajal, 1928). Subsequentto axotomy, both proximal (still attached to the neuronal cell body) and distal(unattached) axon stumps undergo immediate traumatic signs of degeneration.Distal axon stumps continue with Wallerian (or secondary) degeneration(Waller, 1852;Ranvier, 1871;Ranvier, 1873). Whereas, proximal axon stumpsgo on to form functional growth cones which then elongate, leading toregeneration proceeding at a rate of 1-2 mm per day (Ramon y Cajal, 1928).Growth cones of regenerating axons follow empty Schwan cell tubes recentlyvacated by the degenerating distal axon fragments (Tello, 1911). Consequently,near normal functional recovery will often occur after a relatively short recoveryperiod of weeks or months, provided the proximal and distal nerve ends arerealigned and there is adequate blood perfusion (Sabra and Adams, 1977).Paradoxically, repair after central nervous system (CNS) injuries, in birdsand mammals, is relatively poor (Ramon y Cajal, 1928;Barnard & Carpenter,1950;Bjorklund et al, 1971). When the axons of neurons projecting within theCNS are damaged, both proximal and distal stumps undergo immediatetraumatic degeneration, and again distal axon stumps undergo Walleriandegeneration (Ramon y Cajal, 1928;Bjorklund et al., 1971). Although, theproximal axon stumps develop new growth cones, these structures usuallyelongate only 1-2 mm before arresting and subsequently degenerating proximallytoward the cell body (Ramon y Cajal, 1928; Bjorklund et al, 1971). In general,axotomized fibers within the adult higher vertebrate CNS fail to regenerate oreven substantially elongate (Ramon y Cajal, 1928;Bamard & Carpenter,1950;Bjorklund, et al, 1971). Thus, functional recovery of CNS injuries islimited to compensatory mechanisms, such as short collateral sprouting of2surviving nearby fibers resulting in local rewiring of neuronal connections(Bjorklund et al, 1971;Sabra & Adams, 1977).Explanations for the difference in CNS and PNS axonal regeneration havebeen a center of controversy for many years. It was suggested that neuronsprojecting to targets within the CNS were unable to regenerate because theywere inherently different from their peripheral counterparts. Thus, the inabilityto regenerate was suggested to be a result of the failure of an intrinsic growthprogram in CNS neurons, even though there was evidence to the contrary in theliterature (Tello, 1911). While investigating the trophic actions of sciatic nervegrafts in rabbit, Tello showed that injured optic nerve fibers could sprout andgrow when sutured onto a piece of grafted sciatic nerve. Although Tello's resultswere published in a journal alongside some of Ramon y Cajal's own work, theobservation was refuted and discredited in the 1930's. Tello's fmdings wereforgotten until similar experiments were performed some fifty years later.In 1981, David and Aguayo confirmed and greatly expanded on thefindings of Tello and his colleagues (1911). Using newly developed retrogradeneuroanatomical tract-tracers, David and Aguayo (1981) showed that some cutCNS axons will regenerate if their proximal stumps are permitted access to agrafted PNS environment. Rats were subjected to a complete transection of thethoracic spinal cord such that an impassable gap was left between the rostral andcaudal spinal cord sections. A piece of sciatic nerve (from a litter mate) wasgrafted to either side of the lesion, forming a bridge around the damaged tissue,and the animal was allowed 6-8 weeks for recovery. The neuroanatomical tract-tracer horseradish peroxidase (HRP) was injected into the CNS, above or belowthe PNS graft. After a further 24-48 hours, the animals were sacrificed and thetissue processed to reveal that neuronal cell bodies within the spinal cord hadgrown axons through the entire sciatic nerve graft and into the spinal cord on the3opposite side. This study established that the lack of adult CNS regenerationwas not entirely due to the irreversible suppression of an intrinsic neuronalproperty, but also the result of a difference in the environment of the adult spinalcord (David & Aguayo, 1981). It is now generally accepted that, ifenvironmental conditions are "favorable", at least some CNS axons are capableof significant axonal regeneration.An interesting corollary to the CNS/PNS regeneration paradox is therecent discovery that embryonic chick spinal cord is capable of substantialaxonal repair (Nelson & Steeves,1987;Valenzuela et al., 1988;Shimizu et al.,1990;Hasan et al, 1991, 1993). The success of regeneration in embryonic spinalcord perhaps results from a more favorable interaction between the growth conesof regenerating axons and the spinal cord environment (Hasan et al., 1991).These findings are not surprising considering that axonal regeneration is, at leastin part, a recapitulation of the original developmental event (Holder & Clark,1988). An environment capable of supporting extensive growth cone migrationsduring spinal cord development should also be supportive of regenerating axons.However, embryonic spinal cord is not always supportive to axonal regeneration;although early embryonic spinal cord is permissive for regeneration, laterembryonic and then adult spinal cord, are both restrictive for regeneration(Nelson & Steeves, 1987;Solomenko & Steeves, 1987;Valenzuela et al,1988;Shimizu et al, 1990;Hasan et al, 1991;Webster & Steeves, 1991).Our laboratory has been actively investigating the parameters underlyingthis transition from permissive to restrictive periods for axonal regenerationwithin the spinal cord of embryonic chick (Nelson & Steeves, 1987;Valenzuelaet al, 1988;Hasan et al, 1991). The organization of brainstem-spinal locomotornetworks has been highly conserved throughout vertebrate evolution such thatthese pathways are virtually identical in birds and mammals (Lawrence and4Kuypers, 1968 a,b;Annstrong, 1986;Jordan, 1986;Steeves et al, 1987;Grillnerand Dubuc, 1988). Briefly, embryos of different developmental ages weresubjected to a complete transection of the upper thoracic spinal cord, in ovo.Following a recovery period of approximately 7 days, anatomical repair wasassessed by injection of retrograde neuroanatomical tract-tracers, such as HRP orrhodamine conjugated dextran amines, caudal to the transection site. Afteranother 1-2 days for retrograde transport of the tracer, subsequent histologicalprocessing and examination of brainstem nuclei revealed labelled brainstem-spinal neurons, confirming their fibers had traversed the injury site and projectedseveral centimeters beyond (Nelson & Steeves, 1987;Shimizu et al, 1990;Hasanet al, 1991). Focal electrical micro-stimulation of brainstem locomotor centersresulted in evoked locomotor patterns in the leg muscles, demonstratingfunctional connectivity between brainstem spinal neurons and locomotor patterngenerators within the caudal spinal cord (Sholomenko & Steeves, 1987;Steeveset al, 1987;Valenzuela et al, 1988;Hasan et al, 1991). Anatomical andphysiological connectivity of descending brainstem-spinal projections in sham-operated controls were virtually identical to embryos transected as late asembryonic day (E) 12 of the 21 day developmental period. Embryos subjectedto transection on E13-E14 demonstrated minimal anatomical or functionalrepair, and those transected after E14 showed none. Using these techniques we,and others, have established that a developmental transition of the thoracic spinalcord occurs around E13-E14 that results in a loss of regenerative ability (Nelson& Steeves, 1987;Valenzuela et al, 1988;Shimizu et al, 1990;Hasan et al, 1991).What are the changes in the micro -environment that bring about thisloss of regenerative capacity? Proper discussion of the possibilities will firstrequire a brief summary of some mechanisms involved in the developmentand/or regeneration of an axon.5Growth Cones.When a differentiating neuron sends out a projection and it cannot yet bedetermined if it will give rise to an axon or dendrite, it is called a neurite(Goldberg and Burmeister, 1989). At the distal tip of all growing neurites is aspecialized structure, the growth cone, which interprets the local environmentand directs outgrowth of the fiber projection (Davenport et al., 1993). Firstdescribed by Ramon y Cajal (Cajal, 1928) as, "club-like structures", neuronalgrowth cones still remain much of a mystery.Growth cones are enlarged compared to the rest of the neurite, with themain part ranging in size from 1-2 gm in diameter. Long, thin, finger-likefilopodia continuously protrude and retract from the distal area of the growthcone. Veil-like, "lamellipodia", occur part-way between filopodia, although theyare seen more readily within in vitro systems, being accentuated by the two-dimensional nature of most cell culture systems.A continuous tubulin cytoskeletal network extends from the cell body intothe growth cone (Strittmatter and Fishman, 1991). This network serves as astructural component and a conduit for fast axoplasmic transport. Bothanterograde (away from the cell body) and retrograde (towards the cell body)transport mechanisms are carried along the tubulin network by the specializedproteins, kinesin and dynein, respectively (Lasek and Hoffman, 1976;Hoffinanand Lasek, 1980;Brady and Black, 1986). As the tubulin network projects intothe growth cone, it ends near an actin mesh that underlies the distal and lateralmembrane of the growth cone (Oblinger and Lasek, 1989). This actin meshserves as the polymerization/depolymerization site for actin bundles projectingoutward forming filopodia.6Polymerization and depolymerization of both tubulin and actin areregulated by small sets of proteins (eg funbrin, fodrin and vinculin), all of whichare vulnerable to large changes of intracellular Ca2(Guthrie et al, 1989). Highintracellular Ca2+ concentrations will spontaneously cause the disruption of actinand tubulin polymers (Bandtlow et al, 1993). Such ion fluxes transientlydamage the growth cones structural integrity and lead to what is called, growthcone collapse. As such, molecules that vary intracellular Ca2+ within the growthcone can modify neurite outgrowth (Guthrie et al, 1989;Davenport et al, 1993).Membranes of the filopodia are continuous with the growth conemembrane and primarily depend upon actin filaments for structure (Popov et al,1993). Within the growth cone plasma membranes are a number of proteins thatare the primary signal conduction mechanisms by which the growth coneinterprets the immediate micro-environment (Guthrie et al., 1989). These"receptors" must be able to interpret signals from other nearby cells, theextracellular matrix and soluble factors (Davenport et al, 1993).As the loss of regenerative ability in developing spinal cord is likely dueto adverse interactions between regenerating growth cones and the environmentprovided by the later embryonic spinal cord, we question what has changed toproduce a more hostile (or less supportive) environment for neurite outgrowth.Empirically, at least two possibilities exist. First, there may be a decrease of"factors" necessary for sustained elongation, either soluble components such astrophic factors, or molecules involved in cell adhesion. Second, factorsdeleterious to neurite outgrowth may begin expression, or increase theirexpression later in development. In either case this study has been conducted toestablish such changes and to establish "cause and effect" relationships betweenfactors and neurite outgrowth.7Growth Promoting FactorsGrowth promoting "factors" can be sub-divided into three majorenvironmental components required for substantial neurite elongation(Millaruello et al, 1988). First, adhesion to other cells, neural, glial, andotherwise. Second, trophic factors and other diffusible cytokines. Third, growthcones, like many cellular protrusions, sometimes use the extracellular matrix(ECM) for anchorage and process outgrowth.Adhesion to other cells is mediated by cell adhesion molecules (CAMs)(Rutishauser et al, 1988;Rutishauser, 1988;Kemler et al, 1989). CAMs aremembrane glycoproteins with homotopic and/or heterotopic binding properties.Briefly, there are two major classes of CAMs, depending upon their calciumbinding requirements. Highly calcium dependent proteins include NCadherin(Kemler et al, 1989), whereas neural cell adhesion molecule (NCAM) is anexample of a calcium independent adhesion protein (Rutishauser et al, 1988).NCAM is an intrinsic membrane protein that can be found with orwithout a long polysialic acid (PSA) glycosylation on the extracellular surface(Rutishauser et al, 1988;Rutishauser, 1988 :Edelman, 1991). Post-transcriptionalregulation of NCAM function is achieved by alternative splicing andglycosylation with PSA residues. The regulation of post-translational eventsremains unclear, but it is certain they have a profound affect on NCAM andconsequently a cell's adhesive properties (Edelman, 1991). Some controversyexists as to which form participates in process outgrowth; PSA glycosylationcauses a cell's membrane to act less adhesively toward CAMs on other cells.This may make it easier for fibers traversing an area to pass quickly, whereaswhen a target area is expecting innervation, slightly altering (eg. reducing) PSA8glycosylation of NCAMs could make those cells more adhesive to projectingfibers (Edelman, 1991).The neurotrophic hypothesis, originally proposed by Ramon y Cajal inthe 1880s (see Cajal, 1928), suggests that neuronal fibers require a substancefrom their target of innervation, for survival. Cajal suggested that a target-derived trophic requirement for neuronal survival and maintenance eliminatessome redundancy of neuronal projections, and neurons with inappropriatesynaptic connections. The first neurotrophic factor, Nerve Growth Factor(NGF;Levi-Montalcini and Hamburger, 1951 ;Cohen, 1960;Levi-Montalcini andAngeletti, 1968), was isolated in the early 1950s and initially shown to berequired for the survival of sympathetic ganglia neurons. In numeroussubsequent studies, NGF has been shown to support the survival of a widevariety of neurons both in vitro and in vivo (Thoenen, 1991). It was not until30 years later that a second neurotrophic factor, Brain Derived NeurotrophicFactor (BDNF), was purified and sequenced (Barde et al, 1982). AlthoughBDNF seems to support different populations of neurons within the CNS, itseffects on their survival are the same as with NGF on sympathetic neurons(Hyman et al, 1991). Several other neuron survival factors have been identified,including Neurotrophin 3 (NT-3;Hohn et al, 1990), Ciliary Neurotrophic Factor(CNTF; Lin et al, 1989) and NT-4/5 (Berkemeier et al, 1991;Hallbrook et al,1991). Some of the other factors found to affect neuronal metabolism andregulation include basic and acidic Fibroblast Growth Factor (a & bFGF),Epidermal Growth Factor (EGF), and some members of the interleukin family(Varon et al, 1988;Bandtlow et al, 1990).Transduction of neurotrophin (NGF, BDNF, NT-3, NT-4/5) signals ismediated by a small family of tyrosine protein kinases (irk, irk B and trkC) inassociation with another membrane protein, p75 (Klein et al, 1990, 1991a,91991b;Cordon-Cardo, 1991;Kaplan et al, 1991;Ragsdale and Woodgett,1991;Lamballe et al, 1991). The subset of proteins phosphorylated by activetrks have yet to be worked out, but as described above, we know theirphysiological consequence is the maintenance of neurons in a healthy state. Asfor axonal repair, it has been proposed that for successful axonal regeneration tooccur, higher amounts of neurotrophic factors may be required to maintain thenecessary metabolic level (Thoenen, 1991;Chao, 1992;Verge et al, 1992). Theroles that neurotrophic factors play during CNS development and specifically inaxonal pathfmding remain unclear; however, it is known that they have afunction in the survival of neuronal populations throughout life, and that theremoval of a specific neurotrophin often results in the death of the affectedneurons (Thoenen, 1991).Components of the extracellular matrix (ECM) can be categorized intothree groups: 1) collagens and related molecules, 2) non-collagenousglycoproteins including fibronectin, laminin, entactin, ligatin and 13-galactosidase, 3) proteoglycans and their associated glycosaminoglycan (GAG)side chains (Carbonetto, 1984).While collagens serve as the major structural component of the ECM,interactions between cells and the ECM are usually facilitated by other proteins,such as laminin, that are attached to the fibronectin of the ECM (Kleinman et al,1981). Laminin is an adhesive glycoprotein of the ECM with high affinity forneuronal membrane receptors, type IV collagen, NGF and heparan sulfateproteoglycan (Dundore and Bunge, 1988;Kleinman et al, 1988;Edgar,1989;Millaruello et al, 1989).Proteoglycans are the main component of the extracellular matrix withinthe extracellular space of the CNS (Carbonetto, 1984). These molecules consistof dozens to hundreds of long carbohydrate chains tied to a central protein core.1 0The chemical composition of the carbohydrate side-chains dictate the biologicalcharacteristics for the complex. Proteoglycans containing many heparan sulfateresidues have been implicated as facilitators of a number of neurite outgrowthprocesses (Dundore and Bunge, 1988;Dow et al, 1993).Inhibitory InfluencesAs the nervous system develops, neurite inhibiting factors play a criticalrole in axonal guidance (Schwab and Thoenen, 1985;Schwab et al, 1993). Therecent identification of several neurite inhibiting molecules suggests theirinfluence may vary from poorly adhesive substrates, such as tenascin(Gundersen, 1987), to proteins that strongly and specifically inhibit neuriteougrowth, such as NI-35 (Carom and Schwab, 1988;Igarashi et al, 1993).Two putative neurite inhibiting proteins have been found in the posteriorhalf of somites in embryonic chick (Davies et al, 1990). These two membraneassociated glycoproteins are 48 kDa and 55 IcDa in size and have not been foundin other tissues. Axons from developing motoneurons (MNs) avoid the area,causing them to bundle into what will become segmented ventral roots of thespinal cord (Tosney, 1992).Within the optic tectum of embryonic chick there is a gradient ofpreference for innervating retinal ganglion cells (Undin and Fawcett,1988;Stirling, 1991). Temporal axons preferentially extend on cells from moreanterior tectal regions, whereas nasal axons show no preference (Bonhoeffer andHuf, 1982, 1985). The gradient of preference observed with temporal axons hasbeen shown to arise from the distribution of a neurite inhibiting component ofthe tectal membranes (Walter et al, 1987a,b).1 1Recently, chondroitin sulfate proteoglycans have been identified as"neurite repulsors" acting in a gradient fashion within the developing retina(Brittis et al, 1992). Outside the CNS, components of the ECM form connectivetissues that serve to create tissue, organ and regional boundaries duringdevelopment (Edelman, 1991). The ECM may serve similar roles during CNSdevelopment, forming anatomical barriers that compartmentalize the CNS as itorchestrates the co-development of many subsystems within one tissue.Perhaps the most widely studied neurite inhibitor is a putative 35 kDaprotein (NI-35), isolated from myelin it specifically inhibits neurite outgrowthboth in vitro (Caroni and Schwab, 1988;Schwab and Caroni, 1988) and in vivo(Schnell and Schwab, 1990) by causing growth cone collapse (Bandtlow et al,1993). Interaction of neuronal growth cones with membranes containing NI-35causes a massive increase of Ca2+, released from intracellular stores (Bandtlow etal, 1993). The inhibitory effects do not occur after the application of anantibody, inhibitor neutralizer 1 (IN-1), that specifically blocks NI-35 activity(Schwab and Caroni, 1988;Bandtlow et al, 1993).None of the protein neurite inhibitors have yet been isolated sufficientlyto allow amino acid sequence determination, and as such no complete cDNAsequence is available for analysis. However, some basic characteristics areavailable such as apparent size determinations for both the somite-derivedinhibitory factor and NI-35, and the fact they are both membrane associated.Less is known about the tectal derived inhibitor from Bonhoeffer's laboratory(Bonhoeffer and Huf, 1982, 1985), except that it is temperature sensitive,suggesting conformational requirements as with a protein.12Rationale for this StudyAs development proceeds, the spinal cord changes from an environmentsupporting the outgrowth of numerous axonal projections into an adult conditionthat efficiently maintains connectivity. It is not known why the spinal cordloses, or more correctly eliminates, the potential for axotomized fibers toregenerate, but inhibitory factors may play a pivotal role in this physiologicaltransition. Although there is mounting evidence that inhibitory influences areinvolved in CNS development, it is possible that their function is perhaps morefor directional guidance than the outright prohibition of neurite elongation.Inhibitory factors used in axonal tract development, such as the chondroitinsulfates or somite and tectal derived proteins, are phylogenetically as old as thepathways they help to shape; as such these proteins are likely to be present inboth lower and higher vertebrates as part of the morphogenic machinery.The differences in spinal repair capacity between lower and highervertebrates could be due to a newly evolved level of ubiquitous neuriteinhibition. As with most evolutionary steps reflected in developmentalsequences, a recent addition/modification would occur relatively late in CNSdevelopment. This evolutionary/developmental correlation is seen with the lossof repair capacity in embryonic chick spinal cord occurring somewhere aroundE13, well after all ascending and descending tracts have been completed (Okadoand Oppenheim, 1985). The introduction of ubiquitous neurite inhibitors at thispoint in development would not adversely affect tract formation. Interestingly,the developmental appearance of myelin begins on E13 within the chick spinalcord (Bensted, 1957;Keirstead et al, 1993), and at E18 in mouse (Foran andPeterson, 1992).13Along with the late appearance of myelin is the myelin-associated protein,NI-35. Although NI-35 has not yet been purified sufficiently for amino acidsequence determination, preliminary work has shown a developmentally lateappearance that would be consistent with a ubiquitous inhibitor, decreasingneurite sprouting over the entire CNS (Carom and Schwab, 1988;Schwab andCaroni, 1988;Schnell and Schwab, 1990;Schwab, personal communication).Preliminary in vivo, activity blocking studies with IN-1 and NI-35 have shownthat it may indeed significantly contribute to the lack of regeneration in adult ratspinal cord (Schnell and Schwab, 1990). As well, NI-35 may be just onemember of a family of proteins involved in neurite inhibition (Schwab, personalcommunication).As the most widely used system in vertebrate embryology, many aspectsof chick development have been addressed. The foundations laid by previousand concurrent work allow for a more comprehensive interpretation of newfindings than would be permissible with a less well characterized model. Inturn, the similarities between spinal cord development in chick and othervertebrates, specifically mammals, allows for further inferences thus increasingour knowledge of spinal cord development in those systems as well. Anevaluation of the developing chick spinal cord around E 13 could provide a goodsystem in which to look for ubiquitous neurite inhibitors. Examination of earlyand late embryonic spinal cord components for changes, including neuriteinhibiting activity should establish whether or not this system will provide anappropriate model for the initial identification and purification of ubiquitousneurite inhibitors of the CNS for higher vertebrates.14CHAPTER 2Changes in Protein Expression Associated with aDevelopmental Transition from Permissive to Restrictive Statesof Spinal Cord Repair in Embryonic Chick15INTRODUCTIONIn an effort to identify some of the factors which may underly CNSrepair, I have investigated developmental changes that temporally correlate withthe E13 transition from early (permissive) to late (restrictive) repair periods.The goal of this initial study was to identify substantial changes in proteinconstitution between early and late embryonic spinal cord. High resolution twodimensional (2D) gel electrophoresis allows the comparison of large numbers ofspinal cord proteins at many different developmental stages. As well, proteins ofinterest identified with these procedures can be then purified for amino acidsequencing and analysis.If a ubiquitous neurite inhibitor began, or increased, expression aroundE13, and was expressed at even a moderate level, then it should be detected withthese techniques. In addition, I was interested in proteins only expressed athigher levels early in spinal cord development. Since some of these proteinscould be identified from the same experimental results, those analyses were alsodone. I surveyed the developing thoracic spinal cord for proteins thatsubstantially increased or decreased their expression in a developmental windowbetween E8 and E18, encompassing the E13 transition. Using these criteria, 10protein spots were clearly identified with 5 increasing and 5 decreasing theirexpression levels over the same developmental period. In consideration of thenumbers of proteins under current and future investigation, I have termedproteins identified in this study using the prefix developmental spinal cordprotein, or DSP. The DSPs have been numbered such that those with the highestexpression level after the E13 transition are odd-numbered, while those withhigher expression levels prior to E13 are even-numbered.16Two of the protein spots identified as substantially increasing over theE8-E18 developmental period (DSP 1 and 7), were purified for partial aminoacid sequencing. Proteins were purified using preparative 2D gelelectrophoresis and fragmented to obtain internal amino acid sequence. For DSP11 selected a method of protein fragmentation employing cyanogen bromide(CNBr) in formic acid (Nikodem and Fresco, 1979;Mahboub et al,1986;Charbonneau, 1989;LeGendre and Matsudaira, 1989) since it produces asmall number of large peptides that can be separated using an SDS-PAGEapparatus. Chemical cleavage with CNBr specifically cuts peptide bonds atmethionine residues. As a relatively uncommon amino acid, methionine mayonly occur in a few places within a protein, or not at all (Charbormeau, 1989).For example, 2-3 methionines would create a small number of large peptides thatcould be separated using SDS-PAGE.An alternative method of fragmentation was employed for DSP 7 since itwas less abundant than DSP 1 and fragments could not be purified in sufficientquantity for amino acid sequencing. Proteolytic enzymes cleave proteins atspecifically recognized peptide sequences (Aebersold et al, 1987;Kennedy et al,1988;Aebersold, 1989). Perhaps the most widely used protease for thisprocedure is trypsin, which specifically cleaves peptide sequences at lysine andarginine residues (Charbonneau, 1989). These two common amino acids causethe production of many small peptide fragments. Separation of these peptidefragments for amino acid sequencing is usually accomplished with reverse-phaseHPLC (Aebersold, 1989).17METHODSProtein isolationFertilized White Leghorn chicken eggs (B & J Farms, Surrey, B.C.) wereplaced in a humid incubator at 38°C and turned four times daily. At E8, E10,E12, E14, E16, E18 and post-hatching day (P)2 hatchlings, embryos wereremoved from the eggs and pithed. Within 1-2 min, the complete thoracic spinalcord was isolated by micro-dissection and frozen on dry ice. Following spinalcord excision, the remaining carcass was carefully staged according to thecriteria of Hamburger and Hamilton (1951). All spinal cords were processed andanalysed individually.Isolated spinal cord samples were thawed in hot sample buffer (0.3 %SDS; 5 % 2-mercaptoethanol; 22 mM tris; 28 mM tris-HC1), and boiled for 10min with occasional vortexing. Samples were cooled on ice for 5 min, thentreated with 1/10 volume of 1 mg/nil DNase I for 10 min. Next, the sampleswere spun at 1000 x g for 10 min and the supernatant carefully drawn off.Supernatants were precipitated with 5 x volumes of cold acetone, and left on icefor 1 hr. Following centrifugation (10 min at 14,000 x g), the acetone wasaspirated off and the pellets dried in a speedvac for 10 min at room temperature.Two-dimensional gel electrophoresisDried pellets were resuspended in fresh isoelectric focusing (IEF) samplebuffer (9 M ultrapure urea; 4 % NP-40; 2 % 2-mercaptoethanol; 0.02 M KOH; 2% ampholytes pH = 3-10). Electrophoresis was run according to O'Farrell (1975)with modifications (Garrels, 1983;Latham et al, 1991), in a MilliporeTMInvestigator 2D gel electrophoresis system IEF was run in 1 mm (ID) threadedglass tubes (9.5 M ultrapure urea; 2 % NP-40; 4 % acrylamide:bis; 2.3 %18ampholytes pH = 3-10; 0.06 % Ammonium persulfate) which had been pre-focused for 1 hr. Based on comparison of preliminary 2D gels, the optimalamount of protein that could be accurately resolved on the 2D apparatus was 25-50 lig. Buffer containing 25-50 jig of protein was applied to each tube under 51.11 of overlay (0.5 M urea; 0.2 % NP-40;5 mM DTT; 5 % 2-mercaptoethanol; 0.1% ampholytes pH = 3-10). The upper reservoir contained freshly degassed 0.1 MNaOH, and the lower reservoir contained 0.05 M H3PO4. IEF was run for 17 hrat 1,000 V and then for 1/2 hr at 2,000 V, for a total run of 18,000 V.hr.Immediately following the first dimension, tubes were placed on ice for20 min, and the gels pressure-extruded using a syringe and adapter. Isoelectricfocusing gels were equilibrated in buffer (0.3 M tris; 0.075 M tris-HC1; 3 %SDS; 50 mM DTT; 0.1 % Bromophenol blue) for 2 mm, then laid on top of theslab gels and gently pressed into place. No agarose was required to hold the tubegels in place. SDS-PAGE gels (12.5 % acrylamide; 25 mM Iris; 192 mMglycine; 0.1 % SDS) were 1 mm thick and pre-cooled, in running buffer, to 15°C.The second dimension run was carried out at 16 Watts per gel until theBromophenol dye was within 1 cm of the lower gel edge (typically 4.5 hr) at aconstant temperature of 15°C.Slab gels were removed from the glass plates and fixed overnight in 50 %methanol (Me0H) and 12 % acetic acid (HAc), for silver staining (Morrissey,1981). The following day, gels were rinsed in 3 x 700 ml dH20 for 20 min eachon a slow shaker table. Immersion in a dilute solution of 3.2x10-5M DTT for 30min was sufficient for reduction of the proteins, followed by 2-5 hr in 700 ml of0.012 M AgNO3. Gels were rinsed for 5 min in 700 ml d1120, then developedwith 2 washes in 0.33 M sodium carbonate and 0.19 % fresh formaldehyde. Thedeveloping reaction was stopped by immersion of the gel in 3 % HAc.19Based on testing of a number of 2D gel apparatuses, including the ProteanII from Biorad, a mini 2D system from LKB and the Phast system fromPharmacia, the Investigator 2D system from Millipore consistently provided gelsthat were far better for resolution and reproducibility.Analysis of 2D gelsGels were soaked in 15 % glycerol for 10 min and dried between 2cellophane sheets. The transparency of this drying system, along with thereproducibility of the 2D migration patterns, allowed easy calibration of any twogels by superimposing them on a light box. The protein spots on different gelslined up with minor variations that could be easily corrected by eye.One representative gel was selected from each of: E8, E10, E12, E14,E16, El8 and P2. Each gel in this set was then compared to gels from the othersets for differences between developmental stages. Subsequently, each gel wascompared with other gels at the same developmental stage to confirmconsistency within a particular age group.Initially, E8 and E18 gels were compared. The majority of proteins werefound on both gels in the same configurations and at approximately the samelevel of abundance. Differences noted between the two gels were either spotspresent on one gel but not the other or noticeable changes in the abundance of aspot that was present on both gels. Next, each spot of interest was examined onrepresentative gels from the intervening developmental stages (E10, E12, E14and E16) and the subsequent P2. Any increasing or decreasing trend, inexpression level, was confirmed by comparing the representative gels with othergels from the same developmental stages. In all, 18 gels were used in theanalysis, with a minimum of two from each age.20Calculation of isoelectric point and apparent molecular weightTwo-dimensional protein standards were purchased from Biorad(Richmond, CA.). A 2D run of E18 thoracic spinal cord mixed with 1.5 pl of 2Dmarker provided a migration pattern reference. Patterning of the E18 fractionproteins was not noticeably affected by the presence of the 2D markers, nor werethe relative positions of the markers affected by mixture with the E 18 fraction aschecked by running a 2D gel containing 1.5 1.11 of marker alone.Positional information was also calculated using image analysis. A digitalimage of the E18 + marker gel was captured using a laser scanner and analyzedusing PDIQuest 2D gel analysis software (Protein Databases Inc., HuntingtonStation, N.Y.). Five protein standards were marked which best encompassedDSP 1-10. Once the relative molecular weight (Mr) and isoelectric point (pI)values were assigned to the standards, corresponding Mr and pI values for DSP1-10 were interpolated (Table I).Identification of Actin and TubulinSpots representing tubulin isoforms were confirmed using standardwestern blotting protocols (Sambrook et al, 1989). Actin isoforms wereidentified using the 2D marker proteins mixed and run with E18 protein sample.The 2D standards package contained bovine muscle actin isoforms as several ofthe reference proteins. The much higher levels and slight offset of the samespots in gels containing E18 + marker compared with those run with E18 samplealone, established their identity. In view of the highly conserved nature of actinacross species and the abundance of these spots even without the standardproteins, there is little doubt that the spots indicated on Figure 1 are actinisoforms.21Amino acid sequencingComplete thoracic spinal cords from 25 E16 embryos and one adultchicken were isolated by micro-dissection, and frozen. The spinal cords weresolubilized and acetone precipitated, individually. Protein fractions were storedas acetone precipitates and spun down before use at 10,000 x g for 10 min.Dried pellets were resuspended in fresh preparative IEF sample buffer (8 Murea;2 % CHAPS;2 % 2-mercaptoethano1;0.5 % ampholytes, pH = 3-10;0.01 MKOH) for 1 hr.Protein fractions from the E16 spinal cords were separated on preparative2D gels on a Millipore Investigator 2D gel system. The methods used werebased on the technique of O'Farrell (1975;Garrels, 1983;Harrington et al, 1991),with the modifications described below. Glass tubes 3 mm wide (ID) by 26 mmlong were soaked for several hours in 10 % nitric acid then washed repeatedlywith distilled water and dried. Nylon fishing line (0.25 mm) was threadedthrough the tubes with both ends firmly tied to each other on the outside of thetube. The line provided a flexible support for the gels to prevent stretching orbreaking during handling.Approximately 30 ml of preparative IEF gel solution (4.4% acrylamide;8M urea;2 % CHAPS;2 % ampholytes pH = 3-10;2.8 x 104M ammoniumpersulfate) were placed into a 100 ml graduated cylinder and 8 threadedpreparative tubes were then placed into the solution at the bottom of thecylinder. Next, distilled water was gently overlayed, forcing the gel solution upinto the glass tubes, until the appropriate level had been reached (21 cm).Following 1 hr of polymerization, the tubes were removed from the cylinder andthe nylon line cut such that 2 cm protruded from the bottom of the tube and nonefrom the top.22Dialysis membrane (3 cm2) was placed over the bottoms of the tubes andheld in place with a piece of surgical tubing, to prevent gels from sliding out ofthe tubes during IEF. Five tubes were placed in the IEF chamber for each runwith 5 L of 0.55 % 113PO4 in the bottom chamber (cathode), and 1 L of freshlydegassed 0.1 M NaOH in the upper reservoir. The tube gels were each overlaidwith 40 pi (8 M urea;0.1 % Ampholytes, pH = 3-10;0.2 % CHAPS;50 mMDTT), and pre-focused for approximately 2.5 hr to a constant voltage of 1000 V.Samples applied to each gel consisted of 1000^of protein from eitherEl6 or adult chicken thoracic spinal cord. Conditions for the IEF were 17.5 hrat 1000V followed by 45 inin at 1500 V for a total run of 18,000 V.hr.Following IEF, tubes were removed from the chamber and placed on ice for 30min. The glass tubes were then broken and the gels rinsed free of the glass.Initial attempts at pressure-extruding the gels resulted in unacceptablecontortions of the gels. Gels retreived from broken tubes did not seem to bedamaged and the nylon fishing line provided ample support for movement of thegels into equilibration tubes. Each tube gel was gently agitated in 2 x 30 ml ofequilibration buffer (0.375 tris, pH = 6.8;3 % SDS;50 mM DTT;0.01 %bromophenol blue) for 15 min each and were then placed onto SDS-PAGE slabgels and held in place with 1 % molten agarose.SDS-PAGE slab gels (1 mm) consisted of a 10 % resolving gel with a 2cm stacking gel (4 % acrylamide:bis;0.48 M tris-HC1;14 mM tris;1.4 x 10-3 Mammonium persulfate;0.05 % TEMED). Running conditions for the seconddimension were the same as those described below, except that it tookapproximately 5.5 hr for the run. Following SDS-PAGE, the slab gels wereremoved from the plates and rinsed in 40 % Me0H for 1 hr to remove Laurylsulfate. The gels were stained with 0.2 % Coomassie Blue R-250 in 40 %Me0H for 1 hr and destained with 40 % Me0H. Spot patterns of the preparative232D gels were remarkably similar to the analytical gels (Fig. 2), except for thereduced spot detection of the Coomassie stain. DSP 1,3,5 and 7 were readilydistinguishable, although the actin isoforms were much closer than in theanalytical gels due to the large amount of protein loaded onto the gels.DSP 1 Peptide fragmentation using CNBrSpots were cut from the gels with a clean razor blade. DSP spots, plusactin and tubulin, from the best 15 of 25 E16 preparative 2D gels were pooledand equilibrated with 175 mM tris (pH = 6.8) for 15 min before being pouredinto a 4 % stacking gel mixture and placed into glass tubes, overlaid with 100p1butanol and left 1 hr for polymerization. The tubes were placed in a Bioradmodel 422 electroeluter equipped to trap the protein as it eluted from the bottomof the gel. The elutions were run for 3 hr at 10 mA per tube.The resulting protein-containing buffer was acetone precipitated andwashed repeatedly. The pellets were dried under vacuum centrifugation(Speedvac) and resuspended in 150-200 pl of 0.15 M CNBr in 70 % Formic acid(LeGendre and Matsudaira, 1989). The eppendorf tubes were placed in a glassjar which was flushed with Argon gas and sealed for 24 hr, in darkness.Following CNBr digestion the sample was dried down in a Speedvac andwashed with 50 pi dH20, dried, resuspended in 100 p.1 dH20, acetoneprecipitated on ice for 1 hr, and then spun to a pellet at 10,000 x g.The pellet was resuspended in 20 pl SDS-PAGE sample buffer (120 mMtris, pH = 6.8;5 % SDS;4 % 2-mercaptoethano1;10 % glycero1;0.01 %bromophenol blue) and run on a 0.75 mm mini-gel (16 % resolving;4 %stacking). The gel was washed in dH20 for 5 min and soaked in transfer buffer(10 mM CAPS, pH = 11;10 % Me0H) for 15 min (Aebersold, 1989). The gelwas then placed next to an equally sized piece of polyvinyldifluoride (PVDF),24sandwiched between six pieces of filter paper, and placed in a Hoeffer mini-transphor cell. Transfer of the proteins from the gel to the PVDF was done at 40V for 30 min followed by 60 V for 30 min After a brief rinse with dH20, thePVDF blot was stained (0.2 % Coomassie;12 % HAc;50 % Me0H) for 5 min,destained with repeated washes of 50 % Me0H, and hung to dry (LeGendre andMatsudaira, 1989). Peptide bands were cut from the P'VDF and sent to theUniversity of Victoria amino acid sequencing facility.DSP 7 Peptide fragmentation with trypsinPreparative 2D gels run with 1000 lig of adult spinal cord were soaked in10 mM CAPS (pH=11) and 10 % Me0H for 15 min and then electroblotted ontonitrocellulose in the same buffer, using a semi-dry transfer unit at 0.8 mA/cm2for 2 hr. The blots were stained with Coomassie blue (0.2 %;40 % Me0H;10 %HAc) for 60 s and then destained for 10 min. DSP 7 from five 2D blots werepooled and trypsin digested by personnel in Dr. Reudi Aebersold's laboratory(Biomedical Research Centre, UBC). Peptides were separated using a massspectrophotometer. Several well resolved mass peaks were selected andcompared with a database of peptide masses calculated for each sequence inGenebankTM, if that protein were completely digested by trypsin, cutting peptidebonds at all lysine and arginine residues.25RESULTSTypically, greater than one thousand proteins were resolved on eachanalytical 2D gel, with a clarity similar to that shown in Figure 1. Over onehundred spots, with separation similar to the analytical gels, were resolved oneach preparative 2D gel or blot as shown in Figure 2. The reproducible qualityof the 2D gel techniques ensured a relative position that was consistent for eachprotein spot, at least on the analytical gels. This allowed qualitativecomparisons to be made between gels, and accurate identification of each DSP.Figure 1 shows a representative 2D gel of thoracic spinal cord proteins isolatedfrom an E14 chick embryo. Calculated molecular weights (M1) and isoelectricpoints (pI) of DSP 1-10 are provided in Table I.The triplet DSP 1 (Mr=43.5;pI=5.2), 3 (Mr=43.8;pI=5.1) and 5(Mr=44.2;pI=5.1) are nonexistent or present at very low levels between E8 (Fig.3A) and E12 (Fig. 3B). The molecular weights cited are median calculations asthe spots are quite elongated. A subsequent dramatic increase in expression byE14 (Fig. 3C) makes the DSP 1,3, and 5 proteins some of the most obvious onany 2D gel from E14 through E18 (Fig. 3D). After E18 the abundance of DSP1, 3 & 5 decrease such that they are not present in the P2 gels (Fig. 3E). Thelarge protein spot mass directly above 1,3 and 5 are isoforms of tubulin, whilethe dark spots to the left of DSP 1 are actin isoforms. The close juxtaposition ofDSP 1,3 and 5, and their consistent co-expression in all samples examined fromE14 to E18 suggests they are isoforms of the same core protein.DSP 7 (M1.=45.2;pI=4.9) is expressed at barely detectable levels in E8spinal cord (Fig. 3A), but steadily increases through E12 (Fig. 3B), E14 (Fig.3C), E18 (Fig. 3D) and reaches its highest detected level in the P2 hatchlingspinal cord (Fig. 3E). The protein spots visible on either side of DSP 7 at E 1826Table 1. Calculated isoelectric point (pI) and relative molecular weight (Mr)values for the proteins identified.SPOT pI Mr1 5.2 43,5002 5.2 30,9003 5.1 43,8004 5.0 29,9005 5.1 44,2006 4.9 30,0007 4.9 45,2008 6.7 32,4009 8.2 29,60010 4.9 39,00027Figure 1.^A representative two-dimensional gel visualized with silver staining.The sample applied was E14 thoracic spinal cord. Positions of spots DSP 1-10 areindicated with arrows and respective numbers, as are actin and tubulin isoforms.Molecular weight (Mr) references are indicated on the left margin, in kiloDaltons(10a). Isoelectric focusing was run from left to right, the acidic end is indicatedwith an "a", in the top right corner.28IEF^ aSDS50 - • go.43-Ale•toactin•36-9•21-29Figure 2.^Preparative 2D gel of adult chicken thoracic spinal cord, blottedonto nitrocellulose and stained with Coomassie blue. DSP 7 is indicated by thearrow.30IC731Figure 3.^The area of representative 2D gels showing DSP 1,3,5 and 7 in E8(3a), E12 (3b), E14 (3c) and E18 (3d) and P2 (3e) thoracic spinal cord.32X7T X 5A 1 31, \Apra •33(Fig. 3D) and P2 (Fig. 3E) may be late appearing isoforms of this protein or attoo low a level to be detected prior to E18.Figure 4 shows the 5 DSP that decreased their expression levels betweenE8 and E18, and one more that increased its expression over the same period.DSP 9 (K=29.6;p1=8.2) is not readily detectable from E8 (Fig. 4K) to E12 (Fig.4L), but is apparent by El4 (Fig. 4M) and even more abundant by E18 (Fig.4N); however, by P2 (Fig. 40) its levels are decreased. Some 2D gelsdeliberately overloaded with protein established that DSP 9 is present in E8spinal cord, but at a very low level (not shown).DSP 2 (Mr=30.9;p1=5.2) is robustly expressed at E8 (Fig. 4A), faint atE12 (Fig. 4B), and absent at E14 (Fig. 4C), E18 (Fig. 4D) and P2 (Fig. 4E).DSP 4 (Mr=29.9;p1=5.0) is expressed at E8 (Fig. 4A), but diminishes thereafter(Fig. 4B-E); whereas, DSP 6 (Mr=30.0;pI=4.9) gradually decreases over theentire developmental period (Fig. 4A-E).DSP 8 (Mr=32.4;p1=6.7) is robustly expressed in E8 thoracic spinal cord(Fig. 4F). Another slightly elongated protein spot, positioned above and to theimmediate right, is slightly fainter than DSP 8, but throughout E12 (Fig. 4G),E14 (Fig. 4H) and E18 (Fig. 41) this and other protein spots remain at arelatively constant level; whereas, DSP 8 is reduced from a distinct, if elongatedspot at E8 to a very low expression level at E18. At P2 (Fig. 4J), DSP 8 seemsto be at a slightly higher level than at E18 (Fig. 41), but it is still well below theoriginal E8 (Fig. 4F) levels. Reasons for the elongation of DSP 8 are not knownbut may be due to heterogeneous modifications, such as glycosylation.At E8 DSP 10 (M1=39.0;p1=4.9) is at approximately the same level as theprotein spot to its immediate left and below (Fig 4P). The relative abundance ofthese two protein spots change over the developmental period studied. By El2(Fig. 4Q) DSP 10 is slightly less than the companion spot. Differences in34Figure 4.^Areas of representative 2D gels showing the expression of DSP2,4,6,7,8,9 and 10 over the developmental period studied. The photographs arearranged with E8 gels in the top row with E12 directly beneath, followed by E14and E18, with P2 in the bottom row. DSP spots are indicated with arrows andlabelling.3510•1, X4^6X s^X9•a0—.MN*0• *ft 4AaMIL.36expression remain relatively unchanged between these two proteins at E14 (Fig.4R); however, by E18 (Fig. 4S) DSP 10 appears at a much lower expressionlevel than its companion, and at P2 (Fig. 4T) DSP 10 is barely detectable. Notehow the abundance of the companion spot relative to the neighbouring spots issimilar from E8 (Fig. 4P) to P2 (Fig. 4T), while the abundance DSP 10decreases.Eleven amino acids of sequence were obtained from DSP 1 (Fig. 5), aperfect match with the ovalbumin sequence from GenebankTM confirms theidentity of DSP 1, and suggests that DSP 3 and 5 could be isoforms.37^Figure 5.^Amino acid sequence of a CNBr fragment from DSP 1 (dougAl) asderived by the sequencing facility at the University of Victoria, residues 2-25 areshown. The corresponding sequence from chicken ovalbumin (OACH) shows 100% homology with residues 43-66, from Genebank. Note how all eleven aminoacids match from both sequences, confirming the identity of DSP 1 as chickenovalbumin. The peptide bond at methionine (M) residue 41 of the ovalbuminsequence was cleaved by the CNBr reaction.OACH - Ovalbumin - Chicken100.0% identity in 24 aa overlap10^20dougAI^ YLGAKOSTRTOINKVVRFOKLPGFX^ XCA OH FOVFKELKVHHANENIFYCPIAIMSALAMVYLGAKOSTRTOINKVVRFOKLPGFGOSIEA20^30^40^50^60^70OACH OCOTSVNVHSSLROILNOITKPNOVY5FSLASRLYAEERYPILPEYLOCW(ELYRG0LEPSO^90^100^110^120^13038DISCUSSIONThis study identified 10 different proteins whose level of expressionchanged around the developmental transition from permissive to restrictiveperiods for spinal cord repair. Five of these proteins increase their expressionover an eleven day period between E8 and E18, encompassing the E13transition. As well, five proteins were found to decrease their expression overthe same developmental period. DSP 1 was identified as ovalbumin, while DSP3 and DSP 5 are likely ovalbumin isoforms, from similar CNBr cleavagefragments, nearly identical 2D migration patterns and identical expressionprofiles, likely a contaminant from either blood plasma or the dissectionprocedure.. The growing chick embryo begins to metabolize ovalbumin (eggwhite) in earnest around El3 (Romanoff, 1967), it could be that increasedplasma levels of ovalbumin were detected. Alternatively, during the harvestingof later embryos the yolk sac was often extensively intertwined, necessitating thedamage of extra-embryonic membranes to free the embryo, ovalburnincontamination may have entered at that point. DSP 7 and DSP 9 showcompletely different expression patterns to DSP 1, 3 and 5, suggesting they arenot from the same source of contamination.DSP 7 has been analyzed for peptide fragment homologies and does notshow similarities to any sequences in Genebank; a partial amino acid sequencefor this protein is currently being determined and will be used for furthercloning. No other DSPs were purified for amino acid sequencing due to theirvery small amounts, making the purification of sufficient quantities verydifficult.High resolution 2D gels provide the opportunity to examine a largenumber of components at the same time. To take full advantage of this39technology requires image analysis software that can calibrate different gels andmeasure spot densities for quantitation. Unfortunately, when these studies weredone adequate 2D gel image analysis systems were still under development andnot available. However, 2D databases are currently being compiled that willallow the accurate comparisons of many different gels and the presentation ofprotein identifications in terms of pI and IEF (Garrells, personalcommunication). It is my sincere hope that we will be able to use thesedatabases, when they become available, to identify DSP 2,4,6,8,9 and 10. Aswell, expression levels of DSPs and other proteins could be assayed in responseto a number of experimental manipulations of the developing spinal cord in vivo,for example transection and/or dysmyelination.Detection limitations of this 2D system and the silver staining techniquesemployed are in the nanomolar range per spot (Garrells, 1983). More sensitiveand quanitifiable values could have been obtained if radiolabelled amino acidswere employed, but this was not done due to the difficulties encountered withthe changing growth rate and nutritional uptake of a developing chick embryo(Romanoff, 1967).40CHAPTER 3Developing Chick and Rat Spinal Cords Change fromPermissive to Restrictive Substrates for Neurite Outgrowth, Invitro.41INTRODUCTIONResults from the first chapter identified a small number of proteins withrobust expression changes in the thoracic spinal cord between the E8 and E18developmental periods. Athough several of these proteins are still underinvestigation, it was obvious that alternative purification protocols and aninhibitory assay were required in the search for potential inhibitors of CNSregeneration. Potent inhibitors could work at very low concentrations, or bemasked by abundant proteins that have similar electrophoretic characteristics.An in vivo assay would be far too complicated for analysis, but introduction of afunctional in vitro assay would allow me to track the activity rather than whatmight just be a coincident protein. Because of the importance these structuresplay in the transduction of cell to cell communication, it was decided tospecifically look for neurite inhibition within plasma membranes isolated fromthe developing thoracic spinal cord.I used an in vitro cell culture system to assess whether spinal plasmamembranes, isolated at different developmental stages from chick and rat, arepermissive or restrictive substrates for a neuroblastoma x glioma hybrid clonalcell line (NG108-15). NG108-15 cells, upon induction of differentiation viafactors that increase intracellular cylic AMP, express neuronal propertiesanalogous to those observed in cultured primary neurons. Specifically, NG108-15 cells have been observed to: 1) generate action potentials in response toelectrical and chemical stimuli, 2) form presynaptic terminals, 3) make synapse-like cell to cell contacts with other NG108-15 cells, and 4) form functionalsynapses with cultured myotubes (Nelson et al, 1976; Nirenberg et al, 1983.;Han et al, 1991).42NG108-15 cells grown on plasma membrane fractions, isolated from earlydevelopmental stages, quickly differentiated into neuronal-like cells withextensive process outgrowth; whereas, NG108-15 cells grown on later spinalcord fractions were not well differentiated or developed only a few shortprocesses over the 48 hr culture period.43MATERIALS AND METHODSChicken embryos were incubated and thoracic spinal cords isolated asdescribed in chapter 2. Thoracic spinal cord tissue was taken from chicks atE10.5, E12, E14, and E18. Following spinal cord excision, the remainingcarcass was carefully staged according to the criteria of Hamburger andHamilton (1951). Fetal (E16) and adult Sprague Dawley rats were obtainedfrom the University of Manitoba animal care facility. Neonatal rats (48 hr) wereobtained from the University of Alberta animal care facility. Rats wereeuthanised by a combination of anaesthesia and decapitation, in accordance withanimal care policies of the respective institutions and UBC.Preparation of plasma membranesThe procedure for the isolation of cell-surface plasma membranes wasadapted with slight modifications from Henn (1980). Briefly, the frozen spinalcord tissue was thawed on ice and homogenized in 0.32 M sucrose in 25 mMtris-HC1 (pH = 7.2), containing protease inhibitors (protease inhibitor cocktail I)1 mM phenylmethylsulfonyl fluoride (PMSF), and 10 pig/nil each of aprotininand leupeptin. The homogenate was centrifuged at 500 x g for 10 min to pelletcrude nuclei and cellular debris, and the supernatant was again centrifuged at54,000 x g for 60 min to pellet crude plasma membranes. The pellet wasresuspended in homogenization buffer, containing 50 % (w/v) sucrose (insteadof 0.32 M sucrose). Purified plasma membranes obtained by centrifugation at54,000 x g for 180 min in a discontinuous step gradient of 19 %, 25.5 %, 35.5 %and 50 % (w/v) sucrose, by resuspending the crude plasma membranes in thehomogenization buffer, containing 50 % sucrose, as the bottom cushion layer ofthe sucrose step gradient (Henn, 1980). The enriched fraction of plasma44membranes was obtained as a major band at the interface between 25.5 % and35.5 % of sucrose, and recovered by diluting the solution to approximately 0.32M sucrose with phosphate-buffered saline (PBS; pH = 7.2), followed bycentrifugation at 60,000 x g for 30 min and stored at -70°C.Cell CultureNeuroblastoma x glioma hybrid NG108-15 cells were maintained inmedium D, containing 90 % Dulbecco's Modified Eagle Medium (DMEM), 10% fetal calf serum (FCS), 1x10-6 M hypoxanthine, lx10-4 M aminopterine, 1x10-5 M thymidine, 100 units/nil penicillin, 100 1.tg/m1 streptomycin and 2 mM L-glutamate, at 37°C in a humidified atmosphere of 10 % CO2-90 % air (Ghaharyand Cheng, 1989). For routine assays, cells were dislodged, collected bycentrifugation at 500 x g for 5 min, and resuspended in appropriate volumes ofprewarmed medium at 37°C. To enhance morphological differentiation, NG108-15 cells were induced by subculturing in the presence of 1 rnM dBcAMP inserum free, chemically defmed medium, consisting of 0.75 % DMEM, and 25 %of F-12 nutrient mixture, supplemented with insulin (25 pg/m1), transferrin (50pg/m1), and oleic acid-bovine serum albumin (101.tg/mg albumin/ml) asdescribed by Wolfe and Sato (1982). Numbers of viable cells were determinedby Trypan Blue staining followed by counting with a haemocytometer.In vitro assays for permissive and non -permissive substrates from chickEffects of pennissive and restrictive substrates were monitoredqualitatively by changes in cell morphology at various time intervals afterplating. For assays, plasma membranes were resuspended in PBS, andsolubilized by ultasonication, followed by centrifugation to obtain solubleplasma membrane proteins. Microplates (24-well) were coated with 200 ill/well45of the solubilized membrane proteins and left overnight at 4°C in a humidifiedatmosphere. Upon removal of the coating protein solution by gentle aspiration,the coated wells were washed 3 x each with 1 ml PBS. Hybrid NG108-15 cellswere plated at 20,000 cells/ml/well in serum-free medium in the presence of 1mM dBcAMP. Changes in cell morphology were observed under a phasecontrast microscope at various time intervals. Flattened cells having neuriteslonger than one diameter of the cell body indicated permissive substrates fordifferentiation, whereas round cells without processes or with neurites less thanone-half diameter of cell body indicated undifferentiated cells on a non-permissive substrate.Effects of permissive and non-permissive substrates were also monitoredsemi-quantitatively by a cell number assay. Briefly, circles of 5-10 mm indiameter were marked with a felt pen at the bottom of the 34 mm or 54 mmculture dishes, and coated with a drop (10-20 pd) each of protein samples ofsolubilized plasma membranes at 4°C overnight in a humidified atmosphere.After removal of membrane samples and 3 x washings, each with 2-3 ml PBS,NG108-15 cells at 20,000 cells/nil were plated in serum-free, chemically defmedmedium in the presence of 1 mM dBcAMP. After 24 hr, attached cells werestabilized by direct addition of a dense fixative solution (1 %, v/v,glutaraldehyde in 5 %, w/v, polyvinylpyffolidone), which displaced the culturemedium together with the unattached cells from the attached cells at the bottom,followed by two washings with excess PBS and staining with Giemsa stain. Thenumber of stained cells within the marked circles precoated with membraneproteins or control with bovine serum albumin (BSA) was counted with the aidof an ocular graticule. The substrate permissiveness was expressed as percent ofthe cell number in the membrane-coated area divided by the cell number in theBSA-coated area.46Quantitation of neu rites on rat membranesTo obtain quantifiable values for the extent of neuritic outgrowth on eachsubstrate, the morphology of cells in defmed areas of the wells were categorizedin the 48 hr cultures. Two transects from each well, each constitutingapproximately 5 % of the well bottom, were randomly selected such that theywere not near the edge of the well nor at the center, because of possible unevencoating effects due to a miniscus.Each cell within a transect was scored into one of three categoriesdepending upon the length of its longest neurite. If a cell possessed no processesor neurites shorter than one cell body diameter, it was counted as short (S).Those cells with at least one neurite between 1-3 times the cell body diameterwere counted as medium (M). Long neurite possessing cells (L) had a neuritethat was at least three times as long as the diameter of the cell body. Figure 6shows examples of cells in each of the categories. In all 1,871 cells growing onthe fetal plasma membranes were categorized, compared with 1,844 cells for theneonatal and 1,118 cells for the adult plasma membranes.Protein quantitationFor general determination of protein concentrations, the method ofBradford (1976) was used. For samples containing interfering substances,protein concentrations were determined by the Lowry method (Lowry et al.,1951) as modified by Bensadoun and Weinstein (1976).47Figure 6.^Two photomicrographs showing examples of how NG108-15 cellsgrown on rat spinal cord plasma membranes were categorized as possessing long(L), medium (M) or short (S) neurites as their longest process after 48hr in culture.Magnification = 200 x.4849RESULTSChickenFigure 7 shows the morphology of NG108-15 cells and the extent ofprocess outgrowth after 2, 8, 24, and 48 hr in microwells coated with eitherE10.5, E18 thoracic spinal cord plasma membrane fractions or BSA as a control.Cyto-differentiation and process outgrowth on BSA-coated microwells weretaken as the standard for the NG108-15.Many of the NG108-15 cells grown in the control BSA-coated microwellsexhibited adhesion to the substrate, but only short neurite outgrowth within 2 hr(Fig. 7a). After 8 hr (Fig. 7b), many of the cells in the BSA-microwells haddifferentiated further, having processes approximately twice as long as the celldiameter and some of the neurites also had a few collateral branches, near thegrowth cone. After 24 hr (Fig 7c), the number of neurite extensions from eachcell increased. Although the processes continued to grow, numerous shortdendrite-like processes (multiple branches near the soma) were now alsoapparent. At 48 hr (Fig. 7d), the control cultures consisted of extensive neuriticfibers cross-linking distant cells and cell clusters. Differentiation and extensiveprocess outgrowth were well established.Plasma membrane fraction substrates, isolated from embryonic chickspinal cord at E10.5 (the permissive period for spinal cord repair), markedlyincreased the adhesiveness of NG108-15 cells onto the precoated membranesubstrate as compared to the BSA control substrate, after only short periods ofexposure. At 2 hr (Fig 7e), most of the neuroblastoma cells had undergonemorphological differentiation with short, stout neurites. By 8 hr (Fig 7f), cells inE10.5 membrane coated micro-wells continued to differentiate but the growingneurites remained shorter than those of comparable BSA-coated NG108-15 cells50Figure 7.^Effects of precoating plates with plasma membrane proteinsextracted from spinal cords of chicks at early (E10.5) and late (E18) embryonicstages on cell morphology of NG108-15 cells in culture. Plasma membraneswere prepared from thoracic cord segments from E10.5 and E18 embryonicchicks by sucrose density gradient centrifugation, and solubilized proteins wereobtained after ultrasonication followed by centrifugation. The permissive andrestrictive substrate effects of membrane proteins were measured by precoatingculture dishes at 10011g/m1. Phase-contrast photomicrographs of NG108-15 cellswere taken at 2, 8, 24 and 48hr after cell plating, respectively, for BSA controlcoating (a-d), E10.5 (e-h) and membrane proteins from E18 cords (i-1). Scale forcells = 50 gm.51BSA E10 %52(Fig 7b). These qualitative differences, however, changed over the next 48 hr.At 24 hr (Fig 7g), the cells growing on E10.5 membrane substrates haddifferentiated further and the degree of neurite outgrowth became dramaticallymore extensive than that of control BSA-coated cells (Fig. 7a). After 48 hr (Fig.7h), cells growing on E10.5 substrates became highly differentiated, consistingof an extensive neurite network, and possessed very long axon-like processesand several short, branched, dendritic processes. Furthermore, these highlydifferentiated NG108-15 cells on E10.5 substrates (Fig 7h) remained asindividually attached cells, as compared to mostly cell clusters of differentiatedcells growing on control BSA substrates (Fig 7d).NG108-15 cells grown on plasma membrane fractions, isolated fromembryonic chick spinal cord at E18 (the restrictive period for spinal cord repair),showed little substrate adhesion, neuronal differentiation or process outgrowth.After 2, 8, and 24 hr (Figs. 7i, 7j & 7k, respectively) of culture on El8substrates, there were no apparent morphological changes signifyingdifferentiation. Almost all neuroblastoma cells were spherical and lackedprocesses, with one or two exception having very short processes (eg. Fig 7k).Only after 48 hr (Fig. 7) did cells become attached and showed some degree ofdifferentiation, as evidenced by the appearance of a few short neurites, althoughmany cells maintained a rounded morphology. However, none of the neuriticprocesses on the E18-treated cultures were longer than the diameter of theirparent cell body (Figs. 71).Similarly, when NG108-15 cells were grown on surfaces precoated withplasma membrane extracts from spinal cords at E14, very little substrateadhesion or neuronal differentiation was observed after 2 and 8 hr of culture.Only after 24 hr did cells become loosely attached in clumps of 5-20 cells ofrounded morphology with very short neurites, demonstrating some degree of53differentiation, identical to those grown on E18-coated plates after 48 hr (Fig 71).Even after 48 hr of culture in the E14-coated wells, NG108-15 cell clumpsdifferentiated just slightly further with neurites of one-half to one diameter of thecell body, similar in appearance to BSA control cells after 48 hr (Fig 7c). Celladhesion and thus differentiation were substantially inhibited and delayed byE14-plasma membranes, though comparably less than with E18-membranes.The permissive and restrictive effects of plasma membranes from chickspinal cords at various embryonic stages on cell adhesion and neurite outgrowthof NG108-15 cells were also examined semi-quantitatively by a cell numberassay. Equal numbers of cells were plated onto surfaces precoated with equalamounts of membrane proteins. Under these conditions, seeded cells were notrestricted to only one precoated substrate, but were free to migrate in the dish toadhere to their preferred substrate area precoated with BSA or plasma membraneproteins. The cells pictured in figure 8 are differentiating cells with shortneuritic processes of one-half to two diameters of the cell body, attaching to theprecoated substrates, while non-adhering (undifferentiated) cells had beenwashed off before fixation. The numbers of differentiating cells on precoatedsubstrates of plasma membrane proteins reduced substantially as the embryonicage of the chick increased (Fig 8). These permissive and nonperrnissivesubstrate effects of embryonic spinal cord membrane proteins were furthercompared semi-quantitatively by expressing the number of cells adhering ontoE12, El4 or El8 substrates as a percentage of those attaching to the adjacentBSA control substrate, as shown in Fig 9. The permissive activity of plasmamembrane for NG108-15 cell adhesion decreased gradually and inversely inproportion to the chick embryonic age, being 90 %, 65 % and 20 % for E12, E14and E18 respectively (Fig 9).54Figure 8.^Effects of precoating plasma membrane proteins from spinal cords ofchicks at embryonic days 12, 14 and 18 on cell adhesion of NG108-15 cells inculture. Plasma membrane proteins from thoracic cord segments of E12, E14 andE18 chicks for precoating culture dishes were prepared as described in Figure 1.Culture dishes were precoated with membrane proteins or BSA (control) atconcentrations of 100 jig/ml. Phase contrast photomicrographs of NG108-15 cellswere taken at 20 hr after cell plating. Areas of precoating with membrane proteinsof E12, E14 and E18 or BSA are indicated by solid bars in the middle. Scale = 50Pm5556Figure 9.^Permissive and restrictive effects of precoating plasma membraneproteins from spinal cords of chicks at E12, E14 and E18. Preparation ofmembrane proteins and precoating culture dishes for cell adhesion of NG108-15cells in culture were carried out as described in Figure 2. Substrate permissivenessis expressed as a percentage of cell number of the adhering cells within markedareas of coatings with membrane proteins of E12, E14 or E18 as compared to BSAcontrol (C), taken as 100%. Data are means + S.D. of two experiments intriplicate.57100 — IMINIMMI01/1••80 -.k.i^60WMzf-I•IIND•-fV^40uer20 -0C^E12^E14^El 8Plasma Membranes58Figure 10. Neurite outgrowth inhibitory effect of plasma membrane proteinsfrom spinal cord of E18 embryos. Phase-contrast photomicrograph of NG108-15cells was taken after 3 days of culture. Cells grown in the permissive substrate atthe border of the non-permissive substrate (E 18) extended neurite processesinitially in the direction along the border of the restrictive substrate. After 3 daysof culture, small neuritic processes (indicated by arrows) were observed, extendingonto the initially restrictive environment. Areas of precoating with E18 membraneproteins or BSA are indicated by solid bars in the middle. Scale for cells = 501.im.59MembraneBSA^Proteins60It is noteworthy that the adhesive NG108-15 cells differentiating onpermissive substrates, BSA and E12 membranes, adjacent to the border ofrestrictive substrates, such as E14 and El 8 plasma membranes, extended neuriticprocesses only in the direction along the border of the non-permissive coating,whereas cells inside the permissive environment extended neurites in differentdirections (Fig. 10). In the first 48 hr of culture, no neurite outgrowth onto therestrictive area of nonpennissive membrane-coating was observed. However,after 2-3 days of differentiation in culture, neuritic processes began to extendover the border onto the initially restrictive environment from some (10-20 %) ofthe already differentiated cells growing at the border of the permissive andnonpertnissive coatings (Fig. 10). This suggests that neurite outgrowth of thesecells had initially been inhibited by the nonpermissive spinal plasma membraneextracts.RatExamples of cells cultured on the three plasma membrane extracts (PM)are shown in Figure 11. After just 2 hr in culture (11A), cells on the fetal (E16)PM-coated plates projected neurites considerable distances, although some of thecells showed relatively little differentiation and maintained a roundedappearance. Following 24 hr in culture, many of these cells showed longneurites (11B). Cells grown on the neonate (48 hr) PM-coated plates alsoshowed considerable differentiation and neurite outgrowth after 2 hr in culture(11C) with many long processes after 24 hr (11D). The cells grown on adult (6wks of age) PM-coated plates show differentiation after 2 hr but only shortneuritic outgrowth (11E). After 24 hr (11F), these same cells do not show muchmore process outgrowth than after 2 hr.61Figure 11. NG108-15 cells cultured on plasma membranes isolated from thespinal cords of rats at different stages of development. Cells cultured on plasmamembranes isolated from the spinal cords of E16 rats are shown after 2 lir (a) and24 hr (b) in culture. Neonate (48 hr) plasma membranes coated plates are alsoshown after 2 hr (c) and 24 hr (d) in culture. Much less neuritic outgrowth is seenwith cells cultured on plasma membranes from adult (6 wk) rat spinal cord after 2hr (e) or 24 hr (f). Magnification = 200 x.6263In all, cells were examined at 2, 6, 12, 24 and 48 hr. In cultures up to 24hr the adult PM cultured cells showed considerably less differentiation andprocess outgrowth. Note the similarity of cells cultured on adult PM after 24 hr(11F), and those cells grown on fetal PM after just 2 hr (11A). It was thoughtthat perhaps the lack of neuritic outgrowth for cells gown on the adult PM wasdue to a slower differentiation response, so the cells were given 48 hr todemonstrate a degree of neuritic outgrowth. The differences between thecultures with respect to neuritic outgrowth remained relatively unchanged afterthe additional time.Photomicrographs in Figure 11 are representative areas of 24-well plates.To obtain a more accurate calculation of neurite outgrowth, cells werecategorized semi-quantiatively, as previously described, and the results areshown in Table 2 and Figure 12. The percentage of long process bearing cellson fetal (E16) rat plasma membrane extracts (8.5 + 1.4 %) was significantly(0.05) higher than on either neonatal (5.9 + 1.4 %) or adult (5.4 + 1.2 %) ratextracts, using a student-t test. These results demonstrate that the fetal derivedsubstrates are significantly better for the growth of longer neurites. There wasno significant difference in the percentage of long process bearing cells betweenthe neonatal and adult substrates. However, when the percentage of long andmedium process bearing cells were pooled, there were significantly more ofthose cells on the neonatal substrate (33.5 + 3.0 %) compared with the adultsubstrate (27.6 + 3.2 0/), but no significant difference between the adult (34.9 +2.5 %) and the neonatal substrates.64Table 2. Numbers of cells categorized as possessing long (>3 cell bodydiameters) medium (1-3 cell body diameters) or short (<1 cell bodydiameter) neurites as their longest process. Means of long processcells and long + medium process cells are calculated as percentages ofcells scored in each transect shown with standard deviations.TRANSECT SHORT MEDIUM LONG^fetal 1^191^73^232^139^46^183^302^124^284^169^72^285^225^114^306^189^72^28mean of cells with long processes = 8.5 + 1.4%mean of cells with either long or medium processes = 34.9 + 2.5 %neonatal 1^231^79^222^174^63^193^201^102^214^171^70^185^277^140^206^165^63^8mean of cells with long processes = 5.9 + 1.4 %mean of cells with either long or medium processes = 33.5 + 3.0 %adult 1^84^22^42^150^45^143^107 35^104^187^50^105^111 45^106^171^50^13mean of cells with long processes = 5.4 + 1.2 %mean of cells with either long or medium processes = 27.6 + 3.2 %65Figure 12. A) Percentage of cells possessing long (>3 cell body diameters)neurites as their longest process when cultured on plasma membrane extracts fromfetal (E16), neonatal (48 hr) or adult (6 wk) rat spinal cords. B) Percentage ofcells possessing either long (>3 cell body diameters) or medium (1-3 cell bodydiameters) neurites as their longest process when cultured on plasma membraneextracts from fetal (El 6), neonatal (48 hr) or adult (6 wk) rat spinal cords, after 48hr in culture.661 0cr)CDcn8a.)c:14eq) 644-1 20Fetal^Neonatal^Adult4025Fetal^Neonatal^Adult67DISCUSSIONThe developmental transition from permissive to restrictive environmentsof the chick spinal cord for in vivo axonal repair has been examined by using invitro cell culture assays. Two major differences were found between theNG108-15 cells cultured on plasma membrane substrates isolated from thepermissive (E10.5 and E12) and on those from restrictive (E14 and E18) periodsin chick. First, the degree of cell adhesion was significantly higher for NG108-15 cells grown on early (permissive) spinal cords compared with late (restrictive)membrane fractions. Second, the extent of newite outgrowth (i.e.differentiation) was dramatically greater for cells grown on membrane extractsfrom the earlier embryonic stages.Plasma membrane extracts isolated from E10.5 and E12 chick spinalcords provided a supportive or permissive substrate for NG108-15 cell adhesionand neuron-like differentiation (Fig. 7h). The cells quickly attached to thesubstrate and neurite outgrowth began within 2 hr. By 48 hr the cells had bothlarge axon-like processes and short, branched, dendrite-like processes. As well,the shapes of the cell soma were characteristic of neurons (Fig. 7h).In contrast, significantly fewer NG108-15 cells adhered or survived whencultured on plasma membrane fractions isolated from spinal cords during therestrictive repair period (>E13). Of those cells that survived and eventuallyadhered, to an E14 or E18 plasma membrane substrate, very few differentiated.The rounded appearance of their cell bodies was unlike terminally differentiatedneurons (Fig. 7i-1). Very few of these cells grew any processes and did so onlyafter longer periods in culture.It is possible that inhibitory factors within the El4 and E18 plasmamembrane extracts interfere with NG108-15 differentiation prior to neurite68formation; for example, the substrate could have interfered with dBcAMPtriggering mechanism for NG108-15 cell differentiation. However, my findingthat differentiating cells grown at the border between permissive and restrictivesubstrates extended processes initially only along the border of thenonpennissive coating and eventually extended neurites over to the restrictiveenvironment (Fig 10) supports a direct neurite outgrowth inhibition by the spinalplasma membranes of late embryonic stages. I am uncertain as to why theneurites eventually grow onto the restrictive substrate, but it may be due to agradual loss of biological activity for the inhibitors.Plasma membrane extracts from fetal (E16) rat spinal cord provide asuperb substrate for the attachment, differentiation and neuritic outgrowth ofNG108-15 neuroblastoma x glioma hybrid cells. Cells grown on this substrateshow signs of neuronal like differentiation within 2 hr and have considerableneuritic outgrowth by 24 and 48 hr. Neonatal (48 hr) spinal cord PM are also avery good substrate for the neuritic outgrowth of these cells; whereas, PMisolated from adult rat spinal cord are a poor substrate for neuritic outgrowth byNG108-15 cells. Plasma membrane extracts from adult spinal cord had asignificantly smaller percentage of long neurite possessing cells than fetalextracts and a significantly smaller percentage of long + medium neuritepossessing cells than either the fetal or neonatal extracts.In mammalian systems such as the rat, greater parental care of thenewborn allows for slower embryonic development, as is illustrated bymyelination of the spinal cord in mouse starting around birth and proceedinguntil around 10 days post-natal (Foran and Peterson, 1992). If processesinvolved in the loss of regenerative ability are among these later occuring events,then the capacity for axonal repair should extend later in development than in theembryonic chick.69This study has established that plasma membranes isolated fromdeveloping chick spinal cord are initially permissive to neuronal differentiationand neurite outgrowth by NG108-15 cells. Later in development (after E13), thespinal cord environment becomes restrictive to neuronal differentiation and/orneurite elongation. These fmdings confirm previous work in our laboratoryshowing a developmental decrease in the ability of embryonic chick spinal cordto repair axotomized brainstem-spinal pathways (Hasan et al, 1991,1993;Keirstead et al, 1993). The developing rat spinal cord shows a similardevelopmental decrease in support for extensive neurite outgrowth. Thesedevelopmental variations have been investigated and are discussed in thesubsequent chapters of this thesis.70CHAPTER 4Spinal Cord Lipids Tested for Neurite Inhibition, In vitro71INTRODUCTIONIn chapter 3 I demonstrated a decrease in the capacity of spinal plasmamembranes to support neurite outgrowth, as age increases. These effects wereshown for both chick and rat spinal cords, in vitro. The decreased neuriteoutgrowth seen with cells grown on later spinal cord plasma membrane extractsseemed to be due to direct neurite inhibition rather than a loss of supportivefactors, as shown by the robust growth of neurites on uncoated areas of thesesame plates. Plasma membranes primarily consist of lipids, proteins andcarbohydrates (Saier, 1983), although structural carbohydrates are usuallyconjugated to proteins (glycoproteins and proteoglycans) or lipids (glycolipids)(Saier, 1983;Vance, 1983). In the next three chapters I will describe thepurification and neurite inhibiting assay of components from each of these threemajor plasma membrane constituents: Lipids in chapter 4; Proteins in chapter 5,and glycosaminoglycans in chapter 6.The objective of the research presented in this chapter was to assayconstituitive CNS lipids for neurite inhibiting effects, in vitro. Lipids werepurified from three different sources: 1) El0 chick spinal cord, 2) hatchlingchick spinal cord, and 3) adult cow (bovine) spinal cord. In addition, purifiedbovine lipids were further fractioned into three groups based on charge polarity.All lipids were presented to cells in culture, as liposomes, and assayed forneurite inhibiting effects. Further, these neurite outgrowth responses of NG108-15 cells were compared with pheochromocytoma (PC 12) and superior cervicalganglion (SCG) cell primary cultures.72METHODSLipid PurificationLipids were purified from El0 and hatchling chick spinal cords, as wellas from adult bovine spinal cord (Intercontinental Packers Ltd., Surrey, B.C.).Tissue was collected, as described in chapter 2, and lipids extracted usingmethanol and chloroform (Kates, 1986). Approximately 600 mg of tissue werebriefly homogenized in 3001_11 of dH20 and then 3 ml of HPLC-grade (Fisher)methanol-chloroform (2:1, v/v) was added and the mixture homogenized againfor 2 min at room temperature. The solution was centrifuged at 5000 x g for 5min, the supernatant removed and the residue re-extracted with 4 ml ofmethanol-chloroform by homogenization for 2 more minutes. Aftercentrifugation, the two supernatants were pooled and 2 ml each of chloroformand dH20 were added. After mixing, the solution was placed in a centrifuge at5000 x g for 10 inin and the lower chloroform layer isolated. Chloroform wasremoved from the fractions in a vacuum chamber. Lipids were solubilized in 18inM CHAPS and 10 /TIM tris (pH = 7.4) to a concentration of approximately 10mg/ml.Fractionation of Bovine LipidTo partially separate lipid subtypes, bovine spinal cord lipids werefractioned using thin layer chromatography (TLC;Stahl, 1969). Approximately30 mg of the bovine extract was solubilized in 30 ml of chloroform (1 mg/m1).A standard of phosphatidylcholine/cholesterol was also made to a total lipidconcentration of 1 mg/ml. A faint line was drawn parallel (3 mm) to the edge ofa 0.5 mm pre-made silica plate (20 x 20 cm). Samples of the bovine lipid wereplaced onto the line at 1 cm intervals, along with one spot of the standard.73Equivalents of 10-50 gg of bovine lipid were placed onto each spot, with a totalof 300 gg loaded onto the silica plate. Two lanes of standard containing 10 pigand 20 pg were also run. The samples were allowed to dry onto the silica plateovernight, and then the plate was placed into a chamber containing 300 ml ofrunning solvent (90 % chloroform,5 % methano1;5 % acetic acid) to a depth of 1cm. The solvent took 2 hr to run up the silica plate, reaching within 2 cm of thetop.After a brief drying period in the fumehood, lipids were visualized withuv light (Fig. 13). Uncharged lipids, such as cholesterol, migrated the greatestdistance, charged lipids such as phospholipids did not migrate, while moderatelypolar triglycerides and free fatty acids migrated in between (Stahl, 1969). Theplate was marked into those three zones and the silica in each area scraped offwith a razor blade. Lipids were eluted from the silica with chloroform and rapidvortexing. The silica particles were spun to the bottom of a centrifuge tube(10,000 x g for 10 min), and the elution process repeated (Mangold, 1969).Chloroform was removed from the fractioned lipids in a vacuum chamber, andthe lipids were then resuspended in 18 mM CHAPS, 10 inM tris (pH = 7.4) toapproximately 10 mg/ml, by weight. Suspensions containing each of the threefractions were dialysed against 10 mM tris (pH = 7.4), to remove CHAPS andform liposomes for assay.In vitro assaysNG108 -15 cellsThe protocols followed were the same as described in chapter 3.Neuroblastoma x glioma hybrid NG108-15 cells were maintained in medium D,containing 90 % Dulbecco's Modified Eagle Medium (DMEM), 10 % fetal calf74Figure 13.^Thin layer chromatogram of bovine lipid, illuminated with uv light.The large arrow indicates the direction of the solvent movement, and lines indicatewhere divisions were made for the three fractions isolated. Cholesterol and otherneutrally charged lipids migrated the most and are at the top, triglycerides andfree fatty acids had moderate movement and are found in the center, whilephospholipids and other highly charged lipids did not migrate and are found in thebottom fraction. Standards to the right contained cholesterol (c) andphosphatidylcholine (p).7576serum (FCS), lx10-6 M hypoxanthine, lx10-4 M aminopterine, 1x10-5 Mthymidine, 100 units/ml penicillin, 100 pg/m1 streptomycin and 2 mM L-glutamate, at 37°C in a humidified atmosphere of 10 % CO2-90 % air (Ghaharyand Cheng, 1989). For routine assays, cells were dislodged, collected bycentrifugation at 500 x g for 5 min, and resuspended in appropriate volumes ofprewarmed medium at 37°C. To enhance morphological differentiation, NG108-15 cells were induced by subculturing in the presence of 1 inM dBcAMP inserum free, chemically defined medium, consisting of 0.75 % DMEM, and 25 %of F-12 nutrient mixture, supplemented with insulin (25 pg/in1), transferrin (50pirg/m1), and oleic acid-bovine serum albumin (10 pig/mg albumin/ml) asdescribed by Wolfe and Sato (1982). Numbers of viable cells was determinedby Trypan Blue staining followed by counting with a haemocytometer.Drops of liposome solutions (100) were placed in a concentric circlearound a 35 mm culture plate, and allowed to adhere overnight at 4°C. Afterrinsing the plates with PBS, cells were plated and cultured for 48 hr at 37°C.PC12 cellsThe clonal pheochromocytoma (PC 12) cell line was obtained from Dr. A.Lozano, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, Toronto.PC12 stock cultures were maintained in RPMI-1640 medium supplemented with10-15 % FCS, at 37°C in a humidified atmosphere of 10 % CO2-90 % air(Goodman, 1982). For assays of non-permissive substrate effects, cells weredislodged by enzymatic digestion with 0.025 % trypsin for 30 min at 37°C,collected by centrifugation, and resuspended in appropriate volumes of pre-warmed serum-free, chemically defined medium, consisting of DMEM and N2supplement of insulin (5 gg/m1), transferrin (100 pg/m1), progesterone (20 nM),putrescine (100 mM), and selenium (30 nM) according to Bottenstein (1985).77Cells were plated at concentrations of 20,000 cells/ml, and cultured at 37°C in ahumidified atmosphere of 10 % CO2-90 % air.Superior cervical ganglia cellsThirty five mm culture plates were pre-coated with a 4:1 suspension ofculture grade dH20 and collagen (Campenot, 1992). Circles were marked on thebottom of the culture dishes as described in chapter 3, and 10 ill of liposomesample was placed into each circle. Sympathetic superior cervical ganglia wereisolated from 48 hr neonate Spague-Dawley rat pups, dissociated and placed intothe culture plates (1 ganglion per plate). Cell and media particulars have beenpreviously described in detail (Campenot and Draker, 1989). After 48 hr,cultures were examined for neurite growth in and around the liposome coatedareas.78RESULTSNG108-15 cells grown on the positive control lipid ofphosphatidylcholine/cholesterol (Fig 14A) showed extensive process outgrowthafter 48 hr. Conversely, no NG108-15 cells grew on adult rat spinal cord plasmamembrane extracts, nor were neurites projected into the area (Fig 14B). Lipidspurified from both permissive (E10; Fig 14B and 14C) and restrictive (hatchling;Fig 14D and 14E) periods for spinal cord repair showed no neurite inhibitoryaffects towards NG108-15 cells. Also, no inhibitory effects were seen withNG108-15 cells grown on lipids purified from adult bovine spinal cord (Fig 14Gand 14H). As well, the same neurite growth was observed on lipid fractionsdiluted to 50 % with phosphatidylcholine/cholesterol (Fig 14D, 14F and 14H)and on 100 % (Fig 14C, 14E and 14H).These same lipid substrates also failed to demonstrate neurite inhibitingaffects on either PC12 cells or SCG cells (Fig 15 and 16). PC12 neuriteoutgrowth after 24 hr on the positive control (Fig 15A) was similar to that seenon all of the above mentioned lipid substrates and dilutions (Fig 15C-D). Fibermeshes from SCG cells were not appreciably different on the lipid coatedsufaces than on the control phoshatidylcholine/cholesterol (Fig 16A) coated areaor the surrounding areas coated with just collagen (Fig 16B-G).When lipids were purified from spinal cords using the methanol-chloroform extraction procedure, no inhibitory neurite effects were seen withany of the three in vitro assays used (NG108-15, PC12 and SCG cells).Hatchling chick, El0 chick and adult bovine spinal cord derived lipids, showedno neurite inhibitory effects.The fractionation of adult bovine spinal cord lipids also failed to furtherconcentrate any neurite inhibitory acitivities (Fig 17, 18 and 19). Neurites79grown by NG108-15 cells on the uncharged (Fig 17C), partially charged (Fig17D) and the highly charged (Fig 17E) lipids were not appreciably different thanneurites grown on the phosphatidylcholine/cholesterol control (Fig 17A). PC12cell neurite growth, on the fractioned lipids (Fig 18C-E), did not differ from thatseen on the positive control (Fig 18A). As well, the extensive SCG fiber mesheson the fractioned lipids (Fig 19B-D) were the same as those observed on thephosphatidylcholine/cholesterol control (Fig 19A).80Figure 14. Photomicrographs of NG108-15 cells grown on combinations ofpurified lipids for 48 hr. Panel A shows cells grown onphosphatidylcholine/cholesterol (phos/chol; at a ratio of 2:1) as a positive control.Panel B shows an area covered with adult rat spinal cord plasma membranes as anegative control, no cells attached or sent processes into the area. Panel C showsNG108-15 cells grown on 100 % E10 chick spinal cord lipid, the same substratewas used in panel D but diluted to 50 % with phos/chol. Panel E shows NG108-15 cells grown on hatchling chick spinal cord lipids, Panel F was the samesubstrate but diluted to 50 % with phos/chol. In panel G cells were gown on adultbovine spinal cord lipid, substrate in panel H is the same but diluted to 50 % withphos/chol. Neurite outgrowth on all of the purified lipid substrates matched that ofthe control phos/chol which was indistinguishable from the uncoated areas of theplate (not shown). Magnification = 200x; scale = 50 pin.8182Figure 15. Photomicrographs of pheochromocytoma (PC 12) cells grown oncombinations of purified lipids for 24 hr. Panel A shows cells grown onphosphatidylcholine/cholesterol (phos/chol; at a ratio of 2:1) as a positive control.Panel B shows an area covered with adult rat spinal cord plasma membranes as anegative control, at that concentration no cells attached or sent processes into thearea. Panel C shows PC12 cells grown on 100 % El0 chick spinal cord lipid, thesame substrate was used in panel D but diluted to 50 % with phos/chol. Panel Eshows PC12 cells grown on hatchling chick spinal cord lipids, Panel F was thesame substrate but diluted to 50 % with phos/chol. In panel G cells were grownon adult bovine spinal cord lipid, substrate in panel H is the same but diluted to 50% with phos/chol. Neurite outgrowth on all of the purified lipid substratesmatched that of the control phos/chol which was indistinguishable from theuncoated areas of the plate (not shown). Magnification = 200x; scale = 50 pm.8384Figure 16. Photomicrographs of superior cervical ganglion (SCG) cells grownon combinations of purified lipids for 48 hr. Panel A shows cells grown onphosphatidylcholine/cholesterol (phos/chol; at a ratio of 2:1) as a positive control.Panel B shows SCG cells grown on 100 % E10 chick spinal cord lipid, the samesubstrate was used in panel C but diluted to 50 % with phos/chol. Panel D showsSCG cells grown on hatchling chick spinal cord lipids, Panel E was the samesubstrate but diluted to 50 % with phos/chol. In panel F cells were grown on adultbovine spinal cord lipid, substrate in panel G is the same but diluted to 50 % withphos/chol. Neurite outgrowth on all of the purified lipid substrates matched that ofthe control phos/chol. Magnification = 200x; scale =50 gin.8586Figure 17. Photomicrographs of NG108-15 cells grown on adult bovine spinalcord lipids separated by thin layer chromatography (TLC) for 48 hr. Panel Ashows cells grown on phosphatidylcholine/cholesterol (2:1) as a positive control,and the area in Panel B is a negative control coated with adult rat spinal cordplasma membrane. Panel C shows NG108-15 cells grown on the least TLC-adsorbent lipids eluted from the top most region of the TLC separation,representing uncharged lipids such as cholesterol. Panel D shows cells grown onlipids from the middle region of the TLC, representing moderately charged lipidssuch as free fatty acids and triglycerides. Panel E shows NG108-15 cells grownon the most TLC-adsorbent lipids extracted from the bottom layer of the TLCrepresenting charged lipids such as phospholipids. On all three lipid fractions(C,D and E) NG108-15 cells were able to grow and project extensive neurites.Magnification = 200x; scale = 50 pm.8788Figure 18. Photomicrographs of PC 12 cells grown on adult bovine spinal cordlipids separated by thin layer chromatography (TLC) for 24 hr. Panel A showscells grown on phosphatidylcholine/cholesterol (2:1) as a positive control, and thearea in Panel B is a negative control coated with adult rat spinal cord plasmamembrane. Panel C shows PC12 cells grown on the least TLC-adsorbent lipidseluted from the top most region of the TLC separation, representing unchargedlipids such as cholesterol. Panel D shows cells grown on lipids from the middleregion of the TLC representing moderately charged lipids such as free fatty acidsand triglycerides. Panel E shows PC12 cells grown on the most TLC-adsorbentlipids extracted from the bottom layer of the TLC representing charged lipids suchas phospholipids. On all three lipid fractions (C,D and E) PC12 cells were able togrow and project extensive neurites. Magnification = 200x; scale = 501.un.8990Figure 19. Photomicrographs of superior cervical ganglion (SCG) cells gownon adult bovine spinal cord lipids separated by thin layer chromatography (TLC)for 48 hr. Panel A shows cells grown on phosphatidylcholine/cholesterol (2:1) asa positive control. Panel B shows SCG cells grown on lipids eluted from the topmost region of the TLC separation, representing uncharged lipids such ascholesterol. Panel C shows cells grown on lipids from the middle region of theTLC, representing moderately charged lipids such as free fatty acids andtriglycerides. Panel D shows SCG cells grown on lipids extracted from the bottomarea of the TLC, representing charged lipids such as phospholipids. On all threelipid fractions (B,C and D) SCG cells were able to grow and project extensiveneurites. Magnification = 200x; scale = 50 gm.9192DISCUSSIONUntil now it was not known if naturally occurring lipids in the CNS actedas neurite inhibitors, although the importance of lipids to CNS development andfunction is demonstrated by the seriousness of the neurological syndromesarising from inappropriate lipid metabolism. For example, Tay Sachs syndrome,caused by an inactive lysozomal enzyme leading to the accumulation of TaySach's ganglioside, has a progressive pathology that includes degeneration of thenervous system (Vance, 1983;Sherwood, 1989).Phospholipids are the major structural lipids of all prokaryotic andeukaryotic cells (Vance, 1983). In eukaryotes, the two most quantitativelyimportant phospholipids are phosphatidylcholine and phosphatidylethanolamine.In the CNS, another type of lipid sphingolipids, characterized by a long chainhydroxylated secondary amine, are an important constituent of myelin. Humanmyelin contains 5 % sphingomyelin, 15 % galactosylceramide, and up to 5 % 3'-sulfate-galactosylceramide. Sphingolipid bases occur as components of morecomplex lipids, such as ceramides. The bases themselves are toxic to cells andtherefore occur only in trace amounts. In turn, the addition of carbohydrates tocerarnides give a class of sphingolipid called glycosphingolipid; glucose andgalactose are the most common of these carbohydrates, in the forms N-acetylglucosamine (glucosylceramide) and INTacetylgalactosamine(galactosylceratnide or galactocerebroside), respectively. Commonly in theCNS one or more molecules of N-acetylneuraminic acid (sialic acid) are addedto ceramides, creating a sub-division of glycosphingolipids, called gangliosides.In all, there are over 50 different classes of glycosphingolipid that differ inoligosaccharide content (Vance, 1983).93In this chapter I have demonstrated that methanol/chloroform purifiedspinal cord lipids do not have inhibitory effects on neurite outgrowth in vitro.Lipids were purified from early embryonic chick (permissive) spinal cord,hatchling chick (restrictive) spinal cord, and adult bovine spinal cord, andassayed for inhibitory affects on two immortalized cell lines (NG108-15 andPC 12) and one primary neuronal culture (SCG). No neurite inhibition wasdetected from any of the lipids purified, indicating that the lipid component ofspinal cord plasma membranes is unlikely to evoke growth cone inhibition.This study has reduced the possible sources of neurite inhibition inplasma membranes from mature spinal cord. As well, until now studiesinvolving potential neurite inhibiting membrane proteins have had to rely oncommercially available lipids for the construction of liposomes because ofuncertainty as to whether or not naturally occurring lipids were inhibitory toneurite outgrowth. A disadvantage to this practice is that many membraneproteins require interaction with specific types of lipids, within the plane of theplasma membrane, for appropriate activity. The artifical lipid mixturespurchased for these studies are often extracted from plants and could easily bemissing important components. As well, in presenting a growth cone with amembrane containing a potential protein of interest, it would be moreappropriate to place it in a membrane as close to the representative CNSstructure as possible. Experiments pertaining to neurite inhibitory effects canutilize endogenous CNS lipids, thus decreasing differences from the in vivostructures.94CHAPTER 5Fractionation of Plasma Membrane Proteins forNeurite Inhibiting Assay, In vitro95INTRODUCTIONRecently a number of neurite inhibiting proteins have been identified.Myelin-associated NI-35 (Caroni and Chwab, 1988), two posterior somite-derived proteins (Davies et al, 1990), and a tectal-derived protein (Bonhoefferand Huf, 1982, 1985) are all purported membrane proteins interacting withgrowth cones during cell-to-cell adhesion. However, none of these neuriteinhibiting proteins have been purified sufficiently for amino acid sequencing,due in large part to difficulties incurred in purifying integral membrane proteins,while retaining biological activity. Most protein purification strategies employthe use of detergents to solubilize membrane proteins such that theirhydrophobic domains can be freed from the lipid bilayer; such detergents alsohave the capacity to interact with the proteins so as to lead to completedenaturation, and loss of biological activity.In this chapter I describe several attempts I have made to isolate a neuriteinhibiting activity from the plasma membranes of mature avian and bovine spinalcord. Membrane proteins, from these plasma membranes, were separated usingtwo distinct strategies. First, in a partial attempt to confum the finding ofSchwab and Caroni (1988), SDS-PAGE was used to separate plasma membraneproteins into size fractions, which were reconstituted into liposomes and assayedfor inhibitory effects. A second approach employing ion exchangechromatography (IEC) using a kinder gentler detergent, CHAPS instead of SDS,in an attempt to retain more biological activity. Proteins separated by TEC werealso reconstituted into liposomes and assayed for inhibitory effects in vitro.At a preliminary stage of the lipid project described in the chapter 4,lipids were purified by direct solubilization of bovine spinal cord tissue inchloroform alone, without methanol. Following reconstitution of these lipid96fractions into liposomes, a strong neurite inhibitory effect was observed whenthey were used as substrates, in vitro. These effects were concentrationdependent, and a small number of proteins were found to co-purify with thisprotocol.97METHODSSDS-PAGEBovine plasma membranes, separated as described in chapter 3, weresolubilized in SDS-PAGE running buffer without glycerol or bromophenol blue(0.3 % SDS;100 inM tris, pH = 6.8;5 % 2-mercaptoethanol). To quantify theprotein concentration, a 100 gl aliquot was taken and precipitated with 7 %trichloroacetic acid, on ice for 10-15 min. The suspension was placed in acentrifuge at 14,000 x g for 10 min, supernatant poured off, and the pelletallowed to dry. The pellet was resuspended in 0.5 % Brij-35/10 mM NaOH andthe absorbance read on a uv spectrophotometer at 205 nm.Glycerol and Bromophenol blue indicator dye were added to otheraliquots of solubilized membrane solution, prior to gel loading. Proteins wereseparated on a 10 % SDS-PAGE gel with a 4 % stacking gel using methodsdescribed in chapter 2. Pre-stained molecular weight marker proteins were runin one of the lanes, and 18 jig of protein were applied to each of the 9 remaininglanes. The gel was run until the dye front almost reached the bottom, when thegel was removed, rinsed with dH20, and cut. Horizontal strips corresponding topre-stained molecular weight markers were cut from the gels, leaving one lane ofprotein and one of the pre-stained molecular weight markers uncut, to be stainedwith coomassie blue (Fig 20).The gel strips were diced into small cubes, placed in 2 ml/each of elutionbuffer (18 mM CHAPS;10 niM HEPES, pH = 7.4;50 jig/nil BSA to preventadsorption), and shaken overnight. The buffer was drawn off, replaced with anadditional 1.5 ml, and then shaken again for 4-5 hr. Buffer from both washeswas pooled and concentrated using Centricon concentrators (3,000 Da MW cut-off) in a centrifuge at 6,000 x g until the protein-containing solutions were about98Figure 20. Plasma membrane extract proteins were separated using SDS-PAGEand stained with coomassie blue. Lines indicate where the gel was cut, and labels(20-25) indicate the sample numbers given to proteins eluted from each slice.Relative molecular weights are indicated on the right, from standards.99200 gal, a 100 gal wash was added for the reciprocal wash-out giving a finalvolume of 300 gal.The protein-containing solutions were added to an equal volume of 20mg/ml phosphatidylcholine/cholesterol (2:1) solubilized in elution buffer (18inM CHAPS;10 mM HEPES, pH = 7.4;no BSA). The solutions were dialysedwith (1000 x vol of 5 mM EDTA;10 niM HEPES, pH = 7.4) with rapid mixingfor 4 days, changing the solution twice daily (Brunner et al., 1978). Thesetechniques reconstitute membrane proteins into small liposomes, containingmembrane proteins. Liposomes containing proteins fractioned shown in Figure20, were each assayed for neurite inhibiting activity using the NG108-15 cell invitro assay system described in chapter 3. Droplets of each liposome fraction(10 gl) were placed into marked circles mid-way around the bottom of 50 mmculture dishes and allowed to adhere overnight in a humid chamber at 4°C. Thesolutions were carefully aspirated, the plates rinsed with 3 ml of PBS, and then 3nil of cells (20,000 cells/m1) were plated in each dish and cultured for 48 hr.Plasma membranes were solubilized in SDS-PAGE sample buffer, lessbromophenol blue and glycerol, and then reconstituted into liposomes as acontrol for the initial solubilization effects of the detergent on neurite inhibitingaffects. All cultures were examined at 2, 6, 12, 24 and 48 hr.Ion Exchange ChromatographyHatchling chick (P2) thoracic spinal cord plasma membranes weresolubilized in IEC buffer (10 inM HEPES, pH = 7.4;18 mM CHAPS;proteaseinhibitor cocktail I) for 35 min with gentle agitation. Samples of 200 IA wereloaded onto a 5 ml Q-pak anion column (Biorad), connected to an HPLC.Proteins were eluted in IEC buffer with a linear gradient of 0.0 M NaC1 (pH =7.4) to 1.0 M NaC1 (pH = 7.0) run at 1 ml/min over 20 min. An online uv101detector (280 nm) allowed individual peaks to be collected as they eluted. In all,10 separate protein fractions were collected from each run (Fig. 21). The samepeaks from five identical runs were pooled. Volumes of the solutions wereconcentrated using a centrifuge column fitted with dialysis membrane(centricons), as before. Final volumes of 500 Al were mixed with equal volumesof phosphatidylcholine/cholesterol (2:1) solubilized in IEC buffer (no proteaseinhibitor cocktail) and dialyzed, as before. Two controls were also run: 1) IECbuffer solubilized plasma membranes were reconstituted into liposomes.2)phosphatidylcholine/cholesterol alone.Crude Lipid PurificationApproximately 5 g of bovine spinal cord were minced with a razor blade,placed in 50 ml of chloroform, and vigorously shaken for 15 min. Thesuspension was allowed to settle and undissolved tissue was drawn off thebottom of a separatory funnel. 100 ml of dH20 was added to the solution andgently mixed for 10 min. The water extraction was repeated twice more and thechloroform drawn off. The chloroform was evaporated in a rotovap at roomtemperature. A waxy remnant was scraped from the flask and weighed prior tosolubilization in 18 mM CHAPS, 10 mM HEPES (pH = 7.4), to a concentrationof 20 mg/ml.In vitro assayThe protocols followed were the same as described in chapter 3.Neuroblastoma x glioma hybrid NG108-15 cells were maintained in medium D,containing 90 % Dulbecco's Modified Eagle Medium (DMEM), 10 % fetal calfserum (FCS), 1x10-6 M hypoxanthine, lx10-4 M aminopterine, lx10-5 Mthymidine, 100 units/nil penicillin, 1001.1g/m1 streptomycin and 2 mM L-102glutamate, at 37°C in a humidified atmosphere of 10 % CO2-90 % air (Ghaharyand Cheng, 1989). For routine assays, cells were dislodged, collected bycentrifugation at 500 x g for 5 min, and resuspended in appropriate volumes ofprewanned medium at 37°C. To enhance morphological differentiation, NG108-15 cells were induced by subculturing in the presence of 1 niM dBcAMP inserum free, chemically defined medium, consisting of 0.75 % DMEM, and 25 %of F-12 nutrient mixture, supplemented with insulin (25 pg/m1), transferrin (50Lig/nil), and oleic acid-bovine serum albumin (10 pig/mg albumin/m1) asdescribed by Wolfe and Sato (1982). Numbers of viable cells were determinedby Trypan Blue staining followed by counting with a haemocytometer.Drops of liposome solutions (10}.d) were placed in a concentric circlearound a 35 mm culture plate, and allowed to adhere overnight at 4°C. Afterrinsing the plates with PBS, cells were plated and cultured for 48 hr at 37°C.103Figure 21. A trace of absorbance from proteins as they are eluted from a Q-palcanion exchange column run on an HPLC. The proteins were eluted with a lineargradient of 0.0 M NaC1, 10 HIM HEPES (pH = 7.4), 18 inM CHAPS to 1.0 MNaC1, 10 mM HEPES (pH = 7.0), 18 mM CHAPS run at 1 ml/min over 20 min.Absorbance was read at 280 mn. Lines indicate divisions between fractions.104iOimRESULTSProteins separated with SDS-PAGE did not demonstrate any inhibitoryactivity when reconstituted into liposomes, (Table 3); although, plasmamembranes denatured with the sample buffer and reconstituted into liposomeswithout electrophoresis still showed a high amount of inhibitory activity (Fig.22). The proteins separated on SDS-PAGE actually demonstrated a greaterability for supporting neurite outgrowth than did the BSA control. Cells growingon the SDS-PAGE sample buffer-solubilized control fraction were few innumber, and none showed any process outgrowth after 24 hr (Fig. 22D). Thiscontrol demonstrated that SDS solubilization did not irreversibly destroy all ofthe inhibitory activity in the plasma membrane fractions and implies that activitywas lost in some other way.Plasma membranes solubilized in the IEC buffer and reconstituted intoliposomes, without IEC separation, still showed neurite inhibiting affects but at amuch reduced level to that seen prior to solubilization or even with the SDS-PAGE sample buffer solubilization and reconstitution control (fraction 26;Fig22C and 22D). Protein fractions eluting from the IEC column on the HPLC alsofailed to show any inhibitory affects. NG108-15 cells grew as well on thesesubstrates as uncoated areas of the same plates. None of the eluted protein peaks(Fig 21) showed any inhibitory activity (Table 3).Crude LipidAn example of the inhibitory effects seen with control adult rat plasmamembrane is shown in Figure 23. Note the border across which very few cellswill adhere and only extremely short processes will grow. Compare this cell106Table 3. Neurite inhibiting effects of plasma membrane protein fractionsseparated by ion exchange chromatography and SDS-PAGE, then reconstitutedinto liposomes for in vitro assay using NG108-15 cells.FRACTIONNUMBERNEURITEINHIBITIONIEC 1 -2 -34567891011 + (CHAPS solubilization control)SDS-PAGE^20212223242526---H-+ (SDS solubilization control)107Figure 22. NG108-15 cells grown on BSA after 4hr (A) and 12hr (B)in cultureare shown. Cells grown on plasma membrane from hatchling chick spinal cordsolubilized in SDS-PAGE sample buffer, reconstituted into liposomes, and used assubstrates for N6108-15 differentiation and neurite outgrowth are shown after 4hr(C) and 12hr (D) in culture.1088(0Figure 23. Fractions of adult rat spinal cord plasma membranes are used as acontrol substrate for NG108-15 differentiation and process outgrowth. At 48 hr,note the boundary between the uncoated plate (with many cells) and the areacoated with the adult PM substrate (few cells) Magnification =200x.110•I •,P•.^• 41i•*'111border with those seen in Figures 24, 26 and 27. Figs 25A and 25B show cellsgrowing at the edge of a control BSA drop. There is no obvious cell border seenanywhere along the edge of where the droplet was placed. They grow on thesubstrate as well as the plastic.Cells grown on 100% bovine lipid are shown in Figure 25C and 25D.Again, the cells have no problem growing on the surrounding (uncoated) plate,but note the obvious cell border, or boundary overwhich the cells cannot growand extend processes. In the higher magnification photomicrograph (Fig. 25D),cells at the edge of the boundary will send processes in all directions except intothe lipid-coated zone. The few cells within the lipid-coated zone show very littledifferentiation and no process outgrowth.Bovine lipids were also diluted with a mixture ofphosphatidylcholine/cholesterol (2:1) prior to reconstitution. Cell boundarieswere present in all lipid combinations containing the bovine lipid (Fig. 25C-Fand Fig. 25A-D); whereas, cells cultured on liposomes containing onlyphosphatidtylcholine/cholesterol did not grow appreciably differently on thedroplet or the surrounding, uncoated, plate area (Fig 26E and 26F). No cellborder was seen anywhere around where the droplet was placed.Cells found within the bovine-lipid coated areas demonstrated no processoutgrowth as did many cells that grew immediately on the border. A smalldistance from the border, there were highly differentiated cells with longprocesses. As well, very high concentrations of cells grew on the permissiveside of these borders. The numbers of cells seen in these higher density areas farexceeded those anywhere else on the plate. As bovine lipids adhered to the plateovernight, the liposomes clustered together, leaving areas that were initiallycovered by the lipid solution, but later did not; these areas demonstrated the112same high cell densities seen near the borders, but evenly distributed over theentire region previously covered.Thin layer chromatography separation of these crudely purified lipids intosub-type, as described in chapter 4, removed all inhibitory effects (not shown).SDS-PAGE gel electrophoresis of the crudely purified lipid fractions showed theco-purification of a small number of proteins, at very low levels (Fig 27).113Figure 24. Photomicrograph of NG108-15 cells growing on lipids crudelyisolated from adult bovine spinal cord and reconstituted into liposomes, after 24hr. The boundary between the uncoated culture plate and the area coated withbovine lipid is shown. Note how the cells project neurites in all directions exceptinto the coated region. Magification = 100x.114115Figure 25. Panels A (50x) and B (100x) show low and high magnifications ofNG108-15 cells differentiating on the border of a BSA coated area and anuncoated area of a plate after 48hr in culture, note that there are no obviousborders. Bovine lipid was crudely purified, reconstituted into liposomes and thenused as substrates for NG108-15 differentiation and neurite outgrowth, the borderof the 100% bovine lipid and the uncoated surrounding dish are shown in C (50x)and D (100x) after 48hr in culture. Panels E (50x) and F (100x) show the bovinelipid that was diluted 1:3 with phosphatidylcholine/cholesterl (2:1) prior toreconstitution. Note the obvious cell borders in the cultures containing purifiedlipid fractions.116117Figure 26. Panels A (50x) and B (100x) show cells cultured on purified bovinespinal cord lipid that has been diluted 1:1 with phosphatidylcholine/cholesterol(2:1), recontituted into liposomes and then coated onto culture plates. Note theobvious cell boundary, the cells will not grow or send processes into the bovinespinal cord lipid-coated area. The boundary is not as obvious as that seen with100% bovine 20C, 20D or 75% bovine 20E, 20F lipid. Panels C (50x) and D(100x) show a further dilution of the bovine lipid to 25% (3:1), the border is evenless striking as it was when higher concentrations were used. Control substrates ofphosphatidylcholine/cholesterol (2:1) reconstituted into liposomes without anybovine spinal cord lipid are shown after 48hr, no cell borders were foundanywhere around the coated area E (50x) and F (100x).118119BFigure 27. The solid line indicates a plot of 35S incorporation into proteoglycanswithin the thoracic spinal cord of embryonic chicks. An isotope containingsolution was applied to the chorio-allantoic membrane for a 24hr incorporationperiod. The dpm are calculated in proportion to the wet mass of the isolated spinalcords. The dashed line indicates an equivalent plot (y-scale not shown) ofembryonic body mass increases over the same time periods, as a percentage of thefinal mass for that period (values from Romanoff, 1955).120121DISCUSSIONIn this chapter I have described several attempts to isolate inhibitoryproteins from restrictive plasma membranes. Using two different strategies forthe fractionation of the proteins, including the one used for the initialpurification of NI-35 (Caroni and Schwab,1988;Schwab and Caroni, 1988), Iwas unable to purify any inhibitory protein activity. Although the question oftechnical short-comings could be raised pertaining to my results, similar resultsfrom several international groups (Schwab and Keynes, personalcommunication) argue against that.There are considerable obstacles to the purification of an enzymaticallyactive membrane proteins. I was not surprised by the lack of neurite inhibitingactivity in proteins separated with SDS-PAGE as this detergent will completelydenature most proteins, and it is unlikely that an inhibitory protein would refoldinto an active conformation upon removal of SDS (lauryl sulfate). However, Iwas surprised by the robust inhibitory effects seen with plasma membraneextracts that were solubilized with the SDS-containing sample buffer, and thenreconstituted into liposomes. This fmding suggests several options: 1) there wasincomplete denaturation of the proteins in the control sample, although SDS-solubilized controls were left at room temperature for 2 hr which should havebeen enough time for the detergent to completely denature most if not all of theproteins, 2) inhibitory proteins were denatured and then renatured in an activeconformation, although proteins separated on SDS-PAGE gels failed to becomeactive upon removal of SDS, 3) inhibitory effects seen in plasma membraneextract fractions were not related to enzymatically active protein structures andcould have been due to structural modifications such as glycosylations or122covalently membrane-bound nonprotein components such asglycosarninoglycans on proteoglycans, addressed in chapter 6.At moderate concentrations, detergents such as CHAPS will notcompletely denature many proteins, leaving the secondary and tertiary structurelargely intact (Alvedano et al, 1991). As such, I was quite surprised when theCHAPS buffer solubilization control retained less inhibitory activity than theSDS-PAGE solubilization control.It is possible that using more refined techniques in protein chemistry Icould purify a neurite inhibiting protein from plasma membranes. Were itpossible for me to run such experiments, I would have likely employed adetergent such as octyl glucoside to solubilize the membrane proteins as it hasbeen reported that this detergent can solubilize myelin proteins while retainingmost of their alpha helical and 13-pleated sheet structures (Alvedano et al,1991;Monreal et al, 1992). As well, I would have again employed IEC for thechromatographic fractionation of the proteins. In addition lipids purified fromspinal cord using a methanol/chloroform protocol, as described in chapter 4,would have been used to create the liposomes for in vitro assay, therebyreducing variabilies with the environments normally containing the fractionedmembrane proteins.The proteins that co-purify with lipids, in the cruder protocol, provide anintriguing avenue for future investigation. In the future I hope to have theopportunity to separate these proteins from the majority of the lipid andreconstitute them into liposomes for in vitro assay. If there are indeed protein-associated inhibitory components in these fractions, then this protocol may proveto be a reliable method for the purification of a neurite inhibiting component.123CHAPTER 6Sulfated Proteoglycan Synthesis During a DevelopmentallyCritical Period for Spinal Cord Repair, in Embryonic Chick.124INTRODUCTIONI would like to clarify that the work presented in this chapter was part of acollaborative effort with Drs. Kimberly Dow and Richard Riopelle from Queen'sUniversity in Kingston, Ont. Although I did not perform all of the proceduresmyself, my contributions were an integral part of the study, and I also wrote themanuscript for the resulting publication (Dow et al, 1993). As I did make asignificant contribution to the study, and it is relevant to my dissertation, I feel itappropriate to include this chapter along with acknowledgement to mycollaborators.Proteoglycans (PG) are synthesized in the developing central nervoussystem (CNS) by both neuronal cells and Type-1 astrocytes (Riopelle and Dow,1990; Johnson-Greene et al, 1991). Heparan sulfate proteoglycans (HSPG) havebeen implicated as facilitators of neurite outgrowth due to their interactions withlaminin (Hantaz-Ambroise, 1987; Skubitz et al, 1989; Riopelle and Dow, 1990),fibronectin (Rogers et al, 1985), cell adhesion molecules (Dow et al, 1988) andgrowth factors (Rapraeger et al, 1991; Yayon et al, 1991; Hondermarck et al,1992). Conversely, chondroitin sulfate proteoglycans (CSPG) have arepulsatory effect on newites both in vivo (Brims et al, 1992) and in vitro (Snowet al, 1990). In a recent study it was reported that combining HSPG and CSPGas substrates resulted in much less neurite outgrowth in vitro than could havebeen accounted for by the dilution of HSPG alone, implying either a synergisticreduction of HSPG avidity for neurites or the active inhibition of neuriteoutgrowth by CSPG (Guo et al, 1993). Althought the full ramifications of125HSPG/CSPG interactions with axonal projections have yet to be determined invivo, it is certainly of consequence to spinal cord injury and repair mechanisms.Recent studies in the embryonic chick spinal cord have identified adevelopmental transition when the spinal cord loses its initial ability to repairand/or regenerate axotomized fibers (Shimizu et al, 1990;Hasan et al, 1991,1993). A complete transection of the thoracic spinal cord prior to embryonicday (E) 13 of the 21 day developmental period can be anatomically repairedsuch that retrograde labelling patterns of descending brainstem-spinalprojections to the lumbar region are indistinguishable from untransected controls(Hasan et al, 1991, 1993). Around E13 this repair capacity rapidly diminishesso that injuries occurring after E14 show no signs of anatomical regeneration orfunctional recovery. Although several studies have examined protein and lipidchanges in the spinal cord that accompany this developmental transition (Ethelland Steeves, 1993; Ethell et al, 1993), they have not addressed the issue ofproteoglycans.Here we describe an examination of HSPG and CSPG synthesis in thethoracic spinal cord of chick embryos over a developmental periodencompassing the E13 transition. Analysis of proteoglycans isolated fromembryonic chick spinal cords of different ages showed that the ratio of HSPG toCSPG synthesis decreases as embryonic age increases. Further, in vitro assayson laminin substrates coated with the purified PGs from different stages ofembryonic spinal cord demonstrate that the ratio of HSPG to CSPG is morefavorable to neurite growth before E13 than afterwards.126METHODSRadiolabelling of tissueFertilized white leghom chicken eggs (B & J Farms, Surrey, B.C.) wereplaced in a humid incubator at 38°C and automatically rotated 6 times daily. Onthe appropriate developmental day (E8,10,12,14,16 & 18) an egg was removedfrom the incubator, placed in a sponge recess, and swabbed with 70 % Et0H.Next, a small window (-0 5 nun2) was etched into the shell, 1/3 to 1/2 of theway from the top of the egg (rounded end where the air cell is located), so as notto cut through the shell membranes. A small amount of phosphate bufferedsaline (pH = 7.4) was placed on the etched marks to wet the membranes andminimize tearing Next, the shell fragment was gently removed, followed by thetwo shell membranes to reveal the vascularized chorio-allantoic membrane,which slowly fell away to leave a small pocket.An amount of isotope-containing solution, corresponding to the totalliving tissue mass (Table 4), was placed into the pocket and the hole sealed withmasking tape, and placed back in the incubator for 24 hr without rotation.Application of solutions onto the chorio-allantoic membrane provided aconvenient and adequate method for radio-isotope incorporation into PGs.Following this incorporation period, each egg was gently broken, the embryoremoved and pithed. The complete thoracic spinal cord was isolated by micro-dissection and immediately frozen on dry ice. Following excision of the spinalcord, the remaining carcass was carefully staged according to the criteria ofHamburger and Hamilton (1951). A minimum of 3 replicates (generally 6-9)were completed for each 24 hr period: E8-9, E10-11, E12-13, E14-15, E16-17and E18-19.127Table 4. Radio-labelling of embryos with inorganic Na235SO4 for the evaluationof proteoglycan synthesis in the thoracic spinal cord. Stock Na2SO4solution contained 10.5 mCi/m135S in phosphate buffered saline (massesobtained from Romanoff, 1967).EMBRYONIC WET MASS OF^(1 x) Na235SO4AGE^LIVING TISSUE ISOTOPE APPLIED(days) (grams)^(mCi)E8 2.96 0.03EIO 5.41 0.05E12 9.20 0.09E14 15.10 0.15E16 20.65 0.21E18 27.27 0.27128Proteoglycan purificationProteoglycans were extracted from the spinal cord with 4 M guanidine-HC1, 2 % Triton X-100, 50 mM Sodium acetate, 50 mM EDTA, 1 mM PMSF(pH = 6.0) for 24 hr (Riopelle and Dow, 1990). The extracts were passed overPD 10 columns equilibrated with 8 M urea, 0.05 M Sodium acetate, 0.15 MNaCl (pH = 6.0). Ion-exchange chromatography was performed on a DEAE-5PW column equilibrated in the above buffer and maintained at a flow rate of 1ml/min. Bound species were eluted with a 30 min linear gradient of 0.15-1.0 MNaC1 (pH = 6.0). Aliquots of fractions were taken for liquid scintillationcounting. Profiles were compared by counting areas under peaks.Purified PGs were digested with pronase and the glycosaminoglycan(GAG) chains separated on a DEAR column, equilibrated in 10 % methanol,using an increasing gradient of 0.15-1.0 M NaCl. Fractions were collected andcounted for 35S. Resulting counts were plotted to show which fractionscontained newly synthesized sulfated PGs. The identity of chondroitin andheparan were confirmed by their susceptibility to digestion by eitherchondroitinase or heparanase (Dow et al, 1990).Biological assayPG fractions eluting from DEAE-5PW were pooled, dialyzed extensively,vacuum concentrated and dissolved in defmed medium. Dilutions were used totreat Terasaki microwells previously coated with poly-D-lysine and laminin (5gg/m1). After 24 hours at 4°C, the medium was removed and the wells washedthoroughly prior to seeding of sensory neurons obtained from dorsal root ganglia(DRG) of E8 chick embryos. Dorsal root ganglion neurons were added inserum-defined medium with 4 pM NGF. Plates were incubated at 4°C for 30min in the upright position and then 30 min in the inverted position followed by129incubation at 37°C for 180 inin in the inverted position. Afterwhich cells werescored for process formation (Dow and Riopelle, 1990).130RESULTSProteoglycan synthesisMean counts for PG fractions from each incorporation period arepresented in Figure 28. The incorporation of 35S relative to wet tissue massgives an approximate value for PG synthesis, provided saturation was reached.A set of three embryos per age group received twice the amount of 35SO4 (Table4), and did not show more labelling in the PG preparations, demonstrating that arate of saturation had been achieved with the lesser amount (1 x). Althoughthere is a 3-fold decrease of isotope incorporation as a proportion of wet tissueweight, this plot closely matches one for increases in body mass (as percentages)over the same periods (Fig. 28, dashed line). This correlation demonstrates thatthe amount of PG synthesized changed at the same rate as the growth rate (ie. anincrease of tissue mass produced over a given incorporation period contains thesame proportion of new PG as other incorporation periods).Areas under the peaks representing HS and CS fractions were used as avalue for the relative abundance of the two GAG chains. Figure 29 shows thatthe ratio of HS to CS decreased as development proceeded, with the largestchange occurring between the E10-11 and E12-13 incorporation periods. AfterE13 the HS to CS ratio continued to decrease, at least to E18-E19, but at aslower rate. The ratio of HS to CS at E8-E9 (1.6 + 0.1) is over twice that at E18(0.7 + 0.1). As these are relative values, it cannot be determined if this change isdue to an increase of CS synthesis, a decrease of HS synthesis, or both.Biological assayA biological assay to quantify the neurite growth-promoting effects ofproteoglycans was carried out on fractions eluting from DEAE-5PW using two131Figure 28. Chicken embryos of six embryonic ages had solutions containing "SO4placed onto their chorio-allantoic membranes for 24 hr incorporation periods. Thesolid plot shows disintegrations per minute (dpm) of SO435 ^incorporated intoproteoglycans (PG) per mg of wet tissue weight for 6 incorporation periods. Thedotted plot shows a scalar equivalent of changes in total body mass over the same24 hr periods (calculated from weights in Romanof , 1967).13220 -8-9 10-11 12-13 14-15 16-17 18-19133Figure 29. PGs were purified from chick embryos following the 24 hrincorporation periods for 35SO4. Heparan sulfate and chondroitin sulfate GAGchains were separated on a DEAE-5PW column connected to a HPLC system,using a 0.15-1 M NaC1 gradient. Fractions were counted for radioactivity and aplot of the eluting radioactivity created. Areas under the peaks corresponding toHS and CS, confirmed with heparanase and chondroitinase sensitivity, werecalculated as were ratios for HS/CS. The plot shows the HS/CS ratios for the sixdevelopmental times periods studied.1340. 10-11 12-13 14-15 16-17 18-19135different substrates, poly-D-lysine and poly-D-lysine/laminin. Laminin wasused as a substrate because it has specific binding sites for HSPGs (Kouzi-koliakos et al, 1989; Skubitz et al, 1988) and does not bind CSPGs (Dow,unpublished observations). While both HSPGs and CSPGs bind to poly-D-lysine, CSPG the latter bind more avidly to this substrate such that when amixture of PGs is immobilized on poly-D-lysine the growth-promoting effects ofHSPG may be inhibited by CSPG (Guo et al, 1993).Titration profiles of neurite-promoting activity associated with PGsisolated from E9 and E17 embryos are shown in Figure 30. E9 fractionssupported neurite growth that was greater than control on both poly-D-lysineand laminin. E9 fractions enhanced neurite growth on laminin at dilutions up to1/1000, but neurite growth on poly-D-lysine was inhibited up to 1/1000. Onlaminin the growth-promoting ability of E9 fractions remained above control at1/10,000, but had returned to control levels at this dilution of E17 fractions.Specific neurite-promoting activity (NPA) was calculated at E9 and E17 byusing as the denominator the maximum dilution giving neurite growth that wasabove control, and correcting for weight of tissue from which the PGs used inthe bioassay was extracted. As shown in Table 5, specific NPA on laminin wasat least 75-fold higher at E9 than at E17, supporting the suggestion that there ismore HSPG-associated NPA present at E9 than on E17. On the poly-D-lysinesubstrate both E9 and E17 extracts contain neurite growth inhibitory activity(NIA). Such activity partially masks NPA in E9 extracts out to dilutions of1/100 but completely masks NPA of El7 extracts to dilutions of 1/1000. In bothcases, at higher dilutions of PG extract NIA titres out and NPA is observed.That the inhibitory activity is associated with the CSPG is suggested by thefinding that the effects were not detected when the substrate contained laminin.136Thus E9 spinal cord is characterized by more HSPG-associated NPA and lessCSPG-associated NIA whereas in the E17 spinal cord the opposite was true.137Table 5. Specific Neurite Promoting Acitivity (NPA) of PG fractions on lamininEmbronic day^E9^E17Wet tissue wt (mg)^1.9 + 0.4^11.3 ±2.8(mean + SD)Max dilution giving^104^1 0-3NPA > controlSpecific NPA *^9210 + 696 122 + 27(mean + SD)* fraction of contolNPA at max dilutionwt of tissue (mg) x dilution138Figure 30. Neurite-Promoting Activity (NPA) of proteoglycans isolated fromthoracic spinal cord of E9 (A) and E17 (B) embryos. Proteoglycans wereimmobilized on a substrate of poly-D-lysine (o) or poly-D-lysine and laminin (•)and DRG neurons were seeded and scored for process formation. Results areexpressed as means + SD (n = 7) of the fraction of control response.1392.01.50.5• — laminino — PDL1.0A1^1^1^I 4^0.010^i 2 103100.• — laminin0 — PDL1^1^1 10^102103104Substrate Coating Concentration(Reciprocal Dilution)DISCUSSIONThese studies have clearly shown a decreasing ratio of heparan sulfate tochondroitin sulfate synthesis in the developing thoracic spinal cord of chick.Between the E8-E9 and E18-E19 incorporation periods, the HS:CS ratiodecreases by more than 50 %. Further, the greatest decrease occurs at E13, thesame developmental time as the loss of functional axon regeneration. Previousstudies have established E13 as the time when embryonic chick spinal cordchanges from a permissive to a restrictive environment for neurite outgrowth(Ethell et al, 1993) and axonal regeneration (Hasan eta!, 1991, 1993). In vitroassay of purified PGs from E9 (permissive) and E17 (restrictive) spinal corddemonstrated a greater degree of HSPG-associated neurite promoting activity inE9 tissue and a greater degree of CSPG-associated neurite inhibiting activity inE17 tissue.Heparan sulfate has non-covalent interactions with a number ofcomponents that affect neurite growth. Laminin has at least four heparin-binding domains, distinct from sites bound by neuronal growth cones (Engvall etal, 1986; Skubitz et al, 1989). As HSPGs bind to both NCAM and laminin atdifferent sites (Cole and Akeson, 1989), they can act as tethers between theextracellular matrix and neuronal plasma membranes (Johnson-Greene et al,1991). HSPGs have also been shown to bind basic fibroblast growth factor(bFGF) and facilitate its binding to bFGF receptors on neurons (Kan et al,1993). HSPG-associated NPA of E9 spinal cord is 75 times higher than at E17.Thus, damaged axons attempting to regenerate may have greater success at E9 inpart due to the higher NPA than is found at El 7 or later.Chondroitin sulfate distribution in the developing retina has beencorrelated with axon boundaries, suggesting a role in pattern formation (Brittis et141al, 1992). In vitro CS inhibits neurite growth even in the presence of NCAMand laminin (Snow et al, 1990). Permissive spinal cord (E9) PGs had a neuriteinhibiting activity that weakly blocked HSPG-associated NPA; whereasrestrictive spinal cord (E17) PGs had over 10 x the blocking potential. Thus aregenerating growth cone that relied on HSPG for some attachment would findE17 spinal cord less hospitable, in part due to the higher CS-associated NIA.In terms of sulfated proteoglycans, permissive spinal cord provides amuch more hospitable environment for axon regeneration due, in part, to both ahigher NPA and a lower NIA. These fmdings may reflect a change around E13when a threshold of facilitatory effects by HSPG are over-ridden by the neuriteinhibiting effects of CSPG. The fmdings of this study suggest that theequilibrium between HSPG and CSPG is changed during the transition frompermissive to restrictive periods for spinal cord repair.142CHAPTER 7GENERAL DISCUSSION143This dissertation has described some initial biochemical investigationsinto the loss of regenerative ability incurred by the developing chick spinal cord.Due to the complexity of this physiological phenomenon I have had to focusprimarily on inhibitory effects at the expense of facilitatory components such asgrowth associated proteins and trophic factors. A comparison of proteins in thedeveloping chick spinal cord identified two unknown proteins that substantiallyincreased their expression, and five that decreased their expression between E8and P2 (Ethell and Steeves, 1993). Partial amino acid sequencing is currentlyunderway for one of the proteins (DSP 7) that substantially increases itsexpression over a developmental period encompassing the transition frompermissive to restrictive periods for spinal cord repair (Ethel! et al, 1993b).Currently, technology is still being perfected that will allow researchers to takefull advantage of comparative high resolution 2D gel electrophoresis.Advancements in image analysis and the construction of 2D gel proteindatabases could soon provide further insights into the roles of the DSPs in spinalcord development and if they are involved in repair processes.I also have shown that components of spinal cord plasma membranesincrease their overall inhibitory effect on neurite outgrowth, as developmentproceeds. Plasma membrane extracts isolated prior to E13, in chick, are goodsubstrates for neurite outgrowth; whereas plasma membrane extracts isolatedafter E13 are poor substrates, showing an inhibitory influence toward neuritegrowth by NG108-15 cells (Ethel! eta!, 1993c). I have also shown a similardevelopmental decrease in neurite support to occur with plasma membranesisolated from fetal, neonatal and adult rat spinal cord. Fetal and neonatal ratspinal plasma membranes provided good substrates for NG108-15 cell neuriteoutgrowth, but adult did not. From these results, I speculate that an analogous144transition from permissive to restrictive periods for spinal cord repair in chick,may be seen in the developing rat spinal cord (Ethell et al, 1993d). The lateronset of a restrictive period for spinal cord repair in rats would be in keepingwith the slower CNS development, compared to chick. As chicks hatch, they areable to stand and feed themselves almost immediately. The precocial nature ofnew hatchlings is not seen with most mammals that completely develop in 21days. A slower developing nervous system, as in rat, would take longer todevelop pathways and "fine-tune" the CNS, resulting in a consequent delay inthe increase of inhibitory effects.I further examined the developmental increase of neurite inhibition withinthe spinal cord in terms of 3 major components of plasma membranes, lipids,proteins and carbohydrates. Purified lipids were thoroughly examined as apotential source of neurite inhibition in both the restrictive and permissiveperiods in chick spinal cord, and also in adult cow spinal cord. Using a wellestablished protocol for the purification of lipids from these various tissues, Iassayed the products for neurite inhibitory effects with NG108-15 cells, PC12cells, and superior cervical ganglion cell primary cultures. As well, lipids fromcow spinal cord were fractionated into 3 lipid subtypes using thin layerchromatography, and subsequently assayed for neurite inhibiting effects. Noinhibitory effects were observed in any of the lipid fractions, with any of thecells used for assay (Ethell et al, 1993a).Two different approaches were used in attempts to purify neuriteinhibitory proteins from mature spinal cord plasma membranes. Althoughinhibitory effects of the plasma membrane extracts were not lost duringsolubilization in detergent-containing buffers, no inhibitory effects were shownby any proteins following separation by either SDS-PAGE or ion exchangechromatography, when assayed in vitro. Many factors can complicate the145isolation of biologically active membrane proteins, not the least of which is thedenaturation of membrane proteins with detergents and then hoping they re-foldcorrectly upon removal of detergent. Although further experiments wereconsidered and discussed, time has not allowed more of these approaches to beattempted, although there are a number of groups using very similar approachestoward the same goal (Schwab, Keynes, Raper and Braun, personalcommunication).In addition to plasma membranes, changes in the ratios of somecarbohydrates also contribute to a more hostile environment for neurite growthas the spinal cord develops. In collaboration with others (Dow et al, 1993), Ihave demonstrated a developmental decrease in HSPG-associated neuritepromoting activity over a period encompassing the E13 transition frompermissive to restrictive periods for spinal cord repair. Concurrently, there is adevelopmental increase of CSPG-associated neurite inhibiting activity. Further,these changes are reflected in the ratio of HS and CS synthesis, with the largestsingle decrease in HS/CS synthesis occurring just prior to E13.Why the mature CNS would actively discourage the regeneration ofdamaged fibers seems perplexing at first, until further insights reveal thesubstantial role that neurite inhibition may play in the development andmaintenance of a complex CNS.During development, the CNS may use inhibitory factors to guide orrestrict fiber formation, as seen with the posterior somite-derived inhibitoryfactors (Davies et a!, 1990) and chondroitin sulfate in the developing retina(Brittis et a!, 1992). Further, inhibitory factors may be used as selective guidesfor the growth cones of different neurons as has been observed with the posteriortectum-derived neurite repulsor having a greater effect on growth cones frommore temporal retinal ganglion cells. Context may be extremely important for146the correct interpretation of an inhibitory signal; that is, how is the moleculepresented to the growth cone and in the context of what other molecules?.Several of the neurite inhibiting factors already identified affect thedevelopment of archaic systems such as retinal formation (Brittis et al, 1992)and somite segmentation (Davies et al, 1990). So although lower vertebrateshave a much greater capacity for CNS repair, their ontogeny must utilize at leastsome neurite inhibiting components. If inhibitory components are indeedpresent in the central nervous system of both lower, are they responsible for thedevelopmental loss of regenerative ability observed in higher vertebrates?Lower vertebrates have a substantial ability to regenerate following CNSinjury. Goldfish can repair spinal cord injuries and optic nerve sections(Steurmer, 1988), as can axolotl (Sperry, 1944). If the loss of regenerativeability in higher vertebrates is due to a greater overall level of neurite inhibitionwithin the CNS, then there must be an evolutionary advantage that overrides theadvantage of a CNS regeneration capacity. How important is it for lower andhigher vertebrates to retain a CNS repair capacity? Phylogenetically highervertebrates, by definition, have a much greater reliance upon CNS function interms of thought as is evidenced by evolution of the neopallium, the mainreasoning center of the brain. A higher vertebrate with a damaged CNS wouldnot likely evade predators, and still obtain food, long enough to repair such aninjury even it were possible; whereas a lower vertebrate such as an axolotl couldeasily burrow into the mud and lie dormant until near normal CNS function hadreturned. So for lower vertebrates there is a distinct advantage to retaining aCNS repair capacity, but not for higher vertebrates.So what evolutionary advantage could there be to a greater neuriteinhibition that could usurp the minimal advantage of a retained CNS repaircapacity? In the CNS of adult higher vertebrates versatility within the system147depends upon stable distant wiring that can be dynamically modified at a verylocal level as would be required for learning. It may be that to better retainlearned concepts and routines requires less interference from nearby systems; assuch, short sprouting would be tolerated within the CNS, but growth cones couldnot be allowed to extend too far as they may form extraneous synapses anddisrupt nearby systems. The opportunistic tendency of axonal projections tosprout when given the chance, as with neighboring denervation (Diamond,personal communication), suggests ongoing close regulation to discourage suchpromiscuity in a Tmely tuned' adult CNS, if for no other reason than to preventthe eventual short circuiting of every major system. Axotomized fibersencountering this tight down regulation would only be able to sprout for shortdistances of one or two millimeters, as is seen in the adult CNS ofphylogenetically higher vertebrates.Indeed, neurite inhibition may prove to be an extremely importantprinciple for the development and maintenance of the CNS. Much moreresearch into the roles and mechanisms of neurite within the CNS of vertebratesfrom fish up to man will be required if we can hope to understand and eventuallymanipulate these factors to aid in the recovery of paralysis in humans.148REFERENCESAebersold, R.H., Leavitt, J., Saavedra, A., Hood, L. and Kent, S.B.H. (1987).Internal amino acid sequence analysis of proteins separated by one- or two-dimensional gel electrophoresis after in situ protease digestion onnitrocellulose, Proc Natl Acad Sci USA, 84:6970-6974.Aebersold, R.H. Internal Amino acid sequence analysis of proteins after "in situ"protease digestion on nitrocellulose, in A Practical Guide to Protein andPeptide Purification for Microsequencing, P.T. Matsudaira ed. AcademicPress, Toronto. pp. 71-88.Alberts, B., Bray, D., Lewis, J., Raff, M., Roberts, K. and Watson, J. (1983).Molecular Biology of the Cell. Garland Publishing, Inc. New York, N.Y.p702-707.Alvedano, M.I., Diaz, R.S., Regueiro, P. and Monreal, J. (1991). J Neurochem,57:250-257.Armstrong, D.M. (1986). Supraspinal contributions to the initiation and controlof locomotion in the cat, Prog Neurobiol, 26:273-361.Bandtlow, C.E.' Meyer, M., Lindholm, D., Spranger, M., Heumann, R. andThoenen, H. 0990). Regional and cellular co-distribution of interleukin 113and nerve growth factor mRNA in the adult rat brain: possible relationship tothe regulation of nerve growth factor synthesis, J Cell Biol, 111:1701-1711.Bandtlow, C.E., Schmidt, M.F., Hassinger, T.D., Schwab, M.E. and Kater, S.B.(1993). Role of intracellular calcium in growth cone collapse evoked by theneurite growth inhibitor NI-35, Science 259:80-83.Barde, Y.-A., Edgar, D. and Thoenen, H. (1982). Purification of a newneurotrophic factor from mammalian brain, EMBO J, 1:549-553.Barnard, J.W. and Carpenter, W. (1950). Lack of regeneration in the spinal cordof rat, J Neurophysiol, 3:223-228.Bensadoun, A. and Weinstein, D. (1969). Assay of proteins in the presence ofinterfering materials, Anal Biochem, 10:241-250.Bensted, J.P.M., Dobbing, J., Morgan, R.S., Reid, R.T.W., & Payling Wright, G.(1957). Neurological development and myelination in the spinal cord of thechick embryo, J Embryol Exp Molphol, 5:428-437.Berkemeier, L.R., Winslow, J.W., Kaplan, D.R., Nikolics, K., Goeddel, D.V.and Rosenthal, A. (1991). Neurotrophin-5: A novel neurotrophic factorthatactivates Irk and trial, Neuron, 7:857-866.Benifield, M. (1989). Extracellular matrix, Current Opinion in Cell Biology,1:953-955149Bonhoeffer, F. and Huf, J. (1982). In vitro experiments on axon guidancedemonstrating an anterior-posterior gradient on the tectum, EMBO J, 1:427-31.Bonhoeffer, F. and Huf, J. (1985). Positional dependent properties of retinalaxons and their growth cones, Nature, 315:409-410Bottenstein, J.E. (1985). Growth and differentiation of neural cells in defmedmedia, In Cell Culture in the Neurosciences, J.E. Bottenstein and G.H. Satoeds. Plenum Press, New York, pp. 3-43.Bjorklund, A., Katzman, R., Stenevi, U. and West, K.A. (1971). Developmentand growth of axonal sprouts from noradrenaline and 5-hydroxytryptainineneurones in the rat spinal cord, Brain Res, 31:21-33.Bradford, M.M. (1976). A rapid and sensisitive method of the quantitation ofmicrogram quantities of protein utilizing the principle of protein-dye binding,Anal Biochem, 72:248-253.Brady, S.T. and Black, M.M. (1986). Axonal transport of microtubule proteins:Cytotopic variations of tubulin and MAPs in neurons, Ann IVY AcadSci,86:199-217.Brittis, P.A., Canning, D.R. and Silver, J. (1992). Chondroitin sulfate as aregulator of neuronal patterning in the retina, Science, 255:733-736.Brunner, J., Hauser, H. and Semenza, G. (1978). Single bilayer lipid-proteinvesicles formed from phosphatidylchohne and small intestinal sucrase-isomaltose, J Biol Chem, 253 :7538-46.Cajal, R.S. (1928). Degeneration and Regeneration of the Nervous System,vol. 1 & 2, Oxford Univ. Press, LondonCampenot, R.B. and Draker, D.D. (1989), Growth of sympathetic nerve fibers inculture does not require extracellular calcium, Neuron, 3, 733-743.Carbonetto, S. (1984). The extracellular matrix of the nervous system, TrendsNeurosci, 7(10):382-387.Caroni, P. & Schwab, M.E. (1988). Two membrane protein fractions from ratcentral myelin with inhibitory properties for neunte growth and fibroblastspreading, J Cell Biol, 1281-1288.Chao, M.V. (1992). Neurotrophin Receptors: A window of NeuronalDifferentiation, Neuron, 9:583-593.Charbormeau, H. (1989). Strategies for obtaining partial amino acid sequencedata from small quantities (<5 nmol) of pure or partially purified protein. in APractical Guide to Protein and Peptide Purification for Microsequencing,P.T. Matsudaira ed. Academic Press, Toronto. pp.15-30Cohen, S. (1960). Purification of a growth promoting protein from mouse alivarygland and its neurocytotoxic antiserum, Proc Nair Acad Sci USA, 46:302-311.150Cole, C.G. and Alceson, R. (1989). Identification of a heparan binding domainof the neural cell adhesion molecular N-CAM using synthetic peptides.Neuron, 2:1157-1165.Cole, G.J. and McCabe, C.F. (1991). Identification of a developmentallyregulated keratan sulfate proteoglycan that inhibits cell adhesion and neuriteoutgrowth, Neuron, 7:1007-18Cord°ln-Cardo, C., Tapley, P., Jing, S., Nanduri, V., O'Rourke, E., Lamballe, F.,Kovary, K., Klein, R., Jones, K.R., Reichardt, L.F. and Barbacid, M. (1991).The irk tyrosine protein lcinase mediates the mitogenic properties of nervegrowth factor and neurotrophin-3, Cell, 66:173-183.Danishefsky, S.J., McClure, K.F., Randolph, J.T. and Ruggeri, R.B. (1993). Astrategy for the solid-phase synthesis of oligosaccharides, Nature, 260:1307-1309.Davenport, R.W., Dou, P., Rehder, V. and Kater, S.B. (1993). A sensory rolefor neuronal growth cone filopodia, Nature, 361(6414):721-4.David, S. & Aguayo, A.J. (1981). Axonal elongation into peripheral nervoussystem "bridges" after central nervous system injury in adult rats, Science,214:931-933.Davies, J.A., Cook, G.M.W., Stern, C.D. and Keynes, R.J. (1990). Isolationfrom chick somites of a glycoprotein fraction that causes collapse of dorsalroot ganglion growth cones, Neuron, 2:11-20.Davis, S., Aldrich, T.H., Valenzuela, D.M., Wong, V., Furth, M.E., Squinto,S.P. and Yancopoulos, G.D. (1991) The receptor for ciliary neurotrophicfactor, Science, 253:59-63.Dione, C.A. Crumley, G., Bellot, F., Kaplow, J.M., Searfoss, G., Ruta, M.,Burgess,,^Jai, M. and Schlessmger, J. (1990) Cloning and expressionof two distincthi -affinity receptors cross-reacting with acidic and basicgrowth factors, E BO J, 9:2685-2692.Dow, K.E., Ethell, D.W., Steeves, J.D. and Riopelle, R.J. (1993a). Biosynthesisand neurite growth modulatory functions of proteoglycans during permissiveand restrictive periods for spinal cord repair m the chick embryo, submitted toExp Neurol.Dow, K.E., Guo, M., Kisilevsky, R. and Riopelle, R.J. (1993b). Regenerativeneurite growth modulation associated with astrocyte proteoglycans, BrainRes. Bull, 30:461-467.Dow, K.E. Mirski, S.E. L., Roder, J.C. and Riopelle, R.J. (1988). Neuronalproteogiycans: Biosynthesis and functional interaction with neurons in vitro. JNeurosci, 8:3278-3289.Dundore, M. and Bunge, R.P. (1988). Heparan sulfate proteoglycan and laminininununoreactivity on cultured astrocytes: Relationship to differentiation andneurite growth, J Neurosci, 8(8):2844-2858.151Edelman, G.M. (1991). Topobiology. Academic Press. Toronto, Ont.Edgar, D. (1989). Neuronal laininin receptors, Trends Neurosci, 12(7):248-251.Eng.vall, E., Davis, G.E., Dickersonm, K., Ruoslahti, E., Varon, S. andManthorpe, M. (1986). Mapping of domains in human laminin usingmonoclonal antibodies: localization of neurite-promoting site. J Cell Biol,103:1457-1465.Ethel!, D.W., Cheng, K.W., Campenot, R.B., Jordan, L.M. and Steeves, J.D.(1993a). Lipids purified from the developing and mature spinal cord do notinhibit neunte outgrowth, in vitro. Submitted to J Cell Biol.Ethell, D.W. Rott, M.R., Aebersold, R.H. and Steeves, J.D. (1993b).Identification and sequence of a protein that is developmentally up-regulatedin the spinal cord at the transition from permissive to restrictive periods forspinal cord repair. In preparation for submission to Mol Brain Res.Ethel!, D.W. and Steeves, J.D. (1993). Changes in protein expression associatedwith the developmental transition from permissive to restrictive states ofspinal cord repair in embryonic chick. Dev Brain Res, in press.Ethel!, D.W., Steeves, J.D., Jordan, L.M. and Cheng, K.W. Developmentaltransition by spinal cord plasma membranes of embryonic chick frompermissive to restrictive substrates for the morphological differentiation ofneuroblastoma x glioma hybrid NG108-15 cell, Dev Brain Res, 78 (1993c) 1-8.Ethell, D.W., Steeves, J.D., Jordan, L.M. and Cheng, K.W. (1993d). Developingrat spinal cord plasma membranes show an increasing level of neuriteinhibition, in vitro. Submitted to Dev Neurosci.Fawcett, J.W. (1993). Growth-cone collapse: too much of a good thing? TrendsNeurosci, 16(5):165-167Feuerstein, G. and Hallenbeck, J.M. (1987). Prostaglandins, leukotrienes andplatelet activating factor in shock, Ann Rev Pharmacol and Toxicol, 27:301-331Foran, D.R. and Peterson, A.C. (1992). Myelin aquisition in the central nervoussystem of the mouse revealed by an MBP-Lac Z transgene, J Neurosci,12(12):4890-4897.Garrels, J.I. (1983). Quantitative two-dimensional gel electrophoresis ofproteins, Methods Enzymol, 100:411-423.Ghahary, A. and Cheng, K.W. (1989). Characterization of muscarinicacetylcholine receptors on intact neuroblastoma X glioma NG108-15 cellupon induced differentiation, Mol Cell Biochem, 86:181-188.Goldberg, D.J. and Burmeister, D.W. (1989). Looking into growth cones, TrendsNeurosci, 12(12):503-506.152Goodman, R. (1982). Growth and differentiation of pheochromocytoma cells inchemically defmed medium, In Growth of Cells in Hormonally DefinedMedium, G.H. Sato, A.B. Pardee, and D.A. Sirbasku eds. Cold Spring HarborPress, Cold Spring Harbor, NY. Book B, pp. 1053-1068.Grillner, S. and Dubuc, R. (1988). Control of locomotion in vertebrates: Spinaland supraspinal mechanisms. In: Functional Recovery in NeurologicalDisease, Adv. Neurol., vol 47, Raven Press, New York, pp.425-453.Gundersen, R.W. (1987). Response of sensory neurites and growth cones topatterned substrate of lammin and fibronectin in vitro, Dev Biol, 121:423-431.Guo, M. Dow, K.E., Kisilevsky, R. and Riopelle, R.J. (1993). Novel neuritegrowth inhibitory properties of an astrocyte proteoglycan. J Chem Neuroanat,m press.Guthrie, P.B., Lee, R.E. and Kater, S.B. (1989). A comparison of neuronalgrowth cone and cell body membrane: Electrophysiological and ultrastructuralproperties, J Neurosci, 9(10):3596-3605.Hallbrook, F., Ibanez, C.F. and Persson, H. (1991). Evolutionary studies of thenerve growth factor family reveal a novel member abundantly expressed inXenopus ovary, Neuron, 6:845-858.Hamburger, V. and Hamilton, H.L. (1951). A series of normal stages in thedevelopment of the chick embryo, J Morph, 88:49-92.Han, H-Q., Nichols, R.A., Rubin, M.R., Bahler, M. & Greengard, P. (1991).Induction of formation of presynaptic terminals in neuroblastoma cells bysynapsin IIb, Nature, 349:697-700.Hantaz-Ambroise, D., Vim'', M. and Koenig, J. (1987). Heparan sulfateproteoglycan and lammm mediate two different types of neurite outgrowth. JNeurosci, 7:2293-2304.Harrington, M.G., Gudeman, D., Zewert, T., Yun, M. and Hood, L. (1991).Analytical and micropreparative two-dimensional electrophoresis of proteins,Methods:A Companion to Methods in Enzymology, 3(2):91-97.Hasan, S.J., Nelson, B.H., Valenzuela, J.I., Keirstead, H.S., Shull, S.E., Ethell,D.W. and Steeves, J.D. Functional repair of transected spinal cord inembryonic chick, Restor Neurol Neurosci, 2 (1991) 137-154Hasan, S.J., Keirstead, H.S. and Steeves, J.D. (1993). Regeneration ofbrainstem-spinal axons after spinal cord transection in the embryonic chicken,J Neurosci, 13(2):492-507.Henn, F.A. (1980). Separation of neuronal and glial cells and subcellularconstituents, in Advances in Cellular Neurobiology, vol. 1. edited by S.Federov and L. Hertz. Academic Press, Toronto, Ont. pp.373-403.153Hoffman, P.N. and Lasek, R.J. (1980). Axonal transprot of the cytoskeleton inregenerating motor neurons: Constancy and change, Brain Res, 202:317-333.Holm, A., Leibrock, J., Bailey and Barde, Y.-A. (1990). Identification andcharacterization of a novel member of the nerve growth factor/brain derivedneurotrophic factor family, Nature, 344:339-341.Holder, N. and Clarke, J.D.W. (1988). Is there a correlation between continuousneurogenesis and directed axon regeneration in the vertebrate nervoussystem? Trends Neurosci, 11(3):94-99.Honderinarck, H., Duedon, E. and Boilly, B. (1992). Embryonic brain-derivedheparan sulfate inhibits cellular membrane binding and biological activity offibroblast growth factor, Dev Brain Res, 68:247-253.Hyman, C., Hofer, M., Barde, Y.-A., Jushasz, Yancopoulos, G.D.' Squinto, S.P.and Lindsay, R.M. (1991). BDNF is a neurotrophic factor for dopaminergicneurons of the substantia nigra, Nature, 350:230-232.Igarashi, M., Strittmatter., S.M., Vartanian, T. and Fishman, M.C. (1993).Mediation by G proteins of signals that cause collapse of growth cones,Science, 259:77-79.Johnson-Greene, P., Dow, K.E. and Riopelle, R.J. (1991). Characterization off2y1c.osaminoglycan production by primary astrocytes m vitro. Glia, 4:314-Jordan, L.M. (1986). Initiation of locomotion from the mammalian brainstem. In:Neural Control of Vertebrate Locomotion, eds. S. Grilhier, R.M. Herman,P.S.G. Stein, H.Forssberg and D.G. Stuart, MacMillan Press, London, pp. 21-37.Kaplan, D.R., Hempstead, B.L., Martin-Zanca, D., Chao, M.V. and Parada, L.F.(1991). The trk Proto-oncogene product: A signal transducing receptor fornerve growth factor, Science, 252:554-557.Kates, M. (1986). Techniques in lipidology:Isolation analysis andidentification. eds R.H. Burdon and P.H. van Knippenberg. Elsevier, London.pp. 103-104;252-254.Keirstead, H.S., Hasan, S.J., Muir, G.D. and Steeves, J.D. (1993). Suppressionof the onset of myelination extends the permissive period for the functionalrepair of embryonic spinal cord, Proc Nail Acad Sci USA, 89:11664-11668.Kemler, R., Ozawa, M. and Ringwald, M. (1989). Calcium-dependent celladhesion molecules, Current Opinion in Cell Biology, 1:892-897.Kennedy, I.E., Wager-Smith, K., Barzilai, A., Kandel, E.R. and Sweatt, J.D.(1988). Sequencmg proteins from acrylamide gels, Nature, 336:449-500.Klagsbrun and Baird, (1991). A dual receptor system is required for basicfibroblast growth factor activity, Cell, 67(2):229-231.154Klein, R., Conway, D., Parada, L.F. and Barbacid, M. (1990) The trkB tyrosineprotein Icinase gene codes for a second neurogenic receptor that lacks thecatalytic kinase domain, Cell, 61:647-656.Klein, R., Jing, S., Nanduri, V., O'Rourke, E. and Barbacid, M. (1991a) The irkproto-oncogene encodes a receptor for nerve growth factor, Cell, 65:189-197.Klein, R., Nanduri, V., Jing, S., Lamballe, F., Tapley, P., Bryant, S., Cordon-Cardo, C., Jones, K.R., Reichardt, L.F. and Barbacid,M. (1991b) The trkBtyrosine protein Icinase is a receptor for brain-derived neurotrophic factor andneurotrophin-3, Cell, 66:395-403.Kleinman, H.K, Klebe, R.J. and Martin, G.R. (1981). Role of collagenousmatrices in the adhesion and growth of cells, J Cell Biol, 88:473-485.Kleinman, H.K., Ogle, R.C., Cannon, F.B., Little, Sweeney, T.M. andLuckenbill-Edds, L. (1988). Laminin receptors for neunte formation, ProcNall Acad Sci USA, 85:1282-1286.Kouzi-Koliakos, K., Koliakos, G.G., Tsilibary, E.C., Furcht, L.T. and Charonis,A.S. (1989). Mapping of three major heparin-binding sites on laminin andidentification of a novel heparin-binding site on the 151 chain. J Biol Chem,264:17971-17978.Laemmli, U.K. (1970). Cleavage of structural proteins during the assembly ofthe head of Bactenophage T4, Nature, 227:680-685.Lamballe, F., Klein, R. and Barbacid, M. (1991) trkC, a new member of the Irkfamily of tyrosine protein kinases is a receptor for neurotrophin-3, Cell,66:967-979.Landmesser, L., Dahm, J.T. and Rutishauser, U. (1990).Polysialic acid as aregulator of intramuscular branching during embryonic development, Neuron,4:655-667.Lasek, R.J. and Hoffman, P.N. (1976). The neuronal cytoskeleton, axonaltransport and axonal growth. In Cell Motility, vol C:Microtubules andRelated Proteins, R. Goldman, T. Pollard and J. Rosenbaum eds. pp.1021-1051, Cold Spring Harbor Press, Cold Spring Harbor, N.Y.Latham, K.E., Garrels, J.I., Chang, C. and Solter, D. (1991). Quantitativeanalysis of protein synthesis in mouse embryos. I. Extensive reprogrammingat one- and two-cell stages, Development, 112:921-932.Lawrence, D.G. and Kuypers, H.G.J.M. (1968a). The functional organization ofbilateral pyramidal lesions, Brain, 91:1-14Lawrence, D.G. and Kuypers, H.G.J.M._(1968b). The functional organization ofthe motor cortex in monkey II. The effects of lesions of the descendingbrainstem pathways, Brain, 91:15-36.LeGendre, N. and Matsudaira, P.H. (1989). Purification of proteins and paptidesby SDS-PAGE, in A Practical Guide to Protein and Peptide Purification155for Microsequencing, P.T. Matsudaira ed. Academic Press, Toronto. pp.49-69.Levi-Montalcini, R. and Hamburger, V. (1951). Selective growth-stimulatingeffects of mouse sarcoma on the sensory and sympathetic nervous system ofthe chick embryo, J Exp Zoo!, 116:321-362.Levi-Montalcini, R. and Angeletti, P.U. (1968). Nerve Growth Factor, Physiol.Rev, 8:534-569.Lin, L-F.H., Mismer, D., Lile, J.D., Armes, L.G., Butler III, E.T., Vannice, J.L.and Collins, F. (1989) Purification, cloning and expression of ciliaryneurotrophic factor (CNTF), Science, 246:1023-1025.Lowry, 0.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951). Proteinmeasurement with the Folin phenol reagent, J Biol Chem, 193:265-275.Macklin, W.B. & Weill, C.L. (1985). Appearance of myelin proteins duringdevelopment in the chick central nervous system, Dev Neurosci, 7:170-178.Mahboub, S., Richard, C., Delacourte, A. and Han, K-K. (1986). Applications ofchemical cleavage procedures to the peptide mapping of neurofilament tripletprotein bands in Sodium Dodecyl Sulfate Polyacrylamide GelElectrophoresis, Anal Biochem, 154:171-176.Mangold, H.K. (1969). Thin layer chromatography of lipids, in Thin LayerChromatography, Egon Stahl ed. Springer-Verlag, New York, pp.374-379.Middlemas, D.S., Lindberg, R.A. and Hunter, T. (1991) trkB, a neural receptorprotein-tyrosine kinase: Evidence for a full-length and two truncatedreceptors, Molecular and Cellular Biology, 11:143-153.Millaruelo, A.I., Neito-Sampedro, M. and Cotman, C.W. (1988). Cooperationbetween nerve growth factor and laminin or fibronectin in promoting sensoryneuron survival and neurite outgrowth. Dev. Brain Res. 38:219-228.Monreal, J., Cannona, P., Regueiro, P. and Diaz, R.S. (1992). Proteinconformational changes and myelin solubilization by anion detergentsolutions, FEBS 309(3):376-380.Morrissey, J.H. (1981). Silver stain for proteins in polyacrylamide gels: Amodified procedure with enhanced uniform sensitivity, Anal Biochem,117:307-310.Nelson, P. Christian, C., & Nirenberg, M. (1976). Synapse formation betweenclonal neuroblastoma X glioma hybrid cells and striated muscle cells, ProcNall Acad Sci USA, 73:123-127.Nelson, B.H. and Steeves, J.D. (1987). Repair of brainstem-spinal pathwaysafter thoracic spinal cord transection of the chick embryo, in ovo, SocNeurosci Abstr,Vol. 13, Part 1,p. .156Nikodem, V. and Fresco, J.R. (1979). Protein fingerprinting by SDS-gelelectrophoresis after partial fragmentation with CNBr, Anal Biochem, 97:382-389.Nirenberg, M., Wilson, S., Higashida, H., Rotter, A., Krueger, K., Busis, N.,Ray, R., Kenimer, J.G. and Adler, M. (1983). Modulation of synapseformation by cyclic adenosine monophosphate, Science, 222:794-799.Oblinger, M.M. and Lasek, R.J. (1988). Axotomy-induced alterations in thesynthesis and transport of neurofilaments and microtubules in dorsal rootganglion cells, J Neurosci, 8(5):1747-1758.O'Farrell, P.H. (1975). High-resolution two-dimensional electrophoresis ofproteins, J Biol Chem, 250:4007-4021.Okado and Oppenheim, R.W. (1985). The onset and development of descendingpathways to the spinal cord m the chick embryo, J Comp Neurol, 232(2):143-161Popov, S., Brown, A. and Poo, M-m. (1993). Forward plasma membrane flow ingrowing nerve processes, Science, 259:244-246.Ragsdale, C. and Woodgett, J. (1991). trking neurotrophic receptors, Nature,350:660-661.Ranvier (1871). De la regenerescence des nerfs apres leur section, Compt rendt.53.Ranvier (1873). De la regeneration des nerfs sectionnes, Compt rend, t.76.Rapraeger, A.C., Kruflca, A. and Olwin, B.B. (1991). Requirement of heparansulfate for bFGF-mediated fibroblast growth and differentiation, Science,252:1705-7.Reier, P.J. & Houle, J.D. (1988). Advances in Neurology, Vol. 47:FunctionalRecovery in Neurological Disease, edited by S.G. Waxman. Raven Press,New York, N.Y. pp.87-138.Riopelle, R.J. and Dow, K.E. (1990). Functional interactions of neuronalheparan sulfate proteoglycans with laminin. Brain Res, 525:92-100.Risau, W. & Wolburg, H. (1990). Development of the blood-brain barrier,Trends Neurosci, 13(5):174-178.Rogers, S., McCarthy, J.B., Furcht, L.T. and Letoumeau, P.C. (1985). Neuron-specific interactions with two neurite-promoting fragments of fibronectin, JNeurosci, 5:369-378.Romanoff, A.L. (1955). Biochemistry of the Avian Embryo: A QuantitativeAnalysis of Prenatal Development. Academic Press, Toronto.Rutishauser, U., Acheson, A., Hall, A.K., Mann, D.M. & Sunshine, J. (1988).The neural cell adhesion molecule (N CAM) as a regulator of cell-cellinteractions, Science, 240:53-57.157Rutishauser, U. (1988). Neural cell-to-cell adhesion and recognition, CurrentOpinion in Cell Biology, 1:898-904.Sabra, F. and Adams, R.D. (1977). Principles of Internal Medicine: 8th ed.,edited by G.W. Thorn. McGraw-Hill, Toronto, Ont. pp.1824-1831.Saier, M.H. (1983). Biological membranes: Structure and assembly, inBiochemistry, Geoffrey Zubay ed. Addison-Wesley Publishing Company,New York, pp. 573-616.Sambrook, J., Fritsch, E.F. and Maniatis, T. (1989). Transfer of proteins fromSDS-Polyaciylamide gels to solid supports: Immunological detection ofimmobilized porteins (Western Blotting), in Molecular Cloning: ALaboratory Manual, 2nd Ed. pp. 18.60-18.74.Schnell, L. & Schwab, M.E. (1990). Axonal regeneration in the rat spinal cordproduced by an antibody against myelin-associatedneurite growth inhibitors,Nature, 343:269-272.Schwab, M.E. and Thoenen, H. (1985). Dissociated neurons regenerate intosciatic but not optic nerve explants in culture irrespective of neurotrophicfactors, J Neurosci, 5(9):2415-2423.Schwab, M.E. & Caroni, P. (1988). Oligodendrocytes and CNS myelin arenonpemfissive substrates for neurite growth and fibroblast spreading in vitro,J Neurosci, 8(7):2381-2393.Schwab, M.E. (1990). Myelin-associated inhibitors of neurite growth andregeneration in the CNS, Trends Neurosci, 13(11):452-456.Schwab, M.E., Kapthanuner, J.P. and Bandtlow, C.E. (1993). Inhibitors ofneurite growth, Ann Rev Neurosci, 16:565-595.Sendtner, M., Kreutzberg, G.W. and Thoenen, H. (1990) Ciliary neurotrophicfactor prevents the degeneration of motor neurons after axotomy, Nature,345:440.Sherwood, L. (1989). Human Physiology from Cells to Systems, WestPublishing Company, New York. pp. 25.Shimizu, I., Oppenheim, R.W., O'Brien, M. & Schneiderman, A. (1990).Anatomical-and functional receovery following spinal cord transection in thechick embryo, J Neurobiol, 21:918-937.Sholomenko, G.N. & Steeves, J.D. (1987). Effects of selective spinal cordlesions on hindlimb locomotion in birds, Exp Neurol, 95:403-418.Skubitz, A.P.N., McCarthy, J.B., Charonis, A.S. and Furcht, L.T. (1988).Localization of three distinct heparin-binding domains of lamimn bymonoclonal antibodies. J Biol Chem, 263:4861-4868.158Snow, D.M., Lemmon, V., Canino, D.A., Caplan, A. and Silver, J. (1990).Sulfated proteoglycans in astroglial barriers inhibit neurite outgrowth, in vitro,Exp Neurol, 109:111-130.Sperry, R.W. (1944). Optic nerve regeneration with return of vision in anurans, JNeurophysiol, 7:57-69.Springer, J.E. (1988). Nerve growth factor receptors in central nervous system, JExp Neurol, 102:354-365.Stahl, E. (1969). Apparatus and general techniques in TLC, in Thin LayerChromatography, Egon Stahl ed. Springer-Verlag, New York, pp. 52-85..Steeves, J.D., Sholomenko, G.N. and Webster, D.M.S. (1987). Stimulation of thepontomedullary reticular formation initiates locomotion in decerebrate birds,Brain Res 401:205-212.Stuermer, C.A.O. (1988). Trajectories of regenerating retinal axons in thegoldfish: I. A comparison of normal and regenerated axons at lateregeneration stages. J Comp Neurol, 267:55-68.Stirling, V.R. (1991). Molecules, maps and gradients in the retinotectalprojection, Trends Neurosci, 14:509-12.Stoscheck, C.M. Quantitation of proteins. In M.P. Deutscher (Ed) Guide toProtein Purification, Methods in Enzymology Vol 182. Academic Press,Toronto, Ont. pp 50-54.Strittmatter, S.M. and Fishman, M.C. (1991). The neuronal growth cone as aspecialized transduction system, BioEssays, 13(3):127-134.Tello, F. (1911). "La influencia del neurotropismo en la regeneracion de loscentros nerviosos." Trabajos del Lab. de Invest. biol., Book 9:124-159.Tetzlaff, W., Alexander, S., Miller, F.D. & Bisby, M.A. (1991). Response offacial and rubrospinal neurons to axotomy: Changes in nana expression forcytoskeletal proteins and GAP43, J Neurosci, 11:2528-2544.Thoenen, H. (1991). The changing scene of neurotrophic factors, TrendsNeurosci, 14(5): 165-170.Thoenen, H. and Barde, Y.-A. (1980). Physiology of nerve growth factor,Physiol Rev, 60:1284-1335.Tosney, K.W. (1992). Growth cone navigation in the proximal environment ofthe chick embryo. In The Nerve Growth Cone, P.C. Letoumeau, S.B. Katerand E.R. Macagno eds. Raven Press, NY, pp. 387-403.Valenzuela, J.I., Ethel!, D.W. and Steeves, J.D. (1988). The capacity of damagedembryonic spinal cord to repair and make functional locomotor connectionshas been studied, in ovo, Soc Neurosci Abstr, Vol. 18, Part 2, p. 964.159Vance, D. (1983). Metabolism of glycolipids, sphingolipids, and prostaglandins,in Biochemistry, Geoffrey Zubay ed. Addison-Wesly Publishing Company,pp. 505-542.Varon, S., Manthorpe, M., Davis, G.E., Williams, L.R. and Skaper, S.D. (1988).Growth factors, Advances in Neurol, 47:493-521.Verge, V.M.K., Merlio, J.-P., Grondin, J, Emfors, P., Persson, H., Riopelle,R.J., Hokfelt, T. and Richardson, P.M. (1992) Colocalization of NGF bindingsites, trk mRNA, low-affinity NGF receptor mRNA in primary sensoryneurons: Responses to injury and infusion of NGF, J Neurosci, 12(10):4011-4022.Udin, S.B. and Fawcett, J.W. (1988). Formation of topographic maps, Ann RevNeurosci, 11:289-327.Wakai, S. and Hirokawa, N. (1978). Development of blood-cerebrospinal fluidbarrier to horseraddish peroxidase in the avian choroidal epithelium, CellTissue Res, 195:195-203.Waller (1852). Sur la reproduction des nerfs et sur la structure et lesfunctions des ganglions spinaux. Mullers Archiv.Walter, J., Henke-Fahle, S. and Bonhoeffer, F. (1987a). Avoidance of posteriortectal membranes by temproal retinal axons, Development, 101:909-913.Walter, J., Kern-Veits, B., Huf, J., Stolze, B. and Bonhoeffer, F. -(1987b).Recognition of position-specific properties of tectal cell membranes by retinalaxons in vitro, Development, 101:685-696.Webster, D.M.S. and Steeves, J.D. (1991). Funicular organization of avianbrainstem-spinal projections, J Comp Neurol, 312:46-7-76.Wolfe, R.A. and Sato, G.H. (1982). Continuous serum-free culture of N18TG-2neuroblastoma, the C6BU-1 glioma, and the NG108-15 neuroblastoma Xg,lioma hybrid cell lines. In G.H. Sato, A.B. Pardee and D.A. Sinbasku (Eds.)Growth of Cells in Hormonally Defined Media, Book B, Cold SpringHarbor Laboratory Press, Cold Spring Harbor, N.Y., pp. 1075-1088.Yayon, A., Klagsbrun, M., Esko, J.D., Leder, P. and Omitz, D.M. (1991). Cellsurface, heparan-like molecules are required for binding of fibroblast growthfactor to its high affinity receptor, Cell, 64:841-848.160


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items