CA2071898A1 - Inhibition of cell growth by keratan sulfate, chondroitin sulfate, dermatan sulfate and other glycans - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE

The present invention is directed to methods of using keratan sulfate, chondroitin sulfate, dermatan sulfate, heparin sulfate, heparin, and hyaluronic acid, and molecules and compositions comprising keratan sulfate, chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin, and hyaluronic acid, to inhibit or prevent neurite outgrowth, i.e., axonal growth, or nerve regeneration (colectively termed herein "nerve growth"), or glial cell migration or invasion, or regeneration, and therapeutically, where the foregoing is desired. In another embodiment of the invention, inhibitors and antagonists of keratan sulfate, chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin, and hyaluronic acid, and molecules and compositions containing the same, may be used to promote nerve growth or glial cell migration or invasion, and can be administered therapeutically.
Such inhibitors and antagonists include but are not limited to antibodies, degradative enzymes, lectins; and disaccharide antagonists of the receptors for keratan sulfate, chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin or hyaluronate.

Description

~/US90/0~189 S~

INHIBITlON OF CELL GROWI H BY KERATAN SULFATE. CHONDROITIN SULFATE. DERMATAN SUL-FATE AND O~HER GLYCANS

1. INTRODUCTION

The present invention relates to compositions comprising keratan sulfate, chondroitin sulfate, or dermatan sulfate, also heparan sul~ate, heparin, or hyaluronic acid (hyaluronate), or any combination of these molecules~ particular, glycosaminoglycans or proteoglycans- and the uses o~ such compositions in inhibition o neuri~e outgrowth and glial cell invasion or migration. The invention also relates to compositions comprising an antagonist of inhibition mediated by keratan sulfate, chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin, or hyaluronate such as antibodies to keratan sulfate, chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin or hyaluronate; enzymes that degrade keratall sulfate, chondroitin sul~ate, dermatan sulfate, heparan sulfate, heparin, or hyaluronic acid;
lectins specific for keratan sulfate, chondroitin sulfate, dermatan sulfate, heparan sulate, heparin, or hyaluronic acid, or disaccharide antagonists of the r~ceptors for keratan sulate, chondioitin sul~at~, dermatan sul~ate, heparan sulfate, heparin or hyaluronic acid; and the uses of such compositions for promotion of cell growth or regen~ration, in particular, n~urite outgrowth or glial cell migra ion, Therapeutic us~s of the compositions of khe invention are provid~d.

;

~ ~ .

WO91/06303 PCT/US9o/0

2. BACKGROUND OF THE INVENTION
2.1. AXONAL GROWTH
Axons grow in stereotyped patterns toward their targe~s during developmen~ of the nerYous system. During this directed elongation, axonal growth cones undergo multiple interactions with components of the environment such as the extracellular matrix (Carbonetto et al., 1982, Science 2160897-899; Hankin, ~. H. and Silver, J., 1986, Mechanisms o~ axonal guidance: the problem o~ intersecting fiber systems, in The Cellular Basi~ of Morphogenesis, (Leon W. Browder, Ed.) Vol. 2:565-599, Plenum Publishing Corp., New York, New York; Mirsky et al., 1986, J. '~
Neurocytology 15(6):799-815; Rogers et al., 1986, Devel.
Biol. 113:429-435: Antonicek et al., 1987, J. Cell Biol.
104(6):15~7~1595; Bork et al., 1987, J. Comp. Neurol.
t5 264:147-158; Liesi, P. and Silver, J., 1988, Devel. Biol.
130:774-7 5; Lumsden, ~. and Keynes, R., 1989, Nature 337:424-42~), the surface~ of neuroepithelial cell endfeet (Silver, J., and Sapiro, J., 1981, J. Comp. Neurol.
202:521-53~; Silver, J., and Rutishauser, U., 1984, Devel.
Biol. 106:485-499; Bastiani, M. J. and Goodman, C.S., 1986, J. Neurosci. 6(12):35~2 3551; Xuwada, J. Y., 1986, Science 233:740-746; Holley, J. A., 1987, Devel. Biol. 123:389-400;
Holley, J. and Silver, J., 19~7, I)evel. Biol. 123:375-388), other axons (Rutishauser et al., ~L978, J. Cell Biol.
79:382 393; Fushiki, S., and Schachner, M., 19~6, Brain Res. 389:153-167) and glia (Silver, J., and Robb, R. ~., 1979, Devel. Biol. 68:175-190; Singer, et al., 1979, J.
Comp. Neurol. 1~5.1~22; Simpson, S., 1983, in Spinal Cord Reconstruction, Raven Press, New York, New York pp. 15i-3D 162; Silver, J., 1984, J. Comp. ~eurol. 223:238-251; Poston et al., 1385, S~ciety ~or Neuroscience Abstract, 11:584;
Bastiani, M. J. and Goodman, C.S., 1986, supra; Cooper, N.
G. F., and Steindler, D. A., 1986, Brain Res. 380:341-348;
Poston et al., 1987, The Making o~ the Nervous System, ~5 WO9]~06303 ~CT/~9~/06l89 ~3~
London: Longmans). It has been suqgested that the adhesive properties of these various surfaces lurQ axons along their approprlate pathways (Letournau, P. C., 1975, Devel. Biol.
44:92-101, Sidma~, R. L., and Wessells, N. K., 1975, Exp.
Neurol. 48:237-251; Constantine-Paton, M., 1983, DevPl.
Biol. 97:239-244; Ro~ers et al., 1983, Devel. Biol.
98:212-220; Gundersen, R. W., and Park, K. H. C., 1984, Devel. Biol. 104:18-27; Silver, J., and Rutishauser, U., 1984, Dev. Biol. 106:485-499; To~aselli, K. J. et al., 1986, J. Cell Biol. 103:2659-2672). Local cues or~anized into gradients may be important in providing directionality as well (Lumsden, A. G. S., and Davies, A., 1983, Nature (London) 306:786-788). In fact, many or all of these factors could be int~racting, in concert, to promote axon growth along a proper tra~ctory at varlous stages of development.
Recent findings indicate that mechanisms exist which inhibit or repel axons and they may be as important for axon guidance as are adhesive or attractive mecha~isms (Silver, J., and Sapiro, J., 1981, J. Comp. Neurol.
202:521-538; Haydon et al., 1984, Science 226:561-564;
Poston et al., 1985, Society ~or Neurosoiences Abstract 11:584; Kapfhammer, J. P., and Raper, J. A., 1987 J.
Neurosci. 7(5):1595-1600; Kap~hammer, J. P., and Raper, J.
A., 1987, J. Neurosci. 7(1):201-212; Perris, R., an~
2E; Johansson, S., 1987, J. Cell Biol. 105(6)2511-2521; Silver et al., 1987, J. N~urosci. 7(7~ :2264-2272; McCobb et al., 1988, J. Neurosci. Research 19:19-26; P~tterson, P. ~., 1988, Neuron 1:263-267; 5chwab, M. E., and Caroni, P., 1988, J. Neurosci. 8(7):2381-2393; Tosney, K., 19~8, Devel.
3D Biol. 127:266-286; Webster et al., 1988, J. Comp. Neurol.
269:592-611; Gurwitz, D. and Cunningham, D. D., 1~88, Proc.
Natl. Acad. Sci. U.S.A. 85.3440-3444). Inhibitory components ~ay take the form o~ cellula~ boundaries or barriers along an axon pathway (Silver, J., 1984, J. Comp.
~5 :

-, :

WO9l/06303 ~7~3~' Neurol. 223:238-251) and they may act by mechanical as well as chemical means (Silver, ~., and Rutishauser, U., 1984, supra: Tosney, K. and Landmesser, L., 1985, Devel. Biol.
109:193-214; Silver et al., 1987, J. Neurosci. 7:2264-2272;
Stern, C. D., and Keynes, R. J., 1987, Development, ;-99:261-272; Webster, M. J. et al., 1988, J. Comp. Neurol.
269:592-611). Axon inhibition can occur between different classes of neurons (Kapfhammer, J. P., and Raper, J. A., 1987, J. Neurosci. 7(1):201-212), in association with glial cells (Silv~r et al., 1982, J. Comp. Neurol. 210:10-29;
Silver, J., 1384, J. Comp. Neurol. 223:228-251; Poston et al., 1985, Society for Neuroscience Ab~tract 11.584; Silver et al., 1987, supra; Steindler, D. A., and Cooper, N. G.
F., 1987, Devel. 8rain Res. 36:27-38; Schwab, M. E., and Caroni, P., 1988, J. Neurosci. 8(7):2381-2393), in response to mesench~mal components (Keynes, R. J., and Stern, C. D., 1984, Nature (London) 310:786-789; Tosney, K., 1988, Devel.
Biol. 127:266-286~ or to soluble factors (Haydon et al., 1984, Science 226:561-564; Verna, J. M., lg85, J. Embryol.
Exp. Morphol. 86:53-70; McCobb et al., 1988, J. Neurosci.
Res. 19:19-26). As y~t, we have only a limited understanding of axon barriers. one would like to know what they are made o~, how they function, and what happens at the cellular and molecular levels when axons encounter them.
One r~gion of the central nervous system that may be a barrier to axon growth is the roof plate, located at th~ dorsal midline of the developing vert~brate spinal c:ord (His, W., 1891, I. Verlangertes Mark, 29:1-74; Ramon y Cajal, S., lgll, Histologie du ~yste~e nerveux de l'homme et des vertebres. ~francaise rev. et mise a jour par l'auteur, Ed.) Vol. 1. A., Maloine, Paris). Thi~ region is compri ed of pri~itive glial cells as determined morphologically (~is, 1891, supra), by the use of tritiated thymidine autoradiography (Altman, ~., and Bayer, S. A.,

3~

WO9l/06303 P~T/V~90/06189 2~ ~ !7 ~

1984, in Advances in Anatomy, Embryology and Cell Biology, Vol. ~5, pp. S3-83, Springer-Verlag, H~idelberg, ~ermany) and with the use o f antibodies RC1 and RC2 which specifically label embryonic radial glial cells (Edwards et al., 198~, 8C7, Society for Neuroscience Abstract 12:182).
The rooP plate contains transient channels which are first observed as a single row of extracellular spaces at E9 (Snow, D., et al., 1987, Society for Neurosciences Abstract 13(1) :1987) . The roof plate undergoes a gradual change in morphology between ~12.5, when it ha~ a wedge shape, and E15.5 when it becomes a long, thin septum-like structure at the dorsal midline of the spinal cord in rat. A dorsal subpopulation of the early ventral commissural axons as well as primary afferents fxom the dorsal root ganglia come in close proximity to the roof plate. Even though both t5 axon systems have potential targets or pathways in the contralateral spinal cord, they do not cross the roof plate to reach themO Figure 1 is a schematic diagram which depicts the relationship of the commissural and sensory axon systems to the roof plate at E13.5 and El5.5 in rat.
However, at a later stage of embryonic development in rat (Smith, C. L., 1983, J. Comp. Neurol. 220:29-43) and in frog (Nordlander, R. and 5inger, M., 1982, Exp. Neurol.
75:221-228), a population of sensory axons do cross the dorsal spinal cord just below the posterior columns to f orm the dor~al commissllre~

2.2. PROTEOGLYC~NS
__ Proteoglycans are molecules found in abundance in connective tissue, which consist of about 50 95%
polysaccharide and about 5-50~ protein. Glycosaminoglycans are the polysaccharide chains of pro~aoglycans, and contain rep~ating units o~ disaccharides which consist of an aminosugar derivative, either glucosamine or galactosamine.

'~ :
~ . .

~ : ' WO91/06303 P~T/U~90/0618 A negatively charged carboxylate or sulfate group is found in at least one of the sugar units of the disaccharide.
Common glycosamino~lycans include hyaluronate (HA), chondroitin sulfate (CS~, keratan sulfate (KS), dermatan sulfate (DS), heparan sulfate (HS), and heparin (HN). The structural formulas for the disaccharide unit of hyaluronic acid, chondrQitin 6 sulfate, keratan sulf~te, dermatan sulfate, hnd heparin are as follows (note that forms exist with various degrees of sulfation greatar than those shown):
coo- CH20H
~ Hs~o--H OH H NHCOCH3 hyaluronic acid 11ynluron~-COO~ CHzOSO3- ~
H/ H~ - ~ \ - O - HO 'H \ -o- chondroitin 6-sulfate ~/H ~,~H

Chondroitin 6 ~ulf~to keratan sulfate CH20H CH20SO,' HO ~ H " 1 Hro\
H~H --~H

l~nr~t~n ~ulfllltl~
dermatan sulfate D~rmntlln ~ulf~t~

WO9l/063~3 PCT/~S~0/~61~9 ~~- c- -7- ~ s''~
\ H h~parin 0_~
~1 OSO3- H
H ~p~in The ylycosaminoglycan chains of proteoglycans ar~ found covalently a tached to a polypeptide backbone called the core protein (Stryer, L., 1981, Biochemistry, 2d ed., W. H.
Freeman ~ Co., New York, pp. 200-203).
Combination~ of proteoglycans and growth-promoting m~le~ules exist in many regions where pioneerin~
axons elongate. Laminin as well as neural c~ll adhesion molecule (NCAM) are present in tandem with protesglycan along the developing optic pathway (Liesi, P., and Silver, J., 1988, Dev. Biol. 130:77~-7~5; Silver and Rutishauser, 198~, Dev. Biol. 106:4~5-499; Bork et al., 1987, J. Comp.
Neurol. 264:1~7-158).
In other localization studies, Tosney and Landmesser (1985, Dev. Biol. 109:193-214) showed that the pos~erior ~clerotome contains high levels of glycosaminoglycans, as det~rminecl by Alcian Blue staining, :~
and that growth cones do not expl.or~ these regions.
Funderburg et al. ~1986, Dev. Biol. 116:267-277) have confirmed th~ presence of keratan sul~ate in the outer epidermis. Studies in chisk ~ore~rain have shown that chondroitin-6-sul~ate is pr~sent in large quantities in channels ben~ath the cortic~l plate region wher~ axons are not ound (Palmert et al., 1986, Society for Neurosciences 3~ Abstract 12(2):1334). A keratan ~ulfate proteoglycan has be~n identifi~d in the rat cerebral cortex (Vit~llo et al., 1978, 8iochim. Biophys. Acta 539:305-314) and in corpora amylac~a of human brain (Liu, H. Mo ~t al., 1987, J.
N~uroimmunol. 14:49-60).
:

,~ , . ,:
: - ~

, ~ . , ' ' ,:

WO91/06303 PCT/US90/06l89 J~

2.3. PROTEOGLYCAN REGULATION OF CELL GROWTH
Different proteoslycans have been shown to exert a wide spectrum of effects on the migratory behavior of a variety of different cell types (Walicke, P.A., 1988, Exp.
Neurol. 102:l44-148; Daman et al., 1988, J. Cell Physiol.
135:293-300). Perris and Johansson (1987, J. Cell Biol.
105:2511-2521) have shown that a form of chondroitin sulfate proteoglycan is inhibitory to the migration of neural crest cells in vi~ro.
Specific glycosaminoglycans/proteoglycans have been shown to have inhibitory effects on neurite outgrowth and cell attachment in vitro. Carbonetto et al. (1983, J.
Neurosci. 3(11):2324-2335) showed that chondroitin sulfate, hyaluronic acid, and heparin inhibit chick dorsal root ganglion and PC-12 axon outgrowth on fibronectin in a three-dimensional ~EMA-gel culture system. Support of neurite outgrowth by fibronectin was significantly reduced by the addition of heparin to a HEMA/fibronectin gel. This proteoglycan, in high concentrations, also inhibited the attachment and ne~rite formation of human neuroblastoma cells on a cholera toxin B/ganglioside GM1-binding substratum (Mugnai et al., 1988, Exp. Cell Res. 175:299-2~7).
Unfractionated cartilage proteoglycan~, though to a lesser extent a puxified ~artilage component, chondroitin sulfate, wer~ found to inhibit fibroblast binding to collagen and fibronectin in vitro (Rich, ~t al., 1981, Nature 293:224-226). Dermatan sulfate proteoqlycan (DS-PG) was observed to inhibit th~ attachment and spreading of 3T3 fibroblast on plasma fibronectin-coat d 3~ culture substrata (Lewandowska et al., 1987, J. Cell Biol.
105:1443-1454). Previously, however, glycosaminoglycans (GAGs), principally heparan ~ulfate and der~atan sulfate, were identified a~ mediators of fibroblast (murine 3T3 cell) attachment to fibronectin. The heparan and dermatan 3~

WO91~06303 .~CT/US90/06l89 g ~ $
GAGs bound to serum fibronectin covalently attached to Sepharose, while other proteoglycans, notably various chondroitin sulfates and under-sulfated heparan sulfate, did not bind to the column (Laterra, et al., 1980, Proc.
Natl. Acad. Sci. U.S.A. 77:6662-6666).
Sulfonated glycosaminoglycans have been reported to reduce the ability of bacteria to adhere to urinary bladder mucosa, as part of an anti-bacterial defense mechanism (Parsons, C. L., 1986, ~rologic Clinics of North ~merica 13(4):563-568; Parsons, C. L., et al., 1981, J.
Infec. Dis. 144(2):180; Parsons, C. L., et al., Science 208:605-607; Hanno, P. M., et al., 1978, J. Surg. Res.
25:324-329; Paxsons, C. L., et al., 1978, Am. J. Pathol.
93(2):423-432).

3. SUMMARY OF THE INVENTION
The present invention relates to the discovery that keratan sulfate (K5), chondroitin sulfata (CS), dermatan sulfate (~S), heparan sulfate (HS), heparin ~HN) and/or hyaluronic acid (HA) can inhibit n~urite outgrowth, i.e., axonal growth, and glial cell migration or invasion.
Neurite outgrowth, i.e., axonal growth, and nerve regeneration herein may be referred to as ~'nerve growth."
Pr~sence of thes~ glycans inhibits neurite outgrowth even in the pre~ence of nerve growth promoting factors such as laminin and NCAM. These glycans prevent glial cell, in particular astrocyte, migration or invasion on laminin. Accordingly, the pr~sent invention is directed to methods o~ using keratan ~ulfate~ and molecules and compositions comprising keratan sul~ate, to inhibit or 3~ prevent neurit~ outgrowth and/or glial cell migration or invasion, or nerve or ylial cell regeneration. The methods to inhibit neuri~e out~rowth or ~lial cell migration or invasion may be used therapeutically, where hat is desired. Such molecules comprising keratan sulfate include 3~

c~ ~ 7 ~

but are not limited to keratan sulfate glycosaminoglycan and kerat2n sulfate proteoglycan, with keratan sulfate proteoglycan most preferred. The invention is further directPd to molecules and compositions comprising chondroitin sulfate, and the therapeutic uses thereof to inhibit or prevent neurite outgrswth or glial cell migration or invasion. ~olecules comprising chondroitin sulfate include but are not limited to chondroitin sulfate glycosaminoglycan, or more praferably, chondroitin sulfate proteoglycan. The invention also encompasses molecules and compositions comprising dermatan sulfate and the therapeutic uses thereof to inhibit or prevent neurite outgrowth or glial cell migration or invasion. Molecules of dermatan sulfate include but are not limited to dermatan sulfate glycosaminoglycan, or more preferably dermatan sulfate proteoglycan.
In another embodiment, inhibitors and antagonists of keratan sulPate, chondroitin sulfate, dermatan sulfate, heparan sulfate, heparin and/or hyaluronic acid, and molecules and compositions containing the same, may be used to promote neurite outgrowth or glial cell migration or in~asion and can be administered therapeutically. Such inhibitors and antagonists include but are not limited to antibodies to KS, CS, DS, HS, HN or HA, ~nd derivativ~s or fragments 1:hereof, enzymes that degrade KS, CS, DS, HS, HN or HA, lectins specific for KS, CS, DS, HS, HN or HA, and disaccharide antagonists of receptors specific for KS, CS, DS, HS, H~ or HA. In this embodiment, promotion o neurite outgrowth and glial cell migration or invasion occurs by r@moving the inhibitory influence of molecules comprisihg KS, CS, ~S, HS, HN or HA, thus allowing promotion of neurite outgrowth or glial cell migration or in~asion by endogenous or exogenously added molecules.

W~9l/06303 PCT/~S90/061~9 In a further embodiment, molecules comprising keratan sulfate can be used together with msl2cules comprising another glycosaminoglycan or the disaccharide unit thereof, preferably chondroitin sulfate, in the methods of the inventlon.
The present invention also provides pharmaceutical compositions comprising effective amounts of the molPcules and compositions comprising keratan sulfate and/or chrondroitin sulfate, dermatan sulfate, heparan sulfate, heparin or hyaluronate.
~s detailed in the examples sections infra, immunocytochemical localization data is presented which indicates that keratan sul~ate, alone or in combination with other molecules ~uch as chondroitin sulfate, may be in part responsible for the inhibition of axon elongation through the roof plate in the embryonic spinal cord. In a further example, we demonstrate in vitro that keratan sulfate/chondroitin sulfate proteoglycan or dermatan sulfate proteoglycan actively inhibits neurite elongation in a concentr~tion dependent mann~r. The examples also indicate that dermatan sulfate ancl keratan sulfate/chondroitin sulfate inhibi.t invasion or migration of glial cells, including astrocytes, in a concentration dependent manner.

3.1. DEFINIl'IONS
As u~ed herein, the following terms shall have the m~anings indica~ed:
ChE: choline~tera~e C5: chondroitin sul~ate 3D DRG: dorsal root ganglion D5-PG: dermatan sulfate proteoglycan E~1.5: embryonic day 11.5 GAG: glycosaminoglycan HA: hyaluronic acid, hyalurona~e WO9~/06303 Pcr/usso/a -12~ 2 ~ 7 ~
HN: heparln HRP: horseradish peroxidase HS: heparan sulfate Ig: i~munoglobin ~S: k2ratan sulfate ~S/CS~-PG: keratan sulfate/chondroitin sulfate proteo~lycan FITC: fluorescein isothiocyanate ~N: laminin NCA~: neural cell adhesion moleGule 10 NGS: normal goat serum P3: postnatal day 3 PBS: phosphate buffered saline PG: proteoglycan RCS: rat chondrosarcoma tumor cell line cartilage chondroitin sulfate proteoglycan RITC: rhodamine isothiocyanate TBS/BSA: Tris-buffered saline/bovine serum alhumin TPA: tetragonolobus purpureas agglutinin (lotus tetragonolobus, lotus lectin)

4. DESCRIPTION OF l'HE FIGURES
Figure l. Schematic di,agram of embryonic day 13.5 (El3.5j and El5.5 ra~ cervical spinal cord. Th~
relationship of the roof plate (RP) to the developing dorsal column (sensory) axons (SA) and the commissural axons (CAl is depict~d. Th~ earliest dorsal popula~ion of ~0 commissural axons ori~inate near the roof plate. The axons elongate dorsolaterally, then travel ventrally near.the periph~ry of the cord to decussate at the floor plate (FP).
The primary sensory afferents (SA) wait in the dorsal root entry 20ne in an oval bundle on El3.5, trav~lling rostrally 3~

WO91/06303 PCT/US90/~6~89 and caudally for a few segments. The dorsal column axons move medially with development and abut the roof plate by E15.5. Like the commissural axons, the dorsal column axons respect the dorsal midline barrier.
Figure 2. (A) Traverse 1 ~m plastic section of El1.5 rat cervical spinal cord. The roof plate (RP) cells are beginning to form the wedge shape which will become most apparent on E13.5. The cells are arranged in an arching pattern in comparison to adjacent n~uroepithelial cells which are more linear. Extracellular space between 0 the presumptive roof plate glial cells is not yet significant in comparison to the spaces seen betwen the cells of the remainder of the cord (compare with Fig. 3A);
cc, central canal, (250X), (B~ A 10 ~m cryostat section of rat cervical spinal cord on day E11.5 labelled with lCl2 antibody which stains the ventral commissural axons. Only a single neurite i5 labelled in the lateral cord at this time of development. Note the absence of labelled processes near the roof plate (RP) in the dorsal spinal cord (200X), (C) Higher mag~ification of lCl2 labellPd neurite in view (B), (arrow), (400X). The roof plate lacks obvious stainin~ with anti-KS antibodies at this stage of development.
Figure 3. The roof plate of E13.5 rat cervical spinal cord. (A) The roof plate (RP) glia extend an apical proce~ to the pial surfase, terminating in an endfoot. Interspersed among thes~ glial cells is an extensiv~ network of large extrac~llular spaces. Together, the cells and spaces form a wedge-shaped region at the dorsal aspect of the spinal cord. Compare the cells of the 3~ roof plate with the surrounding region of cells which are c105~1y appossd to one another. The surrounding cells and their processe~, som2 of which are commissural neurons, arch dorsolaterally along the perimet~r and then away from the roof plate, (630X). (B) Transverse frozen section of 3~

W~91/06303 PCT~US90/06189~
2~7~

the same age and cord level as in (A), labelled with an anti~keratan sulfate (a-KS) monoclonal antibody. Keratan sulfate epitopes are specific to the roof plate (RP) at this stage of development. Note that the labelling pattern coincides directly with the wedge-shaped region of glial cells and interspersed extracellular spaces of the roof plate seen in (A); cc, central canal, (630X).
Figure 4. Transverse frozen section of El3.5 rat cervical spinal cord labelled with antibody lC12. The commissural axons (ca) take a stereotypical path away from 0 the roof plate (RP) as they course from the dorsolateral wall of the spinal cord along the periphery to the floor plate (fp) where they cross the midline and turn to travel in the ventral funiculus. Ankibody lCl2 also labels the oval bundle (ob), the dorsal root (dr) and the dorsal root ganglia (drg), (180X).
Figur~ 5. Transmission electron micrograph o~
the boundary (arrows) between the roof plate (RP) glia and neighboring neurons and neurites in El3.5 rat cervical spinal oord. No neurites cross the roof plate. ~n example f one of the large extracellular 'spaces occurs just below the "RP", (7,000X).
Figure 6. Relationship of the commissural axons to the roof plate (RP) glia in E13.5 rat cervical spinal cord. (A) Th~ commissural axons (ca), localized with monoclon~l antibody lC12, arise from cell bodies (not 5tained~ along the dorsolateral cord and travel away rom th~ roof plateO Commis3ural axons do not cross the dorsal midline, ~250X). (B) An adjacent ~pinal cord section (low magnification o~ Fig. 3B) shows the roof plate labelled with an anti-keratan sulfate antibody (a-KS). Note the a~sence of reaction product anywhere else in the spinal cord. Superi~position of these two views de~onstrates the 3~

WO9~ 303 ~CT/~90/06189 ~ ?~

location of the roof plate glia between, but not overlapping with, the commissural axons; cc, central canal, (250X).
Figure 7. Differential expression of keratan sulfate epitopes in the roof plate of El5.5 rat cervical spinal cord, localized with various anti-keratan sulfate monoclonal antibodies, (A) a-XS, (B) 4-D-l and (C) ~-C-2 (see Material~ and Methods for d~scrip ions). In views (A) and (B), the roof plate is labelled from the pial surface to the central canal, whereas in (C), only the dorsal-most O portion of the roof plate is immunoreactive. Further, antibodies 4-D-l and 8 C-2 recognize epidermis (e), (B and C) and antibody 8-C-2 recognizes the basal lamina (bl) surrounding the cord, (C~, (A,B,C, = 400X).
Figure 8. Endo-B-galactosidase and keratanase t5 digestion of the roof plate~ (A) An El3.5 rat cervical spinal corcl section treated with chondroitinase ABC for over 2 hours at 37 C, then stained for keratan sulfate with antibody (a-RS). The labelling pattern is unchanged from that seen when sections are not pre-treated with chondroitinase A~C (compare to Fi5~ure 6B). Skin is also normally stained with this antibody, (400X). (B) El3 rat cervical spinal cord pre-treated with two keratan sulfate-specific enzymes: endo-B-galactoc;ida~e and keratanase.
Almost no Xeratan sulfate labelling is observ~d with anti-KS antibodie~ following this treal:~ent. These results wereduplicated in an El5 ~nimal using antibodies 40D-l and 8-~-2, 1630X).
Figure 9. Other antibodies and a lectin also recognize the roof plata glia ~ut are not unique to this region. These include (A) L2, (B) 5A5 (highly sialylated N-CAM), (C)a~SSEA-l and (D) lotus lectin (TPA). Vi~ws (A) and (B) are El3.5 and views (C) and ~D) are El5.5 rat spinal cord. In (A), compare the roof plate (RP), which is L2 immunoreactive ~"V"-shaped pattern) to the floor plate ~

-16- 2~rJ - ~,t~
(fp), which is entirely devoid of reaction product. Just below this olear region lie the decussating commissural axons. This antibody also labels the epidermis (e) and the commissural axons (ca). Antibody 5A5 (B) labels the commissural axon~ (ca) as well as the roof plate (RP), among others; a-SSEA-l (C) labels the roof plate (RP) and the ~loor plate (not shown)0 Note that this antibody, like 8-C-2 (Fig. 7C) only recognizPs the dorsal-most portion of the roof plate. However, some sections show light labelling with this antibody in the lower portion of the t~ roof plate as well. Lotus lectin (D) labels the roof plate (RP) along the dorsal midline; (A, 160X; B, 180X; C, 400X;
D, 400X).
Figure lO. Localization of cholinesterase in El3.5 and El5.5 rat cervical spinal cord. (A) The pattern of expression of cholinesterase in El3 spinal sord resembles the pattern of immuno~taining ~or keratan sulfate in the roof plate. Cholinesterase is present in other locations in the csrd as well, for example, in the ventricular portion of ~he basal neuroepithelial cells (be) and the o~al bundle of His (ob) (250X). (B) on El5.5, the developmentally regulated change in the roof plate morphology coinoides with a change! in cholinesterase expression. The pattern of expres~sion of cholinesterase is again similar to that of keratan sulfate. Also stained on 25 E15 . 5 are the dorsal column axons (àc), the basal ~pithelial cells (be)~ the sulcus limitans (sl) and motor neurons (mn), ~250X~. (C) High magnification of the roof plate ~hown in (B) demonstrating the close proximity of the dorsal column axons to the roo~ plate glia at this stage of 3~ developmentO It appears that cholinesterase is localized to a subpopulation of the dorsal column axons, (630X).
Figure ll. TransY~rse sections of El5.5 rat spinal cord (compare the rooP plate with the anti XS
labelled section in FigO 12). The plastic section shows 3~

W~9~/06303 PCT/USgO/0~189 2 ~

the apical and basal processes of the glial cells and their relationship to the pia and central canal (cc). Note the proximity of the dorsal column (dc) axons to the roof plate glia, (630X).
Figure 12. Cryostat sections (10 ~m) of rat cervical spinal cord (E15.5) labelled with an anti-KS
antibody (a KS). In (A) note that the dorsal portion of the roof plate is flared out with the ventral portion being thinner; dc, dorsal columns (400X~, (B) A similar section at higher ~agnification, shows the dorsal column axons as they approach the dorsal roof plate cells (arrows) and abut them. As the dorsal column axons continue to enter, they fill the spaoe along the remainder of the roof plate, (650X).
Figure 13. Xeratan sulfate epitopes are expressed by other non-innervated regions. Labelling with numerous anti-keratan sulfate antibodies is found ~A) and (B) on cells which surround developing rib cartilage in E15.5 rat, and (C) by the outer layer of the epidermis; (A, 250X; B and C, 630X)~
Figure 14. The roof plate at E17.5 no longer expresses keratan sulfa~e epitopes, but labelling is still present in the surrounding car~ilage, (150X).
~igure 15. Immunocytochemical labelling of the dorsal midline of optic tectum in hamster mesencephalon with antibodies to keratan sulatP. (A) Labelling with antibody 4-D-l occurs solely along the tectal midline as shown by horseradi~h peroxidase reaction product, (400X).
(B) The tectal midline is also labelled with antibody 8-C-2 shown h~re in dark~ield with im~uno~luorescence. Note the intensity of reaction product just above the roof plate in the basal lamina with this antibody, (400X). Other antibodies to keratan sulfate also stain this region (not shown). Note the dense staining at the ventricle.

~5 WO9l/06303 PCT/US90/0618 Figure 16. Substrate preparation technique: 60 mm petri dishes are coated with a mixture of methanol and nitrocellulose and air dried in a laminar flow hood.
Cellulose strips (350 ~m wide) are soaked in the desired protein solution (e.g. proteoglycan, PG, plus LN or NCAM) +
RITC label and transerred to the c nter of the petri dish in vertical strips (shown by hatched lines). Laminin is then applied to the entire dish with a bent glass pipet followed immediately by media. Dishes are stored in the dark to preserve ~luoroescence until DRGs are dissected and 1~ ready for seeding. After 24 hours, the plates are fixed, coverslipped and photographed.
Figure 17. Controls. (A) Strip~ contain 1 ~g/ml laminin + RITC and 100 ~g/ml laminin is spread over the cntire dish. (B) Strips contain 10 ~/ml laminin +
RITC and 100 ~g/ml laminin is spread over the entire dish.
Nitrocellulose only binds the first reagent transferred, thus laminin strips result in alternating concentrations of 1 and lOG ~g/ml (A) or 10 and 10 ~g/ml (B). Arrows denote boundary o~ lanes in (A) and RITC fluorescence denotes location of lane in (B). In each case, neuri~es freely cross the lanes without any signs of inhibition, indicating that neither idiosyncrasies of the protocol nor toxicity of the RITC are problematic in this assay, 250X.
Figure 1~. Bovine KS/CS-PG (1 mg/ml) + RITC is ~rans~erred in strips with lOO ~g~ml laminin spread over the entire dish. Dorsal root ganglia, gently scattered over the center of the dish adhere to the strips of laminin (areas in betwe~n strips of XS/CS-PG + RITC) nd send out neurites. While abundant ~rowth occurs on laminin, 3D complete inhibition of neurites and support c~lls occurs when the neurites encounter the KS/CS-PG; 250X.
Figure 19. Neurite outgrowth .inhibition by KS/CS-PG is concentration dependent. In this protocol, strips of the proteoglycan were placed from left to right ' W~91/06303 PCT/U~0/06189 2 ~ ù~ ~

1~ lncreasing concentrations from 0.2 mg~ml to l.O mg/ml.
This figure shows 0.2 mg/ml (le~t) and 0.4 mg/ml (center).
Arrows denote boundary of lanes. Compare with Figure 18 which shows complete inhibition at 1.0 mg/ml; 40X.
Figure 20. To determine whether neurites were actively inhibited by 1 mg/ml bovlne KS/CS-PG or merely stopping due to the lack of a condusive molecule, we mixed laminin with the proteoglycan. In (A), 1.0 mg/ml KS/CS-PG
is mixed with 10 ~q/ml laminin + RITC (fluorescence shows location of lanes). Neurites are s ill inhibited by XS/CS-PG, even though a concentration of laminin is prese~t which alone allows for abundant outgrowth (~ee control in Fig. 17). (B) When the concentration of laminin is raised to 100 ~g/ml neurites are able ~o ~ross the KS/CS-PG
containing strip; 250X.
Figure 21. Response of DRG neurites to a mixture sf 1 mg/ml KS/CS-PG with polysialylated NCAM at two concentrations: at 10 ~g/ml NCAM, neurites are inhibited by the RS/CS-PG. However, unlike higher concen~rations of laminin, 100 ~g/ml NCAM is still inhibitory ~or all but a 2~ few neurites (not shown); 40X.
Figure 22. Control for NCAM mixture. Strips containing 10 ~g/ml polysialylated NCAM provide a conducive substrate for neurite outgrowth. Neurites growing from 100 ,ug/ml laminin to NCAM in stripæ show no pat~ern change or 2S change in fasciculation; 160X.
Figure 23. Enzyme digestiGn assay I. (A) DRG
neurites are inhibited by 1 mg/ml c:hick KS/CS-PG in the sa~e manner as seen for bovine XS/CS-PG at the same concentration; vertical arrows denote location of lane boundary; (B) When ~S/CS PG is tr~ated with keratanase, some n2urites cross, w~ile many are Still inhibited.
Vertical arrows denote lane boundary: horizontal arrows point out neurites whiCh have elongated onto the lan2;
250X.
3~

,..
.

W~1/06303 PCT/US90/061~s Figure 24. Enzyme digestion assay II. (A) Treatment of KS/CS-PG with chondroitin ABC lyase allows many neurites to cross the stxips, although some inhibition is still quite eviden~; vertical arrows denote one of the lanes; 40X. (~) If KS/CS-PG is treated with bot keratanase and chondroitin ABC lyase, l~aving only the protein core and LN, neurite inhibition is no longer seen.
This experiment also serves as a control for the presence of a neutral molecule which shows no inhihitory effect (see Discussion). Arrows and fluorescence denote position of lane; 100X.
Figure 25. A rat chondrvsarcoma cartilage proteoglycan (RCS; 1 mg/ml~ contains chondroitin sulfate, but not keratan sulfate chains. The chondroitin is in the form of C~4-s and not C-6-S like the bovine and chick 1~ KS/CS-P~ above. This reagent is no~ as effective in the inhibition of neurites as i5 bovine and chick XS/CS-PG, although partial inhibition can bQ seen. Vertical arrows denote lane boundary; horizontal arrows point out neurites which have elongated onto the lan~; 40X.
Figure 26. In vitro assay for C6 glial cell invasion. (A) Inhibition of invasion. No cells are found on the strip. (B) Slight inhibition of invasion. Note the pre~ence of a few cells on the strip, but the cells are not confluent. ~C) No inhibition o~ invasion. Cell migration and confluence are evident on the strip.

5. DETAILED DESCRIPTION OF THE INVENTION
The pr~sent invention relates to the discovery that kerata~ sulfate (KS), chondroitin sulfate (CS), 3D dermatan sulfate (DS), heparan sulfate (HS), heparin (HN), and/or hyaluronic acid (hyaluronate, HA) can inhibit neurite outgrowth i.e., axonal growth, or nerve regeneration ~herein "nerve growth") or gli21 cell, in particular astrocyte, migration, inva~ion or regeneration.
3~

..

WO91/06303 ~CT/~S90/06~89 2 ~

Inhibition of neurite outgrowth, i.e., nerve growth, results from XS and/or CS, DS, HS, HN or HA even in the presence of n~rve growth promoting factors such as laminin and NCAM. Inhibition of glial cell, in particular astrocyte, migration or inva~ion results from XS and/or cS~
DS, ~S, HN OR ~A even in the presence of laminin.
Accordingly, the pre~ent invention is directed to methods of using KS, and molecules and compositions comprising KS
to inhibit or prevent neurite outgrowth and/or glial cell migration or invasion, or nerve or glial cell regeneration, 1~ and therapeutically, where the ~oregoing is desired. Such molecules comprising XS include but are not limited to KS
glycosaminoglycan and KS proteoglycan, with keratan sulfate proteoglycan most preferred. The invention is ~urther directed to molecules and compositions comprising CS, and the therapeutic uses thereof to inhibit or prevent neurite outgrowth, glial cell migration or invasion, or nerve or glial cell regeneration. Molecules comprising CS include but are not limited to CS glycosaminoglycan and CS
proteoglycan, with chondroitin sulfate proteoglycan preferred. The invention is also directed to molecules and compositions comprising dermatan sulfate and therapeutic uses thereof to inhibit or prevent neurite outgrowth, glial cell miyration or invasion, or nerve or glial cell regeneration. Molecules comprising DS include but are not limited to DS glycosaminoglycan and DS proteoglycan, with dermatan sulfate proteo~lycan pr~ferred. The invention is further directed to molecules comprising heparan sulfate, heparin, and hyaluronate, and therapeutic uses thereof to inhibit or prevent neurite outgrowth, glial cell migration ~0 or invasio~, or nerve or glial cell regeneration.
In another embodi~ent, inhibitors and antagonists of KS, CS, DS, ~S, HN or N~, and molecules and compositions containing the sa~, may be used to promote neurite outgrowth or nerve reqeneration, i.e., nerve Wo 91/~6303 PCT/lJS90/06189 growth, or glial cell, in partlcular astrocyte, migration, invasion or regeneration and can be administered therapeutically~ Such inhibitors a~d antaqonists include but are not limited to antibodies to KS, CS, DS, HS, HN or HA and derivatives or fragments ther20f containing the binding domain, enzymes that degrade KS, CS, DS, HS, HN or HA, lectins specific for KS, CS, DS, HS, HN or HA, and disaccharide antagonists of receptors specific for RS, CS, DS, HS, HN or HA. In this embodiment, promotion of neurite outgrowth, i.e., axonal growth, or ylial cell, in 1~ particular astrocyte, migration or invasion, or nerve or glial cell regeneration occurs by removing the inhibitory influenGe of molecules comprising XS, CS, DS, HS, HN or HA, thus allowing promotion of neurite outgrowth, i.e., axonal growth, or glial cell, in particular astrocyte, migration ~5 or invasion, or nerve or glial cell reg~neration by endogenous or exogenously added molecules.
In a further embodiment of the invention, molecules comprising KS can be used together with molecules comprising another glycosaminoglycan or the disaccharide unit thereof, preferably chondroitin sulfate, in the methods of the invention.
The present invention also provides pharmaceutical compositions comprising effective amounts of molecules and compositions comprising KS, CS, DS, HS, HN
2S and/or HA-As detailed in the examples sections infra, immunocytochemical localizatlon data is pr~sented which indicates that RS, alone or in combination with other molecules.such as ohondroitin sulfate, may be in part responsible ~or the inhibition of axon elongation through the roof plate in the embryonic spinal cord. In a fur~her example, we demonstrate in vitro that keratan sul~ate/chondroitin sulPate proteoglycans are actively inhibitory to neurite elongation in a concentration 3~

, , ~. .

WO9]/06303 PCT/U~90/0~1~9 q~~ ~

dependent manner. In other examples, we show that dermatan sulfate inhibits outgrowth of a neuronal cell line and neuronal cells in vitro. A further example in vitro demonstrate~ that KS/CS-PG and DS-PG inhibit migration and invasion of glial cells and astrocytes.

5.l. THE INHIBITORY COMPOS T_ONS OF THE_INVENTION
Compositions which are envisioned for use in the presen~ invention to inhibit or prevent neurite outgrowth, or glial cell, in particular astrocyte, ~igration or invasion, or nerve or glial cell regeneration (termed herein "inhibitory compositions") comprise an ~ffective amount of a molecule consisting of at least the disaccharide unit of KS, CS, DS, HS, HN or HA. Thus, for example, and not as a limita~ion on the scope of this invention, the molecule can be XS disaccharide, KS
~lycosaminoglycan, XS proteoglycan, CS disaccharide, CS
glycosaminoglycan, CS prot~oglycan, DS disaccharide, DS
glycosaminoglycan, DS proteoglycan or a compound containing any of the foregoing. In a specific embodiment of the 2~ invention, such inhibitory compositions include, in addition to such molecules comprising KS, another glycosaminoglycan or proteoglycan or disaccharide unit thereof, selected from the group ~onsisting of such molecules which comprise chondroitin sulfate (CS) and such molecules which comprise dermatan sulfate (DS). In a particular embodi~ent, a proteoglycan containing both KS
and C5 (XS/CS-PG~ can be used. Both C-4-S and C-6-S sulfur linkage forms of chondroitin sulfate are envision~d as within ~he ~cope o~ th~ invention, with ~he C-6-S ~orm 3D being preferred for the inhibitio~ of neurite outgrowth and nerve growth. The ~S for use in the pres~nt invention includes but is not limited to Type I (corneal) KS (which is unbranched a~d highly ulfated, and most easily and completely degraded by endo-b-galaetosida~e and keratanase .
- ~

- .

~, .

PC~U~;90/06 1 89 ~ O rS~ L

used sequentially (Melrose and Ghosh, 1985, Anal. Biochem.
170-293-300~) and Type II (skeletal) KS. In another particular emodim~nt of tha invention, the molecule may comprise DS.
~'he molecules comprising KS, CS, DS, HSj HN or HA, and other proteoglycans/glycosaminoglycans for use in the present invention, ~ay be obtained by standard procedures known in the art. For example, KS/CS-PG may be isolated from the cartilage matrix of cell cultures, such as those of limb mesenchymal cells, by published procedures (see Carrino, A. and Caplan, A. I., 1985, J. Biol. Chem.
260:122-127). In another embodiment, KS-PG can be isolated from shark fin, a rioh source of KS-PG. In an ~mbodiment where KS disaccharide or KS glycosaminoglycan is desired, KS disaccharide or KS glycosaminoglycan can be isolated after digestion of KS-PG with endo-b-galactosidase or keratanase, respectively (endo-b~galactosidas~ specifically cleaves between the KS disaccharide residues; keratanase specifically cl~aves at the glycosaminoglycan bond to the protein)~ Alternatively, KS disaccharides and ~0 glycosaminoglycans can be chemically synthesized, or purchas~d ~rom commercial sour~es.
In brieP, and as but one specific example, proteoglycan can be extracted from the cartilage matrix with 4 M guanidinium chloride containing protease inhibitors and puri~ied by CsCl equilibrium density gradient centrifugation and Sepharose CL-2B chromatography ~see Haynesworth et al., 1987, J. Biol. Chem. ~62:10574-105~

3D 5.2. THE GRO~TH-PROMOTING COMPOSITI_NS OF THE INVENTION
The present invention also provides methods of using compositions which pro~oke neurit~ outgrowth, or glial cell, in particular astrocyte, migratiQn or invasion, or nerve or glial cell regeneration (termad herein 3~

P~T/US~0/06189 "growth-promoting compositions"). Such growth promoting compositions comprise inhibitors or antagonists or agents which are otherwise destructive (collectively termed herein "growth promoting factors") of the neurite outgrowth and glial cell migration or invasion, or nerve or glial cell regeneration inhibitory activity of keratan sulfate (as exhibited by KS disaccharide, KS glycosaminoglycan, XS
proteoglycan, or molecules containing the foregoing). Such growth promoting factors include but are not limi~ed to antibodies which recogni2e keratan sulfate, and derivatives and ~ragments thereof which contain the binding domain, enzymes which degrade keratan sulfate, lectins specific for keratan sulfate, and disaccharide antagonists of the keratan sulfate receptor (see Sections 7.l., 6.3. and 7.2.4.4., infra).
In another embodiment, the qrowth promoting compositions comprise inhibitors or antagonists or agents which are otherwise destructive o~ neurite outgrowth, or glial cell migration or invasion, or nerve or glial cell regeneration inhibitory activity of CS (as exhibited by CS
disaccharide, CS glycosaminoglycan, CS proteoglycan, or molecules containing the foregoing). Thus, the invention is also directed to antibodies to CS (and ~ragments thereof), en2ymes which d~grade CS, lectins specific for CS, and disacchaxide antagonists of the CS receptor.
In a further embodiment, the growth promoting compositions comprise inhibitors or antagoni~ts or agents whioh are oth~rwise deætructive of neurite outgrowth, or glial cell migration or invasion, or nerve or glial cell regeneration inhibitin~ activity of DS (as exhibited by DS
disaccharide, DS glycosaminoglyca~, DS prot~oglycan, or molecules containing the foregoing). Thus, the invention i~ also directed to antibodies to DS (and ~ragments thereof), enzymes which degrade DS, lectins specific for DS, and disaccharide antagonists of the DS receptor.

~ " . .~, .

P~ 90/061 2 ~

In yet another embodiment, the growth promoting compositions comprise inhibitors or antagonists or agents which are otherwise destructive of neurite outgrowth, or glial cell mi~ration or invasion, or ~erve or glial cell regeneration inhibiting activity of HS, HN or HA (as exhibited by HS, HN or HA disaccharide, HS, HN or HA
ylycosaminoglycan, or HS, HN or HA proteoglycan, or molecules containing the foregoing). Thus the invention is also directed to antibodie~ to HS, HN or HA (and fragments thereof), enzymes which degrade HS, HN or HA, lectins specific for HS, HN or HA, and disaccharide antagonists of the HS, HN or HA recPptor.
5.2.1. ANTIBODY COMPOSITIONS
Antibodies which recogniz~ keratan sulfate, CS, DS, HS, HN or HA, and which may be used, include previously isolated known antibodies as well as antibodies which can be newly generated.
XS disaccharide, KS glycosaminQgly&an, KS-PG, or compositions comprising the same, may be used as an immunogen to generate anti-KS ~ntibodies. CS disaccharide, :~
CS glycosaminoglycan, CS-PG, or compositions comprising same, may be used as an immunogen to qenerate anti-CS
antibodies. DS disaccharide, DS s71ycosaminoglycan, DS-PG, or compositions comprising the same may be used as an immunoge~ to generate anti DS an~ibodi~s. Various 2S procedures known in the art may be used for the production of polyclonal antibodies to epi~opes o~ KSI CS, DS, HS, HN
or HA. For the production of antibody, ~ario~s host animal~ can be immunized by injection with XS, CS, DS, HS, HN or RA-containing composition , including bllt not limited 3a to rabbits, mice, rats, etc. Various adjuvants may be used to increase the immunological response, depending on the host species, and including but not li~it d to Frsund's tcompl~te and incomplete), ~ineral gels such as aluminum hydroxide, surface active substances such a lysolecithin, 3~

, ~CT/US~0/~6l89 2 ~ 7 1 ~ ~ 3 pluronic polyols, polyanions, peptides, oil emulsions, keyhole limpet hemocyanins, dinitrophenol, and potentially us@ful human ad3uvants such as BCG (Bacille Calmette-Guerin) and Corynebacterium parvum.
In a preferred embodiment, a monoclonal antibody to KS, CS, DS, HS, ~N or HA is produced.
For preparation of monoclonal antibodies directed to ~S, CS, DS, HS, HN or HA, any technique which provides for the production of antibody molecules by continuous c~ll lines in culture may be used. For example, O the hy~ridoma technique originally developed by Kohler and Milstein (1975, Nature 2S6:495-497), as well as the trioma technique, the human B-cell hybridpma technique (Rozbor et al., 1983, Immunology Today 4:72), the EBV-hybridoma technique to produce human monoclonal antibodies (Cole et 1~ al., 1985, in "Monoclonal Antibodies and Cancer ~herapy,"
Alan R. Liss, Inc. pp. 77-96) and the recombinant E. coli '`
library technique 9~ LRrner and coll~agu0s (Sastry et al., 1989, Proc. Natl. Acad. Sci. U.5.A. 86:5728-5732) and the like are within the ~cope of the present invention.
The monoclonal antibodies for therapeutic use may be human monoclonal antibodie~ or chim ric human-mouse ~or other species) monoclonal antibodies. Human monoclonal antibodies may be made by any of numerou~ techniques known in the art (e.g., Teng et al., 1983, Proc. Natl. Acad. Sci.
25 U.S.A. 80:7308 7312; Kozbor et al., 1983, Immunology Today 4:72-79; Olsson et al., 1982, Meth. Enzymol. 92:3-16).
Chimaric antibody molecules ~ay b~ prepared containin~ a mouse antigen-binding domain with human constant r~gions (Morrison et al., 1984, Proc. Natl. Acad. Sci. U.S.A.
30 81:6851; Takeda et al., 1985, Nature 314:452).
Previously prepared monoclonal antibodies to KS
which may be used according to the present invention include but ~re not limited t9 antibodies MZ15 ~Zanetti et al., 1985, J. Cell Biol. 101:53-59) (specific for sulfated 3~ :

.

.

. . I

Wosl/o~i3o3 PCr/US90/a6lB~

-2~ 8 ~ 8 poly N-acetyllactosamine domains on KS): 1~20/5-D-4 (Caterson et al., 1983, J. Biol. Chem. 258:8848 8854);
4/8/1-B-4 (Caterson et al., 1985, Fed. Proc. 44:386-393) : (both 1/20/5~D-4 and 4/8/1-B-4 recognize epitopes that o~erlap wilth MZ15, but were raised against human articular ,_artilage and steer nasal cartilage, respectively); 4-D-l;
,and 8-C-2 (both 4-D-1 and 8-C-2 were generated against ~hicken bone marrow and recognize two different epitopes of the highly sulfated form of keratan sulfate~; and a-Ks Monoclonal antibody 1/20/5-D-4 is commercially available O ~ICN ImmunoBiologicals, Lisle, Illinois, Cat. No. 696251).
Monoclonal antibody 3-3-3 (commercially available) specifically rPcogniz2s the C-4-S form of chondroitin sulfate after digestion of the C-6-5 form by chondroitin ABC lyase (Couchman, J. R., 1984, Nature : 15 307:65o-652).
A molecular clone of an antibody to a KS, CS, I)S, HS, HN or HA epitope can be prepared by known 1:echniques. Recombinant DNA methodology (see e.g., Maniatis et al., 19~2, Molecular Cloning, A Laboratory 2~ Manual, Cold Spring Harbor Laboratory, Cold Spring Harbor, ~Jew York) may b~ used to construct nucleic acid sequences :~
; which encode a monoclonal antibody molecule, or antigen binding region thereof.
r An~ibody molecul~s may be purii'ied by known 2 t:echniques, a.g., immunoadsorp ion or immunoa~inity c:hromatography, chroma~ographic methods sUch aæ HPLC (high performance liquid chroma~ography), or a combination :~ t:hereof, etc.
Antibody ~ragments which contain the binding clomain o~ the molecule can be generated by ~nown t:echniques. For exampl~, such fragments include but are not li~ited to~ the F(ab')~ fragmen~ which can be produced by pepsin digestion of the antibody molecule; the Fab' i.`ragments which can be generated by reducing the disulfide . .

., ~,.
,- ~ :

_ 9~

bridges of the F(ab')2 fragment, and the.Fab fragments which ~an be generated by treating the antibody molecule with papain and a reducing agent.

5.2.2. ENZYME COMPOSITONS
Enzymes that degrade keratan sulfate can be used in the practice of the instant invention, and include but ar~ not limited to endo-b-galactosidase and keratanase. In a speci~ic embodiment, both endo-b-galactosi~ase and k~ratanase can be used, simultaneously or sequentially, to degrade KS. In a further embodiment of the invention, enzyme(s) that degrade KS can be used simultaneou~ly or seguentially with enzyme(~) that degrade another proteoglycan/glycosaminoglycan, e.g. enzymes that degrade chondroitin sulfate or dermatan sulfate. In a particular aspect, the enzyme that degrades chondroi~in sulfat~ is chondroitin ABC ly~se.
Endo-~-galactosidase, keratanase, and chondroitin ABC lyase are commercially available (e.g., Miles Scientific).
In another a~pect, enzymes that degrade C5 can be used in the practice o~ the invention, and include but are not limited to chondroitinase and chondroitin ABC
lyase.
In a ~urther aspect, enzymes that degrade DS can be used in the practice o~ the invention, and include but are not limted to chondroitin ABC lyas~.
In yet another aspect o the growth promoting composiSions of th~ invention, enzymes that degrade heparan sul~ate, heparin, or hyaluronat~ can be used. These 3~ enzymes includ2, but are not li~ited to, heparanase and hyaluronidase.

3~

WO91/06303 PCT~US90/06l8~

5.2.3. OTHER COMPOSITIONS
Lectins, also referred to as agglutinins, specific for KS, CS, DS, HS, HN or HA comprise another aspect of the growth promoting compositions of the invPntion. Lectins that bind to keratan sulfate can bP
used in the prac ice of the ins~ant invention.
Furthermore, lectins that bind chondroitin sulfate can be used in the practice o~ the invention. In another aspect, lectins that bind dermatan sulfate can be used in the pr~ctice of the invention. In yet another aspect, lectins specific for heparan sulfate, heparin or hyaluronate can be used in tha practice of the inv~ntion.
In a sp~cific embodiment, the lectin from triticum vulgaris (wheat germ) specific for N-acetyl-D-glir osamine can be uced. The triticum vulgaris lectin 1~ binds to KS, CS and DS. In another embodiment, th~
tetrogonolobus purpureas agglutinin ~TPA) may be used.
Tetragonolobus purpureas is known as a~paragus pea, winged pea, and lotus agglutinin or lectin. Other lectins useful in the practice of this invention include, but are by no means limited to, lectins from ab:rus precatorius (Jequirity bean agglutinin), arachis hypogea (peanut agglutinin~, bandeiraea simplicifolia, erythrina corallodendron (Coral tree agglutinin), helix pomatia ('Roman snail agglutinin) and helix aspersia (garden snail aglutinin), limulus polyphemus (limulin or horseshoe crab agglutinin~, maclura pomiera, (o~a~e orange agglutinin~ momordica charantia, phaseolus limensis (lima bean agglutinin), phaseolus vulgaris (red kidney bean agglutinin), psophocarpus tetragonolobu~ (wing~d bean agglutinin), ~ophora japonica 3D (pagoda tree l~ctin), ulex europas (gor~ ayglutinin), vicia villosa (hairy vetch agglutinin), vagna radiata ~mung bean agglutinin~, and wi~t~ria ~loribunda (Jap~nese wis eria agglutinin), to nam~ but a fewO Any lectin that 3~ :

. ': .

W~9~/06303 ~CT/VS90/~6189 binds to the disacchaxide, glycosaminoglycan or proteoglycan comprisiny XS, CS, DS, HS, HN or HA may be used in the growth promoting compositions of the invention.
Disaccharide antagonists tha block receptors on nerve or glial cells ~pecific for KS, CS, DS, HS, HN or HA
can also be used in the practice of the instant invention.
Suitable disaccharide antagonists bind to the receptor for, but do no~ e~fect the inhibitory func~ions of RS, CS, DS, HS, HN or HA.
Metabolic blockers of proteoglycan synthesis may 0 also be used in the practice of the invention. Drugs or agents that inhibit or prevent synthesis or secretion of proteoglycans or glycosaminoglycans prevent the synthesis or secretion of XS, CS, DS, HS, ~N or HA, and thus preclude the inhibitory e~fects of ~S, CS, DS, HS, ~N or HA.
5.3. ~HERAPEUTIC USES
5.3.1. THE INHIBITORY COMPOSITIONS OF THE INVENTION
The inhibitory compositions of the invention can be therapeutically useful where an inhi~ition of neuri~e outgrowth, glial cell migration or invasion, or nerve or glial cell regeneration is desirable. For example, an inhibitory composition can be used in the treatment of pati~nts with qliomas or tumors of n~rve tis-~ue, e.g., malignant tu~ors such as a neuroblastoma. In another embodiment, an inhibitory composit:ion can be used for the 2~ treatment of a neuroma (undirected aXon growth associated with situations where th~ axon is ~issing either its approp~iate target or substrate pathway for neural development), For example, tr~atment of neuroma as osiated with amputation, lesion, or congenital deformities, etc.
can be treat~d. Disorders resulting from an overproduction of nerve growth-promoting factors, including but not limited to nerve growth factor, ciliary neurotrophic factor, brain-derived growth factor, laminin, NC~M, L2, and SSEA-1, can also be treated by administration of an 3~

WO91/063~3 PCT/US90/0618Y

32- 2 ~
inhibitory composition. The inhibitory compositions can be used to treat disorders of the c~ntral and/or peripheral nervous systems.
In another embodiment, the products of this invention can be used as barriers to glial cell migration or invasion caused by trauma, surgery, infection (viral or bacterial), m~tabolic disease, malignancy, exposure to toxic agents, or other hyperplastic situations. They may be used specifically to protect an organ or tissue from the previously mentioned conditions through a coating procedure. For example, dorsal root ganglia, optic nerve, and optic chiasma may be coated with prot~oglycans in order to protect against uncontrolled cell invasion and adhesion.
This may be useful as a preventitive or prophylactic treatment or may be applied as a treatment in patients where a disorder has already been manifested.
In one embodiment, compositions including keratan sulfate, in any molecular form in which it may be made or foundr either alone or with chondroitin sulfate, which also may be in any molecular form in which it may be found, or dermatan sulfate, which also may be in any molecular form in which it may be found, can b~ used to preferentially inhibit neurite outgrowth. In another embodiment, keratan sulfate and/or chondroitin sulfate or dermatan sulfate, in any ~olecular form in which it may be found, may be used to preferentially inhibit ~lial cell, in particular astrocyte, migration or invasion.

5~3O2~ THE GROWTH~PROM~TING COMPOSITIONS OF THE INVENTION
The growth-promoting compositions of the invention can be used therap~uti~ally in reqimens where neurite outgrowth is inhibited and an increa~e in neurite outgrowth or nerve regeneration is desired, e.g., in patients with nerve damage, or in regimens where glial cell, in particular astrocyte, migration, invasion or 3~

,. : ..

,.,':, .:

:' ' WO91/06303 PCT/US90/061~9 _33_ ~7~3~8 regeneration is deslred. The growth-promoting compositions can be administered to patient~ in whom nerves or glial cells have been damaged by trauma, surgery, ischemia, infection, metabolic disease, nutritional deficiency, malignancy, toxic agents, paraneoplastic syndromes, stroke, degenerative disorders of the nervous system, etc.
Examples of such disorders include but are not limited to Alzheimer's Disease, Parkinson's Disease, Huntington's chorea, amyotrophic lateral sclerosis, progressive supranuclear palsy, and peripheral neuropathies. In a specific embodiment directed to the treatment of Alzheimer's disease or systemic amyloidoses, in one particular aspect, the growkh-promoting compositions can be therapeutically applied so as to allow access to the sites of amyloid plaques (see Selkoe, D. J., 1989, Cell 58 611-1~ 612)o In another particular embodiment, the growth-promoting compositions of the invention can be used to promote nerve growth through an existing scar or a scar in the process of formationO The growth-promoting compositions may be used in the central and/or peripheral nervous systems, e.g., to preven~ the inhibition of and thus promote the regeneration of nerve pathways, fiber systems and tractsO
In a particular embodiment, growth promoting and/or inhibitory compositions of the invention may be used 2~ to appropriately direct axon growth along desired paths.
In a ~urther embodiment, th~ growth promoting compo~itions o~ the invention may be used t~ promote the migration or invasion o~ glial cells, in particular astrocytes.

5.~.3. PH~RMACEUTICAL COMPOSITIONS
The present inv~ntion also provides pharmaceutical compositions which comprise an e~fective amount of an an inhibitory composition or a growth-W09~/06303 ~C~/US90/0618' -34~
promoting composition, as the case may be, and a pharmaceutically accep~able oarrier. Such pharmaceutically acceptable carriers include sterile biocompatible pharmaceutical carriers, including, but not limited to, saline, buffered saline, dextrose, and water.
The a~ount o~ inhibitory or growth-promoting composition which will be effective in the treatment of a particular disorder or condition will depend on the nature of the disorder or condition, and can be determined by standard clinical techniques. In an aspect involving the use of an inhibitory composition, a high concentration of the molecul~ comprising KS, CS, DS, HS, HN or HA relative to the concentration of factors which promote neurite outgrowth or adhesion (e.g. laminin, NCAM)or glial cell, including astrocyte, migration or invasion at the desired site o~ therapy, i~ preferred for use.

5.3.4. MODES OF AD~INISTRATION
Methods of introduction of the pharmaceutical compo~itions o~ the invention include methods known to those skilled in the art. It may be desirable to introduce the pharmaceutical compositions of the invention into the central nervous system by any suitable route, including intraventricular and intrath~c~l in~ection;
intr~ventriçular injection may be facilitated by an intraventricular catheter, for example, attached to a re~rvoir, such as an Ommaya reservoir.
Further, it may be desira~le to administer the pharmaceutical co~position~ o~ the invention locally to the area in need of treat~ent: this may be achieved by, ~or example, and not by way of limitation, local infusion during surgery, by injection, by mean of a catheter, or by m~ans of an implant, said implant being of a porous, non-porous, or gelatinous material, including membranes, such as sialastic ~embranes, or fibers. Polymer implants coated ' '': "' :.

.

WO91/06303 ~CT/US90/06189 -35- ~7~ ~98 with the pharmaceutical composition can be applied or inserted at the desired site of treatment. Such polymers can have various comp~sitions, pore sizes, and geometries.
Polymers which can be used include but are not limited to those made of nitrocellulose, polyanhydrides, and acrylic polymers.
~ he invention also provides for the pharmaceutical compositions to be administered via liposomes, microparticles, microcapsules, or other semipermea~le membranes. In various embodimsnts of the invention, it may be useful ~o employ such compositions to achieve sustained release of the inhibitory or growth-promoting compositions.
It is also envisioned that one may introduce cells actively producing an inhibitory or growth-promotin~
composition into areas in need of such. For example~ a recombinant cell secreting an enzyme that degrad~s RS, CS, or DS can ~e administ~red where a growth-promoting composition is indicated. In a different embodiment, a hybridoma cell secreting an anti-lC5, anti-CS or anti-DS
monoclonal antibody can be admini~;tered where a growth-promoting composition is indicated~ The c~lls may be encap~ulated in a suitable biolog:ical membrane and implanted in the patient.

6. MOLECULAR AND CELLULAR C1~RACTERIZATION OF THE
GLIAL ROOF PLATE OF THE SPINAL CORD AND OPTIC
TECTU~: A ROLE FOR XERATAN SULFATE PROTEOGLYCANIN THE DEVELOPMENT OF AN AXON BARRIER
Certain types of glial structures, located at strategic positions along the edges o~ axon pathways, may provide the mechanical and/or ch~mical elem~nts for the construction of barriers which can grossly direct the elongation of axons during developmentO The roof plate, a putative axon barrier, is located along the dorsal midline of th2 d~veloping spinal cord and may be impor~nt for the 3~

WO91/06303 PCT/U~90/061 -3~ 3J ~ $
guidance of the commissural and dorsal column axons. We examined the roof plate to determine the developmental morphology of the region and to determine which ~olecules were correlated with the barrier function when axons were growing nearby. Light and electron microscopic 6 observa~ions of the roof plate revealed that this glial domain undergoes a dramatic chan~e in shape from a "wedge"
with large e~tracellular spaces between the cell apices at El2.5, to a thin, dense septum with reduced extracellular space at El5.5. Immunocytochemical techniques demonstrated O that highly sialylated neural cell adhesion molecule (N-C~M), the carbohydrate recognized by L2 monoclonal antibody, cholinesterase, stage specific embryonic an~igen 1, and a ligand that binds tetragonolobu~ purpureas agglutinin are expressed by th~ roof pla~eO These moleculesr however, were also found in other regions of the spinal cord which are permissive or attractive to axon gro~th. A molecule which is unique to the roo~ plat~ when axons grow close to, but do not cross, the dorsal midline is a glycosaminoglycan (GAG), keratan ~ulfate. Xeratan 2~ sulfate is also present in th~ tectal midline and in other non-innervated regions such as the outer epidermis and developing cartilage. Our data suggest the possibility that keratan sulfate, alone or in combination with other molecules expressed by the roof plate, may be re~ponsible, in part, for the inhibition of axon elongation through the roo~ plate in the embryonic spinal cord.

6.l. ~Y~ A~5 ~Y~ ~er~o--6.1.1. PL~STIC SECTIONS AND ELECTRON MICROSCOPY
_ _ _ _ The trunk region of Sprague-Dawley rat embryos, day ll.5 (El1.5~, El2.5, El3.5, E14.5 nd El5.5 were fixed by immersion in 4% parafor~aldehyde/1% glutaraldehyd in 0.l~ M phosphate huffered saline (PBS) overnight at 4-Co The tissue was trimmed, washed in 0.15 M PBS ~or 1 hour and 3~i . .

W~g1/06303 PCT/US90/06189 _37_ 2~7~3~
post-ixed in 1% osmium tetroxide for 2-3 hours on ice.
The sections were washed for an additional hour in 0.15M
PBS, dehydrated in ethanol and embedd d in Spurr's resin.
One micron sections were stained with 1% toluidine blue.
~hin sections were stalned with uranyl acetate and lead citrate for electron microscopy. In some tissue preparations, the salt concentration of the buffer was varied by 1.5 2.0 to obs~rve the effect on the extracellular spaces between the roof plate glia.

6.1.2. IMMUNOCYTOC~EMISTRY
Embryonic day 11.5, 12.5, 13.5, 15.5 and 17.5 Sprague-Dawley rats were decapitated (except E11.5-12.5 which were small) and fixed by immersion in 4%
paraformaldehyde on ice for 2 hours and cryoprotected with 30% sucrose in PBS overniyht. Cryostat sections (10-15 ~m), collected on gelatin subbed slides, were incubated in 10 mM containing 1% normal goat serum (NGS) (ICN
Immunobiologicals) and 0.1% Tritoll X-100 (Fisher Scientific Co.), pH 7.2 for 30 minutes at room temperature. Following washing in PBS/NGS, sections were incubated in primary antibody diluted in PBS/NGS/Triton X-100 overnight at 4~ C.
Sections were washed and incu~ated with ~RP-conjugated goat-anti-mouse IgG or IgM (~oehr:i~ger-Mannheim Biochemicals) secondary antibody, or fluorescein-conjugated secondary antibody (Boehringer-Mannhein Biochemicals~
overnight at 4- C. For sections treated with the HRP-conjugated secondary ant~body, the tissue was then washed and incubated in 3',3'-diamino~enzidine (Sigma Chemical Co.; final concentratio~ = 0.003%) activated with hydrogen 3D peroxide, for approximately 5-10 minutes. The reaction was stopped with buffer and the sectiolls washed with double distilled water, air dried, then dehydrated in increasing ooncentrations of ethanol and coverslipped in Permount mouting medium (Fisher Scientific CoO). Fluorescein W091/063~3 P~T/~90/061~9 2 ~ 3 ~
t~eated sections were washed with buffer and coverslipped with N~propyl galate to preserve immunofluorescence.
Controls did not receive primary antibody, but all subsequent steps were unchanged.
Postnatal day 0 tP0) and P3 Syrian hamster neonates were dissected, decapitated, immersion fixed and cryostat sectioned as describ~d above for the rat embryos, except that sections through the optic tectum were collected instead of ~pi~al cord. The tissue sections were reacted with a series o~ monoclonal antibodies directed toward a variety of keratan sulfate epitopes described b~low.
Numerous antibodies to various cell surface and extracellular molecules were tested. ~onoclonal antibody lC12 (Dodd et al., 1988, Neuron, 1:105-116) recognizes a glycoprotein on commissural axons. Ankibody 5A5 binds to th~ polysialic acid moietie on NCAM. It has been shown by oth~rs that NCAM may be the only source of polysialic acid in chick brain. Endo-N-treatment of NCAM to remove sialic acid results in a lack of NCAM reoognition by antibody 5A5.
2~ Further, 5~5 labels a band on a Wlastern blot of brain at 250 kD which corresponds to the molecular weight of highly sialylated NCAM. Anti-SSEA-l (So:lter, D., and Knowles, B.
B., 1978, Proc. Natl. Acad. Sci. U.S.A. 75(11):5565-5569) recogniz~s a staye-specific embryonic antigen which is 2~ first expressed in blastomexes of 8~cell stage mouse embryos. The monoclonal antibody MZ15 (Zan~tti et al., 1985, J. Cell Biol. 101:53-59), a gift of Ten F~izi of ~he Clinical ~esearch Cent~r in Harrow, England, has been well characteriz~d using por~ine chondrocytes and -qhown to be 3D highly speci~ic ~or sul~ated poly N-acetyllactosamine domains on k~ra~an sulfate oligosac~harides (Mehmet et al., 1986, Eur. J. Biochem. 157:3B5-391). ~onoclonal ankibodies 1/20/5-D-4 ~Caterson et al., 1983, J. Biol. Chem.
258:8848-8854) and 4/8/1-B-4 (Caterson ek ~1., 1985, Fed.
3~

.

WO9l/06303 ~C~/US90/06189 _39_ ~7~
Proc. 44:386~393) recognize epitop~s that overlap with MZ15, but were rai~ed against human articular cartilage and steer nasal cartilage, respectively. Antibodies 4-D-l and 8-C-2 were generated against chicken bone marrow and recognize two different epitopes of the highly-sulfat~d form of keratan sulfate. Monoclonal antibody a-KS is specific to an epitope of keratan sulfate. L2 (Kruse et al., 1984, Nature (London) 311:1S3-155), a gift from Melitta Schachner (Swiss Federal Institute o~ Technology, Zurich), recognizes a carbohydrate moiety ~ommon to several l~ neural cell adhesion molecules and myelin-associated glycoprotein. Immunocytochemistry for the localiæation of dermatan sul~ate, chondroitin-4- and chondroitin-6-sulfate was used, with chondroitinase digestion (Couchman, J. R., 1984, Nature 307:650-652).

6.1.3~ ENZYME DIGESTION ASSAY
E~bryonic day 13.5 and 15.5 rats were immersion fixed in 4% paraformaldehyde ~or 2 hours on ice and cryoprotected in 30% sucrose overnight at 4 C. Cryostat sections (15 lm) were collected on gelatin-subbed slides and air dried. Enzyme treatment consisted of incubating the tissue sections for 20, 40 or 60 minutes with a concentration of 0.001, 0.01 or 0.1 unit/ml of either endo-B-galactosidase or keratanase or both in sequ~nce.
26 Endo-B-galactosidase, isolated ~rom Esoherichia freundii, was purchas~d from ~iles Scientific, and was diluted in 0.1 M sodium acetate bu~far, pH 5.8. Ker~tanase, isolated from Pseudomona~ sp., was purchased from ~iles Scienfific and was diluted in Tris hydrochloride buf~er, pH 7.4. Controls consisted of: (1) incubatin~ spinal cord sections in chondroitinase ABC enzyme for 20, 40 and 60 minutes at a concentration o~ 0.1 unitJml in Tri~ acetate bu~fer, pH

7.3, ~ollowed by immunostainins ~or keratan sulfate with antibodie~ aKS, 8-C 2 and 4-D-l according to the above ~5 .

WO91/06303 PCT/US90/0618~

~ ~ 3~

protocol, (2) incubating a known chondroitin-containing tissue, chick femur (Stage 36) with chondroitinase ABC, keratanase and endo-B-galactosidase, then staining for chondroitin sulfate with antibody 3-B-3 (commercially available) according to the protocol of Couchman, J. R., 1984, Nature 307:650-652. Incubations with chondroitinase were conducted at 37 C and those for keratanase and endo-B-galactosidase were carried out at room temperature.

6.1.4. LECTIN STAINING
Embryonic day 13.5 and E15.5 rats were imm~rsion fixed in paraformaldehyde/glutaraldehyde (either 1:1 or 4:0.1) for 3 hours on ice and cryoprotected in 30~ sucrose overnight at 49 C. Cryostat sections (15 lm) of spinal cord were collected on chrom-alum sllbbed slide~. The 1~ tissue was then blocked in TBS/~SA (Tris-buffered saline/bovine serum albumin) for 15 minutes at room temperature. The block was removed and the sections were incubated in an HRP conjugate of the winged or asparagus pea lectin, tetragonolobus purpureas agglutinin (TPA), also known as lotus tetragonolobus or :lotus lectin (Steindler, D. A., and Cooper, N. G. F., 1987, Dev. Brain Res. 36:27 38) overnight at 4- C. at a concentration of 1075 or 1:100.
TBS/BSA with cations was added to the primary incu~ation to facilitate TPA binding. The tissue was washed and the TPA
visualized with diaminobenzidine (see protocol in Sec~ion 6.1.2). Following a final wash, the tirsue was dehydrated and coverslipped as above.

6.1.5. CHOLINESTERASE ASSAY
Embryonic day 13.5 and E15.5 rats were immersion fixed in 4% paraformaldehyde in 0.1 M PBS and the spinal cords wer~ cryo5tat section~d (10~15 lm). The tissue sec~ions were processed for cholinesterase~ using a modificat~on of Xoelle, G. B., and Fri~denwald, J. S., ' WO9l/0~303 ~CT/US90/~6~89 : -41-1949, Proc. Soc. Exp. Biol. Med. 70:617-622. The sections were rlnsed in distilled water numerous times and incubated overnight at room temperature in the dark in a mixture of 0.05 M sodium acetate, 4 mM copper sulfate, 16 mM glycine and acetylthiocholine iodide. The sections were rinsed and incubated in 1% sodium sulfide for 5-10 minutes, rinsed again and incubated in 4% formalin buffer overnight at 40 C. The tissue was rinsed a final time, dehydrated and coverslipped as above.

1~ 6.2. RESULTS
The roof plate undergoes morphological changes between embryonic day 11. 5 (Ell .S) and E12.5. At E11.5, the cells of the roof plate are arrang~d in an arching pattern in comparison to adjacent neuroepithelial cells which are mor~ radial (Fig. 2A). The extracellular space between the roof pla~e cells is minimal and comparable to that between the adjacent cells. In comparison, on E13.5, large extracellular spac~s, about 2~10 ~m in diameter, can be seen between the primitive roof plate glia but not 2~ between the adjacent calls (Fig. 3A). The large size and shape of the spaces are consist~nt in all animals and they are locat~d preferentially along the apical region o~ the roof plat~. At E13.5, the roof plate cells are arranged in the shape o~ a "wedge". With the electron microscope, we have observed that the apical processes of the roof plat~
c~lls terminate at the pial surface in endfeet and the ba~1 processes appear to end at ~he dorsal central canal.
Not every cell in the roof plate spans fro~ the pial surface to the cen~ral canal, as these c~lls are dividing until E14 (Altman, J., and Bayer, S. A., 1984, in AdvancPs in Anatomy, Embryoloqy and Cell Biology, Vol. 85, Springer-Verlag, Heidelberq, Germany, pp. 53-83). The roof pla~e is approximat~ly 70 ~m long from the pia to the central canal at E13.5 and about 100 ~m wide at the WO9l/06303 P~T/US90/0~189 -42- 2 ~
midpoint. Rostral-caudal analyses o~ 1 ~m plastic sections indicate that the extracellular spaces are present in the roof plate throughout the cervical and thoracic spinal cord. The spaces appear to be actual ~nd not due to processing of the tis~ue since various perturbations, e.g.
varying the salt conc~ntrations of the buffers by a factor of 1.5-2.0 does not alter the spaces r~latively more or less then those in the surrounding tissue. Further, Altman and ~ayer (1984, supra) have also observed large caliber extracellular spaces in the roof plate in tissue prepared differently.
Two axon systems are present in the dorsal region of the spinal cord and travel near the roof plate at timeG which seem appropriate for the roof pla~e to act as a barrier to them. The first of the~e, a sub-population of 1S the ventral commissural axons, arise from second-order neurons in the dorso-lateral wall of the spinal cord. The processes of these neurons can b0 visualized with antibody lC12 (Dodd et al., 1988, Neuron 1:105-116) (Fig. 4).
Neurogenesis occurs in a ventrodorsal gradient and one sees commissuxal axons arising from lateral cell bodies on E11.5 with very few, if any, axons more dorsally at this time ~Fig. 2B and C). However, by E12.5 the commissural axons are located in the dorsal-most cord as evidenced by their lC12 labelling. The unstained cell bodies of some of the neurons in this group then are located adjacent to the roof plate and their proces~es arch along parts of the perimeter of thQ roof plate before turning ventrally. A second neural population, the dorsal colu~n axons, travel very close to the roo~ plate as they enter the spinal cord and turn in the dorsal funiculus to travel ro trally. On E13.5, these axons reside in the oval bundle about 150 ~m from the dorsal midline (Fig. 13. However, by E15 they abut the roo~ plate at the dorsal midline of the spinal cord. Ultrastructural observations show that neurites do ~ ~ 7 ~

not cross the roof plate but are found in close apposition to i~ all along its perimeter ~Fig. 5). Thus, it appears that all processes from cells adjacent to the roof plate are excluded from this dorsal midline structure.
In order to characterize the surface or matrix molecul~s o~ the roof plate, we examined a variety of antigens and compared their distribution in the roof plate to that in the re~t of the spinal cord. Most of the molecules that were expressed by the roof plate were found elsewhere in the spinal cord or in other regions of the O embryo (discussed below). However, the distribution of keratan sulfate was restricted solely to the roof plate and waQ first detectable at about E~2.5, but more pronounced at El3.5 (Fig. 3~). Immunocytochemical localization showed that this glycosamino~lyean epitope is present in or on the primitive glial cells. The pattern of expression of this molecule directly coincides with the shape of th~ roof plate cells and their extracell~lar spaces revealed in the l lm plastic sections of El3O5 en~ryos (Fig. 3B). Figures 6a and 6B depic~ the relationship of the keratan sulfate labelling of the roo~ plate cells to nearby commissural axons labelled with lCl2. Thus keratan sulfate expression appears well before the arrival of the dorsal column axons and se~ms not to be present prior to, but rather, at about the same ti~e that the dor~al-~o~;t commissural population 2S is extending axons. In addition, the~e markers demon~trate that keratan sulfate epitopes are spscific to the roof plate and are found nowhere else in the sp~nal cord at this stage of developme~k (Fig. 6B~. Antibodies a-~S and 4-D-l label the dorsal midline ~rom the dorsal ~ost to the 3Q ventral-most portion (Figs. 7A and B), while antibodies 8-C-2 (Fig. 7C) and l-B-4 better recognize the dorsal-most portion. The difference in staining patt~rns ~uggests that these antibodies recognize slightly different keratan sulXate epitopes, since keratan sulfate occurs in various 3~

~: .

.

W09l/06303 PCT/U~90/06189 levels o~ sulfation and chain length (Heinegard, D., and Paulsson, M., 1984, Structure and metabolism of proteoglycans, in K. ~. Piez & A. H. Reddi (Ed.), Extracellular Matrix Biochemistry, pp. 277-322, New York:
Els~vier Science Publishiny Co.), and the roof plate contains a mixture of these epitopes.
Some of the anti-keratan sulfate antibodies also label other structures in the sections we studied. For example, many o~ the anti-keratan sulfate antibodies label epidermis (Figs. 7B and C), which has been shown by 1~ Funderburgh et al. (1986, Dev. Biol. 116:267-277) to contain this glycosaminoglycan. Antibody 8~C-2 additionally labels the basal lamina surrounding the spinal cord (Fig. 7C).
In order to tes~ whether the antibodies that we were using to localize keratan sulfa~e to the roof plate were specific for this glycosaminogly~an, we used two keratan sulfate speci~ic enzymes to digest this molecule from the roof plate: endo-~-galactosidase and keratanase.
It has been demonstrated by Melrose and Ghosh (1985, Anal.
Biochem. 170:293:300) that keratanase alone does not completely digest purified Type I keratan sulfate (corneal KS). However, complete degradation could be achieved when both endo-B-galactosidase and keratanase were used sequentially. They showed that Type II KS (skeletal KS) could not be completely degraded using these two enzymes, but substantial improvement wa observed in comparison to the U5~ of either enzy~e alone. With the sequential use of endo-B=galactosidase and keratanase in our assay, the intensity o~ staining in the roo~ plate, using antibodies specific for keratan sulfate, was largely decreased (Fig.
8B), in GompariSOn to untreated sections as analyzed by eye. Some incubations resulted in imperceptible staining levels, but there was subtle variation from section to section. Other keratan sul~ate-containing tissues, such as ~CT/US90/06189 cartilage and basal lamina retained much of their keratan sulfate e~pression following enzymatic digestlon, although skin showed some observable decrease in staining intensity.
The controls showed that chondroitinase digestlon had no ~isible effect on the intensity of keratan sulfate i~munostaining of the roof plate (Fig. 8A) and that keratanase and endo-b-galactosidase had no ef~ect on the staining inten~ity of chondroitin sulfate antibody on chick femur cells, a tissue known to contain large quantities of chondroitin-6-sul~ate (Carrino, A., and Caplan, A. I., 1985, J. Biol. Chem. 260~122-127). Thus, each enzyme appears to be specific for it5 appropriate substrate. The second control reduces the concern tha non-specific proteases contributed to the diminution of labelling in the roof plate by the anti-keratan sulfate antibodies.
The primitive roof plate glia express a number o~ other characteristic molecules on E13.5, but in contrast to keratan sulfate, these are seen elsewhere in the spinal cord. The carbohydrate recoynized by monoclonal antibody L2~HNK 1 (glucuronic acid 3-sul~ate) is expressed by the roof plate cells (Fig. 9A). In contrast, the floor plate is entirely devoid of labelling w:ith L2. Antibody 5A5 localizes highly sialylated NCAM l:o only the midline portion of the roof plate (Fig. 913). Both L2 and 5A5 label the DRG, the dor~al and ven~ral roots, the dorsal root entry zone and the entire margina:L zone o~ the spinal cord.
A histochemical as~ay for cholinesterase (ChE) showed that the roof plate glia ~press this ~olecule as well on E12.5 and E13.5 (Fig. lOA3, Th~ pattern o ChE expression on E12.5 and E13.5 in the roof plate is like that of the anti-keratan sulfate antibodies at this age, i.e. in a wedge-shaped distribution. ChE st~ining is also present in the dorsal root entry zone, on glial cells of the sulcus limitans and the ventricular portion of the basal plate neuroepithelia at this time (Fig. lOA).

Wog1/06303 PCT/US90/~1~9 ` -46- ~7 ~
The roof plate undergoes a second and more dramatic morphological alteration by El5.5. The presumptive glial cells become transformed into a long, thi~ septum-like structure in the dorsal midline (Fig. ll).
The extracellular spac~s are greatly diminished, resulting in a denser construct than that seen before this age. The roof plate spans approximately 160 ~m from the pial surface to the top of the central canal, but is o~ly l0-l5 ~m wide, except at its dorsal aspect where it widens. Thus, the roof plate has undergone about a two-~old increase in 0 l~ngth and approximately a ten-fold decrease in width, Developmental changes occur in the distribution of keratan sulfate epitopes as well as El5.5 and heir localization reflectg th~ morphological pattern change described above (Fig. 12A). With the use of differential int~rference contrast (DIC) microscopy, ons can appreciate that the medial-most dors~l oolumn axons (arrows) a~ut the ~lared out, dorsal portion of the keratan sulfate-labelled roof plate cells (Fig. 12A and B)o Keratan sulfate epitopes ware expressed by developing cartilage. Label is present surrounding groups of chondrocytes ~Fig. 13) and in some cases around the individual ohondrocytes themselvesO
Other markers expressed by the roof plate also change their distribution during development. on El5.5, the roof plate no longer demonstrates positive labelling with antibodies ~2/~NK-~ or SA5, but does become SSEA l positive (Fig. 9C). The ~xpression of this molecule res~bles that of 8-C-2 and 1-~-4 antigens in that it is present in or on the dorsal~o t but not the ventral-most 3~ portion of the roof plate glial cells. SSE~-l antigen is not uni~ue to the roof plate. It is expressed by the floor plate glia as well. The lectin TPA also binds to the roof plate glia and many other radial cells on El5.5. TPA
labels the doral midline in it~ entirety from the pial 3Ei ''. ` .~ :, . ' :, , ~ ~ .
.- ~ -:
. :

WO91/06303 ~CT/US90/06189 -47- ~ 3 ~ ~
sur~ace to the dorsal central canal (Fig. 9D). ChE is present along the dorsal midline from the pial surface to the top of the central canal on ~15.5 (Fig. lOB and C). At this stage, it i5 also expresssd by a subpopulation of sensory axons, in the ventricular portion of the basal neuroepithelia, the motor cells, the ventral root and in the sulcus limitans, as well as in the dev~loping limb bud cartilage of the upper trunk.
By El7.5, the roof plate occupies a minimal area of the dorsal midline laterally but still spans from the pial surface to the dorsal central canal. ~eratan sulfate epitopes are no longer detectable with immunocytochemistry in the roof plate (Fig. 14) nor in the basal lamina surrounding the spinal cord at El7.5, but they persist in cartilage and epide~mis.
In an effort to find any similarity between the roof plate and other putative dorsal midline axon barriers in the central nervous system, we tested normal hamster optic tectum for the presence of keratan sulfate. The hamster was chosen in view of work by Schneider (Schneider, 2~ G. E., 1973, Brain Behavioral Evolution 8:73-lO9) and Poston et al. (Posto~ et al., 1988, Society for Neuroscience Abstract 14:594) which has shown that neonatal lesions that are focused on the tectal midline allow for the abnoxmal development of a recrossed optic pro~ection (sse Discussion). Ind~ed, the dorsal midline of the ham~ter tectum is keratan sulfate positive with antibodies 4-D-l (Fig. 15A), 8-Co2 (Fig. 15B) and l-B-4 ~not shown) on the ~irst day of birth (PO) and on P3, a time when this midline region c~uld function normally to disallow optic 3D axons from passing betw~en the tecta. As in the rat spinal cord, kera~an sulfate label~ solely the midline o~ the m~sencephalon. Labellin~ can be seen along the ~ntirety of the dorsal ~ldline from the pial ~urface to the dorsal cerebral aquaduct of Sylvius.

WO91/06303 PCT/US90tO6l89 -48~ s9 ~
6.3. DISCUSSION
We have shown that the roof plate of the spinal cord undergoes morphological and molecular changes during early embryonic development. On E13.5, a network of large extracellular spaces develops near the pial surface between the glial cells of the roof plate and contributes to the roof plate's wedge shape. By E15.5, the shape of the roof plate has changed to a long, thin septum at the midline and the amount of extracellular space is significantly reduced.
Although th@ roof plate cells express a number of molecules O which are also present in other regions of the spinal cord, a particular glyco~aminoglycan, keratan sulfate, is expr~ssed solely by the roof plate glia beginning on ~12.5 and is no longer detectable by E17.5. ~y following the pattern of keratan sul~ate stainin~ in rela~ion to developing fiber systems in the spinal cord, we have obs~rved that the roof plate is never invaded by axons when keratan sulfate is present and that it eventually hecomes the midline boundary of the dorsa].-most portion of the ventral commissural pathway and of` the dorsal columns. The dorsal gray commi~sure, which crosses the dorsal midline at about E17.5 (Smith, C. L., 1983, J. Comp. Neurol. 185~
22), develops at a time when kerat:an sul~ate is no longer detectable with immunocytochemistry. It is likely that the roof plate interacts specifically with, exerting its inhibitory influence during development on, axon5 that elongate near the ~idline, i.e. a small subpopulation of the ventral co~missural ystem and a larger number of axons which constitute the medial-most (i.e. gracile tract) fibers of the dorsal columns.
It is likely that ker2ta~ sulfate does not occur alone in vivo but rather as a keratan sulfate/chondroitin sulfate proteoglycan (XS/CS-PG). ~e did not find that the roof plate glia express chondroitin sulfate using immunocytochemical techniques, but this could be due to the .
.. .
: ' ' ': ' ;

:
.
~:

WCI 91 /{~ti303 PCI /US90/06189 ~ ~ rj1 fact that glycosaminoglycans can be difficult to ~ix or stain. It is possible that chondroitin sulfate as well as other glycosaminoglycans and/or proteoglycans may be present in the roof plate and may be acting in combination with other molecules such as keratan sulfate glycosaminoglycan/proteoglycan to generate axon i~hibition.

6.3.1. GLIAL CELLS AS AXON BOUNDARIES AND BARRIERS
It has been suggested that glial structures may act a~ axon ~arriers or boundaries in regions of the nervous system other than the roof plate. A glial structure at the diencephalic/telencephalic junction near the front edge of the optic chiasm in mouse (Silver, J., 1984, supra) and chick (Silver et al., 1987, supra; Poston et al., 1985, supra) appears to act as a barrier to developing optic fibers. The "knot-like" structure, which is suggested (Silver, J., 1984~ supra) to be comprised of the progenitor cells of the 02A lineage (Raff et al., 19~3, Nature, 303:390-396), may ensure that the migrating fibers choose one of the ~unctionally advantageous pathways toward the midbrain and dien~ephalon instead of turning rostrally to enter the olfacto~y region of the telencephalon.
Studies on the d~veloping optic tectum in the Syrian hamster have revealed that a dorsal midline barrier comprised of glial cells may serve an important function for est~blishing normal axon-to-t~rget connections.
Classic studies o~ Schneider, G. E., 1973, ~upra and So, X., 1979, Journal of Comparative Neurolo~y, 186:241-258 have shown that early unilateral lesions of th optic tectu~ (superior colliculus) result in abexrant crossi~g of the optic tract axons and ~unctionally maladaptive synaptic connections within the inappropriate tectal lobe.
Re~en~ly, we have learned that crossing occ~rs if damage is focused on the midline alone and the dorsal basal lamina remains intact (Poston et al., 1988, Society for ~5 `:

WOgl/06303 P~T/US9~/06189 :: 2~7~

Neurosci~nce ~bstract 14:594). Intarestingly, this tectal midline boundary consists of glial cells which are much like those of the roof plate of the spinal cord. They maintain a primitive radial morphology and also express kera~an sulfate ~pecifically during development. This suggests that th~ dorsal midline of the mesencephalon may play an important rol~ in the normal maintenance of side restriction for migrating optic axons and that a glial structure can act functionally as a barrier to growth cones. If the roof plate of the developing spinal cord functions similarly to the ~idline glial structure of the tectum, it may constitute an essential blockade to aberrant axon elongation. Xn g~neral, barriers at the dorsal midline of the central nervous system may be instru~ental in separa~ing right v~rsus left side s~nsory information.
1~ Glia may also serve to compartmentalize regions of axonal arbsrization. ~ type of boundary glia identified by anti-glial fibrillary acidic protein (~FAP) has been observed in neonatal cortex (Coop~r, N. G. F., and Steindler, D. A., 1986, Brain Res. 380:341-348). The parcellation by these calls of the somatosensory barrel fields of cortical Layer IV in mouse correlated directly with patterning o~ the mystacial vibrissae. The glia are delineated by their dense expression of glycoconjugates that are specifically recognized by certain lectins (Steindler, D. A., and Cooper/ N. G. F., 1987, Dev. Brain Res. 36:27-38~. Glial cells of the barrel wall domains app~r~ntly r~lect the mature patt~rning o~ he thalamic ter~inal arbors related to vibrissal function. However, this form of boundary differs from the roof plate. It 3~ appears to be more plastic sinc~ the intense matrix producing-glia of the barrel walls are able to shift their position gsometrically in respon~e to an activity-dependent signal associated with the afferant axons. Cells that may play a similar role in cordoning synaptic territories have ' , WO9l/06303 .'CT/US90/0~189 2 ~ ~ ~ 3 ~ ~

also been observ~d by Oland et al. (1988, J. Neurosci.

8(1):353-367~ in the olfactory reyion of the moth, Mand~ca sexta.
The roof plate cells of the spinal cord de~erve discussion on three separate but interrelated aspects which may provide evidence about ~heir shape and mechanism of ~xon repulsion: (1) the po~sible ~tructural contribution of the extracellular space between the glial cells, (2) the absence of growth of axons through the extracellular spaces and (3) the inhibitory functions of th~ ~olecules ~xpressed by these cells.
6.3.2. EXTRACELLULAR SPACE IN THE ROOF PLATE
Developmental changes in the amount of extracellular space may be important for creating the shape and density of the roof plate. At E13.5, it saems likely that the placement of th~ extracellular spaces, preferentially along the apical region of the roof plate, could play a role in the construction of a wedge (Fig. 3A).
By E14.5, sinc~ the cells are no longer dividing (Altman, J ~ and Bayer, S. A., 1984, in Advances in Anatomy, Embryology and Cell Biology, Vol. 85, Springer-Verlag, Heidelberg, Germany, pp. 53-83~ and the spinal cord is expanding laterally, the extracel:Lul~r spaces may be contributing to or allowing ~or thi~ expansion. The 2~ ~eratan sulfate and probably also the chondroitin sulfate chains of the KS~CS-PG could posslbly help play a role in th~ creation o~ the extracellular spac~s since it is known tha~ glycosaminoglycans in g~neral bind water preferentially (Margolis et al., 1975, Biochem. 14(1):85-3D ~B).

, , ' '~
,. . .

WO9lt06303 PCT/US90/06189 2 ~ $

6.3.3. AXON OU~GROWTH IN ~ LATION ~o THE ROOF PLATE
_ _ _ _ _ It was somewhat surprising that axons do notgrow in the roof plate, since in this (Altman, J., and Bayer, S. A., 1984, supra) and other developing neural systems, it has been suggested that large, ma~rix-filled extracellular ~paces between qlial cells, when aligned into channels, may act in a permissive or instructive manner to guide axons toward their targets ~Silver, ~., and Robb, R.
M., 1979, Dev. Biol. 68:175-190; Singer, et alO, 1979, J.
CompO Neurol. 185:1-22; Silver, J., and Sidman, R. L., 1980, J. Comp. Neurol. 189:101~111; Nordlander, R. and Singer, M., 1982, Exp. Neurol. 75:221-228). Evidence exists that the migra~ion of neural crest cells is guided by their channel-liXe extracellular environment and ~hat the extracellular ma~rix i~ directly involved tBronner, M.
E., and cohen, A. M., 1979, Proc. Natl. Acad. Sci. USA
76:1843-18~7: Bronner-Fraser, M., and Cohen, A. ~., 1980, Dev. Biol. 77:130-141; Erickson et al., 1980, Dev. Biol.
77:14~-156; L~Lievre t al., 1980, Dev. Biol. 77:362-378;
Bronner-Fraser, M., 1982, ~ev. Biol. 91:50-63). However, 2~ our observations of the roof plate show that the mere physical presence of large extracellular spaces is not necessarily predictive of a futurc axon pathway, since the spaces in the dorsal ~idline define a region which appears to be rsfractory to axon growth. Our data indicates that keratan sulfate glycosaminoglycans/proteoglycan, alone or in ~ombination with other glycosaminoglycans/proteoglycans, in th~ spaces or perhaps on the surface of the glial cells, extending into the spaces, may function to modify the environment in order to generate the refractory characteristics of a barrierO

:

WO9l/06303 ~CT/US90/06~89 20r~ l~rA8 6.3.4. MOLECULES EXPR~SSED BY THE ROOF PLATE GLIA
Xeratan sulfate epitopes may function, in part, in the creation of a molecular barrier in the roof plate.
Also, we have observed with light microscopy that the region surrounding d~veloping cartilage and the matrix around individual chondrocytes express keratan sulfate like immunoreactivity. Cartilage is not innervated. Further, we have shown that outer epidermis expresses keratan sulfate epitopes during development.
Importantly, tissue culture experimen~s demonstrate that keratan sulfate glycosaminoglycans can directly inhibit axon growth (see Sectisn 7 ~nfr2).
An obvious consideration is whether the antibodies we used were recognizing keratan sulfate in the roof plate or rath~r were binding to another antigen which 5 i5 keratan sulfate-like in molecular compo~ition. A factor in favor of speci~icity was that so many different well-characterized antibodie~ to various keratan sulfate epitopes s~.ain the roof plate. St:rong evidence, however, resulted from our enzyme degradation studles. The fact 2~ that in~ubation of spinal cord sec:tions with two keratan sulfate specific enzymes (while controlling for non-specific effects) reduces the int~nsity of antibody staining suggests~ that kerat:an sul~ate is indeed expressed on or in the roof plate glia, and (2) that the roof plate may contain Type I (corneal) XS. Corneal keratan sulfat~ is a molecule which is unbranched and highly sulfated and one that is most ~asily and co~pletely degraded by endo-8-~alactosidasa and keratanase used sequentially (Melrose and Ghosh, 1985, Anal. 8iocbem.
170:293-300).
our present studies indicat~ that there are mol~cular di~ferences between the dorsal-most and ventral-most roof plate since antibodies anti SSEA~ B~4 ~nd 8-C 2 label only the dorsal portion. This heterogeneity may WO91/06303 PCT/US90~06189 be important with respect to the temporal and regional variation in axon inhibition in the roof plate, since, at late stages, the dorsal commissure develops through the central portion of the dorsal midline (Smith, 1983, J.
Comp. Neurol. 220:29-43). Importantly, the formation of the commissure occurs near the time when the expression of a keratan sulfate diminishes below detectable level~.
The presence of cholinesterase in a putative axon barrier region, the roof plate and in cartilage, along with evidence from the above studie~, suggest that cholinesterase could potentially influence neurite outgrowth.
While L2, sialylated NCAM and SSEA-1 have been implicated in cell-cell adhesion, their possible role in the roof plate may be to generate morphogenetic density which appears to increase between the roo~ plate glia during devalopment. Lacking any other influences, these molecules could potentially attract axons into the roof plate. A negative or inhibitory factor may be required to override this potentially axon-attractive milieu.

7. KERATAN SULFATE/CHONDROITIN
SULFATE PROTEOGLYCAN (KS/CS-PG) INHIBITS NEURITE OUTGROWTH IN VITRO
As detailed in Examples Section 6 herein, in vivo studies of the roof plats and optic tectum in rodent and the developing subplate in the telencephalon of the chick ~howed that two glycosaminoglycans, keratan sulfate, alone or in combination with chsndroitin sulfate, possibly in the proteoglycan ~orm (KS-PG, CS-PG, or KS/CS-PG) were pr~sent within the~e regions at ti~es when axons approach closely but do not invade these territories. In ~rder to determine if XS~CS/PG actively inhibits growth cone elongation and to determine which component(~ of this macromolecule ~ay be critical to this phenomenon, we used a technique employinq nitrocellulose-coated petri dishes onto 3~

W~91/06303 PCT/US90/06~89 8 ~ 8 which strips of varic;s purlfled mol~cules were transPerred. We grew E9 chick dorsal root qanglia on lanes of KS/CS-PG in alternation with lanes o~ the growth-promoting molecules laminin (LN) or NCAM. Neurites grew abundantly along stripes of LN or NCA~. In contrast, upon encountering a stripe containing XS/CS-PG, neurites either stopped abruptly or turned and travelled along the KS/CS-PG
strip border in a concentration dependent manner. To determine whether the inhibitory e~fect was due to the presence of KS/CS~PG or merely to the absence of LN or NCAM
in the axon~free lan2, we mixed LN or NCAM with the KS/CS-PG, in concantrations which alone support luxurious outgrowth, and observed that the ~S/CS-PG was ~till inhibitory when the at ractive molecules were present.
KS/C5-PG plus NCAM remained inhibitory even at very high concentr~tions of NCAM. Howev~r, high concentrations of LN
were able to overcom~ the inhibitory eff~ct of the KS/CS-PG. Enzymatic digestion of the KS or CS from the KS/CS-PG
permitted various degrees o neurite outgrowth to occur across the previously inhibitory lanes, and digestion of both glycosaminoglyoan moieties, le~ving only the protein core o~ the molecule, resulted in a co~plete lack of inhibition. A proteoglycan containing a different glycosaminoglycan, dermatan sulfate ~DS-PG~, and a rat chondrosarcoma cartilage CS PG, wer~ not as e~fectiY~ in axon inhibition as KS/CS-PG at the same concentrations.
These as~ays demonstr~ted that KS/CS-P~ is actively inhibitory to embryonic dorsal root ganglia neurites in vitro. Complete inhibition required contributions from both KS and CS moieties.

, ~

WO9l/06303 PCT/US90/06189 ~7~L~38 7.1. MATERIALS AND MET~ODS
7.1.1. SUBSTRATE PREPARATION
Tissue culture substrates were prepared by coating 60-mm Petri dishes evenly with 0.5 ml of a mixture of 5 cm2 nitrocellulose (Schleicher & Schuell, Type BA 85) 5 dissolved in 6 ml ~ethanol and allowing them to air dry in a laminar flow hood (Lagenaur, C. and Lemmon, V., 1987, Proc. Natl. Acad. Sci. USA 84:7753-7757). Cellulose (Whatman Filter paper, #1) was cut into 350 ~m strips and used to blot various protein substa~cPs down onto the nitrocellulose substrate. Each protein solution contained either rhodamine isothiocyanate (RITC) or fluorescein isothiocyanate (FITC) as a marker which could later be d~tacted to deter~ine the exact position of the strips.
The cellulose strips were soaked in 20 ~l of the desired l~ protein solution, transferred to the nitrocellulose-coated lA
dish in a vertical pattern (FIG. 1), allowed to set for 30 :.
seconds then removed. Following drying of the test molecules onto the nitrocellulose, a thin coat of 100 ~g/ml laminin ~LN) (Gibco, Inc.) was spread evenly across the dish with a bent glass Pasteur pipette. Media was immediately added to the dish (DMFM/F12 ~ 10% fetal calf serum ~ 5% chick embryo extract + 1% antibiotics), and the dish was stored in an incubator until needed. Importantly, the experiments were repeat~d using serum-free media, DMEM/F12 + N2 (1:100, Bottenstein, J. E. and Sato, G. H., 1979, Proc. Natl. Acad. Sci. USA 76:514-Sl7) ~
antibiotics, si~ce it i~ k~own that s~rum may give inaccur~te results with respect to its e~fect on proteoglycan binding. In our hand~, however, the presence or absence of s~rum in the media wa~ not critical to the results of the assays.

3$

WO91/06303 ~ 9~6~9 7.l.2. DORSAL ROOT GANGLIA PREPARATIONS
Chick E9 dorsal root ganglia (DRG) were dissected in a calcium-magnesium free buffer by decapitating the chick, eviscerating, then carefully removing the vertebral column and ~pinal cord. The DRGs 5 were th~n cleaned free of surroundi~g tissue and plucked out using fine forceps. The media in the test culture dishes was removed and r~placed with fresh media containing 100 ng/ml nerve growth factor. With a finely drawn glass Pasteur pipette, the DRGs were picked up and scattered gently around the center of the dish containing the patterned stripes. Approximat~ly 20 DRGs were seeded onto each dish. ~he dishes were then incubated ~or 24 hours followed by fixation with 4% paraformaldehyde/O.l~
glutaraldehyde for l hour. The dishes were coverslipped in Mowial and and observed with a Leitz Orthoplan 2 fluorescent microscope, ~quipped with a Variolume which allows une to mix phase optics with fluorescence so as to observe the neurit0s and the loca~lon of the stripes simultaneously. Each experimen~ wa~ repeated at least 3 ~0 times.

7.l.3. DOT BLOT IMMUNOASSAY
To assur~ ourselves thal: the KS/CS-PG was being bound to the nitrocellulose, we cond~cted a dot blot 2~ immunoassay using various anti-k~ratan sulfate antibodies (an alternatiYe method was to us~ 35S-labelled proteoglycan; see below). Small squares of nitrocellulose pap~r were placed into 4-chambered di~hes. Each chamber was spotted wi~h 1 ~l of th~ KS/CS-PG and air dried ~or 5 minutes. Th~ paper was then blockad with straight normal goat serum (NGS) for lO minutes then washed 5X with buffer.
Each well wa~ filled with a diferent anti-keratan sulfate antibody: 8-C-2 and 4-D-l, a-KS, 5-D-4 and l-B-4 (Cat~rson et al., l985, F~d. Proc. 44:386-393) or MZl5 (Zanetti, et ;

WO~l/063~3 P~ /0 al., 1935, J. Cell Biol. 101:53-59), 1:100 in a ~ixture of 10 mM PBS ~ 3% NGS + 0.2% Triton X-100 and incubated at 37'C ov~rnight. The wells were rinsed 5X with buPfer, and a goat anti-mouse HRP-conjugated IgG or IgM secondary antibody was added and i~cubated overnight at 37 C. The nitrocellulose paper was then reacted with 0.01%
diaminobenzidine in PBS ~ 0.003~ hydro~en p~roxide. All dots showed a reaction produc indicating that the KS/CS-PG
was bound to the paper.
~5 7.1.4. _ S-LABELLING OF XS~CS-PG
To determine the amo~nt of proteoglycan which binds to the culture substrate, labelled mat~rial was used to coat the ~ubstrate by the same procedure as that used ~or the DRG cultures. Labelled proteoglycan wa originally isolated from the cartilage matrix of day 8 chicX limb mes2nchymal cell cultures a~ter the cultures were labelled with 50 ~Ci/~l ~35S] sulfate (DeLuca et al., 1977, J. Biol.
Chem. 252:6600 6608). The prot~oglycans were extracted with ~ M guanidinium chloride contai~ing protease inhibitors and purified by CsCl ~quilibrium density gradient centrifugation and Sepharose CL-23 chromatography tHaynesworth et al., 1987, J. Biol. Chem. 262:10574-10581).
The appropriate fractions fro~ th~ Sepharose CL-2B column were pooled, dialyzed again~t distill~d water at 4-C and 2 lyophilized to dryness so that the number of 35S cpm/mg dry weight could be determined. Aft~r the labelled proteoglycan was bound to th culture substrate, the area of ~he culture dish containing the bound proteoglycans was excised and trans~erred to a 20-ml glass scintillation vial. Cytoscint cintillation cocktail (ICN) was added to the vial and th~ amount of bound 35S was determined by sci~tillation spectrometry on a Beckman LS 6800 counter.

3~

WO91/06303 .~CT/US90/061~9 _59~
From the amount of bound 35S, th~ amount of bound proteoglycan could be calculated based on the number of 35S
cpm~mg dry weight.

7.l.5. IMMUNOCYTOCHEMISTRY FOR THE DETECTION OF LAMININ
_ To detect the presence of laminin in the strips, we used a polyclonal anti-l~minin antibody (BRC, VO Lemmon) to lab~l strips of the KS/CS-PG ~ LN. The 60 mm nitrocellulose-coated dishes were blotted with bovine KS/CS-PG (l mg/ml) + LN (lO ~g/ml). To block non-speGi~ic binding, we incubated the strips ln lO mM PBS with 10%
normal goat serum (PBS/NGS) for 30 minutes at room temperature. The dish was then incuhated with the primary antibody in PBS/NGS, washed 5X with buffer, incubated another hour with a goat anti-rabbit IgG ~onjugated with FITC, washed 5X with buff~r and coverslipped with N-propyl galate to preserve fluore~cenc~. The presence of LN was easily visualized in strips which mimioXed exactly the lanes formed originally with the LN/PG mixture.

7~l.6. MODIFICATIONS OF THE DQR5AL ROOT GANGLIA ASSAY
7.l.6.l. KS/CS-PG:L~ININ MIXTURES
We mad2.stripes on nitrocellulose plates which were strips containing ~ixtures o:f laminin and the KS/CS-PG, varying the laminin concentration among four plates:
25 l00 ~g/ml, 50 ~g/ml, 25 ~g/ml, andl l0 ~g/ml. The concentration of the RS/CS-PG was maintained at l mg/ml, which we knew ko produce maxi~u~ inhibitlon of n~urites.
The r~ainder of the experi~ent proceed~d as a~ove.
Control consi~t~d o~ strips of lO ~g~l laminin + RITC and ~D 10 ~g/Dl laminin spre~d over the whole di~h ~in the basic experim~nt lOO ~g/ml was used), ~ollow~d by seeding of DRGs and 24 hours incubation to determine whether this concentration of laminin is s~f~icient ~or normal out~rowth.
3~

WO~1/0~3 P~T/US90/~6189 f~ ~'J ~

7.1.6.2. KS/CS-PG:NCAM MIXTURES
Polysialylated NCAM i5 present i.n the roof plate during development (see Section 6, supra). In order to test the efect of a combination of the prote~glycan with NCAM, we mixed the PG and NCAM (gift of P. Yang and U.
Rutishauser) in lanes in alternation with LN.
Polysialylated NCAM was prepared by immunoaffinity purification in milligram quantities from the 0.5% NP40 extracts of E14 chick brain vesicles. NCAM in the extracts binds to anti chick NCAM monoclonal antibody ~5E) IgG
conjugated to Sepharose 4B beads which are activat~d by the cyanogen bromide method, and NCAM i5 eluted with 0.57%
diethylamine~ pH 11.5 (Hoffman et al., 1982, J. Biol. Chem.
257(13):7720-7729). The resulting NCAM is polysialylated and runs above 200 kD on SDS polyacrylamide gels. NCA~ (10 or 100 ~g/ml) was used in combination with 1 mg/ml KS/CS-PG
and the incubations conduct~d as abovec These dishes were compared to the XS/CS-PG + LN mixture~ and to the XS/CS-PG
alone.

7.1.6.3. E~ZYME DIGESTION OF :KERAT~N SULFATE CHAINS
2~
Two degrading enzymes s;pecific for Xeratan sulPate ch~ins, endo-B~galactosidase and keratanase (Melrose, J., and Ghosh, P., 1985, Analytical Bicchemistry, 170:293-300~ (Miles Sci.), and/or an enz~me specific for chondroitin sul~ate chains, chondroitin ABC lyase (Miles Sci.), were added at a conc~ntration of 100 U/ml to dishes with stripes that contained a ~ixture o~ laminin (10 ~g/ml) and bovine or chick XS~CS-PG ~1 ~g/ml). With respect to endo-B~galacto~idas~ and keratanase, this concentrakion was 3D observed to significa~tly degrade the keratan sulfate in 10 lm ~ro~en sections of rat spinal cord. The DRGs were seeded and the cells were incubated for 2~ hours as done previouRly. Dishes containin~ the enzyme tr~ated stripes were incuba~ed simultaneously with control dishes which did WO9l/06303 ~T/~S90/06l89 2 ~
~61-not receive enzyme and neurite outgrowth was compared.
Certain control dishes included prot ase inhibitors (1 mg/ml each of apoprotin, leupeptin and pepstat~n in 10 mM
Tris~acetate buffer, pH 7.2).
Further enzy~e digestion prcducts were prepared in solution starting with chick KS/CS-PG, originally isolated from day 8 chisk cartilage limb ~esenchymal cell cultures (Carrino, A., and Caplan, A. I., 1985, J. Biol.
Chem. 260:122-127), a~ follows: (1) keratanase only to remove keratan sulfate chains ~rom the proteoylycan, (2) D chondroitin A8C lyase only to remove chondroitin chains, and (3) a combination of both keratanase and chondroitin ABC lyase to produce a pure protein core. Immuno dot blots showed that all of the chondroitin chains were removed by this treatment, bu~ approximat~ly 10% of the kera~an t5 sulfate chains remained attached to the protein core after digestion. The protein core + LN dishes served as a control to test for the possible effect of ma~s action.
All were reacted at 37-C for 1 hour. Each of the four reagents, mixed with LN, were a~sayed for their inhibitory effect on E9 chick DRG neurites as done previously~

7.1.6.4. RAT CHONDROSARCOMA CARTILA~E PROTEOGLYCAN
A chondrosarcoma carti:Lage proteoglycan (RCS) (Carrino and Caplan, 1985, J. Bio:L. Chem. 260:122-127) was used in the ba~ic assay. This proteoglyc~n is much like the bovine or chick KS/CS~PG except that it lack~ the RS
chain region and the chondroitin ~ulfate is in the C-4-S
for~ rath~r than th~ C-6-S form found in bovine and chick KS/CS-PG.
~0 7.1.6.5. DE~MATAN SULFATE PROTEOG YCAN ASSAY
In order to f ind out i~ neurit~ inhibition with bovine and chi~k ~S/CS PG was co~mon to other proteoglycans, we used dermatan sulfate proteoglycan (DS- :~
:

' '~ -'~ ~";

,~` . , W091/06303 PCT/US90/06189`
8 ~ ~

PG~, as one of many othPr possible glycosaminoglycans, in the baslc protocol described above. The DS-PG was used as before at the same concentration and mixed with the laminin, as previously done with chick and bovine KS/CS-PG.
Chick (E9) DRG neurons were seeded on the test culture and 6 analyzed.
7~2. RESULTS
Our goal was to develop an assay which could de~onstrate whether keratan sulf~te~chondroitin sulfate proteoglycan (KS/CS-PG), alone or in combination with other molecules present in or around the ~o~f plate, could actively inhibit neurite outgrowth in vit.o. To do this, we used modifications of a cel~ culture tæchnique developed by Lagenaur and Lemmon,. (198~/, Proc. Natl. .~cad. Sci. USA
1~ 84:7753-7757~ which utilizes ni~rocellulose-coated p~tri dishes onto which purified molecules can be applied in specific geometric patterns.

7.2.1. ASSAY FOR INHIBITION OF NEURITE QUTGROWTH
Nitrocellulose is a substrate easily adhered to petri dish plastic, which allows for the noncovalent attachment of proteins for culture studies. For this reason, we employed nitrocellulosla-coated culture dishes as a ubstrate by which to attach such proteins as laminin (LN) and neural cell adhesion molecule (NCAM), which are known to ~acilitate cell att~chment and/or allow for the elongation of neurites (Rutishauser et al., 1978, J. Cell Biol. 79:382-393; Letourneau~ P. C., 1975, Dev. Biol.
44:92^101: Nanthorpe et al., 1983, J. Cell ~iol. 97:1882-30 lR90; Liesi et al., 1984, J. Neurosci. Res. 11:241-251;
Lander et al., 1985, Proc~ Natl. Acad. Sci. USA 82:2183-2187: Cohen et al., 1986, Nature 322:465-467; Mirsky et al., 1986, J. Neurocytol. 15(6):799-815, and present results). We also attached the pro~eoglycans of interest, i.e. keratan sulfate/chondroitin sulfate proteoglycan (KS/CS-PG) or various portions of this macromolecule (i~s protein core, or KS-PG, or CS~PG). In addition, we bound dermatan sul~ate proteoglycan (DS-PG) and a rat chondrosarcoma cartilage proteoglycan (~C5) ~or comparisoa.
5 Whole chick (E9) dorsal root ganglia (DRGs) were seeded onto the culture dish~s which contained vertical stripes of the purified molecules (Fig. 16). The assays wer~
conducted in serum-con~aining media and then repeated in serum-free media for compar1son. The rasults of each assay 0 were unchanged regardless of the presence of serum.
Analysis of neurite outgrowth was conducted by separating our observations of those DRGs which adhered to the LN-coated surfaces away from the PG lanes from those that set partially or fully on the proteoglycan containing stripes.
1~ It was not possible to quantitake exactly the number of inhibited neurites in comparison to those that were not inhibited, i.e. thos~ that crossed the PG lanes, because the variable fascicle thicknesses and the anastomotic nature of the neuritee prevent~d an ~ccurate count. In all cases considered, the result was obvious as to whether the bundl~ of neurites was completely inhibited or not and was considered in a "yes-or-no mannerl' with respect to inhibition. When intermediate patterns resulted, a representative photo was taken of the average case.

7.2.2. THE EFFECT OF G OWTH-PROMOTING ~OLECULES
Test cultures consisting o ~ITC labelled stripes of l-l00 ~g/ml LN alone supported neurite outgrowth from the E9 chick DRGs. Culture dishes with l or l0 ~g/ml 3~ LN strips placed centr~lly with RITC added and with l0 or 100 ~g/ml LN spread over the entire dish stimulated the production of a symmetric halo of neurltes around each DRG
with no dif~erence in growth at the lines between the LN/RITC laid down in stripes and the LN spread over the 3~

:,.. . .

W09~/06303 PCT/US90/061~9 t~

dish to occupy the remaining stripes (Figs. 17A and B).
This control demonstrates that technical idiosyncrasies of the culture set-up or toxicity problems from labelling the stripes with RITC are no~ a factor in ~his paradigm.
Controls with DRGs grown on nitrocellulose alone showed little or no cell attachment or neurite outgrowth.

7.2.3. THE ROLE OF INHIBITORY MOLECULES
When KS/CS-PG labelled with RITC was blotted in stripes onto ~he nitrocellulose-coated culture dish and LN
coated over the entire dish to occupy the remaining lanes, the neurites extending from the DRGs either stopped at the border of the XS/CS-PG lane en masse or turned in one direction or the sther and grew along the border of the lane. Importantly, support cells also moving from the DRGs did not correlate in any way with the front of gr~wing or inhibited axons. Some DRGs had zPro or only a few support cells, some had many and some had only support c~ll outgrowth. In each case of neurite contact with the lanes of KS/CS-PG, with or without support cells, there was ~" complete neurite inhibition at 1 mg/ml o~ the proteoglycan (Fig. 18) (see below for analysis of actual concentration of KS/CS-PG bound to the nitrocellulose). Simultaneous observation of the RITC~labelled KS/CS~PG stripes with fluorescence microscopy and the DRG neurites with phase microscopy allowed us to see that the neurites responded distinctly at th~ proteoglycan stripe border (Fig. 18).
Interestingly, cell bodies of DRG neurons which migrated out o~ the ganglia tended to stop alonq th~ border of the PG lane as did ~any o~ the support cells.
3D A concentration gradient of XS/CS-PG, ranging fxom 0.2 ~g/ml to 1~0 mg/~l wa~ t~t~d within sinqle culture dishes. We observed that numerous neurites crossed the lane if the concentration was low but fewer and fewer neurites crossed at the PG concentration of 0.2 mg/ml (left ~6 :

, W~9l/06303 PCT/U~90/0618g -65- ~7~
sid~ of photo) to 0.4 mg/ml (right side of photo) (compare to Figure 18 whioh shows complete inhibition at 1.0 mg/ml).
The pattern of crossing at the lower concentrations of the PG was intermittent with expanses of complete inhibition in between the crossing points as analyzed from top to bottom along the border of the strip.

7.2~4. MODIFICATIONS OF THE DORSAL ROOT GANGLION ASSAY
7.2.4.1. KS/CS-P~:LN MIXTURES
We wished to further investigate the role that LN was playing in these assay cultures. The KS/CS-PG was bound to the nitrocellulose in the dishes, based on positi~e results of a dot blot immunoassay for keratan sulfate, how~v~r, it was unknown whether all of the stripe region was covered with the proteoglycan, or whether there was space left for the laminin to bind ~either to the nitrocellulose coating or perhaps even to the proteoglycan itself) when the dish was covered sequentially first with the PG then with LN.
To be certain that laminin was directly :
incorporated into the stripe, we mixed laminin with KS/CS-PG b~fore blotting onto ~he nitrocellulos~.
Immunocytochemistry using a polyclonal anti-laminin antibsdy showed that laminin was indeed present within the stripe. Additionally, we determin~d the maximum concentration of LN in combination with the proteoglycan which would still elicit the inhibitory e~ect. We learned that a ratio of 25 ~g/ml laminin ~or less) to 1 mg/ml KS/CS-PG was able to completely inhibit outgrowth (Fig.
20A), whereas, 50 or 100 ~g/ml LN allowed increasing numbers of ~eurites to cro s the XS/CS-PG + LN strip~ (Fig.
20B). As a control, we tested the outgrowth abilities of the lowest test conc~ntration o~ LN, in this ca~e 10 ~g/ml, , ,.. . :~

PCT/US90/061~9 -66- ~ ~ 7 ~
and observed that neurite outgrowth was extensive on thls amount of LN wh~n the KS/CS-PG ~as not present (Fig. 17B).

7.2.4.2. ACTUAL KS/CS-PG CONCENTRATION
Although it was known that the KS/CS-PG was adhered to the nitrocellulo~ from the dot blot immunoassay (data not shown), as well as from the behavior of the DRG
neurites in th~ test dish~s and controls, we were not certain as to the amount of proteoglycan that remained adsorbed to the nitrocellulose-coated petri dish during our inhibition assay. To determine the concentration of KS/CS-PG in the stripe, we prelabelled the sulfur o~ the proteoglycan with the radioisotope 35S. By comparing the counts received from our culture plates with that of pre calculated standards of nitrocellulos~ alone and with buffer only, we determined that there was 0.25-l.0 ~g/ml KS/CS~PG in the stripes following transferrance of the proteoglycan at l mq/ml in ^~olution to the di~h, including washes with media. We repeated this test with KS/CS-PG
mixed with increasing concentrations of LN (KS/CS-PG + LN
mixtures discussed below) from 0-l00 ~g/ml, to determine whether and at what rate the concentration o~ the proteoglycan might decrease as the LN concentration increase~. We found omewhat un~xpectedly that the concen~ration of the proteoglycan remained essentially unchanged regardless of the concerltration of IN used. The binding characteristics o~ LN and KS/CS-PG to nitroc~llulose may be complex and as yet undetermined interactions between tha KS/CS-PG molecules and the LN
molecules and the nitro~xllulo~e ~y be occurring. Perhaps 33 with increasing concentr~ti3ns, the ~a molecules may be packing ~ore and ~ore dens~ly around.th~a KS/CS-PG
molecule~. Also, since the LN in solution is pr~sent in such a low concentration in comparison to the PG, an effect on th~ PG concentration may not be observable until the LN
3~

W091/0~303 PCT/~S90/~189 -67- ~ $ ~ ~
concentration is considerably increased. Whatever the rPason, the result is fortunate, in that, we are able to evaluate n~urite responses to a constant amount of KS/CS-PG
whil~ varying the concentration of hN in our assay.

7.2.4.3. KS~CS-PG:NCAM MIXTURES
In the above assay, LN was used as a stimulatory molecule for adhesion and elongationO However, based on immunostaining by us and-oth~rs, we believe that, although LN in its extracellular form is present adjacent to the roof plate in the lateral walls of ths spinal cord, it may be present only in very low concentrations or only in the cytoplasmic form within the roof plate itself. Therefore, we tested a ~olecular oombination in the stripe assay using KS/CS-PG + polysialylated NCAM, shown previously to be expressed by the roof plate cells (Section 6, ~upra). The in vitro combination of RS/C5-PG + NCAM lanes alternating with LN lane~ approximates the patterning o~ these particular molecules ~ound in vivo (but possibly not the CS). We mixed lO or lOO ~g/~l polysialylated NC~M and 1 ~0 mg/ml KS/CS-PG, bl~tted these strlps to the nitroce~lulose, dried and applied lOO ~g~ml LN eve.nly over the dish. We observed that the RS/CS-PG was sis~niPicantly inhibitory with the addition o~ lO ~g/ml NCA~[ (FI~. 21) and that as much as lOQ ~g/ml NCAM still allowed essentially no crossing of neurites-. Controls ol` lO-lOO ~g/ml NCA~ alone (i.~. no XS~CS-P~) in the strip~s alternated with LN
demon6trated that this conc~ntxatlon of NCAM alone is a conducive sub~trate for DR~ ~eurite outgrowth (Fig. 22.
Analysis of the neurite pattern within and between he 30 lan~s did not rev~al any differences between outgrowth on this form and NCAM ver~us outgrowtil on LN. IntePestingly, we did not find any d1fferences in the neurite patterning at th~ interface wh~re axons elongated from LN to NCAM or vice versa.
3~

: .

WO91/~6303 PCT/US90/06189 7.2.~.4. ENZYME DIGESTION ASSAYS
We analyzed the effect of individual components of the KS/CS-PG molecule. To do this, we digested various portions of the KS/CS-PG molecule using specific enzymes.
An immediate problem was that the standard enzymes for this purpose are insufficient to completely degrade keratan sulfate or chondroitin sulfate chains from the protein core of bovine KS/CS-PG. In chick, however, a substantial digestion can be obtained. We therefore used chick KS/CS-PG, isolated from day 8 chick limb bud cartilage cultures for our digestion assays.
Chick RS/CS-PG was treated with either keratanase, chondroitin ABC lyase or both in solution, prior to blotting on~o the dishes. Once digested, these reagents were mixed with 5-l0 ~g/ml LM. Cul~ures using 1~ these four reagents showed that (l) the undigested chick KS/CS-PG significantly inhibitc neurite outgrowth (Fig.
23A) when used at the same concentration as bovine KS/CS-PG, (2) CS as well as KS chains are necessary for complete inhibition sinca some crossing o~curs with the use of 2~ keratanase (Fig. 23B) or chondroitin ABC lyase alone (Fig.
24), and (3) the protein core o~ this proteoglycan molecule has no inhi~itory effect on chick DRG neurites which is demonstrated by the use of both keratanase and chondroitin ABC lyase together ~Fig. 24B). The fact that the protein core as~y did not r~sult in neurite inhibition argues that a mass action effsct is not a signi~icant ~actor which governs ~eurit~ behavior in this protocol.
In a second type of dig~stion assay, we compared the bahavior of the DRG neurites in a dish in whioh the 3~ stripes contain~d an undigest~d KS/CS PG + LN mixture with one which had keratana6e or chondroitinase treatment, just prior to seeding khe DRGs. Thus, these digestions were done in the culture dish, rathex than in solution prior to blotting. This permutation confirmed the above findings 3~

WO9l/06303 P~T/US90/06189 that a significant amount of outgrowth occurs across the lanes when keratanase or chondroitinase treatment preceded DRG seeding in comparison to controls. This also indicated that laminin was indeed present on the dish, and in concentration high enough to promote axon outgrowth when uninhibi~ed.

7.2.4.5. CHONDROS~RCOMA TUMOR CELL LINE C~RTILAGE
CHONDROITIN SULFATE PROTEOGLYCAN ~CS-PG~
We t~sted a rat chondrosarcoma tumor cell line cartilag~ proteoglycan (RCS) (a type of CS-PG) mixed with 1~ 10 ~g/~l LN, which is structurally lik~ the bovine KS/CS-PG
except for two alteratio~s: (1) it lacks the XS chain region (Hascall, V. C., 1981, in Biology of Carbohydrates, ;~
Vol. 1, V. Ginsburg, ed., John Wiley & Sons, Inc., pp. l-49~ and (2) th~ CS chains exist in the C-4-S ~orm (the l~ bovine and chick KS/CS-PG's discussed above ar~ C-6-S).
Many of the DRG n~urites observed in this modification stopped at th~ CS-PG border, but a considerable portisn of the neurit~s oft~n crossed the strip at a concentration of 1 mgJml (Fig. 25) which, importantly, was sufficient for virtually complete i~hibition with the bovine and chick KS/CS-PG. Thus, this CS-PG ~oleculs is l~ss e~fective than the bovine or chick XS/CS-PG in achieving DRG neurite inhibition. Since partial inhibition was observed, the da~a indicated ~hat no~ only raay the KS chains play a role in complete inhibition, but that CS can also be a major contributor to the r~pulsion of neurites. Consideration mu~t be given here to the fact that neurite oukgrowth in r~pons~ to C-4-S in the RCS and C~6-S of th~ bovin~ and chick KS/~S-P~ cannot be directly compared.

3~ :

. ~ i W~91/06303 PC~/US90/06189 ..L

7.2.4.6. THE ROLE OF OTHER GLYCOSAMINOGLYCAN-CONTAINING
PROTEOGLYCANS ON NEURITE INHIBITIOM
To test whether the inhibitory effects observed were specific to ~he glycosaminoglycans keratan sulfate and chondroitin sulfate, we used a ~ovine glycosaminoqlycan, dermatan sulfate, in the proteoglycan form (DS-PG) as one choice of many other possible glycosaminoglycans. We found that this pro~eoglycan was much less inhibitory to DRG
neurit~s than that observed for bovine or chick KS/CS-PG at the same concentration of l mg/ml, and was much like the response to the rat chondrosarcoma cartilage proteoglycan ~ (RCS) shown in Figure 25.

7.3. DISCUSSION
We hav~ shown that dorsal root ganglia neurites are inhibited in a concentration-dependent manner by keratan sulfate/chondroitin sul~ate proteoglyran (KS/CS~PG) mixed wi h eith~r laminin (LN3 or neural cell adhesion molecule (NCA~), even though the LN or NCAM are pr~sent in concentratlons which, by themselve , normally allow abundant neurite growth. Removal of either KS or CS by glycosaminoglycan-specific enzyme digestions allowed neurites to cross into the P~ lanes to various degrees.
When both glycosaminoglycans were removed from the proteoglycan, leaving only the protein core ~ixed with LN, neurite growth a~ross the lanes wa9 totally unimpeded, as oGCurs in controls containing only LN or NCAM. These results suggest that the XS and CS ~hains possess a neurite inhibitory charactex but the protein cor~ does not. Use of the protein coxe alon~ also serves as a suitable c~ntrol to axgue that neurite inhibition is not simply due to a mass action effect.

WO91/06303 PCT/US90/06~9 7.3Ol. GLIAL CELLS ARE CAPABLE OF EXPR~S~
OR SE~UES~ERING AXON ATTRACTIVE OR
REPULSI~E MOLECULES SIMULTANEOUSLY _ Our previous data (Sectio~ 6) showed that the glial cells o~ the xoof plate express adhesive molecules such as SSEA-l, L2 (HNK-l) and NCAM on their surface.
Since axons do not travel through th~ roof plate, these molecul~s may function to hold the glial cells together or to help teth2r them to their surrounding attachment points at the pial sur~ace or the dorsal portion of the ventricle.
The roof plate al50 expresse~ KS which we have shown in this in vitro study to be inhibitory to DRG neurites. Thus the glial cells of the roof plate in vivo are capable of simultaneously producing or sequestering molecules for cell cell attachment as well as for cell repulsion. This scenario may also occur at other axon refractory sites in th8 CNS 5uch as the chick sub-plate where large extraoellular spaces bordered by glial cell processes and filled with CS proteoglycan-containing extracellular matrices have been described (Palm~rt et al., 1986, Society for Neurosci. Abst. 12:1334).

7.3.2. ALL GLYCOSAMINOGLYCANS ARE NOT FUNCTIONALLY
EQUI~ALENT WITH RE5PECT~TO NEURITE ELONGATION
OU7` results show that inhibitory differences exist between proteoglycans within the same species since bovin~ dermatan sul~ate pr~teoglycan (DS-PG) produces a reduced amount of neurite inhibition in comparison to bovine KS/C5-PG at the same concentration. Although proteoglycan~ aff~ct axons in a variable manner, taken together, our prQ~ent studies using keratan sulfa~e-3~ chondroitin ~ulfate-=, and dermatan sulfate proteo~lycan, combined with the results of other laboratories using other glycosamino~lycans, show that these molecules, in ~eneral, are inhibitory to neurite outgrowth and cell attachment.

3~

$

A character of the glyco~aminoglycan portion of the proteoglycans that could be responsible, in part, for the difference in their biological effect is the varia~ility in the level and pattern of sulfation. The RCS
proteoglycan consists of chondroitin sulfate in the form of C-4-S, as does the DS-PG (the iduronic acid of the DS
chains can epimerize to glucuronic acid to make C-4-S
chains and the DS-PG o ten consists of a large number of these chains (Hascall, V. C., 1981, Biology of Carbohydrates, Vol. 1:1~49; Heinegaard and Paulsson, 1984, Ex~-racellular Matrix Biochemistry, 277-322)). The CS
chain~ o~ bovine and chick KSJCS-PG, however, are in th~
form of C-6-S. Since the RCS proteoglycan and DS-PG
produced similar results with respect to the degree of neurite inhibi~ion and this degree o~ inhibition differed from the chick and bovin~ KS/CS-PG, it is feasible that this slight conPigurational change can bs detected by growth cones, causing them to be inhibited to differant extents.

2~ 7.3.3. ALL GLIAL CHANNEI~ ~RE NOT ALIKE.
The literature shows that in numerous developing axon systems in vert~brates, large matrix-filled extracellular chann~ls surrounded by glial cell processes, like those of th2 roof plate and chick subplate boundary, usually ~oreshadow the route o~ the pioneering axons whose growth cones closely associat~ with ~he glial cell ~mbrane (His, W., 1887, Arch. Anat. Physiol. Leipzig ~nat. Abt.
92:368-378: Silv~r, J. and Robb, R. H., 1979, Dev. Biol.
68-175-190: ~rayaneX, S., 1980, The Anatomical Reoord 197:95-lO9; Silv~r, J., and Sidman, R. L., 1980, J. Comp.
Neurol. 189:101-111; Nordl~nder, R., and Singer, M., 1982, Exp. Neurol. 75:221-228: Nakanishi, S., 1983, Dev. Biol.
95:305~316; Simpson, S., 1983, in Spinal Cord Reconstruction, C. C. K~o et al., eds., Raven Press, New 3~

PfT/US90/061~9 2~7~3~8 York, N.Y., pp. 151-162; Bork et al., 1987, J. Comp.
NeurolO 26~:147 158). Wide bored channels are likely to be formed by the hydratPd glycosaminoglycans bound to ~he proteoglycans of the matrix (Margolis, et al., 1986, Ann.
N.Y. Acad. Sci. ~81:46-54; Rutka, et al., 1988, J.
Neuro~urg. 69:~55-170). ~t i5 intriguing that this same type of pre~ormed glial channel for axons is present in invertebrates as well (Jacobs, J. R., and Goodman, C. S., 1989, J. Neurosci. 9(7):2402-2411), suggesting that the basic molecular mechanism for building axon "highways" has been widely conserved during evolution. Clearly, however, the large ex~.racellular spaces made by the glial cell processes of the rodent roo plate and chick subplate do not ~acilitate axonal elongation. Anatomically, the glial cells of axon pathways and barriers and their associated channels look alike, but functionally, they are clearly quite different. Why do axons grow alon~ the glial walls of som~ channel and not others? ~hy are glycosaminoglycans used at all to create spaces in regions where axons grow if glycosaminoglycans are innately inhibitory to axo~ outgrowth? It seems that th~ a~swer lies not in the physical presence o~ the extracellular space itself, but in the molecular nature of the surface of the cell~ which border the channe:L or the extracellular matrix within the channel.

7~3O4~ PROTEOGLYCANS CAN BE IT~ER INHIBITORY OR
Alt~ough we have shown khat glycosaminoglycans produce varying a~ounts of inhibition, dependent upon concentr~tion and ~ype, we have also learned that glycosaminoglycans can b~ mad~ ~o be growth per~issive, i.e., the inhibitory effect o~ the proteoglycan can be reduced or completely ma~ked if accompanied by an appropriate concentration of the growth-promoting molecule, 3~ laminin. NCAM was far less effective in counteracting the ~ ~ ri ~ S~ ~3 proteoglycan mediated inhibition. These results suggest that a growth cone can sample chemical differences in its environment and make motile "decisions" based on summation of its sampling (see Letourneau, P. C., 1975, Dev. Biol.
44:92-101). It appears then, ~hat growth-promoting molecules can modify the effect of those molecules which normally function ~o inhibit neurite outgrowth and vice versa. Therefore, by varying the ratio of attractive or adhesive molecules to inhibitory molecules on or around glia or modifying their temporal appearancP, a wide range of neurite patterns can be eli.cited. The ran~e extends from complete separation between the glial border and all adjacent axons (e.g. the roo~ plate, chick subplate, and dorsal optic stalk), to partial separation (a pattern of intermittent crossing like that of the right lane of Figure 19, and which occurs in vivo in th~ commissural and spinothalamic axons of the floor plate, or thalamic afferents to the barrel fields of somatosensory cortex) to totally undeflected/rectilinear pattarns of axons (like that of the proximal op~ic nerve).

8. DERMATAN SULFATE AND KERATAN SULFATE/
CHONDROITIN SULFATE PRO~EOGLYCAN INHIBIT NEURITE
OUTGROWTH OF A NEURON-L:tKE CELL LINE IN_VITRO
Dermatan sulfate proteoglycan (DS-PG) and keratan sulfate/chondroitin sulfate proteoglycan (KS/CS-PG) were found to inhibit neurite outgrowth from cells of the neuronal~like cell line, PC-12. The inhibition of neurite out~rowth was demons~rated at a concentration o~ 0.1 mg/ml (1.2S ~M) DS-PG and 0.25 mg/ml (0.31 ~ S/CS-PG.

33 8.1. M~TERIALS ~ND METHODS
8.1.1. SUBSTRATE PREPARATION
Tissue culture Petri dishes t60 mm) were coated with nitrocellulose as described in Section 7.1,1., supra.
Cellulose filter paper (Whatman #l) was cut into 350 ~m W~91/06303 PCT/US90/0618 strips and used to blot various proteoglyoans onto the nitrocellulose substrate. The strips were soaked in 20 ~1 of the desired proteoglycan mixture. A solution of 1 mg/ml laminln (LN) was then spread evenly across the dish with a bent gla55 Pasteur pipet. Quantitation of these procedures, and suita~le controls, are described in detail in Section 7., supra.
Stripes were made on the nitrocellulose-coated culture dishes with mixtures of LN (40 ~g/ml), and KS/CS-PG, or DS-PG at various concentrations.

8.1.2. PC-12 NEURON-LIXE CELL LINE PREPARATIONS
~ he PC-12 cells used for the experiment were grown in media composed of DMEM plus 10% Horse Serum, 5%
F~tal Calf Serum and 30 ~g/ml gentamycin, final concentration. Confluent plates were disaqgregated with 0~25% trypsin.

8.1.3. ASSAY FOR INHIBITION OF NEURI~E OUTGROWTH
Plates us~d for experimental procedures were seeded at approxima~ely one-million cells per 60 mm plate.
Media used in the experiments was supplemented with Nerve Growth Factor (NGF) at a Pinal concentration of 50 ng/ml.
Stripes coated with protcogylcan which were completely inhibitory to neurite outgrowth were evaluated as (-), khose allowing sliyht outgrowth (+/-), and those permissive to neurite outgrowth as (+).

8.2. RESULTS: EFFECT OF RS/CS-PG ON PC-12 CELL NEVRITE _ TGROWTH _ _ PC-12 cells were plated on a substratum 3D containing different concentrations of proteoglycans and grown in the presenc~ of NGF. Neurite outgrowth was evaluated 24, 48 and 96 hours later, and is reported in Table 1.

~ ?~

TAB~E l.

BY KS~CS-PG AND DS PG

KS/CS-PG Concentration 24h 48h 95h 2.7 mg/ml (3.37 ~M) - - -1.0 m~/ml (1.25 ~M) +/- - -0.5 ~g/ml (0.62 ~
0.25 mg/ml (0.3I ~M) ~ +/ _ DS-PG Concentration 24h4~h 96h 0.~ mg/ml (10 ~M) - - -0.4 mg/ml (5 ~) +/-0.2 mg/ml (2.5 ~M) +/- +/-0.1 mg/ml (1.25 ~ /- +/-0.05 mq/ml (0.62 ~M) + + +

ao 8.3. DISC_~SION
These data suggest that both DS and KS/CS-PGs are inhibitory to neurite outgrowth o~ PC-12 c~lls. PC-12 cells displayed a ~reater sensitivity to RS/CS-PG than to DS-PG on a molar basis, although on a per weight basis, the 2~ different protoglycan compositions show comparable inhibition. This latt~r observation corresponds with the co~parativ~ inhibit~ry effects Q~ XS/CS-PG and DS-PG on neuri~e ou~growth o dorsal root ganglia neurons. DS-PG
appeared to be less inhibito~y on a molar basis than KS/CS-PG or DR5 neuron~ (Section 7.2.4.6., supra, Section

9.2, in~raj. Thus, ~he growth cones of PC-12 neuron~like cell5 appear to share the same specificity for pxoteoglycan inhibition that DRG n~uron~ demonstrat~. As discussed in 3~

~:
: , .~ : :

.~, , ::

. ~ : : . ~

U~O ~1/06303 PCI/US90/û6189 Section 7 . 3 . 2 ., supra , growth csnes from both cell types may demonstrate a configurational specificity for C-4-S
over C-6-S.

9. EFFECT OF DS-PG ON DRG NEURITE OUTGROWrH
KS/CS-PC; was shown to inhibit neurite outgrowth of dorsal root ganglia (DRG from chick E6, Section 7.2.1., supra). In this example DS~PG inhibition of neuri'ce outgrowth was assayed.

9.1. MATERIALS AND METHODS
DRGs were prspared as described in Section 7 . 1 . 2 ., supra. DS-PG stripes were prepar2d on laminim-coated nitrocellulose as described in Sections 7.1.1. and 8.1.1., supra.
1~ The assay for the inhibition of outgrowth was performed as described in Section 8.1.3., supra. DRGs (Chick E 6 dorsal root ganglia) w~re plated on a substratum containing different ooncentrations of dermatan sulfate proteoglycan on 100 ~g/ml laminin and grown in the presence f Nerve ~rowth Factor (NGF).
Stripes coated with DS~-PG which were completely inhibitory to neurite outgrowth wiare ~valuated as (~
those allowing slight out~rowth (~/-), and those permissive to n~urite outgrowth as (~).

9.2. RESULTS
E~ dorsal root ganglia (DRG) cells were cultured on nitrocsllulose treated with a dermatan sulfate proteoglycan (DS-PG) strip. The results of the growth 3D assay arP shown in Table 2.

~, ~

$

TABLE 2.
INHIBITION OF DRG OUTGROWTH

DS-PG Conc~ntration Neurite and Assay Conditlons Outgrowth B

0.1 mg/ml DS-P&

0.2 mg/ml DS-PG +/-1~
0~4 mg/ml DS-PG

0.8 mg/ml DS-PG

DRG neurite outgrowth was completely inhibited by as little as 0.4 mg/ml (5 ~M) DS-PG, and partly inhibited by 0.2 mg/ml (2.5 ~) DS-PG.

9.3. DISCU';SION
These results suggest that D~G cells are : .
slightly more sensitive to inhibition by DS-PG than the neuron-like cell line PC-12. DRG cells are partly inhibited ~rom outgrowth by as little as 0.2 mg/ml DS PG, whereas PC-12 oells were partly i.nhibited by 0.4 mg/ml DS-PG.
On a molar basis, th~ DRG cells are more sen3itive to R5/CS-PG than to DS~PG, as reported in Section 7.2.4.6., supra. These results correspond to the results with the PC-12 cell line, which was also more sensitiYe to KS/CS PG than to DS~PG on a molar basis.

3~
:

,,, . , ,:
. :. ., , .,.

~091/06303 PCT/US90/061g9

10. DERMAT~N SULFATE AND KERATAN SULFATE/
CHONDROITIN SULFATE
INHIBIT GLIAL CELL INVASION IN VITRO
Dermatan sulfat~ proteoglycan (DS-PG) and keratan sulfate/chondroitin sulfate proteoylycan (KS/CS-PG) were found to inhibit glial cell and astrocyte invasion.
C-6 rat glial tumor cells and MCG-28 young immortalized mouse astrocytes were unable to invade the proteoglycan coated sub~tratum for up to 96 hours.

10.1. MATERIALS AND METHODS
10.1.1. SUBSTRATE PREPARATION
Tissue culture Petri dishes (60-mm) wer~ coated with nitrocellulose (5chleichar & Schuell, Type BA 85 : 0.5 ml of a 5 cm2 section dissol~ed in 6 ml methanol~.
Cellulose filter paper (Whatman ~1) was cut into 350 ~m strips and us~d to blot various proteoglycans onto the nitrocellulose substrate. The strips were soaked in 20 ~1 -of the desired proteoglycan mixture. A solution of 1 m~/ml laminin (LN) was then spread evenly across the dish with a b~nt glass Pasteur pipet. These methods are also described in Section 7.1.1, supra.
Stripes were made on the nitrocellulose-coated culture dishes with mixtures of LN (40 ~g/ml), and XS/CS-PG
or DS-PG at various concentrations.

10.1O2. PREP~R~TION OF CELL LINES
The different cell linlas used ~or the experiment were grown in media compased of DMEM plus 5% Fetal Bovine Serum, 5% Calf Serum and 30 ~g/ml gentamycin. Confluent plate~ were disaggregated with 0.25% trypsin. Plates used 3~ for ~xperimental procedures were seeded at a ratio o~ 1:6 from confluent plate~. C-6 rat glial tumor cells (Paganetti et al, 1988, J. Cell Biol. 107:2291-2291) and W~91/06303 PCT/US90/06189 ~ 3J

~CG-28 youny immortalized mouse astrocytes (a murine neonatal astrocyte line immortalized with SV-40), were utilized for these experiments.

l0~l.3. ASSAY FOR INHIBITION OF OUTGROWTH
Stripes coated with proteoglycan which were completely lnhibitory to cell invasion were evaluated as (-), those ~llowi~g slight invasion ~ ), and those permissive to invasion as (+) (Figure 26).

l0. 2 . RESUhTS
1~
10.2.l. EFFECT OF DS-PG ON C-6 AND

C-6 glial c@lls were plated on different concentrations of DS-PG. Plates were evaluated after 3, 24 and 48 hours, and 5 days~ The results ars shown in ~able 3.

_~_ _ _ TABLE 3.
INHIB~TION OF C-6 CELL XNVASION ON DS-PG
DS-P~ Concentration 3 hrs 24 hrs 4~ hrs 5 days ~ ..... . .. _ .

0.8 mg/ml (l0 ~M) 2S 0.4 mg/ml (5 ~M) O.2 mg/ml l2.5 ~M) +/- + + +

0~l mg/ml (l.25 ~M) ~ + + +

2~CG 28 immortalized young astrocytes wer~ plated on di~ferant concentrations of DS-PG. Plates were evaluated 24, 48, and 72 hours, and s days later. The results are ~hown in Table 4.

. .
-:
.. .
....

~'O 91/~6303 1Clr/lJ~90/06189 _ _ _ . . . ................ _ _ TABLE 4~
INHIBITION OF MCG-2 8_ CELL INV~SION ON DS-PG

DS PG Concentration 24 hrs 48 hrs 72 hrs 5 days O . 8 mg/ml ( l0 ~M) 0 ~ 4 mg/ml ( 5 ~M) - - - +/

tQ 0 ~ 2 mg/ml (2 . 5 ~M) +/- ~ + +
0.1 mg/ml tl. 25 ~M) + ~ + +
.

1~ At the concentration of between 0.4 and 0.2 mg/ml DS-PG, the cell lines are no longer inhibited and can invade the introcellulo~e strip. DS-PG inhibited outgrowth of both cell lines comparably; no obvious enhancement of inhibition of glial cells (C-6) or astrocytes (MCG-28) was observed.

10 . 2 . 2 . EFFECT OF KS/CS-PG ON C-6 AND
MCG 28 CELL_MIG~TION AND INtlASION
C-6 and MCG-28 cells w~3re grown on various concentrations of RS~CS-PG, and evalua~ed at several ~ime points later. The results are shown in Table 5.

' .:
. . ' , , ,. ., ., : : : ., .: i ,,.,,. .,: :. , WO 91/~6303 PCI/US90/~61~9 TABLE 5.

CELL INVAS I ON BY KS/ CS ~PG

a. C 6 RS/CS-PC; Concentration 24 hrs 48 hrs 5 days 2 . 7 mg/ml ( 3 . 3 7 ~ 2~ ) +/ - +/ - +

l . 0 mg/ml ( l . 2 5 ~M) + ~ +

O.5 mg/ml (O.62 ~M) + +

b. MCG-28 XS/CS-PG Concentration 24 hrs 48 hrs 5 days 2.7 mg~ml (3.37 ~M) +/- +/ +

1.0 mg/ml (1.25 IIM) +/- + +

O . 5 mg/ml ( O . 62 ,uM) +/- + +

0.25 mg/ml (0.31 ~M) + ~ +
_ _ _ _ _ _ 10 . 2, 3 . COMPARISON OF XS/CS-PG AND DS-PG
ON CELL MIGRATION AND INVASION
Th~ di~ferent cell lines wsre grown on either KS/CS-PG or DS-PG coated stripes. Plates were evaluatad for the extent of cell inva~iorl after 24 and 48 hours in cultur2. Th~ result~ for grc~wth on XS/CS-PG are shown in 30 Table 6. The results for growth on DS-PG are shown in Table 7.

3~

WO9l/06303 PCr/US90/0618~
2~8~

_ _ TABLE 6.
COMPARISON OF INHIBITIO~ OF C-6 AND MCG-28 CELL INVASION BY KS~CS-PG

Cell Line KS~CS-PG 1.0 m~ml ~1.25 ~M~
S
24 hrs 48 hrs c-6 ~ +

MCG-28 +/- +
... . . . ... _ _ TABLE 7.

Cell Line 0.8 mq/ml (10 ~ 0.1 mq/ml ~1.25 ~) 24 hrs 48 hrs 24 hrs 48 hrs C-6 - - + +
2~
MCG-28 - - + ~ ~-10.3. DISCUSSION
DS-PG is a more potent inhibitor th~n X5/CS-PG
when compared at equal dry weight concentration. C-6 and MCG-28 cells grown on 1.0 mg/ml (1.25 ~) KS/CS-PG
exhibited invasion by 24 hour~ (Table 6), where~s 3D inhibition of inv~ ion on 0.8 mg/ml (10 ~M) of DS-PG was maint~ined for at least 4 days (Tables 3 and 4). At equal molar concentrations the KS/CS- and DS PG5 were ~ound to inhibit outgrowth comparably. This result contrasts with the observations for dorsal root ganglia neurons (Section ,: :
.

WO91/06303 PCT/VS90/~18 9.2., supra) and the neuron-like cell line PC-12 (Section 8.2., supra), in which KS/CS-PG acted as a more potent inhibitor on a molar basis than DS-PG. On a dry weight basis, KS/CS-PG and DS-PG show comparable inhibition of neurite outgrowth.
These results indicate that glial cells, including astrocytes, have a slightly different specificity for proteoglycan inhibition than neurons. Presumably, neurit~ outgrowth and glial cell migration or invasion will be inhibited to different degrees, possibly depending on t0 the composition of proteoglycan in the local matrix~ Thus, proteoglycans may exert a fine regulatory action on the growth of neurons and non-neuronal cells in vivo, thus spacially regulating cell growth.

Various references are cited herein, the disclosures of which are incorporated by reference herein in their entireties.

:

35 '

Claims (150)

WHAT IS CLAIMED IS:

1. A pharmaceutical composition comprising an effective amount of a molecule, which molecule comprises keratan sulfate disaccharide; and a pharmaceutically acceptable carrier.

2. The pharmaceutical composition of claim 1 in which the molecule comprises keratan sulfate proteoglycan.

3. The pharmaceutical composition of claim 1 in which the molecule comprises keratan sulfate glycosaminoglycan.

4. A pharmaceutical composition comprising an effective amount of a molecule, which molecule comprises chondroitin sulfate disaccharide; and a pharmaceutically acceptable carrier.

5. The pharmaceutical composition of claim 4 in which the molecule comprises chondroitin sulfate proteoglycan.

6. The pharmaceutical composition of claim 4 in which the molecule comprises chondroitin sulfate glycosaminoglycan.

7. A pharmaceutical composition comprising an effective amount of a molecule, which molecule comprises dermatan sulfate disaccharide; and a pharmaceutically acceptable carrier.

8. The pharmaceutical composition of claim 7 in which the molecule comprises dermatan sulfate proteoglycan.

9. The pharmaceutical composition of claim 7 in which the molecule comprises dermatan sulfate glycosaminoglycan.

10. The pharmaceutical composition of claim 7 in which the dermatan sulfate has a C-4 sulfer linkage.

11. The pharmaceutical composition of claim 1, 2 or 3 which further comprises a second molecule comprising a compound selected from the group consisting of chondroitin sulfate disaccharide, chondroitin sulfate glycosaminoglycan, and chondroitin sulfate proteoglycan.

12. The pharmaceutical composition of claim 1, 2 or 3 which further comprises a second molecule comprising a compound selected from the group consisting of dermatan sulfate disaccharide, dermatan sulfate glycosaminoglycan, and dermatan sulfate proteoglycan.

13. The pharmaceutical composition of claim 4 in which the chondroitin sulfate has a C-6 sulfur linkage.

14. The pharmaceutical composition of claim 1, 2 or 3 in which the keratan sulfate is Type I (corneal).

15. The pharmaceutical composition of claim 1, 2 or 3 in which the keratan sulfate is Type II (skeletal).

16. A pharmaceutical composition comprising an effective amount of a molecule which antagonizes or destroys the growth inhibitory function of keratin sulfate disaccharide, keratan sulfate glycosaminoglycan or keratan sulfate proteoglycan; and a pharmaceutically acceptable carrier.

17. A pharmaceutical composition comprising an effective amount of a molecule which antagonizes or destroys the growth inhibitory function of chondroitin sulfate disaccharide, chondroitin sulfate glycosaminoglycan or chondroitin sulfate proteoglycan; and a pharmaceutically acceptable carrier.

18. A pharmaceutical composition comprising an effective amount of a molecule which antagonizes or destroys the growth inhibitory function of dermatan sulfate disaccharide, dermatan sulfate glycosaminoglycan or dermatan sulfate proteoglycan; and a pharmaceutically acceptable carrier.

19. A pharmaceutical composition comprising an effective amount of an antibody to keratan sulfate, or a fragment or derivative thereof containing the binding domain; and a pharmaceutically acceptable carrier.

20. A pharmaceutical composition comprising an effective amount of an antibody to chondroitin sulfate, or a fragment or derivative thereof containing the binding domain; and a pharmaceutically acceptable carrier.

21. A pharmaceutical composition comprising an effective amount of an antibody to dermatan sulfate, or a fragment or derivative thereof containing the binding domain; and a pharmaceutically acceptable carrier.

22. The pharmaceutical composition of claim 19 in which the antibody is a monoclonal antibody.

23. The pharmaceutical composition of claim 20 in which the antibody is a monoclonal antibody.

24. The pharmaceutical composition of claim 21 in which the antibody is a monoclonal antibody.

25. The pharmaceutical composition of claim 22 in which the monoclonal antibody is selected from the group consisting of MZ15, 1/20/5-D-4, 4/8/1-B-4, 4-D-1, and 8-C-2.

26. A pharmaceutical composition comprising an effective amount of an enzyme which degrades keratan sulfate; and a pharmaceutically acceptable carrier.

27. A pharmaceutical composition comprising an effective amount of an enzyme which degrades chondroitin sulfate; and a pharmaceutically acceptable carrier.

28. A pharmaceutical oomposition comprising an effective amount of an enzyme which degrades dermatan sulfate; and a pharmaceutically acceptable carrier.

29. The pharmaceutical composition of claim 26 in which the enzyme is selected from the group consisting of endo-b-galactosidase and keratanase.

30. The pharmaceutical composition of claim 27 in which the enzyme is selected from the group consisting of chondroitinase and chondroitin ABC lyase.

31. The pharmaceutical composition of claim 28 in which the enzyme is selected from the group consisting of chondroitin ABC lyase.

32. The pharmaceutical composition of claim 26 which further comprises an effective amount of an enzymP
which degrades chondroitin sulfate.

33. The pharmaceutical composition of claim 29 which further comprises an effective amount of an enzyme which degrades chondroitin sulfate.

34. The pharmaceutical composition of claim 26 which further comprises an effective amount of an enzyme which degrades dermatan sulfate.

35. The pharmaceutical composition of claim 29 which further comprises an effective amount of an enzyme which degrades dermatan sulfate.

36. The pharmaceutical composition of claim 33 which further comprises an effective amount of an enzyme which degrades dermatan sulfate.

37. A method for treatment of a patient in whom inhibition of nerve growth is desired comprising administering to the patient an effective amount of a molecule comprising keratan sulfate disaccharide.

38. A method for treatment of a patient in whom inhibition of glial cell migration or invasion is desired comprising administering to the patient an effective amount of a molecule comprising keratan sulfate disaccharide.

39. The method according to claim 37 in which the molecule comprises keratan sulfate proteoglycan.

40. The method according to claim 37 in which the molecule comprises keratan sulfate proteoglycan.

41. The method according to claim 37 in which the molecule comprises keratan sulfate glycosaminoglycan.

42. The method according to claim 38 in which the molecule comprises karatan sulfate glycosaminoglycan.

43. A method for treatment of a patient in whom inhibition of nerve growth is desired comprising administering to the patient an effective amount of a molecule comprising chondroitin sulfate disaccharide.

44. A method for treatment of a patient in whom inhibition of glial cell migration or invasion is desired comprising administering to the patient an effective amount of a molecule comprising chondroitin sulfate disaccharide.

45. The method according to claim 43 in which the molecule comprises chondroitin sulfate proteoglycan.

46. The method according to claim 44 in which the molecule comprises chondroitin sulfate proteoglycan.

47. The method according to claim 43 in which the molecule comprises chondroitin sulfate glycosaminoglycan.

48. The method according to claim 44 in which the molecule comprises chondroitin sulfate glycosaminoglycan.

49. The method according to claim 38, 40, 42, 44, 46, or 48 in which the glial cell is an astrocyte.

50. A method for treatment of a patient in whom inhibition of nerve growth is desired, comprising administering to the patient an effective amount of a molecule comprising dermatan sulfate disaccharide.

51. A method for treatment of a patient in whom inhibition of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of a molecule comprising dermatan sulfate disaccharide.

52. The method according to claim 50 in which the molecule comprises dermatan sulfate proteoglycan.

53. The method according to claim 51 in which the molecule comprises dermatan sulfate proteoglycan.

54. The method according to claim 50 in which the molecule comprises dermatan sulfate glycosaminoglycan.

55. The method according to claim 51 in which the molecule comprises dermatan sulfate glycosaminoglycan.

56. The method according to claim 51, 53, or 55 in which the glial cell is an astrocyte.

57. The method according to claim 37, 39, or 41 in which the patient has a glioma.

58. The method according to claim 37, 39, or 41 in which the patent has a tumor of nerve tissue.

59. The method according to claim 58 in which the tumor is a neuroblastoma.

600 The method according to claim 37, 39, or 41 in which the patient has a neuroma.

61. The method according to claim 37, 39, or 41 which further comprises administering to the patient an effective amount of a second molecule comprising a compound selected from the group consisting of chondroitin sulfate disaccharide, chondroitin sulfate glycosaminoglycan, and chondroitin sulfate proteoglycan.

62. The method according to claim 37, 39, or 41 which further comprises administering to the patient an effective amount of a second molecule comprising a compound selected from the group consisting of dermatan sulfate disaccharide, dermatan sulfate glycosaminoglycan, and dermatan sulfate proteoglycan.

63. The method according to claim 61 in which the chondroitin sulfate has a C-6 sulfur linkage.

64. The method according to claim 37, 39, or 41 in which the keratan sulfate is Type I (corneal).

65. The method according to claim 37, 39, 41 in which the keratan sulfate is Type II (skeletal).

66. The method according to claim 50, 52, or 54 in which the patient has a glioma.

67. The method according to claim 50, 52, or 54 in which the patient has a tumor of nerve tissue.

68. The method according to claim 67 in which the tumor is a neuroblastoma.

69. The method according to claim 50, 52, or 54 in which the patient has a neuroma.

70. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the nerve growth inhibitory function of keratan sulfate disaccharide, keratan sulfate glycosaminoglycan or keratan sulfate proteoglycan.

71. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the glial cell migration or invasion inhibitory function of keratan sulfate disaccharide, keratan sulfate glycosaminoglycan or keratan sulfate proteoglycan.

72. The method of claim 71 in which the glial cell is an astrocyte.

73. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the nerve growth inhibitory function of chondroitin sulfate disaccharide, chondroitin sulfate glycosaminoglycan or chondroitin sulfate proteoglycan.

74. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the glial cell migration or invasion inhibitory function of chondroitin sulfate disaccharide, chondroitin sulfate glycosaminoglycan or chondroitin sulfate proteoglycan.

75. The method of claim 74 in which the glial cell is an astrocyte.

76. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroyes the nerve growth inhibitory function of dermatan sulfate disaccharide, dermatan sulfate glycosaminoglycan or dermatan sulfate proteoglycan.

77. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroyes the glial cell migration or invasion inhibitory function of dermatan sulfate disaccharide, dermatan sulfate glycosaminoglycan or dermatan sulfate proteoglycan.

78. The method of claim 77 in which the glial cell is an astrocyte.

79. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of an antibody to keratan sulfate, or a fragment or derivative thereof containing the binding domain.

80. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of an antibody to keratan sulfate, or a fragment or derivative thereof containing the binding domain.

81. The method of claim 80 in which the glial cell is an astrocyte.

82. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of an antibody to chondroitin sulfate, or a fragment or derivative thereof containing the binding domain.

83. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of an antibody to chondroitin sulfate, or a fragment or derivative thereof containing the binding domain.

84. The method of claim 83 in which the glial cell is an astrocyte.

85. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of an antibody to dermatan sulfate, or a fragment or derivative thereof containing the binding domain.

86. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of an antibody to dermatan sulfate, or a fragment or derivative thereof containing the binding domain.

87. The method of claim 86 in which the glial cell is an astrocyte.

88. The method according to claim 79, 80, or 81 in which the antibody is a monoclonal antibody.

89. The method according to claim 88 in which the monoclonal antibody is selected from the group consisting of MZ15, 1/20/5-D-4, 4/8/1-B-4, 4-D-1, and 8-C-2.

90. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of an enzyme which degrades keratan sulfate.

91. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of an enzyme which degrades keratan sulfate.

92. The method of claim 91 in which the glial cell is an astrocyte.

93. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of an enzyme which degrades chondroitin sulfate.

94. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of an enzyme which degrades chondroitin sulfate.

95. The method of claim 94 in which the glial cell is an astrocyte.

96. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of an enzyme which degrades dermatan sulfate.

97. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of an enzyme which degrades dermatan sulfate.

98. The method according to claim 97 in which the glial cell is an astrocyte.

99. The method according to claim 90, 91, or 92 in which the enzyme is selected from the group consisting of endo-b-glactosidase and keratanase.

100. The method according to claim 93, 94, or 95 in which the enzyme is selected from the group consisting of chondroitinase and chondroitin ABC lyase.

101. The method according to claim 96, 97, or 98 in which the enzyme is selected from the group consisting of chondiotin ABC lyase.

102. The method according to claim 90, 91, or 92 which further comprises administering to the patient an effective amount of an enzyme which degrades chondroitin sulfate.

103. The method according to claim 99 which further comprises administering to the patient an effective amount of an enzyme which degrades chondroitin sulfate.

104. The method according to claim 90, 91, or 92 which further comprises administering to the patient an effective amount of an enzyme which degrades dermatan sulfate.

105. The method according to claim 99 which further comprises adminsistering to the patient an effective amount of an enzyme which degrades dermatan sulfate.

106. The method according to claim 102 which further comprises administering to the patient an effective amount of an enzyme which degrades dermatan sulfate.

107. The method according to claim 70, 79, or 90 in which the nerve damage is caused by a disease or disorder selected from the group consisting of trauma, surgery, ischemia, infection, metabolic disease, nutritional deficiency, malignancy, exposure to toxic agents and degenerative disorders of the nervous system.

108. The method according to claim 71, 80 or 91 in which the glial cell damage is caused by a disease or disorder selected from the group consisting of trauma, surgery, ischemia, infection, metabolic disease, nutritional deficiency, malignancy, exposure to toxic agents and degenerative disorders.

109. The method of claim 108 in which the glial cell is an astrocyte.

110. The method according to claim 73, 82, 93 in which the nerve damage is caused by a disease or disorder selected from the group consisting of trauma, surgery, ischemia, infection, metabolic disease, nutritional deficiency, malignancy, exposure to toxic agents and degenerative disorders of the nervous system.

111. The method according to claim 74, 83, or 94 in which the glial cell damage is caused by a disease or disorder selected from the group consisting of trauma, surgery, ischemia, infection, metabolic disease, nutritional deficiency, malignancy, exposure to toxic agents and degenerative disorders.

112. The method of claim 111 in which the glial cell is an astrocyte.

113. The method according to claim 76, 85, or 96 in which the nerve damage is caused by a disease or disorder selected Prom the group consisting of trauma, surgery, ischemia, infection, metabolic disease, nutritional deficency, malignancy, exposure to toxic agents and degenerative disorders of the nervous system.

114. The method according to claim 77, 86, or 97 in which the glial cell damage is caused by a disease or disorder selected from the group consisting of trauma, surgery, ischemia, infection, metabolic disease, nutritional deficency, malignancy, exposure to toxic agents and degenerative disorders.

115. The method of claim 117 in which the glial cell is an astrocyte.

116. The method according to claim 70, 79, or 90 in which the nerve damage is caused by a degenerative disorder of the nervous system selected from the group consisting of Alzheimer's disease, Parkinson's disease, Huntington's chorea, amyotrophic lateral sclerosis, progressive supranuclear palsy, and peripheral neuropathies.

117. The method according to claim 73, 82, or 93 in which the nerve damage is caused by a degenerative disorder of the nervous system selected from the group consisting of Alzheimer's disease, Parkinson's disease, Huntington's chorea, amyotrophic lateral sclerosis, progressive supranuclear palsy, and peripheral neuropathies.

118. The method according to claim 76, 85, or 96 in which the nerve damage is caused by a degenerative disorder of the nervous system selected from the group consisting of Alzheimer's disease, Parkinson's disease, Huntington's chorea, amyotrophic lateral sclerosis, progressive supranuclear palsy, and peripheral neuropathies.

119. The method according to claim 38, 40, or 42 in which the glial cell migration or invasion is caused by a disease or disorder selected from the group consisting of trauma, surgery, viral infection, bacterial infection, metabolic disease, malignancy, exposure to toxic agents, and hyperplastic situations.

120. The method according to claim 44, 46, or 48 in which the glial cell migration or invasion is caused by a disease or disorder selected from the group consisting of trauma, surgery, viral infection, bacterial infection, metabolic disease, malignancy, exposure to toxic agents, and hyperplastic situations.

121. The method according to claim 51, 53, or 55 in which the glial cell migration or invasion is caused by a disease or disorder selected from the group consisting of trauma, surgery, viral infection, bacterial infection, metabolic disease, malignancy, exposure to toxic agents, and hyperplastic situations.

122. A method to protect an organ or tissue from glial cell invasion comprising coating the organ or tissue with an effective amount of a molecule selected from the group consisting of keratan sulfate disaccharide, keratan sulfate proteoglycan, and keratan sulfate glycosaminoglycan.

123. A method to protect an organ or tissue from glial cell invasion comprising coating the organ or tissue with an effective amount of a molecule selected from the group consisting of chondroitin sulfate disaccharide, chondroitin sulfate proteoglycan, and chondroitin sulfate glycosaminoglycan.

124. A method to specifically protect an organ or tissue from glial cell invasion comprising coating the organ or tissue with an effective amount of a molecule from the group consisting of dermatan sulfate disaccharide, dermatan sulfate proteoglycan, and dermatan sulfate glycosaminoglycan.

125. The method according to claim 122, 123, or 124 in which the organ or tissue is selected from the group consisting of dorsal root ganglia, optic nerve, and optic chiasma.

126. The method according to claim 122, 123, or 124 in which the glial cell is an astrocyte.

127. The method according to claim 125 in which the glial cell is an astrocyte.

128. A pharmaceutical composition comprising an effective amount of a molecule, which molecule comprises heparan sulfate disaccharide; and a pharmaceutically acceptable carrier.

129. The pharmaceutical composition of claim 128 in which the molecule comprises heparan sulfate proteoglycan,

130. The pharmaceutical composition of claim 128 in which the molecule comprises heparan sulfate glycosaminoglycan.

131. A pharmaceutical composition comprising an effective amount of a molecule, which molecule comprises heparin disaccharide; and a pharmaceutically acceptable carrier.

132. The pharmaceutical composition of claim 131 in which the molecule comprises heparin proteoglycan.

133. The pharmaceutical composition of claim 131 in which the molecule comprises heparin glycosaminoglycan.

134. A pharmaceutical composition comprising an effective amount of a molecule, which molecule comprises hyaluronate disaccharide; and a pharmaceutically acceptable carrier.

135. The pharmaceutical composition of claim 126 in which the molecule comprises hyaluronate glycosaminoglycan.

136. A pharmaceutical composition comprising an effective amount of a molelcule which antagonizes or destroys the growth inhibitory function of heparan sulfate disaccharide, heparan sulfate glycosaminoglycan, or heparan sulfate proteoglycan.

137. A pharmaceutical composition comprising an effective amount of a molelcule which antagonizes or destroys the growth inhibitory function of heparin disaccharide, heparin glycosaminoglycan, or heparin proteoglycan.

138. A pharmaceutical composition comprising an effective amount of a molelcule which antagonizes or destroys the growth inhibitory function of hyaluronate disaccharide, or hyaluronate glycosaminoglycan.

139. A method for treatment of a patient in whom inhibition of nerve growth is desired comprising administering to the patient an effective amount of a molecule from the group consisting of heparan sulfate disaccharide, heparan sulfate proteoglycan, or heparan sulfate glycosaminoglycan.

140. A method for treatment of a patient in whom inhibition of nerve growth is desired comprising administering to the patient an effective amount of a molecule from the group consisting of heparin disaccharide, heparin proteoglycan, or heparin glycosaminoglycan.

141. A method for treatment of a patient in whom inhibition of nerve growth is desired comprising administering to the patient an effective amount of a molecule from the group consisting of hyaluronate disaccharide, or hyaluronate glycosaminoglycan.

142. A method for treatment of a patient in whom inhibition of glial cell migration or invasion is desired comprising administering to the patient an effective amount of a molecule from the group consisting of heparan sulfate disaccharide, heparan sulfate proteoglycan, or heparan sulfate glycosaminoglycan.

143. A method for treatment of a patient in whom inhibition of glial cell migration or invasion is desired comprising administering to the patient an effective amount of a molecule from the group consisting of heparin disaccharide, heparin proteoglycan, or heparin glycosaminoglycan.

144. A method for treatment of a patient in whom inhibition of glial cell migration or invasion is desired comprising administering to the patient an effective amount of a molecule from the group consisting of hyaluronate disaccharide, or hyaluronate glycosaminoglycan.

145. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the nerve growth inhibitory function of heparan sulfate disaccharide, heparan sulfate proteoglycan, or heparan sulfate glycosaminoglycan.

146. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the nerve growth inhibitory function of heparin disaccharide, heparin proteoglycan, or heparin glycosaminoglycan.

147. A method for treatment of a patient with nerve damage or in whom promotion of nerve growth is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the nerve growth inhibitory function of hyaluronate disaccharide, or hyaluronate glycosaminoglycan.

148. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the glial cell migration or invasion inhibitory function of heparan sulfate disaccharide, heparan sulfate proteoglycan, or heparan sulfate glycosaminoglycan.

149. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the glial cell migration or invasion inhibitory function of heparin disaccharide, heparin proteoglycan, or heparin glycosaminoglycan.

150. A method for treatment of a patient with glial cell damage or in whom promotion of glial cell migration or invasion is desired, comprising administering to the patient an effective amount of a molecule which antagonizes or destroys the glial cell migration or invasion inhibitory function of hyaluronate disaccharide, or hyaluronate glycosaminoglycan.