AU6734190A

AU6734190A - Inhibitor of laminin neurite promoting activity

Info

Publication number: AU6734190A
Application number: AU67341/90A
Authority: AU
Inventors: Eva Engvall; Marston Manthorpe; David Muir; Silvio Varon
Original assignee: University of California; Sanford Burnham Prebys Medical Discovery Institute; La Jolla Cancer Research Foundation
Current assignee: University of California; Sanford Burnham Prebys Medical Discovery Institute
Priority date: 1989-10-31
Filing date: 1990-10-29
Publication date: 1991-05-31
Also published as: JPH05504557A; WO1991006573A1; EP0498856A1; CA2066658A1

Description

1 INHIBITOR OF LAMININ NEURITE PROMOTING ACTIVITY

This work was supported by National Science Foundation grant BNS 86-17034 and National Institutes of Health grants NS 17034, NS 25011, and DK 30051. The U.S. Government may have rights in the invention.

BACKGROUND OF INVENTION

Components of the basal lamina promote neuritic outgrowth from several types of embryonic neurons in vitro. Several extracellular matrix macromolecules including fibronectin, laminin, collagens type I and IV, and a heparan sulfate proteoglycan have been reported to possess neurite-promoting activity. In vitro, cells release into their culture media neurite-promoting factors that express their activity when bound to polycationic substrata. Of these, laminin appears to be the most potent factor and it is often found complexed with other extracellular matrix constituents such as proteoglycans and entactin/nidogen. These associated molecules affect the behavior of laminin during purification and may alter its character and interaction with extracellular matrix and cell surface components. Also, preparations of laminin have been described that have different molecular compositions, immunological properties, heparin binding properties, and proteoglycan affinities.

The significance of laminin's association with entactin and proteoglycans remains unclear. Epidermal cell attachment to laminin was reportedly significantly improved due to entactin in a laminin-entactin matrix. Proteoglycans and their component glycosa inoglycan chains have generally not been found to promote neurite outgrowth, and may even inhibit outgrowth on other substrates. Also, inhibition of cell attachment to extracellular matrix proteins by extracellular proteoglycans has been demonstrated.

Analysis of conditioned medium from rat RN22 Schwannoma cells shows that low buoyant density proteoglycans and entactin copurified with laminin neurite- promoting activity. Laminin has been shown to be essential for the neurite-promoting activity of these complexes.

While laminin's neurite-promoting activity is well documented, there have been no reports of inhibitors of this activity. Such an inhibitor would be useful to study laminin activity as well as a means to control neurite growth. The present invention satisfies this need by providing an inhibitor of the neurite promoting activity of laminin.

SUMMARY OF THE INVENTION

The invention provides a purified proteoglycan- associated inhibitor of the neurite-promoting activity of laminin and a purified compound which:

a) inhibits the neurite-promoting activity of purified rat, mouse, and human laminin;

b) is active whether presented to laminin in solution or after either the inhibitor or laminin is first bound to the culture substratum;

c) does not act by displacing laminin from the substratum;

d) can be prevented from binding to neurite- promoting laminin substrates by polyclonal and some monoclonal anti-laminin or polyclonal anti-entactin antibodies if the laminin contains associated entactin;

e) can be purified from rat RN22 Schwannoma cells; and

f) neurite-promoting activity can be abolished by proteases or glycosaminoglycan lyases but not by heat at 90° for 15 minutes.

BRIEF DESCRIPTION OF THE FIGURES

Figure 1 shows DEAE ion-exchange chromatography of ³⁵Sθ -labeled RN22 Schwannoma conditioned medium. Eight liters of unlabeled plus 80 ml of S-labeled conditioned media were applied to a 100-ml DEAE column from which bound material was eluted by a 600-ml 0.15-0.6 M-linear NaCl gradient (represented by a line connecting the onset and completion of the elution) . Fractions (60 X 10 ml) were assayed for total protein ( ■ ) , radioactivity (S ) , and laminin immunoactivity in an ELISA using polyclonal rat laminin antiserum ( • ) . The late-running peak of laminin immunoactivity (fractions 30-42, indicated by bar) which overlapped with the SO^ peak was collected and pooled for further fractionation.

Figure 2 shows further fractionation of the laminin and ³⁵S-labeled DEAE fraction by a second DEAE ion-exchange step. Fractions 30-42 from the first DEAE step (see Figure 1) were diluted to 0.15 M NaCl and applied to a 6-ml DEAE column from which bound material was eluted in 1-ml fractions by a 60-ml 0.15-0.6 M linear NaCl gradient (represented by a line connecting the onset and completion of the elution) . (A) fractions were assayed for radioactivity (Q ) and laminin immunoactivity in an ELISA using a polyclonal antiserum (• ) and monoclonal antibody No. 2E8 (Δ ) • Bars indicate POOL I (fractions 4-12) and

POOL II (fractions 20-32) . (B) Western blots from reducing

SDS gel profiles of every fifth fraction (20 μl) were immunostained for laminin with monoclonal No. 2E8. Arrow indicated a relative molecular mass of 200 kD. (C) The western blot shown in B was exposed to x-ray film for 3 d and developed. The arrows indicate molecular mass markers: rat yolk sac laminin subunits (400 and 200 kD) and entactin

(150 kD) . (D) An aliquot (20 μl) of each of the fractions from the second DEAE column elution was examined for neurite-promoting activity using embryonic day 8 chick ciliary ganglion (CG₈) test neurons and scored as containing

(+) or not containing (-) such activity.

Figure 3 shows a bioassay of substratum-bound neurite promoting and neurite-inhibiting activities derived from Schwannoma-conditioned medium. Purified CG₈ neurons were cultured on polyornithine-coated tissue culture wells pretreated with samples from DEAE ion-exchange (A) POOL I, (B) POOL II, and (C) POOL I and POOL II combined, each diluted 1:10 in a volume of 50 μl. Phase-contrast photomicrographs were taken after 4 h in culture. (D) A maximal neurite-promoting response by CG₈ neurons was achieved by incubating the polyornithine-wells with 50 μl of 500 ng/ml (25 ng/well) of purified rat laminin (Q). A titration curve for the inhibitory activity of POOL II sample ( • ) indicated a titer of 65 neurite inhibitory units (NIU)/ml. CG₈ neurons were seeded and maintained for 4 h before assessing the percentage of cells bearing neurites greater than four cell diameters. Assays counting 50-100 neurons per well were performed in duplicate and the data points represent the mean from four experiments. Standard deviations were <5%. Bars, 50 μm.

Figure 4 shows isopycnic centrifugation of the neurite inhibitor. For buoyant density analysis, POOL II (from the DEAE fractionation step shown in Figure 3) was made 0.4 M in guanidine-HCl, adjusted to 1.35 g/ml with crystalline CsCl, and then centrifuged. Gradient fractions were assayed for ³⁵S-radioactivity (EJ)_# laminin immunoactivity in an ELISA using polyclonal antibodies (# ) , and for the presence of neurite-promoting activity (POOL B) , and inhibitory activity (POOL A) (top horizontal bars) .

Figure 5 shows SDS-PAGE of CsCl inhibitory fraction (POOL A) . (Lane 1) Aurodye protein staining of the inhibitory material electrophoresed on 4-12% acrylamide gels under reducing conditions and then transferred to nitrocellulose. The inhibitory material was examined by autoradiography before (lane 2) and after digestion with heparitinase (lane 3) or chondroitinase ABC (lane 4) . The arrows indicate molecular mass markers: rat yolk sac laminin (400 and 200 kD) , entactin (150 kD) and phosphorylase b (97 kD) .

Figure 6 shows the inhibition of substratum-bound laminin neurite-promoting activity and recovery of activity after inactivation of inhibitor with heparitinase. (1) Polyornithine wells were coated with laminin (50 ng/well) which resulted in maximal neuritic outgrowth by CG₈ test neurons. (2) Treatment of the laminin substratum with heparitinase (heparin lyase II) did not alter the activity of laminin. (3) The neurite-promoting activity of laminin was inhibited by subsequent treatment with the Schwannoma- derived inhibitor (16 NlU/well) . (4) The neurite-promoting activity of laminin was recovered when the inhibited laminin substratum was treated with heparitinase (0.1 U/well for 2h) before initiating the neurite outgrowth assay. The percentage of CG₈ test neurons bearing neurites was determined as previously described. Data represent the mean of quadruplicate assays from two separate experiments.

Figure 7 shows that inhibition of a laminin-substratum is localized and persistent. The substratum was prepared by treating a polyornithine-coated tissue culture dish with RN22 laminin (500 ng/ml) . A narrow strip of inhibitor (CsCl POOL A, 1:100) (arrows) was applied by allowing the inhibitor solution to be drawn across the substratum by capillary action under a fine glass fiber. All incubations were performed in water saturated air to prevent drying. The substratum was washed and CG₈ neurons were seeded and maintained under serum-free conditions. Phase-contrast photomicrograph was taken after seven days. Bar 50 = μ .

Figure 8 shows the effect of application sequences on inhibitory titer. Serial dilutions of inhibitor (CsCl, POOL A) and 25 ng of laminin were presented to polyornithine-coated wells in various sequences. Inhibitor was presented (a) after laminin, (b) simultaneously with the laminin, (c) preincubated in tubes before their combined presentation to the wells, or (d) before addition of laminin. For each step the applied samples were allowed to bind the polyornithine wells for 2 h at 25°C, the wells were washed, the CG₈ neurons seeded and after 4 h the percentage of neurons with neurites greater than four cell body diameters in length was determined. 50-100 neurons were scored per well and two wells were counted per experiment; the data•points represent the mean from four such experiments. Standard deviations were <5' .

DETAILED DESCRIPTION OF THE INVENTION

By "proteoglycan-associated inhibitor" is meant any molecule or compound which has substantially the same structure and activity as the inhibitor of neurite promoting activity purified from rat RN22 Schwannoma cells as described herein. This inhibitor is likely a proteoglycan but can be any molecule or compound which copurifies with the proteoglycan so long as the neurite inhibitory activity is maintained. Modifications in the structure and function of the inhibitor are also meant to be included as a proteoglycan-associated inhibitor so long as the essential structure is maintained.

By "purified¹¹ is meant the inhibitor is free of many compounds normally associated with or occurring with the inhibitor in its native environment.

The invention provides a purified proteoglycan- associated inhibitor of the neurite promoting activity of laminin and a purified extract having high affinity for cationic resin, high buoyant density, large heterodisperse appearance on electrophoretic gels, ability to label with inorganic sulfate, sensitivity to trypsin and glycosaminoglycan lyases, heat stability and the ability to inhibit the neurite-promoting activity of alminin.

The invention also provides a purified compound which:

c) does not act by displacing laminin from the substratum;

d) can be prevented from binding to neurite- promoting laminin substrates by polyclonal and certain monoclonal anti-laminin or polyclonal anti-entactin antibodies if the laminin contains associated entactin;

e) can be purified from rat RN22 Schwannoma cells; and f) can be abolished by proteases or glycosaminoglycan lyases but not by heat at 90° for 15 minutes.

A compound which has substantially the same structure and activity as the compound with these characteristics is also provided since the essential structure and function can readily be determined and altered by one skilled in the art to, for example, modify the activity.

The inhibitors provide a method of inhibiting the neurite-promoting activity of laminin. This method comprises contacting laminin with an inhibitory amount of the inhibitor or the extract.

Moreover, this invention specifically provides an inhibitor that can bind to laminin and interfere with its ability to promote neurite outgrowth from cultured neurons. This inhibitor has been isolated from medium conditioned by RN22 Schwannoma cells by ion-exchange chromatography followed by isopycnic centrifugation. The inhibitory material contains proteoglycan based upon its high affinity for cationic resin, high buoyant density, large heterodisperse appearance on SDS gels, ability to become labeled with inorganic radiosulfate, sensitivity to trypsin and certain glycosaminoglycan lyases, and heat stability. The proteoglycan-containing inhibitor obtained from the CsCl density gradient also contained a substantial amount of the sulfated, 150-kD laminin-binding protein, entactin. The fact that the entactin sediments in the high density fraction, which is devoid of laminin, shows that entactin retains an association with the inhibitory proteoglycan- containing material.

In ion-exchange chromatography, much of the total RN22-conditioned medium laminin co-eluted with proteoglycans under high salt from DEAE and was active in promoting neurites (cf. Figure 1) . When this laminin- proteoglycan pool was refractionated on a second DEAE column, only half of the loaded laminin antigen eluted with the proteoglycans and a substantial amount of laminin eluted with less salt independent of proteoglycan (cf. Figure 2) . Only the proteoglycan-associated laminin lacked neurite-promoting activity. Subsequent analysis of this material (POOL II, Figure 2) indicates the presence of an inhibitor of the associated laminin. The inhibitory activity can be destroyed by glycosaminoglycan lyase, which results in the reexpression of laminin neurite-promoting activity. Conversely, the laminin can be selectively denatured and its neurite promoting activity destroyed by heat, which apparently causes release of the inhibitor. Thus, by fractionation or treatment that selectively destroys one activity, neurite-promotion or inhibition can be expressed independently.

Chromatography and enzyme digestions indicate that the proteoglycan-associated inhibitor is degraded and inactivated by heparitinase or chondroitinase (Figure 5 and

Table II) , suggesting the involvement of both heparan and chondroitin sulfate glycosaminoglycans in the inhibitory activity. The inhibitor is heat stable but trypsin sensitive, suggesting an intact protein to be important in the manifestation of the inhibitory activity. Additional evidence for the role of an intact proteoglycan in the inhibitory activity was obtained by growing RN22 cultures in the presence of 4-methylumbelliferyl-β-D-xyloside, a competitive inhibitor of glycosaminoglycan addition to proteoglycan core protein. Conditioned medium from xyloside treated RN22 cells contained only a small fraction of the inhibitory activity of that of untreated cells.

A membrane fraction from RN22 cells contains a proteoglycan with properties very similar to inhibitor isolated from the conditioned medium (i.e., similar fractionation, appearance by autoradiography, and lyase inactivations) . The regulation of proteoglycan synthesis and deposition can dictate whether laminin expresses its inherent neurite-promoting activity. The degradation of proteoglycan by specific lyases can remove the inhibition and restore neurite-promoting activity to a previously suppressed substratum. Thus, laminin neurite-promoting activity can depend as much on a fine tuning of proteoglycan synthesis and degradation as on laminin production and its cell surface or extracellular deposition. This discovery provides a method of enhancing the neurite-promoting activity of laminin which is bound to a neurite-promoting inhibitor comprising disassociating laminin from the inhibitor or binding a laminin neurite- promoting inhibitor with a compound which prevents the binding of the inhibitor to laminin. The compound which prevents binding can comprise a portion of laminin which is reactive with the inhibitor.

Laminin preparations contain variable amounts of entactin, a sulfated protein that binds to a region near the cross region of laminin. Since entactin itself can bind laminin, but also copurifies with the inhibitor even through high ionic strength conditions, entactin may have the ability to interact with both laminin and the inhibitor and thus affect inhibitor binding. Anti-entactin antibodies prevent the inhibitor from blocking laminin neurite-promoting activity. Also, the monoclonal antibody 2E8, deposited with the Hybridoma Bank, is known to bind to the cross region of rat laminin and, without diminishing neurite-promoting activity, will protect the neurite- promoting activity of laminin from the inhibitor. Thus, the ability of the inhibitor to interact with laminin may depend on its association with the cross region of the laminin tertiary structure, a region different from the postulated neurite-promoting domain at the end of the 400- kD long arm. Through interaction with the cross region of the laminin molecule, the inhibitor may influence the distant neurite-promoting site either directly, by additional binding to another portion of the laminin molecule, or indirectly, by inducing laminin conformational changes. The ability of the 2E8 and anti-entactin antibodies to protect the neurite-promoting activity of laminin suggests that entactin or other components may mediate the formation of laminin complexes which modulate the potential neurite-promoting activity of laminin in basement membranes. These discoveries provide a method of preventing the suppression of the neurite-promoting activity of laminin comprising contacting laminin with a reagent which prevents the binding of a neurite-promoting inhibitor to laminin. The reagent can be an antibody or peptides analogous to laminin or entactin or the like.

While the following examples utilize Schwannoma cells to isolate the neurite inhibitor, it has been established that conditioned media from rat sciatic nerve Schwann cells also contain activity which inhibits neuritic outgrowth in response to laminin. The Schwann cell inhibitory activity has properties identical to the RN22 inhibitor.

EXAMPLE I

Isolation of ³⁵S-labeled Schwannoma Neurite-promoting and Inhibitory Activities

Growth conditions for the rat RN22 Schwannoma cells and the methods for labeling with ³⁵S0₄ have been described within the art. See for example, Davis, G.E. et al., Neurochem. Res. 12:909-921 (1987) for typical growth conditions. Eight liters of RN22 conditioned medium containing 80 ml ³⁵S-labeled RN22 medium was first applied to a 100 ml bed volume DEAE (DE52; Whatman, Inc., Clifton, NJ) ion-exchange column (2.5 cm-diam) and eluted with a 600 ml linear salt gradient from 0.15-0.6 M NaCl in 10 mM phosphate buffer, pH 8.0. Laminin immunoreactivity began to elute at 0.3 M NaCl in the gradient and overlapped with the elution of radiosulfated material. Fractions eluted between 300 and 420 ml of the gradient were pooled and diluted to 0.15 M NaCl with phosphate buffer (10 mM, pH 8.0). This diluted pool was applied to a 10 ml bed volume DEAE column and bound material was eluted using a 60 ml NaCl gradient (0.15-0.6 M) . The laminin immunoreactivity which eluted at 0.2-0.3 M NaCl was collected and pooled (POOL I) . Also, the inhibitory activity began to elute at 0.3 M NaCl and was also collected and pooled (POOL II). Pool II was concentrated by ultrafiltration using an XM100 membrane (Amicon Corp., Danvers, MA) and dialyzed against 0.4 M guanidine-HCl in 50 mM Tris-HCl, pH 7.5. Solid CsCl was added to give density of 1.35 g/ml and sample were subjected to isopycnic centrifugation at 180,000 g for 48 h. High density fractions (>1.5 g/ml) were pooled (Pool A) and dialyzed against PBS, pH 7.2. Laminin antigen was measured by ELISA as described by Engvall, E., Methods Enzymol. 70:419-439 (1980). Polystyrene, 96-well plates were used for coating substrates and the polyclonal and monoclonal antibodies were those described by Davis, G.E. et al., J. Neurosci. 5:2662-2671 (1985). Fractions were assayed for neurite-promoting activity using embryonic day 8 ciliary ganglion neurons. Nurite-promoting assays are well established within the art such as, for example, those described by Manthorpe et al., J. Neurochem. 37:759-767 (1981) and Manthorpe et al., J. Cell Biol. 97:1882-1890 (1983) . Radioactivity was monitored by liquid scintillation. Protein was measured by the method of Lowry et al., J. Biol. Chem. 193:265-275 (1951), using BSA as a standard. Glycosaminoglycan content was measured by Alcian blue binding using heparan and chondroitin sulfates (Sigma Chemical Co., St. Louis, MO) as standards according to Bartold, P.M. and Page, R.C., Anal. Biochem. 150:320-324 (1985) . To address whether proteoglycan is involved in the inhibitory activity, the RN22 cells were cultured in the presence of an inhibitor of proteoglycan assembly, 4-methyl umbelliferyl-β-D-xyloside. Conditioned medium from RN22 cultures grown in the presence of 1 mM β-D-xyloside was collected, heat-treated, and assayed for activity.

Gel Electrophoresis, Immunoblotting, and Autoradiography

SDS-PAGE (4-12%) was run under reducing conditions using the buffer conditions of Laemmli, U.K., Nature (Lond.) 227:680-685 (1970). Unstained gels were electroblotted to nitro-cellulose sheets by the method of Towbin et al., Proc. Natl. Acad. Sci. USA 76:4350-4354 (1979) , and the blots were immunostained for laminin as described by Davis et al, (1985) supra. or stained by colloidal gold using Aurodye (Janssen Life Science Products, Piscataway, NJ) as recommended by the manufacturer. Autoradiography was performed on the immunoblots by exposing them to XAR-5 x-ray film (Eastman Kodak Co., Rochester, NY) for 3 days at ^"70°C.

Enzyme Digestion

Some test samples were treated with proteoglycan lyases or proteolytic enzyme before assay for activity or analysis by SDS-PAGE. For analysis of proteoglycans, samples were incubated for 3 h at 37°C with heparitinase (Heparin lyase II, Sigma Chemical Co.) at 1 U/ml in 50 mM Tris-HCl, pH 7.5, containing 25 mM sodium acetate and 5 mM calcium acetate and/or chondroitinase ABC (Miles Laboratories, Inc., Naperville, IL) at 0.2 U/ml, under the same conditions except that calcium acetate was excluded. Digestions were terminated by boiling. Protease digestion was performed using 100 μg/ml trypsin (Irvine Scientific, Santa Ana, CA) in PBS, followed by the addition of 200 μg/ml trypsin inhibitor (Sigma Chemical Co.) Induction and Inhibition of Neurite Outgrowth

Neurite promotion and inhibition were assayed as substratum-bound activities on polyornithine-coated tissue culture wells (Manthorpe et al., (1983) supra) . For neurite promotion assays, polyornithine wells were treated for 2 h with purified rat yolk sac laminin purified according to Engvall et al., Arch. Biochem. Biophys. 222:649-656 (1983), or rat RN22 Schwannoma cell conditioned medium fractions in PBS (50 μl) and then washed with sterile PBS. Embryonic day 8 chick ciliary ganglion neurons were seeded (10³ neurons per well) in serum-free Nl medium (Bottenstein and Sato, Proc. Natl. Acad. Sci. USA 76:514-5517 (1979), containing 1% albumin and ciliary neuronotrophic factor (40 Trophic units/ml) (Barbin et al., J. Neurochem. 43:1468-1478 (1984)), and incubated in 5% C0₂ in humidified air. Neurite outgrowth was scored by phase- contrast microscopy after 4 h by counting the percentage of neurons bearing processes greater than four cell body diameters. When serial dilutions of purified rat laminin were presented to polyornithine-coated tissue culture wells and ciliary neurons cultured on the resulting substratum, the maximal neurite outgrowth occurred when approximately 25 ng laminin antigen was applied (i.e., 500 ng/ml X 50 μl/well) . Over 95% of the laminin antigen had adsorbed to the polyornithine-coated wells determined by ELISA of the unadsorbed solution.

Neurite-inhibitory activity was assayed by mixing rat laminin and samples to be tested for inhibitory activity (50 μl total volume) and incubating the mixture in polyornithine wells for 2 h at room temperature, followed by three substrate washes with sterile PBS. In some experiments, samples were added as follows: (a) laminin was first bound to polyornithine wells for 2 h followed by washing (the wells were blocked with 1% serum albumin solution in some cases) and then inhibitory sample was added; (b) the inhibitory sample was first incubated in polyornithine wells for 2 h followed by washes and addition of laminin; (c) the inhibitory sample and laminin were preincubated for 1 h in polypropylene tubes before being applied to and incubated in polyornithine wells for 2 h. Ciliary test neurons were seeded, cultured, and scored as described above. We define the half maximal inhibition as 1 neurite inhibitory unit (NIU) and the titer in NlU/ml as that dilution of test sample (total volume = 50 μl) added to the well during substratum treatment eliciting 1 NIU.

Fractionation of RN22-derived Neurite-promoting and Inhibitory Activities

Fractionation of RN22-conditioned medium on DEAE cellulose is shown in Figure 1. Over 95% of the recovered protein eluted as one early peak before 0.25 M NaCl while two peaks of laminin antigen eluted with approximately 0.3 and 0.4 M salt. The first and second laminin peaks represented about one-third and two-thirds, respectively, of the total recovered laminin antigen measured by an ELISA using rabbit polyclonal anti-rat laminin antiserum. Over 95% of the recovered radioactivity eluted in one peak immediately after, and overlapping with, a second laminin peak. Essentially these same elution profiles were obtained using the previous smaller scale fractionation.

The previous study (Davis et al., (1987) supra) indicated that the overlapping laminin and ³⁵S0₄ peaks from the DEAE column contained laminin-proteoglycan complexes possessing neurite-promoting activity. In an attempt to resolve free laminin, free proteoglycan, and laminin- proteoglycan complexes, the overlapping laminin and ³⁵S0₄ fractions were pooled as indicated in Figure 1, diluted and fractionated on a second, smaller DEAE column (Figure 2) . The radioactivity eluted in a slightly broader and earlier peak with 0.3-0.5 M NaCl (Figure 2A) as compared with the first column (cf. Figure 1) . The laminin antigen (measured by ELISA using polyclonal anti-rat laminin) eluted as two broad and slightly overlapping peaks with approximately 0.2-0.3 M and 0.3-0.4 M NaCl and these peaks contained about one-and two-thirds, respectively, of the recovered antigen. Thus, the single laminin ³⁵S0₄ pool from the first column resolved in the second column into two laminin peaks, one of which was associated with nearly all the recovered ³⁵S-radioactivity. The elution of two laminin peaks from the second DEAE column occurred with a little less salt concentration as that of the laminin from the first column. This result suggests that the laminin-³⁵S0₄ peak from the first column contained laminin-proteoglycan complexes from which the laminin component can be released with additional fractionation.

When these same fractions from the second DEAE column were examined for laminin antigen using a mouse monoclonal antibody, termed 2E8, the majority of immunoactivity was found in one peak which eluted before the peak of radioactivity. In fractions containing most of the ³⁵S- radiosulfated material, laminin antigen could only be detected by the polyclonal antibodies. Since the 2E8 antibody is known to bind to an epitope located in the cross region of laminin on the 200 kD, B subunit (Engvall et al., J. Cell Biol. 103:2457-2465 (1986)), these results indicate that in the radioactive peak fractions the 2E8 epitope of the laminin B chain is either missing or masked. However, 2E8 immunoactivity was demonstrated on western blots of the same column fractions (Figure 2B) showing that the 2E8 epitope was present in fractions containing the peak of ³⁵S-radioactivity but apparently was prevented from interacting with the 2E8 antibody in the ELISA.

Autoradiography on the immunoblots revealed two bands of ³⁵S-radioactivity, neither of which comigrated with laminin polypeptides (Figure 2C) . One radioactive band was broad, diffuse, and in a region corresponding to a relative molecular mass of 300-900 kD and is likely to contain radiosulfated proteoglycan. The other band was focussed at 150 kD and comigrated with the sulfated laminin-binding protein, entactin. Immunoblots of these same DEAE fractions using anti-entactin antibodies showed staining of this 150-kD radiosulfated band, thus supporting its identification as entactin.

Each fraction of the second DEAE column was examined and scored as containing (+) or not containing (-) neurite- promoting activity using the CG₈ test neurons. The results are depicted in Figure 2 D. At the concentration tested (20 μl of sample plus 30 μl buffer per well) all fractions expressed neurite-promoting activity except in those associated with the peak of ³⁵S-radioactivity. Thus, the lack of neurite-promoting activity correlated with the presence of proteoglycan and entactin and the absence of monoclonal antibody 2E8 binding. However, significant amounts of laminin antigen are detected by immunoblotting in these same peak radioactive fractions (cf. Figure 2) , suggesting that the neurite-promoting activity, like the 2E8 epitope, was masked by material within the peak ³⁵S- radiolabeled fractions, e.g., by proteoglycan and/or entactin.

EXAMPLE II Inhibition of Laminin Neurite-promoting Activity

Next, we explored more directly whether the peak of ³⁵S-radioactivity in POOL II contained a material that could inhibit the neurite-promoting activity of the laminin in POOL I. Pools I and II were tested for neurite-promoting activity separately and after combination. Results are shown in Figure 3. When POOL II was combined with the neurite-promoting POOL I, this activity was no longer detected (Figure 3 C) . Thus, POOL II contains an activity capable of eliminating the neurite-promoting activity in POOL I. Purified mouse, rat, and human laminin, all previously shown to possess potent neurite-promoting activity, were similarly inhibited by POOL II.

To monitor and characterize the inhibitor in more detail, we used a quantitative bioassay as described in Example I. When serial dilutions of POOl II were incubated in wells treated with 500 ng/ml X 50 μl/well (25 ng/well) of laminin, a concentration-dependent inhibition of the laminin activity was observed. The results are shown in Figure 3D.

EXAMPLE III

Characterization of the Schwannoma-derived

Neurite Inhibitor

The inhibitory material eluted relatively late from the DEAE column (Figure 2) and was associated with ³⁵S0₄- labeled material, suggesting that it may contain proteoglycan. To isolate this component, POOL II was submitted to fractionation by isopycnic centrifugation on a CsCl gradient. The CsCl density gradient fractions were examined for radioactivity, laminin immunoreactivity, neurite-promoting, and inhibitory activities. The results are shown in Figure 4. Most of the radio-sulfated material had a high buoyant density (>1.5 g/ml) although a small proportion of radioactivity appeared at a lower density

(1.35 g/ml). The high density fractions lacked laminin immunoreactivity (using polyclonal antibodies in ELISA) and contained inhibitory activity (Pool A) . Laminin antigen was presented as a distinct peak (1.35 g/ml) which contained neurite-promoting activity (Pool B) . The high ionic strength of this gradient (0.4 M guanidine-HCl and 8 M CsCl) effectively separated neurite-promoting activity from inhibitory activity. Thus, when POOL II samples containing laminin but no neurite-promoting activity were subjected to isopycnic centrifugation, neurite-promoting activity was recovered.

The neurite-promoting and inhibitory activities of fractions obtained during various fractionation steps are shown in Table I. The inhibitory activity recovered in the high density CsCl fractions (POOL A) represented approximately four times (i.e., 3,800 vs. 1,000 NIUs loaded) the activity found in the native second DEAE POOL II starting material, apparently because laminin (with its neurite-promoting activity) was separated away from the higher density inhibitor. This result shows that the interaction between laminin and inhibitor is reversible.

Given that laminin is heat sensitive while proteoglycans are often heat stable, we attempted to destroy selectively the laminin activity by heat treatment and thereby possibly release inhibitor from laminin- inhibitor complexes. The results are shown in Table I. Heat treatment (90°C, 15 min) of conditioned medium from RN22 cells and of the laminin-radiosulfated pool from the first DEAE step resulted in the entire loss of neurite- promoting activity and the appearance of inhibitory activity. Heat treatment of POOL I samples from the second DEAE step eliminated its neurite-promoting activity but generated no inhibitory activity in accordance with the view that this material contains laminin but no inhibitor. When POOL II samples were heated, the titer of the inhibitory activity increased approximately fourfold. Thus, exposure of POOL II to heat (to inactivate laminin neurite-promoting activity) increased its inhibitory activity to about the same extent as does the isopycnic centrifugation (which separated inhibitor from laminin neurite-promoting activity) . The specific activity of the inhibitor from the CsCl gradient pool (Figure 4) was approximately 17,000 inhibitory units/mg protein, representing a 90,000-fold increase over the serum- containing conditioned medium starting material (after heating) .

The inhibitory fractions from the CsCl gradient were pooled (POOL A) and analyzed by SDS-PAGE and electroblotting. The results are shown in Figure 5. Sensitive protein staining revealed a major band with an apparent molecular mass of 150 kD and a few minor bands

(lane 1) . The 150 kD band reacted with anti-entactin antibodies but no laminin immunoreactivity was detected.

Autoradiography of the radiosulfate labeled material showed a broad band, that was most dense in a region corresponding to a relative molecular mass of 300-400 kD, and a narrow 150-kD band (lane 2) . The electrophoretic mobility of the high molecular weight substance was increased after digestion by heparitinase (lane 3) and by chondroitinase ABC (lane 4) . Lyase that digest different glycoaminoglycans had no effect. These results indicate that the components in the inhibitory CsCl sample are heparan/chondroitin sulfate proteoglycan and entactin.

Table I

Neurite-promoting and Inhibiting Activity Levels at Various Fractionation Steps or After Heat Treatment

EXAMPLE IV Characterization of the Inhibitory Activity

The above fractionation studies suggested that the inhibitory activity may involve proteoglycan and entactin. To analyze further the properties of the inhibitory activity, we submitted the CsCl preparation to several treatments followed by bioassay. The results are shown in Table II. The inhibitor appears to be heat stable but trypsin sensitive suggesting that protein is important in expression of the inhibitory activity. The ability of heparitinase or chondroitinase ABC to eliminate the inhibitory activity, suggests the involvement of both heparan and chondroitin sulfate glycosaminoglycans in the inhibitory activity. The products of lyase digestion (i.e., glycosaminoglycan fragments plus core protein) were not active. Also, purified heparin, heparan sulfate, chondroitin sulfate (up to 20 μg/ml) , or the chondroitin sulfate proteoglycan decorin, which inhibits cell attachment to fibronectin (Brennan et al., Cancer Res. 43:4302-4307 (1983)), did not decrease the percentage of neurons bearing neurites in response to a laminin substratum.

Table II

Effects of Selected Treatments on the Inhibitory Activity

Percent of Remaining Treatment Inhibitory Activity*

None 100 ± 2

Heat: 95°C, 15 min 100 ± 2

Trypsin 0 ± 2

Chondroitinase ABC 5 ± 5

Heparitinase 10 ± 5

Inhibitory sample of CsCl preparation (POOL A, Fig. 4) was treated with heat or with either trypsin (100 μg/ml) , chondroitinase ABC (0.2 U/ml), or heparitinase (1 U/ml) for 3 h at 37°C. Inhibitory titer (cf. Fig. 3), was determined after treatments. Data are averages of quadruplicate assays in three separate experiments, presented ± Sd of the means. * Starting activity = 160 NlU/ml.

To address further whether proteoglycan is involved in the inhibitory activity, conditioned medium from RN22 cultures grown in the presence of methyl-umbelliferyl β-D- xyloside was collected, heat treated, and assayed for activity. Compared to heat-treated medium from control cultures, medium from β-D-xyloside-treated cultures (having the same number of cells) contained <20% of the inhibitory activity (1.7 and 0.3 NlU/ml, respectively), suggesting that this activity involved an intact proteoglycan. Although we were unable to assay directly the expression of proteoglycan core protein, assays of these conditioned media for glycosaminoglycan content suggested that xyloside treatment did not alter the synthesis (or release) of this major proteoglycan component. Several experiments were undertaken to test whether the inhibitory activity in our assays was due to displacement of laminin from the polyornithine-coated substratum. In one experiment, an inhibited substratum, created by adding the inhibitor to a defined amount of laminin, was treated with heparitinase and the treated substratum then seeded with CG₈ neurons to test for neurite- promoting activity. The results are shown in Figure 6. The enzyme treatment eliminated the inhibitory activity and restored the neurite-promoting activity of laminin. This shows that the inhibitor does not displace laminin from the substratum but binds to and blocks the substratum-bound laminin. In addition, when a completely inhibited substratum (4 neurite-promoting units [NPU]/well of laminin plus 8 NlU/well inhibitor) was then treated with a high amount of laminin (16 NPU/well) , a neurite-promoting substratum was created (64% ± 14 neurons bearing neurites) . These results suggest that the inhibitor can be saturated with additional laminin and that the presence of inhibitor is not toxic to neurons.

The inactivation of substratum-bound laminin can be localized and will persist for many days in vitro. When a laminin-substratum was treated with localized inhibitor, neurons seeded onto the inhibited area did not extend neurites within 24 h and exhibited nominal, if any, neurite outgrowth even after 5 d (Figure 7) . Neurons that had attached adjacent and immediately proximal to the inhibited area of the laminin-substratum showed extensive outgrowth of neurites which, however, generally did not enter the inhibited region.

To test whether a particular sequence of substratum additions was critical for inhibitory activity, dilutions of the inhibitor were presented to wells before addition of the laminin, simultaneously with the laminin, or after preincubation with laminin in tubes. The results are shown in Figure 8. The inhibitor retained its inhibitory effects regardless of the presentation sequence. However, the inhibitor exhibited the greatest potency (i.e., had the highest titer) when allowed to bind to the substratum before the addition of laminin and when preincubated with the laminin in solution before addition to the polycationic substratum.

EXAMPLE V Antibodies Bound to Laminin Protect Its

Neurite-promoting Activity from Inhibition

The inhibitory pool from the CsCl centrifugation (POOL A, Figure 4) was shown to be free of laminin but yet retained the capacity to interact with laminin and inactivate its neurite-promoting activity. To define further the nature of this association, we examined whether antibodies which bind to laminin would prevent its inactivation by the inhibitor. This approach used purified RN22 laminin and antibodies that bind to laminin but do not block its neurite-promoting activity. The results are shown in Table III. When RN22 laminin was preincubated with polyclonal antibodies raised against rat yolk sac tumor laminin, which .do not block RN22 laminin activity, its neurite-promoting activity was minimally decreased by the inhibitor. Similarly, pretreatment of the RN22 laminin with 2E8 anti-human laminin monoclonal antibody also prevent inhibition. Since the 2E8 epitope on laminin is located at the cross-region of the laminin molecule, this region of laminin might participate in the interaction between the inhibitor and laminin resulting in inhibition.

The partially purified inhibitor contains relatively high levels of the laminin-binding protein, entactin, which copurified with the inhibitory proteoglycan even under conditions which separated inhibitory activity from laminin neurite-promoting activity (i.e., isopycnic centrifugation) . Like the 2E8 antibody, entactin is known to bind to laminin near the cross-region of laminin (Paulsson, M.M. et al., Eur. J. Biochem. 166:11-19 (1987)). Thus, we examined whether anti-entactin antibodies could interfere with the ability of the inhibitor to interact with laminin. The results are presented in Table III. Entactin antibodies, which do not by themselves interfere with the neurite-promoting activity of RN22 laminin (a preparation containing entactin) , did prevent the inhibition from occurring.

In related experiments, anti-entactin antibodies were shown to be immunoreactive with the inhibitory material as well as with various tested laminin preparations. To further examine the role of entactin in the inhibitory activity, the isolated inhibitory fraction (CsCl Pool A, Figure 4) was fractionated by isopycnic centrifugation run under dissociating conditions (in the presence of 2 M guanidine-HCl) . A high buoyant density fraction containing heparan/chondroitin sulfate proteoglycan, but no entactin, was obtained that contained approximately 60% of the original inhibitory activity. However, the laminin preparation used in these assays contained entactin immunoreactivity. All preparations of laminin tested contained substantial entactin immunoreactivity and conditions required to dissociate entactin from laminin compromised the neurite-promoting activity of isolated laminin.

Table III Antibody Interference of the Inhibitors Action on the Neurite-promoting Activity of RN22 Laminin

RN22 laminin (25 ng) was incubated in 6-mm-diam polyornithine-wells for 2 h. The laminin-polyornithine substratum was treated with PBS containing 1% BSA followed by a 30-min incubation with PBS alone or PBS containing 1:50 dilutions of the following antibodies: polyclonal anti-human fibronectin (FN) , monoclonal anti-neurofilament (NF) , polyclonal anti-rat laminin, monoclonal anti-rat laminin No. 2E8, or polyclonal anti-mouse entactin. Without removing the antibodies, inhibitor (160 NlU/ml: CsCl inhibitory fraction) was added for 2 h and then the wells were washed. CG₈ neurons (10³/well) were seeded and the percentage of neurons bearing neurites was determined 4 h later by scoring 50-100 neurons per well. Conditions were replicated eight times in four separate experiments. Standard deviations were <10%.

Although the invention has been described with reference to the presently-preferred embodiments, it should be understood that various modifications can be made without departing from the spirit of the invention. Accordingly, the invention is limited only by the following claims.

Claims

WE CLAIM:

1. A purified proteoglycan-associated inhibitor of the neurite-promoting activity of laminin.

2. A purified extract having high affinity for cationic resin, high buoyant density, large heterodisperse appearance on electrophoretic gels, ability to label with inorganic sulfate, sensitivity to trypsin and glycosaminoglycan lyases, heat stability and the ability to inhibit the neurite-promoting activity of laminin.

3. A purified compound which:

c) does not act by displacing laminin from the substratum;

e) can be purified from rat RN22 Schwannoma cells; and

4. A compound which has substantially the same structure and activity as the compound of claim 3.

5. The inhibitor of claim 1 bound to laminin.

6. The compound of claim 3 bound to laminin.

7. The compound of claim 4 bound to laminin.

8. A method of inhibiting the neurite-promoting activity of laminin comprising contacting laminin with an inhibitory amount of the inhibitor of claim 1.

9. A method of inhibiting the neurite-promoting activity of laminin comprising contacting laminin with an inhibitory amount of the extract of claim 2.

10. A method of inhibiting the neurite-promoting activity of laminin comprising contacting laminin with an inhibitory amount of the compound of claim 3.

11. A method of inhibiting the neurite-promoting activity of laminin comprising contacting laminin with an inhibitory amount of the compound of claim 4.

12. A method of enhancing the neurite-promoting activity of laminin which is bound to a neurite-promoting inhibitor comprising disassociating laminin from the inhibitor.

13. A method of enhancing the neurite-promoting activity of laminin comprising binding a laminin neurite-promoting inhibitor with a compound which prevents the binding of the inhibitor to laminin.

14. The method of claim 13 wherein the compound comprises a portion of laminin which is reactive with the inhibitor.

15. A method of preventing the suppression of the neurite- promoting activity of laminin comprising contacting laminin with a reagent which prevents the binding of a neurite- promoting inhibitor to laminin.

16. The method of claim 15, wherein the reagent is an antibody.

17. The inhibitor of claim 1, wherein the inhibitor is a proteoglycan.

18. The compound of claim 3, wherein the compound is a proteoglycan.