Effects of /i-xylosides on biosynthesis and morphology of PC12 pheochromocytoma cells and primary cultures of rat cerebellum

R. K. MARGOLIS*, B. GOOSSEN, H. TEKOTTE, L. HILGENBERG and R. U. MARGOLIS

Department of Pharmacology, State University of New York, Health Science Center, Brooklyn, New York 11203, and Department of Pharmacology, New York University Medical Center, New York, New York 10016, USA • Author for correspondence at: Department of Pharmacology, Box 29, State University of New York, Health Science Center, 450 Clarkson Avenue, Brooklyn, NY 11203, USA

Summary We have examined the effects of /J-xylosides, which ively). /J-Xyloside inhibition of proteoglycan biosyn- act as exogenous acceptors for thesis was accompanied by significant morphologi- chain initiation, on the morphology and proteogly- cal effects in NGF-treated PC12 cells, consisting of an can biosynthesis of PC 12 pheochromocytoma cells, increase in length and decrease in the branching, and on monolayer, aggregate and explant cultures of diameter and adhesion to the collagen substratum of early postnatal rat cerebellum. PC 12 cells cultured the PC 12 cell processes. p-Nitrophenyl- and for 13 days in the presence of nerve growth factor 4-methylumbeUiferyl-/S-D-xylosides produced similar (NGF) and /J-xyloside, and labeled during days 11-13 effects, which were not seen with p-nitrophenyl-/J-D- with sodium [3SS]sulfate, showed an 8- to 11-fold galactoside. /J-Xylosides also produced distinct alter- increase in [3BS]sulfate-labeled macromolecules re- ations in the adhesion and morphology of monolayer, leased into the culture medium. Most of the increase aggregate, and explant cultures of early postnatal rat was accounted for by , which was cerebellum, which occurred together with inhibition in the form of free glycosaminoglycan chains, which of chondroitin sulfate proteoglycan biosynthesis and were not acid-precipitable. The presence of /J-xylo- a decrease in glycosaminoglycan chain size. These side also led to a 65-115% increase in [36S]sulfate studies indicate that chondroitin sulfate (and prob- incorporation into cell-associated glycosaminogly- ably also ) play a cans and glycoproteins of untreated and NGF- significant role in modulating cell-cell and cell- treated PC 12 cells, respectively. /J-Xyloside treatment matrix interactions in nervous tissue development reduced the size of the chondroitin sulfate chains in and differentiation. both the cells and medium from approximately 34 000 to 10 000 Mr, but had much less effect on heparan sulfate, which decreased in size from 16000 to Key words: proteoglycans, chondroitin sulfate, /S-xylosides, 13 000-14 500 Mr (in the medium and cells, respect- PC12 cells, cell interactions.

Introduction cells with fibronectin (Carbonetto et al. 1983) and the migration of amphibian neural crest cells on fibronectin There has recently been increasing interest in the roles of substrata (Perris and Johansson, 1987), respectively. It proteoglycans in various aspects of the development and has also recently been shown that sulfated glycosamino- function of nervous tissue (for a review, see Margolis and glycans modify growth factor-induced neurite outgrowth Margolis, 1989). In developing brain, there is a progressive in cultured PC12 pheochromocytoma cells (Damon et al. change in the localization of chondroitin sulfate proteogly- 1988). In one of the few studies of this type that did not cans and hyaluronic acid from extracellular to predomi- employ , a chondroitin sulfate proteo- 3 nantly intracellular (cytoplasmic, intra-axonal and nu- glycan with a 280xl0 Afr core protein, which binds to clear) sites in specific types of neuronal and glial cells tenascin (cytotactin), has been implicated in cell—cell (Aquino et al. 1984a,6; Ripellino et al. 1988, 1989). Retina, interactions in developing chicken brain (Hoffman et al. peripheral nerve, and neural tumor cell lines are also 1988). known to synthesize chondroitin sulfate and heparan Since significant levels of free sulfated glycosaminogly- sulfate proteoglycans, which appear to be involved in can chains do not occur in most tissues, we have evaluated processes such as cell interactions, the anchorage and the effects of inhibiting proteoglycan biosynthesis in an localization of cell-surface acetylcholinesterase, and the attempt to obtain additional information concerning the binding of growth factors (Margolis and Margolis, 1989). role of proteoglycans in cell interactions during nervous Most of these studies have concerned basement membrane tissue development. Exogenous ^-xylosides have been heparan sulfate proteoglycans, although it has been widely employed for examining the contribution of reported that chondroitin sulfate, or a chondroitin sulfate proteoglycans to biological processes. Chondroitin sulfate proteoglycan, inhibits the adhesive interaction of nerve and other sulfated glycosaminoglycans are O-glycosidi- Journal of Cell Science 99, 237-246 (1991) Printed in Great Britain © The Company of Biologists Limited 1991 237 cally linked to serine residues in a core protein through a Immunobiologicals, Costa Mesa, CA; 1:300 dilution), rabbit anti- characteristic-glucuronosyl-galactosyl-galactosyl-xylosyl- mouse IgG (Sternberger-Meyer Immunocytochemicals, Jarretts- O-serine linkage region. /J-Xylosides competitively inhibit ville, MD; 1:50), and clonal mouse peroxidase-antiperoxidase the synthesis of glycosaminoglycan chains on xylosylated (Sternberger-Meyer; 1:100). Color development was carried out using diaminobenzidine/hydrogen peroxide. Cultures were coun- core protein, resulting in carbohydrate-deficient proteo- terstained using Mayer's hematoxylin solution (Sigma, St Louis, glycans that are usually accompanied by large amounts of MO) as described by Hockberger et al. (1987). free glycosaminoglycans in which chain initiation occurs on the exogenous /S-xyloside acceptor (Galligani etal. 1975; Analysis of glycosaminoglycans Schwartz, 1979). In the present study we have biochemi- Cells were washed, extracted with chloroform-methanol, and the cally demonstrated a significant inhibition by /J-xylosides lipid-free protein residue was digested with Pronase (after of chondroitin sulfate (and, to a much lesser extent, addition of unlabeled brain 'carrier') as described previously heparan sulfate) proteoglycan biosynthesis in nerve (Margolis et al. 1983). Cell protein was determined by the method growth factor-treated PC 12 pheochromocytoma cells and of Lowry et al. (1951) after dissolving a sample of cell suspension primary cultures of early postnatal rat cerebellum, and in 0.5 M NaOH. To determine the proportion of macromolecular that these biochemical effects are accompanied by distinct radioactivity in the media of xyloside-treated cells that was present as free glycosaminoglycan chains, samples of dialyzed morphological alterations. media were made 10% in trichloroacetic acid to precipitate proteoglycans. Pronase digests of cells and media were desalted by gel Materials and methods filtration on Sephadex G-15, and glycosaminoglycans were precipitated with cetylpyridinium chloride (Margolis et al. 1975). Cell culture The proportion of labeling in chondroitin sulfate and heparan Rat PC12 pheochromocytoma cells were cultured on collagen in sulfate was determined after digestion with chondroitinase ABC Eagle's Basal Diploid Medium with 0.2 nu or no added sodium (ICN ImmunoBiologicals, Costa Mesa, CA), followed by ultrafil- sulfate. Collagen preparation and coating and cell culture tration on a Centricon-10 membrane (Amicon). The molecular conditions were as described by Greene et al. (1987). Cells sizes of heparan sulfate and chondroitin sulfate were determined 1 cultured for 13 days with 2.5 S nerve growth factor (60 ngml" ) by gel filtration on Sepharose CL-6B, using the Kav values were grown in medium containing 1% heat-inactivated (30min, determined by Wasteson (1971). In the case of chondroitin sulfate, 56°C) horse serum, whereas in the absence of NGF, the medium the elution profile was calculated by subtracting heparan sulfate contained 3% horse serum and 1.5% fetal calf serum. /J-Xyloside radioactivity from that of the combined heparan sulfate+chon- treatment was for a total of 13 days in the presence of 1 mM droitin sulfate present in the total glycosaminoglycan fraction 4-methylumbelliferyl-/3-D-xyloside or p-nitrophenyl-/S-D-xyloside. (determined by two separate Sepharose CL-6B elutions under Cultures were labeled for 48 h with sodium [3*S]sulfate identical conditions). (lOO/fCiml"1), and the medium was dialyzed against 50 mM sodium sulfate followed by distilled water. PC 12 cell process Results diameters were measured on X670 photographic enlargements using an optical comparator-magnifier (Edmunds Scientific Co., Barrington, NJ) fitted with a 0.05 mm calibrated reticle. Effects of /3-xylosides on proteoglycan biosynthesis and morphology of PC 12 cells Monolayer and reaggregate cultures of early postnatal rat cerebellum were prepared in 35 mm dishes according to the /J-Xyloside treatment produced an almost 8-fold increase general procedure described by Trenkner and Sidman (1977). in [ S]sulfate-labeled macromolecules released into the Briefly, the meninges were removed from 2- to 3-day postnatal medium by NGF-treated PC12 cells cultured in 0.2 mM cerebella, which were chopped several times with fine scissors and sodium sulfate (11-fold in cells labeled without added immersed in calcium- and magnesium-free buffered Tyrode's sulfate), and chondroitin sulfate accounted for over 80 % of solution. After several washes, single cell suspensions were this material (Table 1). The increased release of [3BS]sul- prepared by incubation in 1% trypsin/0.1% DNAase, residual fate-labeled macromolecules was in the form of free enzyme was removed by further washes in Tyrode's solution, and glycosaminoglycan chains, as demonstrated by the fact the tissue was thoroughly triturated with flame-polished Pasteur pipete of sucessively decreasing bore sizes in Eagle's glucose that all of the radioactivity in the dialyzed medium of medium containing 0.05% DNAase. Remaining tissue clumps xyloside-treated cells remained soluble in 10 % trichloro- were allowed to settle by gravity, and suspended cells in the acetic acid, whereas the [^SJsulfate-labeled macromol- supernatant solution were sedimented by brief centrifugation. ecules (primarily proteoglycans) released by untreated The cells were resuspended in culture medium (BME containing cells were almost quantitatively precipitated under these glucose, glutamine, penicillen/streptomycin, and 10% horse conditions. Although chondroitin sulfate accounted for serum) and diluted for plating at a concentration of lOxlO6 or 6 most of the increased release of glycosaminoglycans in 2 xlO cells/dish for aggregate cultures on plastic and monolayer xyloside-treated cells, there was also an increase in cultures on poly-D-lysine, respectively. heparan sulfate (whose proportion, however, decreased Cerebellar explant cultures were prepared in polylysine-coated considerably in relation to chondroitin sulfate and sul- 96-well plastic tissue culture plates (Falcon) by the general fated glycoproteins; Table 1). method described by Hockberger et al. (1987), with a modification in the procedure used for obtaining the tissue suspension. Tissue In NGF-treated cells cultured in the presence of 0.2 mM was cut coarsely with a scalpel, suspended in buffer, and passed sodium sulfate, /J-xyloside treatment produced a 115% twice through a 380 /

238 R. K. Margolis et al. Table 1. Effects of j3-xylosides on the labeling of glycosaminoglycans and glycoprvteins in the culture medium of NGF-treated PC12 cells Ctsmin"1, xlO^mg-1 Chondroitin Heparan Glycoproteins cell protein sulfate (%)* sulfate (%)• - /J-xyloside 2.8 (4.8)t 68(69) 26(25) 6(6) + 6-xyloside 18.4 (52.8) 81 (73) 12 (19) 7(8) • Corrected for chondroitinase-digestible material that was not precipitated with cetylpyridinium chloride (20 % in the media of /3-xyloside-treated cells, and 6-7 % in untreated cells), but not for chondroitin sulfate disaccharide-size material present in the cetylpyridinium chloride precipitate (6-7 % in /J-xyloside-treated cells, and 0 3 % in untreated cells). t Cells were routinely grown and labeled in medium containing 0 2mM sodium sulfate Values in parentheses are for cells labeled with carrier-free sodium [^Slsulfate in the absence of added sulfate.

Table 2. Effects of (3-xylosides on the cell-associated glycosaminoglycans and glycoproteins of PC 12 cells Ctsmin~l, Chondroitin Heparan Glycoproteins cell protein sulfate (%)• sulfate (%)• (%)* NGF-treated PC12 cells - /3-xyloside 333 (876)* 40 (41) 38(39) 22 (21) + /3-xyloside 713 (1539) 38 (44) 27 (42) 35(15) Untreated PC12 cells - /3-xyloside 223 (570) 46 (46) 29 (28) 25(26) + /3-xyloside 381 (1010) 26 (15) 54(63) 21 (22)

•Cells were routinely grown and labeled in medium containing 0.2nra sodium sulfate. Values in parentheses are for cells labeled with carrier-free sodium [^Isulfate in the absence of added sulfate.

These effects were very similar in sulfate-free medium, much less effect on heparan sulfate, which decreased in which also did not affect the relative labeling of glycos- size from 16000 to 13 000-14 500 Mr, in the medium and aminoglycans and glycoproteins of NGF-treated or cells, respectively (Table 3). However, the cell-associated untreated PC12 cells in the absence of/3-xyloside (Table 2). heparan sulfate eluted from Sepharose CL-6B as a broad However, cells treated with both NGF and /3-xyloside and peak, the larger component of which was significantly grown in sulfate-free medium showed a labeling pattern reduced in /3-xyloside-treated cells (Fig. 2). It is therefore (Table 2) that differed from that of cells grown in possible that although there was relatively little overall physiological levels of sulfate (i.e. a relative decrease in effect on heparan sulfate chain size, a small proportion of glycoprotein labeling and a small increase in the labeling the heparan sulfate may have been significantly affected of glycosaminoglycans). Since cells grown in sulfate-free by the /3-xyloside treatment. Chondroitin sulfate also medium have, in certain cases, been shown to synthesize decreased in size by over 70% in xyloside-treated PC 12 undersulfated glycosaminoglycan chains (Silbert et al. cells that were cultured in the absence of NGF (Table 3). 1986; Tyree et al. 1986; Humphries et al. 1986), it would The decrease in chondroitin sulfate chain size is probaby appear that such effects in PC12 cells may occur only in attributable to limitations in the rate at which the the presence of combined NGF and /3-xyloside treatment. glucuronosyl and/or iV-acetylgalactosaminyl transferases /3-Xyloside treatment reduced the size of the chondroitin are capable of adding disaccharide repeating units to the sulfate chains in both the NGF-treated cells (Fig. 1) and medium from approximately 34 000 to 10 000 Mr, but had Table 3. Effects of P-xylosides on the size of nervous

1 1 1 1 1 tissue glycosaminoglycans v, Chondroitin Heparan sulfate sulfate -0 O NGF-treated PC12 cells' o f \ Cells without /3-xyloside 34 000 16000 X 14 • 6 Medium without /J-xyloside 33000 16000 Cells with ^-xyloside 9000 14 500 o Medium with j3-xyloside 10000 13 000 10 - i - Untreated PC12 cells' c J V Cells without /J-xyloside 29000 15000 b Cells with /3-xyloside 8000 15 000 i 6 - Rat brain

G / 6 \ b Control culture mediumt 33 000 28000 2 /3-xyloside culture medium 16000 13000 Reaggregate cells (untreated) 30000 26 000 s X-O 3-day-old cerebellum^ 28 000 26 000 20 24 28 32 36 40 44 Adult brain 21000 16000 Effluent volume (ml) • Glycosaminoglycan chain sizes were not affected by the presence or Fig. 1. Molecular size of chondroitin sulfate from NGF-treated absence of sulfate in the growth medium. PC12 cells grown in the presence (O O) or absence t Reaggregate cultures of cerebellum from 3-day-old rats. t Chondroitin sulfate and heparan sulfate chain sizes in forebrain (• •) of /3-xyloside, as determined by gel nitration on were both approximately 10% greater than in cerebellum. Sepharose CL-6B (0.9cmx67cm) eluted with 0.2 M NaCl.

Nervous tissue proteoglycans 239 24 28 32 36 40 44 Effluent volume (ml) Fig. 2. Molecular size of heparan sulfate from untreated (A) and NGF-treated (B) PC12 cells grown in the presence (O O) or absence (• •) of /S-xyloside, as determined by gel filtration on Sepharose CL-6B (0.9 cm x 67 cm) eluted with 0.2 M NaCl. large number of newly initiated chondroitin sulfate chains. It is of interest that NGF treatment alone led to a small increase in the size of the PC 12 cell chondroitin sulfate chains (from 29 000 to 34000Mr), since we have previously reported that NGF increases the proportion of 2 - the larger molecular size component of heparan sulfate in PC12 cell membranes (Margolis et al. 1987). 20 26 32 38 44 Xyloside treatment also had an unexpected effect on the Effluent volume (ml) size of sulfated glycopeptides obtained by Pronase diges- tion of PC 12 cell glycoproteins. (These glycopeptides Fig. 3. Molecular size of [^Sjsulfate-labeled glycoprotein- remained in the supernatant after precipitation of glycos- derived glycopeptides from NGF-treated PC 12 cells grown in aminoglycans with cetylpyridinium chloride, and were not the absence (A) or presence (B) of /S-xyloside, and from normal rat brain (C), as determined by gel filtration on Sephadex G-25 affected by treatment with chondroitinase ABC.) In NGF- (0.9 cm x 66 cm) eluted with 0.1M NaCl. CSD and Glc indicate treated PC12 cells, gel nitration on Sephadex G-25 the elution positions of chondroitin sulfate disaccharides and demonstrated the presence of three populations of sulfated glucose, respectively. The three elution curves are normalized glycopeptides; a major peak that elutes in the void volume, to lOOOOctsmin"1 of original sample radioactivity applied to a second peak eluting between chondroitin sulfate di- the column. saccharide and glucose markers, and a small peak eluting just before the total volume of the column (Fig. 3A). In /S- xyloside-treated cells the proportion of the large molecular /3-Xyloside-treated and untreated cultures could be easily size glycopeptides eluting in the void volume is very much distinguished by inspection of dishes (at least 10 of each decreased (Fig. 3B), with almost all of the radioactivity type) coded only by number, and on gentle tapping of the eluting with the second peak (Kav=0.6). For comparison, it dishes it was also observed that in /S-xyloside-treated can be seen that all of the sulfated glycopeptides from rat cultures the PC12 cell processes adhered poorly to the brain glycoproteins elute in the void volume of a Sephadex collagen substratum. For example, processes of untreated G-25 column (Fig. 3C). These results suggest that the cells grown for 9-10 days in the presence of NGF had quite larger glycopeptides may be derived from chondroitin heterogeneous diameters and showed extensive fascicu- sulfate proteoglycans, and that their synthesis is in some lation. Numerous fascicules had diameters of > 1.5/an way linked to the normal addition of chondroitin sulfate (ranging up to 3/mi), whereas in ^3-xyloside-treated cells chains to acceptor sites on the core protein. no fascicule or process was found to exceed 0.75/an in NGF-treated PC 12 cells that were cultured for 1 to 2 diameter. The relative scarcity of processes in /3-xyloside- weeks in the presence of 1 mM /3-xyloside showed a treated cells is also a reflection of their fragility, since they consistent increase in process (or fascicule) length and a were frequently unattached. Xyloside-treated cells often decrease in their diameter and branching (Fig. 4). appeared to grow at a somewhat lower density than

240 R. K. Margolis et al. »\ T»

•*>

B fa*-

Fig. 4. PC12 cells grown for 13 days with nerve growth factor in the absence (A) or presence (B) of 1 mM 4-methylumbelliferyl-/3-D-xylo8ide. Bar, 30/an.

3 untreated cells, although accurate cell counts could not be decreased by over 50 % (from 33 x 10 Mr in control cells to 3 made due to the typical clumping of PC12 cells. However, 15xlO Mr in xyloside-treated cultures; Table 3). Heparan the xyloside-induced morphological effects are not merely sulfate, which represents only a minor component of the a reflection of a lower cell density, since they were not secreted glycosaminoglycans (7 % and 12 % in the medium observed in untreated cells even at densities considerably of /3-xyloside-treated and control cultures, respectively) 3 lower than those seen in any xyloside-treated cultures. also decreased significantly in size, from 28xlO Mr to 3 Culture for similar time periods in the presence of 13 x 10 Mr after /3-xyloside treatment. O.lmgml"1 or O.Smgml"1 chondroitin sulfate had no Whereas control reaggregate cultures contained irregu- effect on the morphology or adhesion of processes in NGF- larly shaped cell aggregates joined by a large number of treated PC12 cells (or on cerebellar explant cultures; see broad 'cables', /3-xyloside treatment led to the formation of below), indicating that the /3-xyloside effects were due to heavy clumps of cells that produced a relatively small alterations in cell surface properties and not merely to the number of thin cables (Fig. 5). Cell migration along these release of free chondroitin sulfate chains. Although there thin cables was observed infrequently, and many of the was a much smaller inhibition of PC 12 cell heparan cell clumps and cables remained unattached to the culture sulfate proteoglycan biosynthesis, addition of heparin to dish. Monolayer cultures (on polylysine) were similarly the culture medium (O.lmgml ) did not produce any affected by /3-xyloside treatment, insofar as the usual /3-xyloside-like effects on PC 12 cell cultures. Moreover, sheets of cells were replaced by small aggregates some- p-nitrophenyl- and 4-methylumbelliferyl-/3-D-xylosides what ressembling control cultures on plastic (data not produced similar effects, which were not seen with shown). In both cases (reaggregate and monolayer cul- />nitrophenyl-/3-D-galactoside. tures) the effect of /3-xyloside treatment was to change the surface properties of cerebellar cells so that they acquired a greater affinity for each other than for the plastic or Effects of fS-xylosides on primary cultures of rat polylysine substratum. This decrease in fasciculation and cerebellum adhesiveness to the substratum is similar to that observed /3-Xyloside treatment of reaggregate cultures of 3-day in the case of PC12 cell processes with respect to collagen. postnatal rat cerebellum produced 3.4-fold and 6-fold In /S-xyloside-treated explant cultures, a dense mass of increases in the release of heparan sulfate and chondroitin cells was surrounded by a tangled network of thick GFAP- sulfate into the culture medium, accompanied by a 24% positive processes on which relatively few cells migrated, and 39% decrease in the amount of cell-associated whereas control cultures had a regular pattern of radially heparan sulfate and chondroitin sulfate, respectively. The oriented processes that surrounded a more spread-out average size of the secreted chondroitin sulfate chains central mass of cells and supported the migration of

Nervous tissue proteoglycans 241 wmm

B ^~

Fig. 5. Aggregate cultures of 3-day postnatal rat cerebellar cells grown for 2 days on plastic in the absence (A) or presence (B) of 1 mM 4-methylumbelliferyl-/J-D-xyloside. Bar, 100 /an. numerous granule cells (Fig. 6). However, in jS-xyloside- terminal morphogenesis and differentiation in the fetal treated cultures the explant itself frequently washed away mouse lung (Smith et al. 1990). Studies employing /S- during the staining procedure, leaving only a halo of xylosides have also specifically implicated a large chon- tangled processes surrounding an empty area where the droitin sulfate proteoglycan in the process of fetal kidney explant was originally located (Fig. 6C). In one exper- branching morphogenesis, while glomerular development iment, 80% of 10 xyloside-treated cultures detached as and heparan sulfate proteoglycan biosynthesis were compared to only 7 % of 29 control cultures, while in a unaffected (Platt et al. 1987; Klein et al. 1989). In other cell second experiment the comparable figures were 61 % of 31 types (Galligani et al. 1975; Stevens and Austen, 1982) and xyloside-treated cultures and none of 30 control cultures. in kidney glomerulus (Kanwar et al. 1984), only chondroi- It would appear that /3-xyloside inhibition of proteoglycan tin sulfate proteoglycan biosynthesis is inhibited by fi- biosynthesis does not prevent the extension of processes by xylosides, whereas they have been shown to initiate the explant cultures, but disrupts their organization, de- synthesis of small amounts of free heparan sulfate chains creases cell migration along these processes (Fig. 7), and in 3T3 cells (Dietrich et al. 1982; Johnston and Keller, favors cell-cell adhesion while greatly decreasing the 1979), hepatocytes (Sudhakaran et al. 1981), mastocytoma adhesion of cerebellar cells to the polylysine substratum. cells (Robinson and Lindahl, 1981), and the basement These results are therefore generally in agreement with membrane producing Englebreth-Holm-Swarm tumor those obtained from monolayer and reaggregate cultures (Ledbetter and Hassell, 1986). In a monoblastic cell line of isolated neurons and glia described above. (U-937) that synthesizes only chondroitin sulfate proteo- glycans, /S-xyloside treatment inhibited both proteoglycan biosynthesis and cell proliferation (Kolset et al. 1990). Discussion Since both chondroitin sulfate and heparan sulfate /3-Xylosides have recently been employed in several proteoglycans appear to contain the same carbohydrate- studies which examined the role of proteoglycans in protein linkage region, the fact that ^-xyloside inhibition differentiation. Inhibition of heparan sulfate proteoglycan of heparan sulfate proteoglycan biosynthesis is observed biosynthesis in cultured rat Schwann cells indicated that rather infrequently, and then usually to only a small proteoglycans play an essential role in basement mem- extent, suggests that this difference is probably due to the brane assembly, but that the integral plasma membrane relative affinities of the chondroitin sulfate and heparan sulfate proteoglycan galactosyl transferases for exogenous proteoglycan is not required for the basement membrane /3-xylosides as compared to their respective xylosylated to exert its effects on Schwann cell differentiation (Carey core proteins, and/or to differences in the sizes of the two et al. 1987). More recently, it has been reported that j8- core protein pools. Our studies indicate that in both PC 12 xyloside inhibition of proteoglycan biosynthesis inhibits 242 R. K. Margolis et al. Fig. 6. Explant cultures of 4-day postnatal rat cerebellum grown for 8 days in the absence (A) or presence (B and C) of 1 mM 4-methylumbelliferyl-/3-D-xyloside, and stained with hematoxylin and antibodies to glial filament acidic protein. In C the explant has washed off during the staining procedure. Bar, 100/on.

Nervous tissue proteoglycans 243 Fig. 7. Higher-power view of processes in control (A) and xyloside-treated (B) cultures similar to those shown in Fig. 6. Bar, 30 ^m. cells and in primary cultures of early postnatal brain, the biosynthesis, and the resultant accumulation of free effect of /S-xylosides is also predominantly on chondroitin glycosaminoglycan chains in these intracellular storage sulfate proteoglycan biosynthesis. However, in both cases organelles. (but to a greater extent in brain) /3-xyloside treatment also It has previously been reported that added glycosamino- produces a significant increase in the synthesis of free glycans or glycosaminoglycan-coated substrata modify the heparan sulfate chains, which have a smaller molecular adhesion and neurite extension of cultured PC 12 cells size than those synthesized by untreated cells. (Akeson and Warren, 1986; Damon et al. 1988) and dorsal Accompanying the 8- to 11-fold increase in [36S]sulfate root ganglia (Carbonetto et al. 1983; Verna et al. 1989). radioactivity in the medium of /J-xyloside-treated PC 12 Recent studies also indicate that chondroitin sulfate or cells, there was also a 1.6- to 2.1-fold increase in cell- chondroitin/keratan sulfate proteoglycans may be com- associated radioactivity. Pertinent to this finding is the ponents of astroglial axon barriers (Snow et al. 1990a,b). fact that PC12 cells contain numerous chromaffin gran- Although there is in certain cases a degree of chemical ules (Tischler and Greene, 1978), which have been isolated specificity for these effects, it is not clear to what extent and partially characterized (Roda et al. 1980; Wagner, they are due to factors such as modification of the surface 1985; Cutler and Cramer, 1990). We have previously charge of the substratum, alteration of the binding of demonstrated that potassium-induced depolarization of growth-promoting or adhesive molecules produced by the PC12 cells leads to a large increase in the release of cultures, or a simple reduction in adhesiveness in chondroitin sulfate proteoglycans (Salton et al. 1983) comparison to a preferred substratum (e.g. collagen). having properties very similar to those of two proteogly- Previous studies of this question also bear directly only on cans that we had earlier characterized in the bovine the glycosaminoglycan portion, although these molecules chromaffin granule matrix (Kiang et al. 1982), and more occur in tissues as covalent complexes with protein in the recently we have also identified a proteoglycan form of form of proteoglycans (Margolis and Margolis, 1989). It is chromogranin A in both PC12 cells and bovine chromaffin therefore important to obtain information concerning the granules (Gowda et al. 1990). Since chromaffin granule effects of alterations in nervous tissue proteoglycans, and proteoglycans probably account for a significant pro- it has, in fact, recently been reported that the glycos- portion of the chondroitin sulfate proteoglycans that we aminoglycan chains do not contribute significantly to the have previously characterized in PC12 cells (Gowda et al. binding properties of a chondroitin sulfate proteoglycan of 1989, 1990), it is likely that the much smaller increase in embryonic chicken brain (Hoffman et al. 1988). cell associated (as compared to secreted) chondroitin Our studies demonstrate that both in PC 12 cells and in sulfate in /S-xyloside-treated PC12 cells reflects the early postnatal rat cerebellum, /J-xyloside inhibition of expected inhibition of chromaffin granule proteoglycan normal chondroitin sulfate (and possibly also heparan

244 R. K. Margolis et al. sulfate) proteoglycan biosynthesis is accompanied by clear Glycosaminoglycan production by bovine aortic endothelial cells alterations in the morphology and adhesive properties of cultured in sulfate-depleted medium. J. bwl. Chem. 281, 9122-9127. the cultured cells. The morphological changes are, by their JOHNSTON, L. S. AND KELLER, J. M. (1979). The effect of /J-xylosides on heparan sulfate synthesis by SV40-transformed Swiss mouse 3T3 nature, difficult to quantitate, but readily apparent and cells J. bwl. Chem. 254, 2575-2578. highly reproducible. Although we have demonstrated that KANWAR, Y. S., HASCAU^ V. C, JAKUBOWSKI, M L. AND GIBBONS, J. T. these effects are not merely a result of the presence of free (1984). Effect of /J-D-xyloside on the glomerular proteoglycans. I. glycosaminoglycan chains released into the culture me- Biochemical studies. J. Cell Biol. 99, 715-722. dium, the relative contribution(s) of a decrease in normal KIANG, W.-L., KRUSIUS, T , FINNE, J., MARGOLIS, R. U. AND MARGOUS, proteoglycan and/or the presence of core protein deficient R. K. (1982). Glycoproteins and proteoglycans of the chromaffin in glycosaminoglycan chains cannot be evaluated at this granule matrix. J. biol. Chem. 257, 1651-1659. KLEIN, D. J , BROWN, D. M., MORAN, A., OEGEMA, JR, T. R. AND PLATT, time. However, the present findings strongly support the J. L. (1989) Chondroitin sulfate proteoglycan synthesis and assumption that chondroitin sulfate proteoglycans, which reutilization of /J-D-xyloside-initiated chondroitin/dermatan sulfate are present in the extracellular space of developing brain glycosaminoglycans in fetal kidney branching morphogenesis. Devi (Aquino et al. 19846; Ripellino et al. 1988, 1989) and have Biol. 133, 515-528. KOLSET, S. O., SAKURAI, K., IVEHD, I., 0VERVATN, A AND SUZUKI, S. also been localized at the cell surface of neurons (Zaremba (1990) The effect of /3-D-xylosides on the proliferation and et al. 1989) and astrocytes (Levine and Card, 1987), are proteoglycan biosynthesis of monoblastic U-937 cells. Biochem. J. 265, involved in cell-cell and cell-matrix interactions in 637-645. nervous tissue. LEDBETTER, S. R. AND HASSELL, J. R (1986). /3-D-Xyloside-mediated alteration in the synthesis of basement membrane proteoglycan. This work was supported by grants NS-09348 and NS-26283 Archs Biochem. biophys. 246, 403-410. from the National Institutes of Health, and a Research Scientist LEVINE, J. M. AND CARD, J. P. (1987). Light and electron microscopic Award (MH-00129) to R.U.M. from the National Institute of localization of a cell surface antigen (NG2) in the rat cerebellum: Mental Health. Association with smooth protoplasmic astrocytes. J. Neurosci. 7, 2711-2720. LOWRY, O. H., ROSEBROUGH, N. J., FARR, A. L. AND RANDALL, R. J. References (1951) Protein measurement with the Folin phenol reagent.

246 R. K. Margolis et al.