Biothlmlca et Biopl!vsica Acts. 1081 (1991) 253-261 '~ 1991 Elsevi~ Science Publishers B V (BiomcdlcMDivision}0005-2760/91/$03.50 dDONIS 000527609100082R

253

Glycolipids and glycosyltransferases in permanent cell lines established from human medulloblastomas J o h a n G o t t f r i e s }, A l a n K . P e r c y '-, J a n - E r i c M h n s s o n i P a r e F r e d m a n i C a r o l J. W i k s t r a r , d 3 H e n r y S F r i e d m a n 3 D a r e l l D . B i g n e r 3 and Lars Svennerholm 1 Department of Psychiatry' and N ~ h e m l s t o ' . UnwerMtvof f;i~tebo~, St. Ji:rgen Hospital. tl:sings Backs (SwedenL : Departmenl of Pediatrics. Bavlor College of Medtclne. Ho~ton, TX md ¢Department of Pa'holo~3~ Duke Univer~i~¢Medical Center. Durha~ NC ¢U~S.A.I

(Received 14 Jolte 19£J0) Key words: Medullob[astoma;Glycosyltransferase: Xe:tograft: Gangliosld¢:Sphlngolipid Medelloblnstoma biopsies are heterogenons and might contain normal brain tissue, which limits the nsefullness of such tumor material for htnebumieal analyses. We have, therefore, examined the gangtiosides and their metabolism using the medulloblastmna ~ell lines, Daoy and D341 Med, cultured both in vitro and as xenogralts in nude mice. The ganglio~side patterns in the Daoy showed a switch from a high GM2, 70% (mol% of total ganglioslde siallc acid) and low lactoseries ganglioaldes (2%) content in monoinyer cultures, to a high proportion of laetnseries ganglinsides (50%) and virtually no GM2 (1%) in xenografts, but an increased proportion of ether a-series ganglinsides. The D341 Med showed a similar change regarding tile lacto-series gangliosides from 1% in suspansion culture to 10% in xenografts. The activity of five glyensyltransfemses, GM3, GD3, GM2, G M I and LA2 syntheses, did not paralleU the ganglioslde parents and could not account for the noted variations therein. In the Daoy cell line the LA2 syntimse as well as the G M 2 synth~se activity was relatively high in both culture systems, despite the marked difference in the expression of G M 2 and the lactnseries gaasUosides. These results suggest that environmental factors play a crucial role for the in vivo activity of the glycosy Itranslerases. lntroduellon The gangliosides are distributed in a~l mammalian plasma membranes, but are most abundant in the brain |1]. The concentration of individual gangliosides changes

Ahbrevtations: OM3 synthase (EC 2.4.99.9),SAT-I [10l. SAT-1 131l; GD3 synlhase (EC 2.4.99.8). SAT-2 110]. SAT-II [31]: LA2 synthase. OIcNAcT-! [lO]: GM2 ~ynth~ (EC 2.4.1.92). GaINAcT-I [10l, OalNAcT [3"i]; GM1 synthase (EC 2.4A.62L GAIT-3 111)],GalT-[I [31]: C'F-B, cholera ioxin B-subunil: Oanglioside nomenclalure according to Svenncrholm [341 IUPAC-IUB: GM3, II:-NcuAcLacCe*; GM2, lI3NeoAc-GgOse~Cer:GM1. ll~NeuAc-GgOse4Cer: 3'-LMI, IV~NeuAc-nLcOseaCer; Y-isoLMl. IV3NeuAc-LcOse4Cer: Fuc,-3'-isoLMl. IV3NeuAc.lll4Fue-LcOse4Cer; GD3, 113(NeuAc)2LacCer; GD2. IIJ(NeuAc)2-GgOse~Cer: GOla. IV3NcuAc, [13NenAc-OgOsc4Cer; GD|b, l[3(NeuAclz-GgOse~Cer; GTIb, IV3NeuAeAI3(NeuAc)z-GgOse4Cer; GQlb, IV3(NeuAe)2 , ll3(NeuAc)2-GgOsc4Ccr;LacCcr, lactosylccr~idc. Correspondence: L. Svennerholm, Department of Psychiatry and Neurochemistry. University of GiSteborg.St. Jbrgen Hospital. 42203 Hislngs Backs, Sweden.

with development [2] and neoplastic transformation 131. Previous studies on human primary brain tumour tissue have shown that gliomas [4] as well as medulloblastomas [5] are associated with quantitative and qualitative differences in ganglioside composition as compared to normal brain tissue. However. biopsies of primary tumours are heterogenous and contain normal brain tissue and reactive astrocytes [6], which limits the possibility to interpret the gangiioside analyses, Permanent cell lines provide a source of uncontaminated tomour cells which might be useful for the characterization of tumour-associated gangliosides and studies of their melabolism. In a previous study [7], the cells of the human glioma cell line, U-I18 MG, showed marked differences in ganglioside composition when cultured in vitro or as xenografts in nude mice, and it was suggested that environmental factors influenced the ganglioside expression. The aim of this study was to investigate the gangliosides in human medulloblastoma cell lines grown in in vitro or as xenografts in nude mice. Medulloblastoma by itself is a neoplasm with uncertain origin,

254

featuring cells of both neuronal and glial phenotype [8], and the gangliosidc composition is a potential marker for these phenotypes [5]. Therefore, two medulloblastoma cell lines representing both phenotypes [9] were chosen for this study. The glycosyhransferases [10] catalyze the biosynthesis of gangliosides and their structure is assumed to depend on the relative in rive activity of these transferases. We have measured the activity of five giycosyltransferases, LA2, GM3, GD3, GM2 and GMI synthases in the two medulloblastoma ceil lines in order to find out if their gangiioside composition parallelled the in vitro activity of these enzymes. Materiah~ and Methods Materzals Glass and aluminum-backed precoated silica gel 60 HPTLC-plates and LiChroprep RP-18 (40-63 #m) were obtained from Merck AG (Darmstadt, F.R.G.) and plastic-backed silica gel precoated TLC-plates from Machery-Nagel (DiJren, F.R.G.). N-2-Hydroxyethylpiperazine-N'-2-ethanesulfonic acid (Hepes), fatty acid free bovine serum albumin and 5-bromo-4-chlore-3-indolylphosphate (BCIP) were from Sigma (St. Louis, U.S.A.). The liquid scintillation fluid, lnsta-Gel was from Packard III. (U.S.A.). Newborn Sprague-Dawley rats were purchased from A-lab (Stockholm, Sweden.] Vibrio cholerae sialidase (EC 32.1.18) was from Behringwerke (Marburg, F.R.G.). The ion-exchange resin Spherosil-DEAE-dextran and cholera tcxin B-subunit (CT-B) were from lnstitut M6rienx (Lyon, France). NEN EN3HANCE spray was from New England Nuclear Research products (Boston, U.S.A.) and Kodak X-Omat AR film was purchased from Kodak (Stockholm, Sweden). All solvents were of HPLC-grade. Monoclonal antibodies, All monoclonal antibodies (MAb) used for the determination of gangliosides were produced by us. Ganglioside 3"-isoLM1 was assayed with the MAb SL-50 (binding epitope, NeuAca23Galfll-3GlcNAcfll-), 3'-LMI with the MAb SL-I (binding epitope, NeuAca2-3Galfll-4GIcNAc/31-), GM2 with DMAb-1 (binding epitope, GalNAcfll4(NeuAc,x2-3)Galfll-) [111, GD2 with DMAb-20 (binding epitope, G a l N A c f l l - 4 ( N e n A c t ~ 2 - S N e u A c a 2 3)Gal,81-) and GD2 and 3'g'-LDI with DMAb-8 (binding epitope. NeuAcct2-8NeuAclx2-3Gah~l-) [121. The MAb raised against CT-B was produced and had the same specificity as d~,cribed previously [13]. Alkaline phosphatase-conjugated goat anti-mouse lgM and lgG (H + L) was purchased from Jackson lmmunoresearch Lab. (West Grove, PA, U.S.A,). Glycolipid accepters and standards. Laetosyleeranlide (LacCer) was prepared from human brain gangliosides by reflux for 90 min in 0.05 M sulfuric acid in water. After neutralization, chloroform and methanol was ad-

ded to the ratio 4:2:1 (v/v). The LacCer was isolated frorc the obtained lower phase by silica gel column chromatography [14]. Ganglioside GM3 was extracted from the spleen of a patient who had died from Gaucher disease [15]. Ganglioside GM2 was isolated from the brain of a patient who had died from Tay-Saeha disease [1(]. Other gangliosides and neutral glycolipids used as star!dards were all isolated from human sources and characterized by fast atom bombardment-mass spectrometry. Gangliosides GM3 and GM1 were tritium labelled in the ceramide portion according to Schwarzmann [17]. Activated sugars. Uridine 5"diphospho-N-acetyl-D-[l14C]galactosamine (UDP-GalNAc, 2.22 GBq/mmol), uridine 5"diphospho-D-[U-14C]galaetose (UDP-Gal, 11.8 G B q / m m o l ) , uridine 5'diphospho-N-acetyl-D-[U14C]giucosamine (UDP-GIcNAc, 10.7 GBq/mmol) and cytidine 5 ' monophospho-N-acetyl[4,5,6,7,8,934Clsialie acid (CMP-NeuAc, 9.69 GBq/mmol) were purchased from Amersham International (Buckinghamshire, U.K.). Corresponding unlabelled nucleotide sugars were purchased from Sigma. Cell culture and xenografls The permanent human medulloblastoma cell line Daoy was obtained from Dr. Jacobsen [18]; D341 Med was established by Dr. Friedman [191. An extensive phenotypic analysis of these lines has been published [91. Daoy grows as an adherent monolayer, requiring 0.25~ trypsin-0.02% EIYFA for passage; DM1 Med grows in suspension. Both lines were grown in 10~ fetal calf serum-Richter's Zinc Option Medium (Gibeo, Grand Island, NY) at 3 7 ° C in 5~0 CO z. The absence of contamination with mycoplasma was confirmed as previously described [20]. Suhenta,aeous routine xenografts of Daoy were established by the inoculation of passage 57 cultured cells with an initial latency to serial passage of 14 days; subcutaneous xenografts of I)341 Med were established by inoculation of passage 68 cultured cells with an initial latency of 35 days. Initial establishment doses were 1 • l 0 t cells per recipient. Growing subcutaneous tumours were measured twice weekly, approximate volume determined and tumours harvested while still in linear growth phase; dissected turnouts were maintained at - 7 0 ° C until use for glyenlipid extraction, or Golgi preparation. Morphological studies of the original biopsies and routine xenografts of both Daoy and D341 Med have been previously published [18,19]; in addition, an extensive immunohistochemical phenotype of these xenografts grown in atbymie rats, necessitated by the use of specific murine monoelonal antibodies, has been published [91. The morphology of all xenografts maintained ni serial transplant is routinely monitored, and has proven consistent throughout the time course of this study.

255

lmmunohistoehemical ana(t'sis of medulloblastoma-associated proteins An extensive phenotypie analysis of cultured Daoy and D341 Med cells and their derived xenografts has been published [9]. The antigen-monoclonal antibody pairs used in this study were selected to highlight the essentially non-neuronal phenotype of Daoy cells, and the essentially neuronal phenotype of D341 Med cells. Daoy cells cultured on 2 × 2 mm coverstips or D341 Med cells spun onto glass slides by the cytospin technique were fixed in 70°C acetone for 15-30 s, and irnmunohistocbemistry using Streptavidin-biotin complex (Zymed Laboratories, S. San Francisco. CA) was performed as previously described [9]. Primary antibodies used in this study included: an anti-NFP 200 Kd Mab (1/100 dilution tissue culture supernatant) [211; anti-synaptophysin Mab SY 38 (5 txg/ml: BoehringerMannheim, Indianapolis, IN) and monoclonal antitenasoin antibody 81C6 (5 ,ttg/ml) [221. Glycolipid isolaaon and determination Cell pellets and tissues were homogenized in water (an aliquot was removed for protein determinations) and lipids were extracted with c h l o r o f o r m / m e t h a n o l / water (4: 8 : 3, v / v ) [14]. The lipid extracts were applied to a silica gel column. Less polar lipids were elt;~e.:l with eight bad volumes of c h l o r o f o r m / m e t h a n o l / w a t e r ( 6 5 : 2 5 : 4 , v / v ) and the crude ganglioside fraction was eluted with ten bed volumes of c h l o r o f o r m / m e t h a n o l / water ( 3 : 6 : 2 , v/v). The neutral glycolipids were separated from the ganglio~ides by anion exchange chromatography on SpherosiI-DEAE-dextran [23] and the gangliosides were eiuted from the resin into two fractions: (a) monosialogangliosides; and (b) oligosialogangliosides with 0.02 and 0.5 M KAc in methanol, respectively. The ganglioside fractions were desalted by dialysis against a 0.02 mM carbonate buffer (pH 8.4) and quantitative determination was made by the resorcinol method [241. The gangliosides were separated by HPTLC with chloroform/methanol/0.25% aqueous KCI ( 5 : 4 : 1, v/v), visualized by the resorcinol reagent, and their proportions evaluated by densitometric scanning at 620 am. The TLC-ELISA method has been described in detail elsewhere [25]. The ganglioside fractions were separated on plastic backed TLC plates using chloroform/ methanol/0.25% aqueous KCI or 2.5 M ammonia ( 5 : 4 : 1 , v / v ) as developing solvent. After drying at least 1 h at room temperature, the plate was dipped in 0.1% polyisobutylmethaerylate in n-hexane. Unspecific background staining of the plate was blocked by preincubation for 30 rain at room temperature using Tris-BSA (50 mM Trls-FICI (pH 8.0), 0.15 M NaCI and 1~ bovine serum albumin). The TLC-plate was in sequential order incubated with an antiganglioside MAb in Tris-BSA for 16 h at 4°C, and the alkaline phosphatase-linked anti-

mouse lg antibody in Tris-BSA for 3 h at room temperature. Bot, rtd antibody was visualized using BC1P as substrate in 0.1 M glyeine (pH 10.4), 1 mM ZnCI., and 1 mM MgCI:. Determination of GM1 and oligosialcgangliosides of the gangliotetraose series, separated on TLC and hydrolysed by slalidase, was performed as above with lbe antiganglioside MAb substituted with CT-B and an additional step with an anti-CT-B MAb in Tris-BSA [25]. Quantification of gangliosides was performed by densitometrie .scanning at 620 nm.

En:yme preparations The preparation of a crude membrane-enriched fraction from in vitro cultured cells was performed as a modification of the procedure of Momni and co-workers [26]. All steps were performed at 4°C. The cells were suspended in two volumes of ice-cold 0.32 M sucrose and sonicated four times for 60 s. on ice, in a Soniprep 150 MSE with an amplitude of 26 microns. The disrupted cell suspension was immediately centrifuged at 1000 × g for I0 rain and the supematant collected. An aliquot w0s taken for protein measurement and the enzyme preparations were kept at - 2 0 ° C until the transferase assays were performed (within 2 weeks from preparation). Golgi preparation of the xenograft tissue was pc:formed as follows, the tissue was homogenized on ice i,': five volumes of ice-cold 0.32 M sucrose with eight strokes of a motor-driven (300 rpm) Teflon glass homogenizer and the suspension was centrifuged at 1000 X g for 5 min (4°C). The supernate was centrifuged at 2 7 0 0 0 X g for 10 rain and the resulting pellet was resuspended in three volumes of 0.32 M sucrose. Vnis suspension was layered on top of 1.3 M sucrose and centrifuged for 60 min at 100000 x g. The interface material was collected, diluted with distilled water and centrifuged at 27000 × g for 30 rain. The pellet was resuspended in 0.32 M sucrose to a concentration of 2 5 mg protein/ml and stored at - 2 0 ° C until the enzyme assays were performed. The assays were performed within 2 weeks from preparation, Glycosyltransferase assays The assay conditions for the different glycosyhransferases are described in Table 1. The detergent, except oetylglueoside in the G M 2 synthase (GaII~AcT) assay, and the glycolipid acceptor were added in the reaction tube dissolved in chloroform/methanol (2:1, v/v) and dried under a stream of nitrogen before 40 ?~1 of the reaction mixture was added. The detergent and the acceptor were allowed to dissolve for 10 rain. with repeated agitation of the tubes. The reactions, performed at 37°C in capped tubes, were started by the addition of 10 .al of the enzyme source (Golgl preparations containing 20-50 /tg of protein, or membrane enriched preparations from cell cultures containing 75-

256

TABLE I Array condJtionsfo# the delermination of gly~syllransferase~ Parameters

Unit

Lipid accepter mM Nucleotide sugar mM dpm/nmol Detergent mg/ml Buffer M pH Mn2+ CDP-Chol Incubation volume Incubation time

mM mM /tl h

Glycosyitransferases GM3 GD3 synthase syntha~ LacCer GM3 0.3 0.9 CMP-NeuAc CMP-NeuAc 0.8 0.6 8000 1000(I Triton CF-54/ Triton X-ILO "rween 80 (2 : 1) 2.0 1.O Hopes Hopes 0.2 0.2 6.25 6.25 50 1

50 l

125 /~g of protein per assay). After incubation for 60 min the reaction was stopped by addition of 0.5 ml ice-cold 0.05 M NaAc buffer (pH 4.4). The reaction mixture was immediately loaded on a column of LiChroprep RP-18 (5 × 10 mm). The radioactive precursor was ehited with 1 ml NaAe buffer, 3 ml water and 2 ml water/methanol (4 : 1, v/v). The glycolipids were eluted with 1 ml methanol, 2 m] chloroform/methanol ( 1 : 2 , v/v) and 2 ml chloroform/methanol ( 2 :1 , v/v). The glycolipid containing eluates were collected directly into a scintillation tube and dried under nitrogen. The residue was redissolv~:l in 0.5 ml water, mixed with 10 ml scintillation liquid and counted in a Packard Tri-Carb 1500 liquid scintillation analyzer. All assays were performed in triplicate with and without glycolipid accepter.

Identification and stability or formed products After the enzyme reaction, 10-50% of the isolated glycolipid fraction was separated on aluminum-backed HPTLC-plates. Developing solvents used were, chlorof o r m / m e t h a n o l / w a t e r ( 6 0 : 3 2 : 7 , v / v ) for the separation of the products from the LA2 synthase, G M 3 synthase and G M 2 synthase assays and chloroform/ methanol/0.25% aqueous KCI ( 5 : 4 : 1 , v/v) for products formed in the G M ] synthase and G D 3 synthase assays. The radioactive glycolipids were visualized by exposing the plates to X-ray film for two weeks at - 8 0 ° C . Their migration was compared to orcinolstained standards separated on the same HPTLC-plate. The degradation of formed products in the incubation mixtures was determined by adding tritium-labelled gangliusides G M 3 and G M 1 (4 nmol, 1 1 0 0 0 0 d p m / n m o l ) in the various glycosyltransferase assays (Table II with labelled nueleotide sugar omitted. 1 / 1 0

(3M2 synthase GM3 0.9 UDP-GalNAc 0.4 10000 Octylglueoslde

GMI synlhase GM2 0.4 UDP-Gal 0.4 11ooo Triton CF-54

LA2 synlhase LacCer 0.3 UDP-GIcNAc 1.0 11000 Triton X-t00

8.0 Hepes 0.2 8.00 6.0 5.0 50 1

1.0 Helms 0.2 7.25 6.0 5.0 50 1

1.0 Hepes 0.2 7.25 5.0 5.0 50 l

of the isolated glycolipid extract was separated on aluminum.backed HPTLC-plates using chloroform/ methanol/0.257o aqueous KCI (5 : 4 : 1, v / v ) as developing solvent and the radioactive glycolipids were visualized by autoradiography. The plates were exposed to X-ray film for 2 weeks at - 80°C. When the autoradiography disclosed bands migrating differently from G M 3 or GM1 standards, the remaining portion of the isolated glycofipid fraction was, together with nonlabeled standards, separated by HPLC [27] for quantitation. One ml fractions were collected and 10 #1 were applied to HPTLC and visuafized by resorcinol. The radioactivity of the remaining part of each fraction was determined by liquid scintillation counting as described above.

Analytical methods Protein was measured using bicinchoninic acid [281 with fatty acid free bovine serum albumin as standard. Results

lmmunochemical results Daoy and D341 Med exhibit two different phonetypic profiles (Fig. 1): Daoy a non-neuronal profile and D341 Med a neuronal-like one [9]. Daoy is unreaetive with monoclonal antibodies to the high molecular mass (200 kDa) neurofilament protein triplet (Fig. I D ) and to synaptophysin (Fig. 1F), which identify cells of neuronal and neuroendocrine phenotype, respectively. Daoy cells bind Mab 81(26 (Fig. 1H), reactive with an epitope on the extracellular matrix molecule tenascin, which has been defined as a component of glioma extracelhilar matrix [22]; thus Daoy presents an essentially non-neuronal picture. Conversely, D341 Med, unreactive with

257

A

B

C

D

E

F

when exogenous acceptor (LacCer) was included (Fig. 2). The D341 MOd xcoograft contained endogenous accepturs for GM3 synthase and LA2 symhase that gave products comigrating with GM3 and LA2. respectively. The amount of product formed from endogenous accepter was approximately half of that induced when the appropriate exogenous accepter was included. When cultured D341 Med cells were assayed for GM2 syuthase activity the same amount of product was found, both with and without exogenous accepter. Cultured ceils and xenografts of both medunoblastoma cell lines formed products from endogenous accepters in variable proportions (30 100% of the activity when exogenous accepter was included) in the GM1 synthase assay. The radioactive products formed from endogenous accepters in all other assays constituted less than 10% of total activity. When transferase activities were calculated (Fig. 3A and 3B), the zero-time incubation value was subtracted from the measured activity. This was don,: unless products. migrating clearly different from the expected product, were found in the respective assay independent of me inclusion of exogenous acxeptor, In those cases the 1 h incubation without exogenous accepter was used to subtract from the measured gross activity. To evaluate the stability of the formed glycosphingolipids, triSated gangliosides G M 3 and G M I were used. For G M 3 symhase and GD3 synthase conditions (Table 1) 5-10% of the GM3 was converted to LacCer. Virtually no ( < 1%) hydrolysis of either labeled gang-

LacCer

9m ~

G

~

H

Globletra : ,In

Fig. 1, Cultured D341 Med cells (A,C,E,GL passage 45, were ~xamined in cytospin preparation; cullured Daoy cells(B,D,F.HLpassage 69, were grown on c~¢rslips, All panels are at 260Xmagnification. A.B: tissueculture medium negativecontrol, hcmatoxylincounterstain: C,D: anti-NFP 200 kDa pllmary muilmdy: F-F: anti-syn~,,tophysin primary; G.H: antt-tenascin primary.

the Mab to tenasein (Fig. IG), binds Mabs specific for synaptophysin (Fig. 1E) and the 200 kDa N F P (Fig. IC), suggestive of a neuronal lineage.

Stability and identification of formed glycosphingolipids The enzyme sources contai,~,.'d endogenous glycosphingolipids which might function, as glycosyltransferase accepters. In the LA2 syathase assay with xenografts from the Daoy cell line as enzyme source, a product comigrating with CA1 standard, was formed from endogenous accepter. This product wag not found

GIobttl GAI

origin 1

2

3

4

5

Ref.

Fig. 2. Autoradiographic dct~tion of glycolipldsformed by the LA2 syntha~ a.~ay. The glycolipids were separated by HPTLC using v/v} a.s developing mlvcnt. Products of endogenous accepter in lane I. Daoy x*nogratt; lane 2, D34l Med mnograft tissues. Products fo~cd troth exogenous accepter (LacCer, see Table [) lane 3, Daoy monolayer cultured cells; lane 4, Daoy xcnograf[; and Ian¢ 5, D341 Med xenograft tissue. Ref,. characterizedreference glycolipidsvisualizedby orcinol reagent. LA2 ~ynthase a.~say,glycollpid isolation and culture conditions are described under Materialsand MelhlxJs.

chloroform/methanol/0.25~ aqueous KCI (5:4:1,

258

nmoUmg~ o ~ : a

0/I

02

o i ~

Qva~

GlJt~

U~m

nca/n~ pcoud~h

0 Fig. 3 Glycogyltransferaseactivities(nm,,I/mg protein/h) of (A) in vilro cultured cells of medullnblastomacell lines, and (B) xeaogfaft tissue of the medulloblastoma cell lines Daoy and D341 Med. respa~tively. Cell and tissue preparations, glycosyltransferase assays ant: c¢11 cultul~ conditions are described under Materials and Methods. Bars indicate $.E,M. (n ~ 3). ]inside was noted in the other assays. These results were consistent for Golgi preparations and the membrane enriched preparations used for cultured cells.

Glycosyltransferases and gangliosides of the human medulIoblastoma cell lines Daoy and 1)341 Med GM3 synthase activity was found in all preparations (Fig. 3A and 3B). The Daoy cell line grown as xenograft

had a low G M 3 synthase activity (Fig. 3B) although G M 3 constituted 22 tool% (Table lI) of total ganglioside sialic acid. The G D 3 synthase activity was in comparison to that of G M 3 synthase relatively low in both xenografts and the Daoy cultures, and higher in the D341 Med grown in suspension, which also had a high proportion of b-series ganghosides (Fig. 4B). The xenografts of Daoy

259 TABLE It Ganghosides of the permanent cell lines Daov and D341 Med

Gangi;oside~ w~e separated on HPTLC by cnlorofo~/methanoi/ 025% aqueous KCI f5:4:1, v/vl. vlsaalized by the r~orcino[ reagent and assessed by densilornetric scanning at 620 am, Gangliosides GM2, GM1, 3'-LMI 3"-isoLM1.GD3 and GD2 were assessed by TLC-ELISA. using MAbs. GangJiosides GDIa, GDIb and GTlb were assayed by stalidase treatment of HFTLC-separated gangllosid¢~and ~ubscquent ELISA as for GM1. The ,klAbs. ganghostdc isolation and qu~titation, xenograft pr~edure and celt cuhure couditioas are described in detag under Materials and Methods. n.d.. not detected. The data a~ presented ~ the m ~ of dupli~te det~minations. Daoy (mol% of total ganglioside sialic acid) cell culture xenograft 13 22 70 1 I 9 6 12 2 2 6 t