238 Biochimica et Biophysrca Acta 834 (1985) 238-248 Elsevier

BBA 51896

G ,,-ganglioside metabolism in cultured human skin fibroblasts: unambiguous

diagnosis of G ,,-gangliosidosis

Srinivasa Raghavan a, * , Allan Krusell a, Timothy A. Lyerla b, Eric G. Bremer a, * * and Edwin H. Kolodny a

’ Department of Biochemistry, Eunice Kennedy Shriver Center-for Mental Retardation, 200 Trapelo Rd., Waltham, MA 02254 and’ Department of Biology, Clark University, Worcester, MA 01610 (U.S.A.)

(Received September 2nd, 1984) (Revised manuscript received January 2nd, 1985)

Key words: Ganglioside metabolism; Tay-Sachs disease; G M2 gangliosidosis; (Human skin fibroblast)

The metabolism of GM,-g an lioside was studied in situ using cultured skin fibroblasts from normal g individuals and patients with different forms of G ,,-gangliosidosis. [ 3H]Sphingosine-labeled G M2 was provided in the culture medium to confluent cells in 6-cm petri dishes. After 10 days, the cells were washed free of radioactivity and harvested by trypsinization. The cellular lipids were extracted and analyzed for radioactivity in G,, and its metabolic products. In fibroblasts from healthy subjects, 5040% of the total cellular radioactivity was found in the neutral glycosphingolipids, ceramide, sphingomyelin and fatty acids. Degradation of the labeled G M2 progressed rapidly via G M3, ceramide dihexoside and ceramide monohexo- side with a build-up of radioactivity mainly in the ceramide pool of the cell. The labeled ceramide is also reutilized for the synthesis of ceramide trihexoside, globoside and sphingomyelin or is converted to fatty acid and incorporated in ester linkages. In contrast, cells from patients with G,z-gangliosidosis representing Tay-Sachs, Sandhoff and AB variant forms of the disease did not metabolize the ingested labeled G,&ke controls. Nearly all of the radioactivity was present in the ganglioside fraction in the lipid extracts from these cells and consisted of unhydrolyzed GM,. High-performance liquid chromatographic analysis of monosialo- gangliosides from cells grown without added labeled GM2 in the medium indicated accumulation of

endogenously synthesized G M2 in cell lines from all patients with G MZ gangliosidosis compared to healthy

controls. This approach provides a reliable tool for pre- and post-natal diagnosis of all forms of GM2-gang- liosidosis without ambiguity.

Introduction

P-N-Acetylhexosaminidase (EC 3.2.1.52) occurs

in two major isoenzymic forms, designated as Hex

* To whom correspondence should be addressed. ** Present address: Fred Hutchinson Cancer Research Center,

University of Washington, Seattle, WA 918104, U.S.A. Nomenclature: The gangboside nomenclature used here is that of Svennerholm: GM,, I13NeuAc-LacCer; GMz, Il’NeuAc- GgOse,Cer; GM,, I13NeuAc-GgOse,Cer. Abbreviation: Hex, hexosaminidase.

A and Hex B in the lysosomes of mammalian cells. Hex A is responsible for the hydrolysis of G,, ganglioside in vivo, since a metabolic defect affect- ing the activity of Hex A results in the accumula- tion of G,, in the lysosomes, causing variant forms of G,, gangliosidosis [l]. The activity of Hex A and Hex B is usually measured by a differential assay technique employing the syn- thetic fluorogenic compound 4-methyl-umbel- liferyl-N-acetyl-P-glucosaminide (4MU-/3- GlcNAc) as the substrate [2]. This substrate pro-

0005-2760/85/$03.30 0 1985 Elsevier Science Publishers B.V. (Biomedical Division)

239

vides great ease and extreme sensitivity for the assay of hexasamidase, but it lacks the total specificity for Hex A as seen in vivo with the natural substrate G MZ ganglioside. Recently Kresse et al. [3] reported that p-nitrophenyl-6-sulfo-2- acetamido-2-deoxy-P-glycopyranoside (PNP- GlcNAc-6-SO,) was specifically hydrolyzed by Hex A and not by Hex B. Fuchs et al. [4] have cautioned that Kresse’s substrate is also not abso- lutely specific for Hex A, since another isoenzyme, Hex I, present in serum exhibited 17% of the PNP-GlcNAc-6-SO, degrading activity of Hex A. Also, this substrate cannot diagnose the activator- deficient AB variant form of this disease [5] and has not been evaluated in the unusual cases of apparently normal individuals who are deficient in Hex A activity measured with 4MU+GlcNAc as substrate [6-lo].

Methods for determining G,, cleaving hexosaminidase activity with radiolabeled natural substrate require the use of detergents [11,12] or the natural activator protein [13] to mediate the interaction between the enzyme and the lipid sub- strate. The detergents, however, alter the substrate specificity of the isoenzymes; i.e., Hex B hydro- lyzes G,, in vitro in the presence of bile salts [14-171, although it does not hydrolyze G,,’ in vivo as evident in Tay-Sachs disease. The assay system employing the natural activator protein [13] truly reflects the in vivo status for G,, hy- drolysis, but this activator protein is difficult to prepare and is available only in a few laboratories.

The ability of cultured cells to ingest exogenous glycosphingolipids from the culture medium pro- vides the means by which cells can be challenged with a labeled sphingolipid and the metabolism of the labeled component followed in situ in the living cell [18-261. We have utilized this approach to examine the kinetics of uptake and intracellular metabolism of labeled G,, and other sphingoli- pids present in cultured fibroblasts derived from a single lo-cm or 6-cm dish of confluent cells. This approach provides a reliable system to detect de- fects in G,, metabolism and offers the potential for both pre- and post-natal diagnosis of G,, gangliosidosis without ambiguity.

Materials and Methods

Preparation of sphingosine-labeled [ ‘H]G,,. Pure G,, ganglioside was isolated from Tay-Sachs brain according to Itoh et al. [27]. Sphingosine- labeled [ 3 H]G M2 was prepared from pure G,, by reacting with sodium boro[ 3H]hydride according to Schwarzmann [28]. The sample was then taken in methanol and loaded onto a DEAE-Sephadex column (1.4 cm i.d. vs. 16 cm height) prepared according to Momoi et al. [29]. After washing the column with 300 ml methanol, pure labeled G,, was eluted completely with 300 ml of 0.025 M ammonium acetate in methanol. The solvent was removed in a rotary evaporator and the sample dissolved in 10 ml of chloroform/methanol/water (3 : 48 : 47, v/v). It was passed three times through a BondElut Cl8 cartridge (Analytichem Interna- tional, Harbor City, CA) containing 500 mg of absorbent that had been equilibrated with chloro- form/ methanol/ water/O.75 M ammonium acetate (3 : 48 : 47, v/v). The cartridge was washed with 100 ml water to remove all the salts. The labeled G M2 absorbed onto the Cl8 matrix was eluted off the cartridge with 25 ml methanol. The radiolabeled G MZ was found to be pure as judged by its mobility on thin-layered chromatography (TLC) and high-performance liquid chromatogra- phy (HPLC) according to Bremer et al. [30]. Acid hydrolysis of the sample in acetonitrile [31] and analysis of the component products indicated that 84% of the radioactivity was present in the dihy- drosphingosine and 16% in fatty acids with virtu- ally none in the sugars or sialic portion of the ganglioside molecule.

The specific radioactivity of G,, was de- termined by counting an aliquot in a Packard Tricarb liquid scintillation spectrometer and by quantitating the spingosine base in another aliquot of the sample as previously described [32], using pure G,, as standard. The specific activity could also be calculated by HPLC analysis of the sample by collecting the eluant under the G,, peak for determining the radioactivity and by quantitating the amount of material from the area under the peak by integration in reference to standard G,,. The radioactive sample was then diluted with non- radioactive G M2 to a specific activity of about 20-80 000 cpm/nmol.

240

Cell culture. Most of the human diploid fibro-

blast lines serially propagated for these studies were grown from biopsies in the Genetics Labora-

tory at the Shriver Center from individuals di-

agnosed as normal or diseased. Some lines were shipped from other laboratories in T-25 flasks. All

lines were tested to be free of contamination by

mycoplasma [33] before being used for experi- ments.

Cells m approximately the fifth to the fifteenth

passage were raised in antibiotic-free modified Ea-

gle’s minimum essential medium (Gibco, Grand Island, NY; Cat. No. 410-1500) supplemented with

15% heat-inactivated fetal calf serum (Microbio-

logical Associates, Walkersville, MD) and 2 mM

L-glutamine. Eagle’s minimum essential medium

was modified in our laboratory to contain 25 mM

glucose and 35 mM NaHCO,, along with 10 ml each of minimum essential medium non-essential

amino acids (X 100) minimum essential medium

essential amino acids (X 50) and minimum essen-

tial medium vitamins (X 100) (all from Gibco,

Grand Island, NY) added per liter of medium, and the pH was adjusted to 7.4-7.6.

For experimental work, 0.5 . lo6 cells of a given

line were plated onto 6-cm petri dishes (Falcon Labware, Cockeysville, MD) and brought to con-

fluency in 3 days in antibiotic-free media contain- ing 15% fetal calf serum. All cells were raised in a

moisturized incubator at 37°C in an atmosphere

of 95% sir/5% CO,. All sterile work was done in a

vertical laminar-flow stainless steel hood (Baker

Co., ME). 3 H-labeled ganglioside medium was prepared in

modified minimum essential medium containing

10% fetal calf serum/2 mM L-glutamine/lOO U/

ml penicillin/100 pg/ml streptomycin. An ap- propriate volume of the labeled ganglioside solu-

tion in chloroform/methanol, (2 : 1, v/v) was

blown to dryness with a sterile N, stream in a sterile lOO-ml bottle. The dried material was taken up in culture medium, sonicated for 30 min in a bath-type sonicator and tested for sterility by in- cubation overnight at 37°C.

To begin the experiment, the unlabeled medium from confluent cells was removed by aspiration and ganglioside medium added, usually at 5 ml/ dish. At the end of the designated experimental times, the labeled medium was removed and cells

washed first with a 3 ml rinse of phosphate-

buffered saline (Ca2+ and Mg’+-free containing

5.6 mM glucose, pH 7.2-7.5) and then with three consecutive phosphate-buffered saline rinses at 2

ml each, so that washings were free of radioactiv- ity. The cells were harvested by trypsinization for

10 min at room temperature with 2 ml of enzyme solution prepared as follows: 10 ml pancreatin, 4 ml minimum essential medium essential amino

acids, 2 ml minimum essential medium non-essen-

tial amino acids, 1 ml minimum essential medium

vitamins and 1.5 ml phenol red solution (all ob-

tained from Gibco, Grand Island, NY) made up to

100 ml with phosphate-buffered saline containing 5.6 mM glucose and 0.54 mM disodium ethylene- diaminotetracetic acid at pH 7.2-7.4. The reaction

was stopped with the addition of 0.2 ml fetal calf

serum. Cell pellets were collected by centrifugation at

approx. 100 x g for 5 min. These were rinsed twice

with 2 ml phosphate-buffered saline and analyzed

immediately or kept frozen ( - 20°C) until analysis

for 3 H-labeled lipids. Extraction and analysis of lipids in fibroblasts.

The fibroblast cell pellets were taken up in 200 ~1 water and sonicated with 15-s bursts of pulsed

(20% cycle at step 3) ultrasound (sonifier Model W185, Heat Systems Sonication Inc., Plain View,

Long Island, NY) in an ice-bath. After removing

an aliquot for protein determination [34], the sonicated homogenate was extracted overnight with

5 ml chloroform/methanol (2: 1, v/v) [35]. The

sample was centrifuged to remove the clear ex-

tract, and the pellet washed with 3 ml chloroform/ methanol (2 : 1, v/v). The combined extracts were

taken to dryness and redissolved in 5 ml of chloro- form/methanol (2: 1, v/v). After the addition of

1 ml water, the sample was vortexed and centri- fuged briefly to separate the phases. The aqueous

upper phase was withdrawn and saved. The lower phase was washed with 2 ml of chloroform/ methanol/water (3 : 48 : 47, v/v), and the re- sultant upper phase combined with the first aque- ous phase for the analysis of gangliosides. The

recovery of G,, and G,, ganglioside is more

than 90% in the aqueous upper phase. An aliquot from the upper and lower phases was taken in minivials for determining radioactivity. After re- moving the solvent in the vials under nitrogen, the

ganglioside sample from the upper phase was dis- solved in 250 ~1 water before the addition of 4 ml Scintiverse (Fisher Scientific Co., Medford, MA). The lipids in the vial from the lower phase were dissolved directly in Scintiverse. Radioactivity was determined in a Packard Tricarb liquid scintilla- tion spectrometer.

The gangliosides contained in the upper phase were purified on a small BondElut Cl8 cartridge containing 100 mg of the absorbent according to Williams and McCluer [36]. The whole sample was perbenzoylated to separate and quantitate monosialogangliosides by HPLC according to Bremer et al. [30].

The total lipids after removal of the solvent from the lower phase were subjected to alkaline methanolysis in 1 ml of 0.6 M methanolic NaOH for 1 h at room temperature. After the addition of 1.3 ml of 0.6 N methanolic HCl, 4.6 ml chloroform were added to make the solution chloroform/ methanol 2 : 1, v/v. It was then mixed with 1.4 ml water, vortexed and briefly centrifuged at low speed to clarify the phases. After discarding the upper phase, the lower phase was washed with 2.8 ml of chloroform/methanol/ water (3 : 48 : 47, v/v) and again centrifuged. The resultant lower. phase was taken to dryness under nitrogen and the lipid sample redissolved in 0.5 ml chloroform. It was applied on a small column containing 180 mg silicic acid (Unisil, Clarkson Chemical Co., PA), slurried and packed in chloroform in a short Pasteur pipette. The sample tube was rinsed three times with 0.5 ml of chloroform for transferring the lipids completely onto the column. The chloro- form eluant (2 ml) consisted of unabsorbed non- polar lipids (fatty acid methyl esters, fatty acids and cholesterol). The column was further eluted with 5 ml acetone/methanol (9 : 1) to remove the neutral glycosphingolipids after which sphingo- myelin was eluted with 3 ml of methanol. Each fraction was taken to dryness, redissolved in a known volume of chloroform/methanol (2 : 1, v/v) and an aliquot taken for determining the radioac- tivity. The glycosphingolipids were further sep- arated and quantitated by HPLC with a solvent gradient of dioxane in hexane according to Ull- man and McCluer [37]. The eluant for each peak was collected in vials for counting the radioactivity present in each peak.

241

Results

Uptake and metabolism of labeled G,, by normal fibroblasts as a function of ganglioside concentration in the medium

[ 3 H]Sphingosine-labeled G MZ at different con- centrations was added to normal skin fibroblasts grown to confluency in 6-cm petri dishes. After 10 days in culture exposed to the radioactive medium, the cells were washed free of radioactivity with phosphate-buffered saline and harvested by tryp- sinization. The uptake of labeled G,, by the cells over a lo-day period, expressed as the amount of radioactivity in the total lipid extract in chloro- form/methanol (2 : l), is shown in Fig. 1. Cell uptake increased progressively as the concentra- tion of G,, in the medium was raised from 2.5

l8OC 4 :

? I’

0 ,’ x 140- ,I’

2 6 IO 14 18 22 26

GM2 Con. in the Medium (PM_)

Fig. 1. Uptake of labeled G,, (21000 cpm/nmol) by fibro- blasts during 10 days in culture varying the concentration of ganglioside in the medium. After 10 days, the cells were washed free of radioactivity and harvested by trypsinization. Total lipids (A- A) were extracted with chloroform/methanol (2: 1) and partitioned with water to obtain the unhydrolyzed ganglioside in the aqueous upper phase (0 - 0) and the metabolized lipid products in the organic lower phase

(.- l ), as described. Each point represents the average of three separate dishes in which cpm/mg protein did not vary over 2%.

242

PM and appeared to plateau at a concentration of about 19 PM G,,. The rate at which the cells

metabolized the ganglioside, measured by the

amount of radioactivity (approx. 60%) in the lower organic phase of the total lipid extract, was con-

stant until the concentration of the medium

reached was 19 PM. Beyond this concentration,

there was a rapid uptake of G,,, but the rate of

metabolism was slower compared with the lower concentrations.

The lower phase lipids were subjected to mild

alkaline hydrolysis and fractionated on a small silicic acid column in chloroform, acetone/

methanol (9: 1) and methanol fractions. About

60% of the radioactivity in the lower phase lipids was associated with neutral glycosphingolipids in the acetone/methanol fraction. The remainder was

almost equally divided between the chlorofrm and

methanol fractions, respectively (Fig. 2). TLC of the chlorcform fraction on a silica gel G plate with

hexane/diethyl ether/acetic acid (90 : 10 : 1) as

56

2 48

s

.c 40 10 b

; 32 a, s c 24

2 6 10 14 18 22 26

GM2 Con. in the Medium (PM)

Fig. 2. Metabolic profile of the labeled G,, taken up by

fibroblasts maintained in culture for 10 days varying the con-

centration of ganglioside in the medium. The metabolic prod-

ucts of GMu12 present in the lower phase (Fig. 1) were subjected

to mild alkaline hydrolysis and fractionated on a small sihcic

acid column into chloroform (0 -0) acetone/methanol

(Acetone-MeOH) (A - A) and methanol (MeOH)

(=- n ) fractions as outlined. Each point is the average of

three determinations as stated in Fig. 1.

the solvent system showed that all the radioactiv-

ity was contained in spots corresponding to free fatty acids and fatty acid methyl esters. Similarly,

TLC of the methanol fraction with chloroform/ methanol/water (65 : 25 : 4) as the solvent system

showed that all the radioactivity was present as

sphingomyelin. It was clear that the cells metabo-

lized the labeled G,, via neutral glycosphingoli- pids and incorporated radioactivity into

sphingomyelin.

Uptake and metabolism of labeled G,, by normal

fibroblasts as a function of time in culture with the

ganglioside

When the uptake was studied as a function of time in days at a fixed concentration of 26 PM G M2 in the medium (Fig. 3), the radioactivity in the total lipid extract until day 5 was found to be

mostly in the upper phase containing unhy- drolyzed ganglioside. There was little metabolic

breakdown of the radioactive G,, taken up by the

cells during this time, since the lower phase con-

tained only very small amounts of radioactivity. After this initial lag period, the ganglioside was

metabolized rapidly as it was taken up so that radioactivity in the lower phase from the lipid

extract sharply increased from day 5 to day 10.

About 70-80% of the radioactivity from the metabolized ganglioside was found in the glyco-

sphingolipid fraction with the rest being divided

almost equally between sphingomyelin and fatty acid in ester linkage in the methanol and chloro-

form fractions, respectively (Fig. 4). These studies

indicate that uptake and increased metabolism into lower phase lipids occurred at a maximal concentration of about 20-26 PM G,, in the

medium with the cells maintained for 10 days in culture. Cell viability remained unaltered during

this experimental period.

Deficient G,, metabolism in fibroblasts from pa- tients with G,, gangliosidosis

Table I shows the results obtained with fibrob- lasts from patients with various types of G,,- gangliosidosis in comparison with control subjects.

It is clear that cells from patients with Tay-Sachs, Sandhoff and AB variant forms of the disease could not effectively metabolize the ingested labeled G M2. In these experiments, 84-93s of the

243

Days in Culture with 26 pM GM2 in Medium

Fig. 3. Uptake over time of radioactive G,, (21000 cpm/nmol) by fibroblasts maintained in culture medium containing G,, at an initial concentration of 26 pM. At the indicated time intervals, the cells were washed free of radioactivity and harvested by trypsinization. Total lipids (A --a) were ex- tracted with chloroform/methanol (2 : 1) and partitioned with water to obtain the unhydrolyzed ganglioside in the aqueous upper phase (0 - 0) and the metabolized lipid products in the organic lower phase (O- l ), as described. The points are the averages of three determinations as in Fig. 1.

total radioactivity in the lipid extracts from these cells partitioned into the aqueous upper phase as unhydrolyzed ganglioside when the extracts were washed in the absence of KC1 with water and chloroform/methanol/water (3 : 48 : 47) accord- ing to the Folch procedure [35]. In contrast, cells from control subjects metabolized this ganglioside with 50-608 of the total radioactivity in the lower organic phase of the lipid extract washed by the same procedure. Cells from the obligate carriers of Tay-Sachs disease also metabolized about 40-50% of the ingested ganglioside as seen from the radio- activity in the lower-phase lipids.

HPLC of the upper-phase gangliosides showed that nearly all of the radioactivity is present in the

36 -

32 -

28 -

24 -

20 -

16 -

12 -

Acetone - MeOH

_I / MeOH Fraction

:I#, Fraction

2 4 6 8 10

Days in Culture with 26 pM G,,,(2 in Medium

Fig. 4. Metabolic profile of the labeled G,, taken up by fibroblasts maintained in culture for various periods of time with 26 gM G,, in the medium. At the indicated time inter- vals, the metabolic products of G,, present in the lower phase (Fig. 3) were subjected to mild alkaline hydrolysis and fractionated on a small silicic acid column into chloroform

@- q ), acetone/methanol (Acetone-M&H) (A --n)

and methanol (MeOH) W- n ) fractions as outlined. The points are the averages of three determinations as in Fig. 1.

G,, peak (Table II). The small amount of radio- . .

activity seen as GM, in G,, gangliosidosis cells could very well represent a shoulder of the major G,, peak. This provides direct evidence for the metabolic defect in G MZ hydrolysis in cells from diseased patients. Even with controls and carriers, only a small amount of radioactivity is seen in the G,, peak, showing that when control fibroblasts metabolized the labeled GMM2, the degradation pro- gressed rapidly with little buildup of radioactivity in the immediate product, GMJ.

244

TABLE I TABLE II

G M2 METABOLISM IN FIBROBLASTS FROM CON-

TROLS AND PATIENTS WITH VARIOUS TYPES OF G,,

GANGLIOSIDOSIS

RADIOACTIVITY IN THE GANGLIOSIDES OF

FIBROBLASTS MAINTAINED IN CULTURE WITH

LABELED G,,

Confluent cells in 6-cm petri dishes were maintained in culture

for 10 days with 20 PM labeled GM2 (approx. 40000 cpm/nmol)

in modified Eagle’s minimum essential medium containing 10%

fetal calf serum. Chloroform/methanol (2 : 1) extracts of these

cells were partitioned with water to remove unhydrolyzed G,,

into the aqueous upper phase while retaining the metabolic

products in the organic lower phase. The results are the mean

of two or three separate experiments for each line. In no case

did the percent upper phase differ over two percentage points.

After purifying the gangliosides present in the aqueous upper

phase from washed lipid extracts, monosialogangliosides were

separated and quantitated by HPLC. Individual peaks were

collected and counted for radioactivity. The results are the

mean of duplicate or triplicate experiments and did not differ

more than two percentage points for each line.

Cell line) % total radioactivity

(No.) of the upper phase

Cell line cpm/mg protein % in the

(No.) Upper phase Lower phase upper phase

Infantile Tay-Sachs

1172 324000

3 339000

5 213800

6 182000

I 175 300

5 791 463 000

Infantile Sandhoff

306 336 200

AB variant

3015 a 95 909

4036 256 900

Controls

2493 172200

1028 69424

Obligate carriers of

infantile Tay-Sachs

248 138000

600 181000

36000 90

26 760 93

26 160 89

18550 91

21090 89

89 200 84

27910 92

14827 87

25 900 91

156500 52

101665 40

121000 53

128000 58

a In experiments with this AB variant line, the culture medium

contained only 4 gM GM1 of specific activity 81000 cpm/

nmol.

When the lower-phase lipids from the cells that metabolized the labeled ganglioside were fractionated on a silicic acid column, 60-70% of the radioactivity was found with the sphingolipids in the acetone/methanol (9 : 1) fraction. Further characterization of the sphingolipids by HPLC indicated that about half of the radioactivity was in the ceramide fraction (Table III). This clearly showed that the labeled G,, was rapidly metabo- lized in these cells via G,,, ceramide dihexoside and ceramide monohexoside with radioactivity

Infantile Tay-Sachs

Infantile Sandhoff

AB variant

Controls

Obligate carriers of

infantile Tay-Sachs

1172

306

3015

2493

1028

248

600

G M3 GM,

12 88

11 89

12 88

13 87

19 81

13 87

10 90

building up mainly in the ceramide pool of the cell. Small amounts of radioactivity were also found in ceramide trihexoside, globoside, sphingomyelin and in ester-linked fatty acids, showing that the labeled ceramide was reutilized

TABLE 111

LABELING PATTERN OF INDIVIDUAL NEUTRAL

SPHINGOLIPIDS IN FIBROBLASTS THAT METABO-

LIZED SPHINGOSINE-LABELED G,, ADDED TO THE

CULTURE MEDIUM

The sphingolipids in the acetone/methanol (9 : 1) fraction were

separated and quantitated by HPLC. Individual peaks were

collected and counted for radioactivity. The results are the

mean of two or three separate experiments for each cell line.

CMH, ceramide monohexoside; CDH, ceramide dihexoside;

CTH, ceramide trihexoside.

Cell line (No.) % of total radioactivity in each component

Ceramide CMH CDH CTH Globoside

Control 1028 50 24 13 8 5

Obligate carriers of

infantile Tay-Sachs 248 63 7 4 18 7

600 66 5 6 17 6

245

for the synthesis of complex sphingolipids or was converted to a labeled fatty acid.

Endogenous glycosphingolipid composition of fibroblasts

Fibroblasts from controls, patients and obligate carriers of Tay-Sachs disease were grown to con- fluency in medium containing 15% fetal calf serum without added ganglioside. Cells harvested from one or two lo-cm petri dishes containing about 1-3 mg protein were routinely used for the extrac- tion of total lipids. Monosialogangliosides and neutral glycosphingolipids were isolated and analyzed by HPLC. The chromatographic pattern of monosialogangliosides from control and Tay- Sachs fibroblasts is shown in Fig. 5. The sample chromatographed from control lines is equivalent to 0.6 mg of cellular protein and that from the Tay-Sachs line is equivalent to 0.16 mg of cell protein. Peaks I, II and III in the control corre- spond, respectively, to Ghi13, GM2 and G M1 gang- lioside, with G,, being the major component. In

Fig. 5. HPLC analysis of the monosiaIogangIiosides of human skin fibroblasts grown to confluency without addition of labeled ganglioside to the medium. Ganghosides present in the cells were extracted and purified as given in Methods. After per- benzoylation, they were separated and quantitated by HPLC according to Bremer et al [30].

Tay Sachs fibroblasts, the G,, ganglioside is in- creased over that of G,,. Although G,, escapes detection in this small quantity of ganglioside analyzed compared to the control sample, the ac- cumulation of G,, in Tay-Sachs compared to control cells is striking. On a protein basis, the accumulation of G,, in this case of Tay-Sachs fibroblasts is about 8-fold compared to the con- trol.

The quantitative distribution of G,, and G,, in fibroblasts derived from variant forms of G,,- gangliosidosis compared to the control is shown in Table IV. G,, is the major component of normal fibroblasts, but very small amounts of G,, can be detected in a few control cell lines. G,, occurs only in traces much below the concentration of G MZ- It can be identified sometimes if larger amounts of cells with at least 6-10 mg protein are used for the extraction and analysis. On the other hand, accumulation of endogenous G,, in all cell

TABLE IV

ENDOGENOUS MONOSIALOGANGLIOSIDE COMPOSI- TION OF FIBROBLASTS GROWN TO CONFLUENCY IN 15% FETAL CALF SERUM WITHOUT ANY ADDITION OF GANGLIOSIDE TO THE MEDIUM

n.d., not detected

Cell line

(No.1

Control 2493 1028

Tay-Sachs 1172

3 5 6 7

5 791

Sandhoff 306

AB variant 3015 4036

Obligate carriers of infantile Tay-Sachs

248 600

nmoI/mg protein

G hi3 G h42

1.42 0.23 0.65 n.d.

1.72 1.93 0.88 0.39 0.77 0.52 1.84 3.43 1.48 3.79 0.80 1.36

1.81 0.95

1.84 0.45 4.80 1.90

0.33 n.d. 0.31 n.d.

246

lines from patients with G ,,-gangliosidosis com- pared to controls can be clearly demonstrated in cells with only l-3 mg protein. Although G,, concentration is also increased in cell No. 4036, the primary metabolic defect in this line is the inability to hydrolyze G,, as shown in Table I. Thus, in all types of GM,-gangliosidosis, a meta- bolic defect for the hydrolysis of exogenous, labeled GM, has been clearly demonstrated along with an accumulation of endogenously synthesized G MZ ganglioside.

The liquid chromatographic profile of neutral glycolipids from controls and Sandhoff fibroblasts grown in 15% fetal calf serum without the addition of labeled ganglioside is shown in Fig. 6. As in visceral organs, the glycolipid pattern of control fibroblasts consists of ceramide monohexoside, caramide dihexoside, ceramide trihexoside and globoside with ceramide trihexoside being the major component. The peak ‘x’ seen between ceramide trihexoside and globoside has not been characterized. Sandhoff fibroblasts exhibit a dis- tinctly different pattern with marked accumulation of globoside. Table V shows the glycolipid con-

Fig. 6. HPLC analysis of neutral glycosphingolipids of human

skin fibroblasts grown to confluency without addition of labeled

ganglioside to the medium. From the total lipid extract, neutral

glycosphingolipids were isolated by fractionation on a silicic

acid column as outlined in Methods. After perbenzoylation,

they were separated and quantitated by HPLC according to

Ulhnan and McCluer [37].

TABLE V

ENDOGENOUS NEUTRAL SPHINGOLIPID COMPOSI-

TION OF FIBROBLASTS GROWN TO CONFLUENCY IN

15% FETAL CALF SERUM WITHOUT ADDITION OF

GANGLIOSIDE IN THE MEDIUM

CMH, ceramide monohexoside, CDH, ceramide dihexoside;

CTH, ceramide trihexoside.

Cell lines (No.) nmol/mg protein

CMH CDH CTH Globoside

Control

2493 1.10 0.22 3.59 1.13

Infantile Tay-Sachs

6 1.23 0.29 2.74 2.84 7 1.20 0.21 2.71 1.99

Infantile Sandhoff

306 0.95 0.45 2.70 11.15

Infantile AB variant

3015 1.11 0.26 3.84 2.05 4036 2.32 0.59 4.25 1.74

Obligate carriers of

infantile Tay-Sachs

248 0.72 0.34 9.70 3.65

600 0.76 0.50 6.70 2.60

centrations in cell lines derived from variant forms of GM,-gangliosidosis and control subjects. In general, the pattern of sphingolipid composition in cells from patients with Tay-Sachs and AB variant type disease and unaffected subjects is very similar except for small differences in their total amount in individual cell lines. On the other hand, in the infantile Sandhoff disease, globoside is markedly elevated in fibroblasts, as has been shown previ- ously in several visceral organs from patients with this disorder. The absence of any accumulation of asialo G,, in fibroblasts from either Sandhoff or Tay-Sachs disease indicates that it is not a normal metabolite in these cells.

This approach has potential application for un- mistakable prenatal diagnosis of Tay-Sachs dis- ease. Amniotic fluid cells from a healthy control efficiently metabolized 50% of the labeled GM, taken in from the medium, whereas amniotic fluid cells from a fetus at risk of Tay-Sachs disease showed that 84% of radioactivity in the total lipid extract was present as unhydrolyzed GM,. This metabolic block for hydrolyzing G,, is further

241

substantiated by the accumulation of endogenous G,, (2.51 nmol/mg protein) synthesized by these cells, while no endogenous G,, could be detected in control cells.

Discussion

This approach to examine the kinetics of uptake and intracellular metabolism of labeled G,, ad- ded to the culture medium of fibroblasts from controls and patients with G,, gangliosidosis pro- vides direct insight into what does happen in situ in the cell to this ganglioside. G,, added in the culture medium was taken up by cultured fibro- blasts as a function of concentration and time of exposure of the ganglioside in the medium. Signifi- cant metabolism of the ganglioside was not clearly evident until after 5 days in culture. This lag period for active metabolism may be due to initial incorporation of the labeled G,, into the cellular membranes before its eventual transfer to lyso- somes or to slow lysosomal G M2 degradation. Nev- ertheless, following a lo-day period of exposure in medium containing the labeled GMz, cells in cul- ture that are able to metabolize the labeled gang- lioside can be clearly distinguished from those lacking this functional ability. Thus, in cells from all types of G,, gangliosidosis, 84-90% of the ingested G MZ remained unhydrolyzed while 50-608 was metabolized in control lines. In par- ticular, activator-deficient AB variant form of the disease, which is associated with normal in vitro activity of Hex A and B against the synthetic substrate, can be clearly identified by this proce- dure. Thus, the biologic approach, by using the living cell to examine the course of metabolism of ingested labeled GMZ, provides the most reliable system to diagnose the metabolic defects in situ. In control cells, the complete lysosomal degradation

of G,, was so rapid that very little radioactivity appeared in the most immediate product G,, or in the subsequent steps forming ceramide mono- hexoside and ceramide monohexoside, respec- tively. Much of the radioactivity was present in the final ceramide product of glycohexoside catabo- lism.

The endogenous concentration of G,, was in- creased over the controls in all cell lines lacking the ability to metabolize this ganglioside even when

no GM, was added to the culture medium. Studies by Dawson et al. [38] (using TLC and GLC) have indicated that traces of GM, may be detectable in fibroblasts from some patients with GM, ganglio- sidosis, while it is absent in control cell lines. They have cautioned against the use of TLC alone for the identification of GM, in cultured fibroblasts because of overlap with heterogenous GM, and G,, present in the cells. Increased amounts of G M2 have been detected in fibroblasts from Sandhoff disease by Fishman et al. [39] using densitometry following TLC, and by Philippart [40] using GLC analysis following TLC identifi- cation. From densitometric analysis of ganglio- sides separated on high-performance TLC plates, Pullarkart et al. [41] have reported that the con- centration of G,, is elevated at late confluency in fibroblasts from Tay-Sachs and Sandhoff disease. We have used a very sensitive HPLC technique for the quantitative analysis of monosialogangliosides to demonstrate increased amounts of CM, in fibroblasts from all types of GM, gangliosidosis. The extent of accumulation, however, is quite vari- able among the different cell lines. This variation could be due to differences in the passage number and proliferative rate of the different cell lines under conditions of cell culture employed. The GM, ganglioside detected in these cells was pre- sumably derived from endogenous syntheses. It did not enter the cells from the culture media because the fetal calf serum used in these experi- ments contained only GM, (1.03 nmol/ml) and no GM, could be detected despite the high sensitivity of our HPLC technique. The fetal calf serum we used contained the following neutral sphingolipids (concentrations expressed in parenthesis in nmol/ ml): non-hydroxy fatty acid-ceramide (1.88), hy- droxy fatty acid-ceramide (0.83) non-hydroxy fatty acid-glucosyl ceramide (0.44), non-hydroxy fatty acid-galactosyl ceramide (0.20), hydroxy fatty acid-glucosyl and galactosyl ceramide monohexo- side (0.49) non-hydroxy fatty acid-ceramide di- hexoside (0.35), non-hydroxy fatty acid-ceramide trihexoside (0.11) and non-hydroxy fatty acid- globoside (0.16).

The experimental approach presented here for studying defects in G,, metabolism in situ in cultured fibroblasts has the obvious advantage over in vitro test-tube assays for enzyme activity em-

248

ploying non-specific unnatural substrates, unphys- iological detergents for the natural lipid substrate and artificial conditions of pH, buffer, cofactors and substrate concentration.

Acknowledgement

This work was supported by a grant from the National Institutes of Health, GM 28207.

References

1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

16

O’Brien, J.S. (1983) in Metabolic Basis of Inherited Dis-

eases (Stanbury, J.B., Wyngaarden, J.B., Fredrickson, D.S.,

Goldstein, J.L. and Brown, M.S., eds.), pp. 945-969, Mc-

Graw-Hill, New York

Kaback, M.M. (1972) Methods Enzymol. 28, 862-867

Kresse, H., Fuchs, W., Glossl, J., Holtfrerich, D. and Gil-

berg, W. (1981) J. Biol. Chem. 256, 12926-12932

Fuchs, W., Navon, R., Kaback, M.W. and Kresse, H. (1983)

CIin. Chem. Acta 133, 253-261

Li, Y.-T., Hirabayashi, Y. and Li, S.-C. (1983) Am. J. Hum.

Genet. 35, 520-522

Navon, R., Padeh, B. and Adam, A. (1973) Am. J. Hum.

Genet. 25, 287-293

Vidgoff, J., Buist, N.R.M. and O’Brien, J.S. (1973) Am. J.

Hum. Genet. 25, 372-381

Dreyfus, J.C., Poenaru, L. and Svennerholm, L. (1975) N.

EngI. J. Med. 292,61-63

Navon, R., Greiger, B., Ben-Yoseph, Y. and Rattazzi, N.C.

(1976) Am. J. Hum. Genet. 28, 339-349

O’Brien, J.S., Tennant, L., Veath, M.L., Scott, C.R. and

BucknaIl, W.E. (1978) Am. J. Hum. Genet. 30, 602-608

O’Brien, J.S., Norden, A.G.W., Miller, A.L., Frost, R.G.

and Kelly, T.E. (1977) Chn. Genet. 11, 171-183

Poulos, A., Holding, J. and Carey, W.F. (1982) Clin. Chim.

Acta 120, 331-340

Erzberger, A., Conzelmann, E. and Sandhoff, K. (1980)

Clin. Chim. Acta 108, 361-368

TaIlmann, J.F., Brady, R.O., Quirk, J.M., Villalba, M. and

Gal, A.E. (1974) J. Biol. Chem. 249, 3489-3499

Sandhoff, K., Conzelmann, E. and Nehrkom, H. (1977)

Hoppe-Seylers Z. Physiol. Chem. 358, 779-787

Li, S.-C., Hirabayashi, Y. and Li, Y.-T. (1981) J. Biol.

Chem. 256, 6234-6240

17 Harger, K. (1983) CIin. Chim. Acta 135, 89-93

18 Porter, M.T., Fluharty, A.L., Harris, S.E. and Kihara, H.

(1970) Arch. Biochem. Biophys. 138, 646-652

19 Barton, N.W. and Rosenberg, A. (1975) J. Biol. Chem. 250,

3966-3971

20 Tanaka, H. and Suzuki, K. (1978) J. Neurol. Sci. 38,409-419

21

22

23

24

25

26

27

28

29

30

31

32

33

34

35

36

37

38

39

40

41

Chen, W.W., Moser, A.B. and Moser. H.W. (1981) Arch.

Biochem. Biophys. 208, 444455

Kudoh, T., Sattler, M., Malmstrom, J., Bitter, M.A. and

Wenger, D.A. (1981) J. Lab. CIin. Med. 98, 704-714

Beaudet, A.L. and Mans&reck, A.A. (1982) Biochem. Bio-

phys. Res. Commun. 105, 14-19

Spence, M.W., Clarke, J.T.R. and Cook, H.W. (1983) J.

Biol. Chem. 258, 8595-8600

Fishman, P.H., Bradley, R.M., Horn, B.E. and Moss, J.

(1983) J. Lipid Res. 24, 1002-1011

Fishman, P.H., Moss, J. and Vaughan, M. (1976) J. Biol.

Chem. 251, 4490-4494

Itoh, T., Li, Y.-T., Li, S-C. and Yu, R.K. (1981) J. Biol.

Chem. 256, 165-169

Schwarzmann, G. (1978) B&him. Biophys. Acta 529,

106-114

Momoi, T., Ando, S. and Nagai, Y. (1976) Biochim. Bio-

phys. Acta 441,488-497

Bremer, E.G., Gross, S.K. and McCluer, R.H. (1979) J.

Lipid Res. 20, 1028-1035

Kadowaski, H., Bremer, E.G., Evans, J.E., Jungalwala, F.B.

and McCIuer, R.H. (1983) J. Lipid Res. 24, 1389-1397

Raghavan, S.S., Gajewski, A. and Kolodny. E.H. (1981) J.

Neurochem. 36, 724-731

Chen, T.R. (1977) Exp. Cell Res. 104, 255-262

Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall,

R.J. (1951) J. Biol. Chem. 193, 265-275

Folch, J., Lees, M. and Sloane-Stanley, G.H. (1957) J. Biol.

Chem. 226,497-509

Williams, M.A. and McCluer, R.H. (1980) J. Neurochem.

35, 266-269

Ullman, M.D. and McCluer, R.H. (1978) J. Lipid Res. 19.

910-913

Dawson, G., Matalon, R. and Dorfman, A. (1972) J. Biol.

Chem. 247, 5951-5958

Fishman, P.H., Bradley, R.M., Brown, M.S., Faust, J.R. and

Goldstein, J.L. (1978) J. Lipid Res. 19, 304-308

Philippart, M. (1976) in Glycohpid Methodology (Witting,

L.A., ed.), pp. 247-294. Am. Oil. Chem. Sot. IL Pullarkart, R.K., Reha, H. and Beratis, N.G. (1980) Bio-

them. Biophys. Res. Commun. 92, 149-154