Eur. J. Biochem. 116, 565-571 (1981) 6 FEBS 1981

Chloroquine-Induced Accumulation of Gangliosides and Phospholipids in Skeletal Muscles Quantitative Determination and Characterization of Stored Lipids

Olle NILSSON, Pam FREDMAN, Georg W. KLINGHARDT, Henri DREYFUS, and Lars SVENNERHOLM

Department of Psychiatry and Neurochemistry, University of Goteborg, and Max Planck Institute for Brain Research, Frankfurt am Main

(Received January 5, 1981)

High doses of the lysosomotropic drug chloroquine result in lipid storage in many organs in animals. We used miniature pigs, type Gottingen, to study the lipid accumulation in skeletal muscle after chloroquine intoxication for more than 200 days. The lipids of the quadriceps muscle in intoxicated and in age-matched control pigs were characterized and determined. The lipid storage was larger in skeletal muscle than in any other organ of the intoxicated pigs. The concentration of phospholipids was increased threefold, acidic phospholipids relatively more than neutral ones. The lysosome-specific acidic phospholipid bis(monoacylg1yceryl)phosphate content was almost %-fold larger in the intoxicated pigs than in the controls. Cholesterol was increased slightly more than the phospholipids, but there was no particular accumulation of cholesteryl esters, which has been shown to occur in the liver. For the first time a storage of gangliosides, relatively more pronounced than of other lipids, was demonstrated in skeletal muscle in drug-induced lipidosis. The concentration of total gangliosides was increased 10-15-fold, and the pattern of gangliosides showed some distinct changes resulting in at least a 100-fold increase in the concentration of ganglioside GMl (I13NeuAc-GgOse4Cer).

Since the first report of drug-induced lipidosis by Yama- mota et al. [I], more than twenty amphiphilic, cationic drugs with different therapeutic indications have been shown to cause a lysosomal storage disease, named drug-induced lipidosis [2].

One of the best known and most thoroughly investigated compounds is chloroquine, 7-chloro-4(4-diethylamino-l- methyl-buty1amino)quinoline. Since World War I1 the com- pound has been widely used as an anti-malarial drug, but it has also proved useful in the treatment of rheumatoid arthritis and certain skin diseases. It has a wide variety of side-effects, the best known being retina lesions [3]. Long-term use of the drug has also been shown to cause a severe reversible vacuolar myopathy. The myotoxicity of chloroquine was first reported as early as 1948 by Nelson and Fitzhugh [4]. They fed rats small doses of chloroquine and observed severe microscopic muscle lesions. Their report seems to have fallen into oblivion until the myotoxic action of the drug was observed in humans [5-71. Similar lesions have now been produced in a number of animal species [8 - 1 I].

Abbreviations. Ganglioside abbreviations follow the nomenclature system of Svennerholm [la]. G M ~ , I13NeuAc-LacCer; G D ~ , I13(NeuAc)z- LacCer ; GM2, I13NeuAc-GgOse3Cer; GMl, I13NeuAc-GgOse4Cer; LMI, 1V3NeuAc-nLcOse4Cer; Fuc-(NeuAc)GMl, IVZFuc,I13NeuAc-GgO~e4- Cer ; Fuc(NeuGc)GMl, IV2Fuc,I13NeuGc-GgOse4Cer; GDI., IV3NeuAc, I13NeuAc-GgOse4Cer; GD1b, I13(NeuAc)z-GgOse4Cer; G T I ~ , IV3- (NeuAc)2,113NeuAc-GgOse4Cer; GTlb, IV3NeuAc,I13(NeuAc)2-GgOse4- Cer ; G Q I ~ , IV3(NeuAc)z,113(NeuAc)2-GgOse4Cer.

Enzymes. Vibrio cholerae sialidase or acylneuraminyl hydrolase (EC 3.2.1.18); a-~-fucoside fucohydrolase (EC 3.2.1.51).

Light and electron microscopic studies [4- 111 have shown severe muscular injury with fibre necrosis and extensive accumulation of numerous membranous bodies in the cyto- plasm. Histochemical investigations suggest the stored material to be glycogen [7] or phospholipids [9,10]. In one case Mastiglia et al. reported a general increase in neutral lipids and phospholipids [12]. To our knowledge this is the only report of a biochemical characterization of the stored material.

The purpose of this study was to isolate and characterize the storage material. With our newly developed extraction and separation procedures [13,14] is has been possible to make a quantitative determination and characterization of the ac- cumulated material, and the results are presented in this paper.

MATERIALS AND METHODS

Chemicals

The anion-exchange resin Spherosil-DEAE-Dextran was a gift from Institute Merieux (Lyon, France). Sephadex G-25 Fine and DEAE-Sepharose CL-6B were purchased from Pharmacia Fine Chemicals (Uppsala, Sweden). The anion- exchange resins, Spherosil-DEAE-Dextran and DEAE-Se- pharose CL-6B, were converted to the acetate from as de- scribed previously [ 131.

Precoated thin-layer plates Kiselgel 60, 20 x 20-cm; high- performance thin-layer plates, Kiselgel 60, 10 x 20-cm; and Kiselgel G, 230 -400 mesh, were obtained from Merck

566

(Darmstadt, FRG). D-Galactose, D-glucose, L-fucose, D-ga- lactosamine-HC1 and D-glucosamine-HC1, used as standards for analysis by gas-liquid chromatography, were all obtained from Pfahnstiehl Laboratories (Waukegan, IL, USA). Vibrio cholerae sialidase was purchased from Behringwerke (Mar- burg-Lahn, FRG) and a-L-fucosidase of bovine kidney was obtained from Boehringer (Mannheim, FRG). Dialysis tubing was obtained from Union Carbide (Chicago, IL, USA). Orga- nic solvents and other chemicals used were of analytical qua- lity. Phosphatidic acid, phosphatidylglycerol and inositol phosphoglyceride were all obtained from Sigma (St Louis, MO, USA). Diphosphatidylglycerol was obtained from PL Biochemicals (Milwaukee, WI, USA). Serine phosphoglyce- ride, ethanolamine phosphoglyceride and choline phospho- glyceride were all isolated from human brain. Bis(monoacy1- glycery1)phosphate was isolated from the spleen of a patient with Niemann-Pick disease typ C. Gangliosides, glycolipids and N-acetylneuraminic acid used as standards were prepared in this laboratory.

Tissue Material

Miniature pigs, mini-pigs, type Gottingen, a cross-breed between the Vietnamese and the Minnesota mini-pig, were used [15]. They were bred at the Neuropathologische Ab- teilung Max-Planck-Institut fur Hirnforschung (Frankfurt, FRG). Chloroquine intoxication was induced by mixing chloroquine diphosphate (Resochina) with the standard diet in a dose of 2.0 gikg diet. The control animals received the same diet without chloroquine. The intoxication period started when the animals were z 120-days old (two animals) and x200-days old, and continued for 184-230 days. The food intake of the animals varied considerably with age. By the beginning of the intoxication, the two youngest animals were just full-grown, with a body weight of about 15 kg and a food intake of approximately 500 giday. The food intake increased with age and varied between 700 - 1000 g/day by the end of the intoxication. Thus, the initial daily dose of chloroquine was about 70 mg/kg body weight and by the end of the intoxication about 30-40 mg/kg body weight de- pending on the food intake. This should be compared with the recommended dose of 250 mg/day (3 -4 mg/kg body weight) in the long-term treatment ofpatients with rheumatoid arthritis. The experimental animals received the chloroquine diet until the day they were killed. The quadriceps muscle was dissected out immediately after death, placed in an airtight polyethylene bag, frozen and stored at -20 “C until analyzed. The animals examined are summarized in Table 1.

Table 3 . Survey of the animals investigated

Treatment Age Body Dose Duration at weight of chloro- of death quine treatment

days kg g/kg diet days

Control 1 352 58 - -

2 359 49 3 306 57 - -

- -

Chloroquine- 1 352 55 2 223 intoxicated 2 360 48 2 230

3 486 48 2 184

Extraction of the Tissue

Quadriceps muscle (100 - 300 g) was homogenized in a scissor homogenizer with 3 vol. of distilled water at +4”C. To the homogenate was added 10 vol. of methanol and, after thorough mixing, 5 vol. of chloroform to give a final solvent volume ratio chloroformimethanoliwater of 4: 8 : 3 [I 31. After centrifugation at 1200 x g for 30 min the supernatant was collected, and the residue was reextracted with 20vol. of chloroformimethanoliwater (4: 8 : 3, by vol.). The combined supernatants were filtered through Celite on a Biichner funnel and used for quantitative analysis of cholesterol, phospho- lipids and gangliosides.

Isolation and Separation of Phospholipids

For the determination of phospholipids, aliquots of the total lipid extract, equivalent to 1 g of tissue, were evaporated to dryness and dissolved in 10 ml chloroformimethanoli water (60 : 30 : 4.5, by vol.). Salt and low-molecular-weight contaminants were removed on a 2 g Sephadex G-25 column [16]. The phospholipids were then separated into neutral and acidic phospholipids in the following way: lipid extract, cor- responding to 400 mg tissue, was dissolved in 4 ml chloroform/ methanol/water (60: 30: 4.5) in a small round-bottomed tube, mixed with 2 g DEAE-Sepharose CL-6B in acetate form, and equilibrated overnight. The mixture was poured onto a small column (7 mm internal diameter) packed with another 2 g of the same resin. The neutral lipids were eluted with 60 ml chloroform/methanol/water (60 : 30 : 4.9, some of the solvent was used for the complete transfer of all the resin from the tube to the column. The acidic phospholipids were eluted with 40 ml 0.05 M potassium acetate in methanol. The salt in the acidic fraction was removed by phase partition by adding chloroform and water to give a final solvent ratio, chloro- form/methanol/water, of 4: 2 : 1. The lipid P was quantified in the total lipid extract and the separated fractions with a modified Fiske-Subbarow method [17].

The phospholipids were identified by thin-layer chromatog- raphy using three different solvent systems : (a) chloroform/ methanol/water (65 : 25 : 4); (b) chloroform/methanol/17 M ammonia (70 : 30 : 5) ; and (c) chloroform/acetone/methanol/ acetic acid/water (50 : 20: 10 : 10 : 5). The bis(monoacylglycery1)- phosphate isolated from the muscle of the chloroquine- intoxicated animals had the same chromatographical be- haviour as bis(monoacylg1yceryl)phosphate isolated from the spleen of a patient with Niemann-Pick disease type C. The distribution of the phospholipids was determined by scanning the plates with a Zeiss KM3 chromatogram scanner at 450 nm, after visualization of the lipids with cupric acetate [18].

The neutral phospholipid fraction was used also for the quantitative determination of cholesterol [I 91.

Isolation qf Gangliosides

To the total lipid extract water was added to give a final solvent ratio, chloroform/methanol/water, of 4.0: 8.0: 5.6. After mixing and partition overnight the upper phase was removed, and the lower phase was repartitioned by addition of 1 .O vol. of methanol and 0.7 vol. of water to 2.0 vol. of the lower phase. After addition of isobutanol to prevent foaming, the combined upper phases were evaporated to dryness. The residue was dissolved in water and dialysed against running water for three days, then evaporated to dryness and dissolved in chloroform/methanol/water (60 : 30: 4.5). The crude gan- glioside extract was assayed for the total sialic acid and

567

system and the sarcoplasmic reticulum [9]. In many fibres the deposits consisted exclusively or predominantly of ac- cumulations of irregularly concentrically layered membranous bodies (Fig. 1 a). In other fibres the deposits showed a fine granular or amorphous ultrastructure, whereas accumulations of the membranous bodies were scanty or absent (Fig. 1 b). Storage material was also present in the vascular pericytes (Fig. 1 c).

ganglioside pattern, and then separated into monosialo, disialo, trisialo and tetrasialo ganglioside with the aid of a Spherosil-DEAE-Dextran column in acetate form. The gan- gliosides were eluted from the column with a discontinuous gradient of potassium acetate in methanol [14].

The gangliosides were quantified with the resorcinol method [20,21]. The composition of gangliosides in the different fractions was determined with thin-layer chromatog- raphy on standard or high-performance plates in three different solvent systems : (1) chloroform/methanol/0.25 %

monia (60 : 40: 9); (3) propanol/0.25 ”/, aqueous KCI (3 : I),

gliosides were visualized with resorcinol reagent, and quanti-

aqueous KCl (60: 35 : 8), (2) chloroform/methanol/2.5 M am- LIPID DETERMINATION

Phospholipids and Cholesterol

Table 2. The concentration of cholesterol increased ratios were by After chromatography the gan- The results of the lipid analyses are in

than fied by scanning Of the plates with a Zeiss KM3 chromatogram Scanner at 620 nm, and integration with a Varian CDS four times in the chloroquine-treated animals, The ratio of

cholesterol to phospholipids increased from 0.16 i n the computer [13]. control group to 0.25 in the chloroquine-intoxicated group. There was an approximately threefold increase in the con- Separation and Characterization of Individual Gangliosides centration of total phospholipids in the chloroquine-in-

The individual monosialo gangliosides from one chloro- toxicated animals. The increase of the acidic phospholipids quine-intoxicated pig and one control pig were isolated, and was relatively larger than that of the neutral phospholipids. their structure partly determined. Ganglioside G M ~ was All the major neutral phospholipids increased to the same isolated by column chromatography on silica gel with chloro- extent, but the composition of the acidic phospholipid fraction form/methanol/water (65: 25 : 4) as eluant. All the other gan- was changed. The proportion of diphosphatidylglycerol gliosides were isolated by preparative thin-layer chromatog- (cardiolipin) decreased from 35 of the acidic phospho- raphy. The gangliosides were purified to homogeneity with lipids in the control animals, to 17% in the intoxicated respect to their carbohydrate moiety, as judged by thin-layer animals. Bis(monoacylglycery1)phosphate was less than 1 7; chromatography before and after sialidase treatment. in the control animals, but represented ~ 1 5 % of the acidic

Component analyses of the isolated gangliosides were phospholipids in the chloroquine-treated animals. performed by determination of the sugar moieties as alditol acetates [22], fatty acids as the methyl esters after hydrolysis Gangliosides with 1 M HCI in methanol/water (82: 18), and the sphingosine

oxidation ,231, All gas chromatographic analyses sialic acid/g wet tissue, was 10-15 times larger in the chloro- were performed on a Perkin Elmer F-22 Gas Chromatograph quine-intoxicated animals than in the controls (Table 2). This connected to a Hewlett-Packard 3352B Lab Data System, The means that the ratio of ganglioside N-acetylneuraminic acid alditol acetates were analysed on a ECNSS-M column, to lipid P increased from 3 : 1000 in the control animals to

10 : 1000 in the chloroquine-treated animals. The proportion the Fatty acids on a 37; OV-1 column, and the sphingosine of N-glycoloylneuraminic acid was only 5% in the control bases on a 15 animals and in the intoxicated animals [25]. Owing to the low

Tris,maleate buffer pH 6.5 containing mM CaC12, The concentration of N-glycoloylneuraminic acid, no corrections hydrolysis was run for 6 h at 37 O c . The reaction was stop- were made in Fig. 2 or Tables 2 and 3 for the difference in ped by addition of 2,4 ml chloroform~methanol (2: viv), molar absorbance of N-acetylneuraminic acid and N-glyco- Released sialic acid and salt were removed by loylneruaminic acid when measured with the resorcinol

method. on Sephadex G-25 [16]. a-Fucosidase hydrolysis was per- formed as recently described [241. The gangliosides before and The distribution of the major monosialogangliosides, and

anol/water (60:32:7), and solvent systems 2 and 3 for Table 3. There was no great difference in the g a n g h i d e

bases as their corresponding aldehydes after hydrolysis and The gandioside concentration, as nmol

DEGS column. Hydrolysis with sialidase was performed in o,O1

after enzymic hydrolysis were examined by high-performance thin-layer chromatography developed in chloroform/meth-

disialo and po1Ysialo gangliosides from the and the chloroquine-treated Pigs is given in Fig. and

gangliosides, The bands were visualized with resorcinol or cupric acetate. Brain gangliosides as well as glucosyl, galac- tosyl, lactosq’l, globotriaosyl, globotetraosyl, neolactotetraosyl

terial.

distribution between the control and the chloroquine-treated animals, except for the changes in the proportions of the two gangliosides G M Z and LM1. The concentration of GMZ was less

48 nmol/g tissue in the three chloroquine-treated animals examined. Ganglioside LM1, sialosylneolactotetraosylcer- amide, was increased only about threefold, and represented only 4 o/, of the gangliosides in the intoxicated pigs, compared to 15% in the controls. The most frequent ganglioside was the ‘extraneural or visceral’ ganglioside, G M ~ . It was increased

and gangliotetraosyl ceramides were used as reference ma- than 0.5 nmol/g in the Pigs, ‘Ompared to 50, 34 and

RESULTS

Histological Examination of the Muscle

The storage process was most accentuated in the type I fibres. There were many disseminated and aggregated deposits of granular material positive to periodic acid/Schiff, which varied in staining intensity. This material was predominantly located ultrastructurally to lysosome-like autophagic vacu- oles, derived from the membranes of the transversal tubular

to almost the same extent as- the other gangliosides and constituted more than 40 % of the total gangliosides. There was also a relative increase of the disialogangliosides, z 13 yd in the control group and ~ 2 0 % in the intoxicated group, but there were only small changes in the proportions of the individual disialogangliosides. Gola was the predominant

568

Fig. 1. Quadriceps muscle of miniature pig , type Gottingen, after 253 days of chloroquine diphosphate intoxication (2.0 glkg diet). Perfusion fixation with formalin, postfixation with 3.9 % glutaraldehyde; Epon embedding; contrastation after Reynolds; magnification 5700 x . (a) Segments of two muscle fibres with accumulations of adjacent membranous cytoplasmic bodies of various size. Their membranes are mostly irregularly concentrically layered. Scattered glycogen granules. (b) Another muscle fibre region with accumulations of vacuoles of different size containing mixtures of amorphous and granular material and some residues of layered membranous bodies. Abundant free and single membrane bound aggregations of glycogen granules. (c) Large amounts of adjacent vacuoles in the pericyte (P) represented on both sides of the capillary lumen in the muscle. Their contents are amorphous or fine granular and show only small residues of myelin figures. The endothelial cells (E) exhibit no similar changes

Table 2. Lipid composition in skeletal muscle from chloroquinone-intoxicated mini-pigs and controls The isolation and quantitation were performed as described in Materials and Methods. The figures are the mean values significance was calculated with the Student's t-test

S.D., n = 3. The statistical

Treatment Cholesterol Phospholipids Ganglioside lo3 x Ganglioside Cholesterol/ _ _ ~ NeuAc NeuAc/lipid P lipid P

neutral acidic total

0.16 f 0.01 Control 1.6 & 0.1 8.5 f 0.4 1.9 f 0.05 10.3 f 0.4 29.0 & 4.9 2.8 f 0.6 Clilcroquinone-intoxicated 8.2 _+ 1.3" 24.4 f 3.8" 8.3 & 1.4" 32.7 f 5.3" 343 & 43' 10.5 2 O.Sb 0.25 f 0.01

~ ~ - -

a P < 0.01 ' P < 0.003

569

disialoganglioside and constituted z 50 % of them, while ganglioside GD3 (the disialoganglioside analogue to GM3) constituted only zi 10 of the disialogangliosides. The other disialogangliosides and the polysialogangliosides were identi- fied only by thin-layer chromatography and compared with known gangliosides. A disialoganglioside, with the same chro- matographic behaviour as GDlb, represented about 20 % of the

monosialo

Fig. 2. Concentration of individual mono.Fialogangliosides, and total disialo and polysialo gangliosides in skeletal muscle from chloroguine intoxicated mini-pigs and controls. The gangliosides were isolated and quantitated as described in Materials and Methods. The bars show the mean value f S.D. (H) Chloroquine-intoxicated animals, n = 3. (0) Control animals, n = 3

disialogangliosides. The most abundant polysialoganglioside was GTlb, which made up ~ 9 5 of them. Also small amounts of G T I ~ and GQlb could be detected by thin-layer chromatog- raphy.

The structure of the monosialogangliosides was deter- mined by component analysis (Table 4), enzymic plus acidic hydrolysis and identification of the degradation products by thin-layer chromatography. The amount of GMZ isolated from control pig muscle was not sufficient for component analysis. The structure was confirmed by enzymic and acidic hydrolysis. Authentic GMZ was hydrolysed under identical conditions and the products obtained had the same chromatographic behaviour, and the same proportions as the G M ~ of skeletal muscle from control pigs. Two fucogangliosides were identified in the fraction called ‘residual monosialo- gangliosides’ (Fig. 2, Table 3). These gangliosides were the dominating gangliosides in that fraction. They were identified as Fuc-(NeuAc)GMl and Fuc-(NeuGc)GMl, respectively. Their chromatographic behaviour was the same as that of Fuc- (NeuAc)GM1 and Fuc-(NeuGc)GMl isolated from cerebrum, spinal cord and dorsal root ganglion, whose structure has been proved by mass-spectrometry [26].

The fatty acid patterns of the major gangliosides from skeletal muscle were determined in two gangliosides, G M ~ and GDla, from the control animals, and six gangliosides from the intoxicated pigs. Since there was no difference in pattern between the two gangliosides from the control animals and the corresponding gangliosides from the chloro- quine-treated animals, only the latter is given in Table 5. There is a striking similarity in the fatty acid patterns of the six gangliosides. They were all characterized by z 50 stearic

Table 3 Ganglioside distribution in ykeletal muscle from Lhloroquinone-intoxicated mini-pig\ cind controls The gdnghoside distribution was obtained after separation of the gangliosides accordig to the number of sialic acid residues, and densitometric analyses of the sepdrated fractions a& described in methods The mean values

Treatment Ganglioside distribution

S D dre given, n = 3

~ ~~ ~~ ~ ~~~ ~

monosialo d 1 s I A 1 o polysialo

totdl GM3 GMZ LMI GM1 ‘residual’ ~ - _ _ _ ~ ~ -~ ~ ~ ~ _ _ _ ~ ~~

x -~ ~ ~- - ~~ - ___ ~~~

Control 8 2 f 1 7 S S k 2 9 < 1 1 S f l S 7 f 0 6 4 k 0 6 3 3 + 1 7 S k O S Chloroquine-intoxicated 78 k I S 46 k 3 S 16 f 2 S b 4 k 1 S b 10 & 1 7 3 f 1 0 1 9 + 0 6 ” 3 k 1 0

a P < O O 1 P < 0 001

Table 4. Carbohydrate composition of gangliosides isolared.from skeletal muscle of mini-pigs The determinations were performed as described in Materials and Methods. Composition is expressed as residues per molecule

GM3 GM2 LMI GMl FUC- FUC- G D 1 ,

(NeuAc)GM1 (NeUGc)GMi

Galactose 1 .0 1.1 1.9 2.1 2.1 2.0 2.1 Glucose 1.1 1 .o 1 .o 1.1 0.9 1 .o I .o Fucose - - - - 1.1 1.1 - Glucosamine - - - - - - 0.9 Galactosamine - 0.9 - 0.9 0.9 1.1 0.9 Sialic acid 1 .o 1 .o 1 .o 1 .o 1 .o 1 .o 2.0

Table 5. Fatty acid composition of gangliosides isolated from skeletal of mini-pigs The analysis was performed as described in Materials and Methods. The figures show the molar percentage of the individual fatty acids

Fatty C i M j GMZ LMI GMI Fuc- GDla

acid (N~uAc)GMI

16:O 4 5 5 5 5 5 18:O 50 48 49 43 42 44 20:o 12 13 12 14 12 12 22:o 9 10 9 10 9 10 22: 1 3 1 1 2 1 2 23:O 4 5 4 6 6 6 24:O 8 9 8 12 13 13 24: 1 7 I 9 6 6 5

acid and similar proportions of 20 : 0, 22: 0 and 24: 0. The similarity between the gangliosides was also found in the sphingosine bases. In all the gangliosides 4-sphingenine re- presented ~ 9 0 % , and the remaining 10% consisted of ap- proximately equal amounts of sphinganine, 4, cis-I 4-sphin- gadienine and 4-icosasphingeine.

DISCUSSION

The aim of this study was to characterize and to quantify the lipids accumulated in skeletal muscles from chloroquine- intoxicated mini-pigs. From previous studies of mini-pigs we were aware that chronic chloroquine-intoxication can lead to diminution of appetite and emaciation [27]. Since the chloroquine was incorporated in the dietary pellets, loss of appetite must have resulted in a diminished intake of the drug, and Gleiser et al. [28] have shown that the severity of the lesions in skeletal muscle and other organs is dose-related. We therefore studied only pigs that had been treated with the same dose of chloroquine for approximately the same time and that did not differ significantly in weight from age-matched controls (Table 1).

The idea of chemical determination of the changes in skeletal muscle tissue accompanying chronic chloroquine intoxication was prompted by certain morphologic pe- cularities of the chloroquine myopathy. These changes have been described in greatest detail in the rat [9,20]. The granulovacuolar type of the muscle fibre lesion [29] and com- plementary electron-microscopic findings suggested some relationship between this process and the storage dystrophy of nerve cells in the chloroquine intoxication [10,27,28].

Three cell types in skeletal muscle exhibited storage phenomena. The major part of the accumulation affected the muscle fibres and consisted of granular material, which was more or less positive to periodic acid/Schiff in frozen sections and dissolved on embedding in paraffin. Considerable storage of lipids occured also in the vascular pericytes - the stored material was only faintly positive to periodic acid/Schiff in frozen sections. The morphologically least important sites of storage changes were the axon segments in the preterminal region and the axon endings of motor nerve fibres and of primary sensory nerve fibres in muscle spindles. The bulk of the accumulated material in skeletal muscle was deposited in the muscle fibres and, though to a less extent, in the pericytes. The positive reaction to periodic acid/Schiff of the deposits and the concentrically layered membranous bodies (Fig. 1 a)

in the muscle fibres suggest that most of the gangliosides are accumulated there.

The accumulation of lipids in the skeletal muscle was much larger than that found in any other organ from chloroquine- intoxicated mini-pigs [25], or in organs from any other species treated with lysosomotropic compounds [30,31]. This greater severity of the affect in the skeletal muscles might be explained by the high affinity of chloroquine for myoglobin, with a relatively higher concentration of the drug [32,33].

Chloroquine belongs to a groups of compounds known to be accumulated and trapped in the lysosomes. The mechanism of this accumulation has been thoroughly described by deDuve et al. 1341, who also coined the name lysosomotropic com- pounds for this group of substances. In certain tissues the concentration of these drugs in the lysosomes can reach 1000-fold that in the plasma [34]. These drugs also induce lipid storage in various organs [2]. Liillman-Rauch et al. have suggested the accumulation to be caused by the binding of the drugs to polar lipids inside the lysosomes with con- sequent formation and gradual accumulation of non-digestible drug-lipid complexes 135 - 371. Another explanation might be that the enrichment of chloroquine in the lysosomes reduces the activities of the lysosomal enzymes involved in the de- gradation of polar lipids, and thus causes the lipid storage. The avtivities of several lysosomal enzymes have been shown to be reduced by lysosomotropic drugs [38 -401. in a recent report Matsuzawa and Hostetler [41] showed a direct inhibitory effect of chloroquine on activities of phospholipase A and C in rat liver lysosomes. This inhibition of the enzymes in- volved in phospholipid catabolism could explain the phos- pholipid storage in the skeletal muscles. The enzyme GM2 - N- acetyl-p-galactosaminidase catalyzing the degradation of ganglioside GM2 to GM3 is the rate-limiting enzyme in ganglioside degradation and thus the most susceptible to disturbances in the lysosomal environment caused by the chloroquine accumulation, e.g. a raised intralysosomal pH [42]. Experiments in vitro with the addition of chloroquine diphosphate to muscle homogenates inhibited the activity of 8-hexosaminidase (unpublished experiments). This reduced activity of the lysosomal enzymes could explain the 100-fold or still larger increase in the concentration of ganglioside GMZ, compared to the 10 - 1 Sfold increase of total ganglioside N-acetylneuraminic acid.

Bis(monoacylglyceryl)phosphate, which was found in the chloroquine-intoxicated animals, is highly enriched in lyso- soma1 preparations, and has been suggested to be a specific marker of lysosomes [43-45). The large portion of this lipid can be interpreted as an increase in lysosomal membranes. The increase in lysosomal membranes has been shown also by histochemical examinations. Thus the myeloid bodies have been shown to possess acid phosphatase activity, indicating them to be of lysosomal origin [9].

In summary, the biochemical characterization of the lipid storage is consistent with the following pathogenesis: a decreased degradation of polar lipids owing to reduced activities of the lysosomal enzymes, and an accumulation of lysosomal residual bodies. The accumulation of lipids is more striking in skeletal muscle than in any other tissue studied and may perhaps be the main or decisive factor in the pathogenesis of the impairment of muscle function.

The costs of the study were partially defrayed by grants from the Swedish Medical Research Council (project 03X-627) and the Medical Faculty, University of Goteborg. The skilful technical assistance of Mrs Birgitla Jungbjer and Mrs Birgitta Dellheden is gratefully acknowledged.

571

REFERENCES

1. Yamamota, A,, Adachi, S., Kitani, T., Shinji, K., Seki, K.-I. Nasu, T. & Nishikawa, M. (1971) J . Biochem. (Tokyo) 69, 613-615.

la. CBN Recommendations (1977) Eur. J . Biochem. 79, 11 -21. 2. Lullmann, H., Lullmann-Rauch, R. &Wassermann, 0. (1973) Dtsch.

3. Hobbs, H. E., Sorsby, A. & Freedman, A. (1959) Lancet, 2, 478-

4. Nelson, A. A. & Fitzhugh, 0. G. (1948) Arch. Pathol. 45, 454. 5. Whisnant, J. P., Espinosa, R. E., Kierland, R. R. & Lambert, E. H.

6. Garcin, R., Rondot, P. & Fardeau, M. (1964) Rev. Neurol. 111,

7. Eadie, M. J. & Ferrier, T. M. (1966) J . Neurol. Neurosorg. Psychiatry,

8. Smith, B. & O'Grady, F. (1966) J . Neurol. Neurosurg. Psychiatry, ZY,

9. MacDonald, R. D. & Engel, A. G. (1970) J . Neuropafh. Exp. Neurol.

Med. Wochenschr. 98, 1616- 1625.

480.

(1963) Mayo Clin. Proc. 38, 501 -513.

177-195.

29, 331 - 337.

255-258.

29,479 - 499. 10. Klinghardt, G. W. (1974) Acta Neuropath. 28, 117-141. 11. Drenckhahn, D. & Lullman-Rauch, R. (1979) Neuroscience, 4, 549-

562. 12. Mastaglia, F. L., Papadimitriou, J. M., Dawkins, R. L. & Beveridge,

B. (1977) J . Neurol. Sci. 34, 315-328. 13. Sbennerholm, L. & Fredman, P. (1980) Biochim. Biophys. Aria, 617,

97 - 1 09. 14. Fredman, P., Nilsson, O., Tayot, J.-L. & Svennerholm, L. (1980)

Biochim. Biophys. Acta, 618, 42-52. 15. Haring, F., Griihn, R., Smidt, D. & Scheven, B. (3967) Zentralbf.

Bakteriol. Parasitenkd. lnfrktionskr. Hyg. I , Abt. Orig. 189, 521 - 532.

16. Wells, M. A. & Dittmer, J. C. (1963) Biochemistry, 2, 1259- 1263. 17. Svennerhol~n, L. & Vanier, M. T. (1972) Brain Ras. 47, 457-468. 18. Fewster, M. E., Burns, B. J. & Mead, J. E. (1969) J . Chromntogr. 43,

19. Crawford, N. (1958) Clin. Chim. Acta, 3, 357-367. 20. Svennerholm, L. (1957) Biochim. Biophys. Acta, 24, 604-61 1. 21. Miettinen, T. & Takki-Luukainen, 1. T. (1959) Acta Chem. Scand.

22. Holm, M., Mlnsson, J.-E., Vanier, M.-T. & Svennerholm, L. (1972)

120-126.

13, 856-858.

Biochim. Biophys. Acta, 28V, 356- 364.

23. MPnsson, J.-E., Vanier, M. T. & Svennerholm, L. (1978) J . Neuro-

24. Svennerholm, L., Vanier, M. T. & MPnsson, J.-E. (1980) J . Lipid

25. Fredman, P. (1979) M.D. Thesis, University of Goteborg, pp. 39 -49. 26. Fredman, P., MPnsson, J.-E., Svennerholm. L., Samuelsson, B.,

Pascher, I., Pimlott, W., Karlsson, K.-A. & Klinghardt, G. W. (1981) Eur. J . Biochem. 116, 553-564.

27. Klinghardt, G. W. (1977) in Neurotoxicology (Roizin, L., Shiraki, H. & Grcevic, N., eds) pp. 171 - 179, Raven Press, New York.

28. Gleiser, C. A,, Bay, W. W., Dukes, T. W., Brown, R. S., Read, W. K. & Pierce, K. R. (1968) Am. J . Pathol. 53, 27-45.

29. Itabashi, H. H. & Kokmen, E. (1972) Arch. Pathol. Y3, 209-218. 30. Yamamota, K., Adachi, S., Matsuzawa, Y., Hiraoka, A. & Seki,

31. Seiler, K.-U. & Wassermann, 0. (1975) Naunyn-Schmiedeher~'.F Arch.

32. Klinghardt, G. W. & Thomas, E. (1970) Proc. 6th Int. Congr. Neuro-

33. Grundmann, M., Mikulikova, J. & Vrublovski, P. (1972) Arch. Int.

34. deDuve, C., deBarsy, T., Poole, B., Trouet, A., Tulkens, P. & van

35. Seydel, J. K. & Wassermann, 0. (1973) Naunyn-Schmiedeherg's

36. Lullmann-Rauch, R. (1974) Acta Neuropathol. 28, 117- 141. 37. Drenckhahn, D. & Lullmann-Rauch, R. (1976) Cell Tissue Rrs. 171,

38. Goldstein, J. L., Brunschede, G. Y. & Brown, M. S. (1975) J . Biol.

39. Lie, S. 0. & Schofield, B. (1973) Biochem. Pharmacol. 22, 3109- 31 14 40. Wibo, M. & Poole, B. (1974) J . Cell B i d . 63, 430-440. 41. Matsuzawa, Y. & Hostetler, K. Y. (1980) J . Biol. Chem. 255, 5190-

42. Ohkuma, S. & Poole, B. (1978) Proc. Nut1 Acad. Sci. USA, 75,

43. Wherett, J. R. & Heuterer, S. (1972) J . Biol. Chem. 247, 4114-4120. 44. Brotherus, J. & Renkonen, 0. (1977) Biochim. Biophys. Acta, 486,

45. Brotherus, J. & Renkonen, 0. (1977) J . Lipid Res. 18, 191 -202.

chem. 30, 273 - 275.

Res. 21, 53-64.

K.-I. (1976) Lipids, 11, 616-622.

Pharmacol. 288, 261 -268.

pathol. pp. 1076- 1077.

Pharmacodyn. Ther. 197,45 - 52.

Hoof, F. (1974) Biochem. Pharmacol. 23, 2495 -2531.

Arch. Pharmacol. 279, 207 -210.

273 - 276.

Chem. 250,7854-7862.

5194.

3327 - 3331.

243 - 253.

0. Nilsson, P. Fredman, and L. Svennerholm, Institutionen for Psykiatri och Neurokemi, S: t Jorgens sjukhus, S-42203 Hisings Backa, Sweden

G. W. Klinghardt, Max-Planck-Institut fur Hirnforschung, DeutschordenstraBe 46, D-6000 Frankfurt am Main-Nlederrad, Federal Republic of Germany

H. Dreyfus, Unite 44 de 1'Institut National de la Sante et de La Recherche Medicale. Centre de Neurochimie du Centre National de la Recherche Scientifique, 5 rue Blaise Pascal, F-67084 Strasbourg Cedex, France