INTRODUCTION

Gangliosides, sialic acid containing glycosphingolipids that are normal components of the plasma membranes (1), are biosynthesized in the Golgi apparatus and degraded in the lysosomes (2). There is evidence that exogenously added gangliosides are taken up by the cells by endocytosis, degraded in the lysosomes, and the formed catabolites, of sugar and lipid nature, partly re-cycled for biosynthetic purposes (3, 4, 5). Little is known about the coordination between the flux of ganglioside degradation, linked to the endocytic vesicular flow, and the flux of ganglioside biosynthesis, linked to the exocytic flow. As well, little is known about the extent of metabolic re-cycling of the individual components of gangliosides produced during degradation, and the contribution of these salvage processes to the overall ganglioside turnover. The present work was undertaken with the aim of improving our knowledge of these events.

The experimental model we adopted consisted of cultured human skin fibroblasts fed with G_M3^1,2 ganglioside isotopically labeled at the level of (a) sialic acid ([Neu5Ac-³H]G_M3) providing information on sialic acid re-cycling, or (b) sphingosine ([Sph-³H]G_M3) giving information on sphingosine re-cycling, or (c) fatty acid moiety ([stearoyl-¹⁴C]G_M3), mainly informing on the ganglioside half-life. G_M3 is known to be the main ganglioside of human fibroblasts (6).

In order to better acknowledge the impact of the coordination between the degradative and biosynthetic routes on the regular course of ganglioside turnover, we made a parallel study of the skin fibroblasts from patients suffering from Salla disease. In this rare inherited storage disorder (7) a gene mutation leads to the lack, or misfunction, of the membrane carrier responsible for the egress from lysosomes of sialic acid liberated during intralysosomal digestion. Since sialic acid in Salla fibroblasts accumulates in the lysosomes, it can be expected that its metabolic recycling is at least partially impaired.

The results obtained show that in fibroblasts: (a) the salvage of radioactive sialic acid produced by catabolism of [Neu5Ac-³H]G_M3 is almost quantitative, this sugar being catabolized to tritiated water in very scant amount; (b) salvage of radioactive sphingosine produced by catabolism of [Sph-³H]G_M3 is relevant, but part of the base is catabolized to tritiated water; (c) ganglioside metabolic half-life is 2-3 days, and (d) in Salla cells not only salvage processes of Neu5Ac are markedly lowered but overall ganglioside catabolism appears to be altered.

EXPERIMENTAL PROCEDURES

The commercial chemicals were the purest available, common solvents were distilled before use and deionized water, obtained by a Milli-Q system (Millipore), was distilled in a glass apparatus. Silica Gel 100 for column chromatography (0.063-0.2 mm, 70-230 mesh, ASTM) and high performance silica gel precoated thin-layer plates (HPTLC Kieselgel 60, 10 × 10 cm) were purchased from Merck GmbH (Darmstadt, Germany); Vibrio cholerae sialidase from Sigma; ceramide glycanase from Macrobdella decora from Boehringer Mannheim (Mannheim, Germany); sodium boro[³H]hydride (8 Ci/mmol), [³H]acetic anhydride (4.6 Ci/mol) and [1-¹⁴C]stearic acid (52 mCi/mmol) from Amersham International (Bucks, United Kingdom).

Ganglioside G_M3 and G_D3, extracted from bovine brain (8) and buttermilk (9), respectively, were purified to over 99% and characterized (8, 9). Lactosylceramide and glucosylceramide were prepared by partial acid hydrolysis of the total ganglioside mixture obtained from bovine brain (10). Ceramide was prepared from G_M3 by treatment with ceramide glycanase (11) and sphingosine by alkaline hydrolysis of ceramide (11).

G_M3 ganglioside was isotopically labeled with tritium at the level of sialic acid ([Neu5Ac-³H]G_M3) or C-3 of sphingosine ([Sph-³H]G_M3), or with ¹⁴C at the fatty acid moiety ([stearoyl-¹⁴C]G_M3). [Neu5Ac-³H]G_M3 was prepared by deacetylation of G_M3 followed by N-acetylation with [³H]acetic anhydride (4, 8, 12). [stearoyl-¹⁴C]G_M3 was prepared by deacylation of G_M3 followed by N-acylation with [1-¹⁴C]stearic acid (8, 12). The preparation of [Sph-³H]G_M3 was accomplished by perfecting (13) the dichlorodicyanobenzoquinone/sodium boro[³H]hydride and reversed phase high performance liquid chromatography purification procedures previously reported (14).

[Neu5Ac-³H]G_M3, [Sph-³H]G_M3, and [stearoyl-¹⁴C]G_M3 were over 99.5% pure, as determined by digital radiochromatoscanning of HPTLC plates developed in three solvent systems (see below), and their specific radioactivity 2.3 Ci/mmol, 2.0 Ci/mmol, and 52.0 mCi/mmol, respectively.

Human skin fibroblasts were obtained by the punch technique. Normal cells were from young subjects and pathological cells from two boys who fulfilled the diagnostic criteria for Salla disease: psychomotor retardation, lysosomal storage demonstrated by electron microscopy in fresh skin biopsy specimens, and increased excretion of free sialic acid in the urine (7). The patients, diagnosed by Professor Martin Renlund (Department of Obstetrics and Gynecology, Helsinki University Central Hospital, Helsinki, Finland) were: patient M., cell line 4 (Table I), is now 11 years old, has never learned to speak or walk, not even to crawl, and is severely spastic; patient L. J., cell line 5 (Table I), is now 18 years old, cannot walk or speak, and has severe spasticity.

Table I. Ganglioside, protein, and DNA content in cultured normal and Salla fibroblasts Gangliosides were extracted from three normal fibroblast lines (1, 2, and 3) and from two Salla fibroblast lines (4 and 5) (3 mg of cell proteins for each fibroblast line), purified, separated by HPTLC and quantified by densitometry. Ganglioside identification was achieved by chromatographic comparison with standard gangliosides. Protein and DNA content were determined harvesting and collecting cells, at confluency, from one 90-mm dish. All the data are mean ± S.D. of three experiments. Component Normal Salla 1 2 3 4 5 nmol/mg protein nmol/mg protein Total gangliosides 3.3  ± 0.4 3.1  ± 0.3 3.7  ± 0.5 4.4  ± 0.5 3.7  ± 0.3 GM3 2.2  ± 0.2 2.6  ± 0.3 3.0  ± 0.4 3.8  ± 0.3 3.2  ± 0.3 GD3 1.1  ± 0.2 0.5  ± 0.1 0.7  ± 0.2 0.6  ± 0.2 0.5  ± 0.1 GM2, GM1, GD1a TR.a TR. TR. TR. TR. Sphingomyelin 21.6  ± 3.5 NDb ND 15.9  ± 2.2 ND µg/dish µg/dish Protein 830  ± 150 790  ± 140 810  ± 145 650  ± 130 695  ± 140 µg/mg protein µg/mg protein DNA 56  ± 3 ND ND 55  ± 3 ND a  TR., traces. b  ND, not determined.

Both pathological and normal cells were cultured and propagated as monolayers in 75-cm² culture flasks at 37 °C in a humidified atmosphere containing 5% CO₂, using Eagle's minimum essential medium supplemented with 10% FCS (15). Subcultures for experiments were made on 90-mm plastic Petri dishes and used at confluency.

Radioactive G_M3 ([Neu5Ac-³H]G_M3, [Sph-³H]G_M3, or [stearoyl-¹⁴C]G_M3) dissolved in 1-propanol/water, 7:3 (v/v), was pipetted into a sterile tube, and dried under a nitrogen stream; the residue was then solubilized in an appropriate volume of pre-warmed (37 °C) Eagle's minimum essential medium to obtain a final ganglioside concentration of 4 × 10⁷ M. We also carried out experiments using simultaneously equimolar amounts of [Neu5Ac-³H]G_M3 and [stearoyl-¹⁴C]G_M3, treated as above, to obtain a final ganglioside concentration of 8 × 10⁷ M. Five ml of the ganglioside mixture were added to each culture dish, after accurate removal of the original culture medium. After a 15-h incubation (pulse), the radioactive medium was removed and the dishes were repeatedly washed first with Eagle's minimum essential medium solution, then with 10% FCS-containing medium (30 min) to remove ganglioside aggregates loosely bound to the cell surface, and finally incubated up to 3 days (chase) with 10 ml of fresh unlabeled 10% FCS/Eagle's minimum essential medium (4). At the end of the chase time the dishes were rinsed three times with Hank's solution, and the cells harvested in water (2 ml) by scraping with a rubber policemen. In some binding experiments with [Neu5Ac-³H]G_M3 the cells were treated, before harvesting, with 2 ml of 0.1% trypsin in phosphate-buffered saline for 4 min to remove the small portion of gangliosides that strongly interacts with proteins protruding from the membrane surface (4).

The cell suspension was subjected to two cycles of freezing-thawing and then centrifuged at 100,000 × g for 30 min. A small portion of the supernatant was used for the analysis of free sialic acid and the remainder, as well as the pellet, was frozen and lyophilized. The residues were subjected to tetrahydrofuran/phosphate buffer lipid extraction and partition (16), resulting in a delipidized pellet, an organic phase, and an aqueous phase. HPTLC analyzes (see below) showed that about 15% of cell associated radioactive G_M3 remained in the organic phase. Thus radioactive G_M3 was quantified using both the aqueous and organic phases. Portions of the total lipid extract from cells treated with [stearoyl-¹⁴C]G_M3 were subjected to alkaline hydrolysis to remove radioactive glycerolipids (17).

Portions of the delipidized pellet, prepared from cells incubated with [Neu5Ac-³H]G_M3, were treated with V. cholerae sialidase (18) to remove sialic acid from sialoglycoproteins.

The radioactivity associated to total cells, sialic acid, proteins (delipidized pellet), and the organic and aqueous phases obtained by lipid extraction and fractionation, was determined by liquid scintillation counting. Radioactivity associated to individual lipids and free sialic acid, separated by HPTLC, was quantified by radiochromatoscanning (Digital Autoradiograph, Berthold, Germany) (11), and in some cases recognized by autoradiography or fluorography.

In the experiments making use of cells fed simultaneously with [Neu5Ac-³H]G_M3 and [stearoyl-¹⁴C]G_M3 the intramolecular distribution of radioactivity in formed G_D3 was determined as follows. Radioactive G_D3 was purified from the total ganglioside mixture extracted from 5 mg of cell proteins by silica gel column (12 × 1 cm) chromatography, using the solvent system chloroform/methanol/water, 60:35:5 (v/v), for both equilibration and elution. The purified G_D3 was mixed with 50 µg of pure standard G_D3 in a final volume of 0.1 ml of 50 mM sodium acetate buffer, pH 5.5, and subjected to partial enzymatic desialylation in the presence of 1 milliunit of V. cholerae sialidase (30 min, 37 °C). The hydrolysis products, G_M3, LacCer, and Neu5Ac, together with residual G_D3, were separated by HPTLC (see below), identified by radiochromatoscanning, scraped off, and analyzed for radioactivity content by liquid scintillation counting using a [³H/¹⁴C] dual-label program. In parallel HPTLC separations the hydrolysis products were identified by treatment with an anisaldehyde or p-dimethylaminobenzaldehyde spray reagents (see below), each spot being quantified by densitometry (19).

Total radioactivity present in the medium was determined by liquid scintillation counting. Tritiated water (20) formed during metabolic processing of the gangliosides was determined as follows. One ml of the cell culture medium was dialyzed overnight against 2 ml of water. The dialysis water (dialysate) was then divided into two portions. One portion was directly submitted to liquid scintillation counting, the other was completely dried at 36 °C and the residue dissolved with the same original amount of water and counted. The amount of tritiated water was determined by the difference in radioactivity content of the dialysate, before and after drying. In parallel experiments the dialysate was distilled at 100 °C and the distillation product, containing tritiated water, counted. Control experiments were performed by incubating [Sph-³H]G_M3 or [Neu5Ac-³H]G_M3 with the cell medium in the absence of cells.

Monodimensional HPTLC was performed with the following solvent systems: (a) chloroform, methanol, 0.2% aqueous CaCl₂, 50:42:11 (v/v), (b) chloroform/methanol/water, 110:40:6 (v/v), and (c) chloroform/methanol, 2:1 (v/v), to assess the homogeneity of radioactive G_M3; chloroform/methanol, 2:1 (v/v), and after plate drying, chloroform/methanol, 0.2% aqueous CaCl₂, 50:42:11 (v/v), to assess the aqueous phase composition (gangliosides) and to separate the products obtained by partial sialidase hydrolysis of G_D3; chloroform/methanol, 2:1 (v/v), and, after plate drying, chloroform/methanol/water, 110:40:6 (v/v), to assess the composition of the organic phase (G_M3, neutral sphingolipids, and sphingomyelin).

Two-dimensional HPTLC was performed for the following purposes and systems. To assess the occurrence of ceramide in the organic phase: 1st run, chloroform/methanol/water, 110:40:6 (v/v); 2nd run, ethyl acetate. To assess the occurrence of sphingosine in the organic phase: 1st run, chloroform/methanol/water, 110:40:6 (v/v); 2nd run, chloroform, methanol, 37% NH₄OH, 40:10:1 (v/v). Colorimetric detection was carried out using the p-dimethylaminobenzaldehyde (21), the anisaldehyde (22), and the molybdate (23) spray reagents. Spot quantification was carried out by densitometry after HPTLC separation.

Gangliosides were assayed as bound Neu5Ac by the resorcinol-HCl method (24, 25), pure Neu5Ac being used as the reference standard. The protein content was determined (26) with bovine serum albumin as the reference standard. Proteins were analyzed by SDS-polyacrylamide gel electrophoresis (18). DNA was assayed according to Burton (27).

RESULTS

Data on gangliosides, sphingomyelin, protein, and DNA content of cultured fibroblasts from three normal subjects and two Salla patients are shown in Table I. Fibroblasts, although displaying some variation from cell to cell line, contained G_M3 as the major compound, followed by G_D3, and traces of G_M2, G_M1, and G_D1a. These results are in agreement with previous data (6).

The feeding experiments with [Neu5Ac-³H]G_M3, [Sph-³H]G_M3, and [stearoyl-¹⁴C]G_M3 aimed to study ganglioside metabolic processing were performed using fibroblasts deriving from lines 1 (normal) and 4 (Salla) (Table I). Metabolic re-cycling of radioactive sialic acid was also studied in fibroblasts deriving from lines 2, 3, and 5 (Table I), fed with [Neu5Ac-³H]G_M3.

The trend of radioactivity association to normal and Salla fibroblasts fed with [Neu5Ac-³H]G_M3 is shown in Fig. 1. In both cell lines after 15 h exposure to 4 × 10⁷ M radioactive GM₃ followed by exhaustive washings with FCS free and FCS containing media, fairly measurable amounts of radioactivity resulted by association to fibroblasts (``serum-stable cell associated radioactivity''). A successive treatment with trypsin reduced the amount of cell associated radioactivity (``trypsin-stable cell associated radioactivity'') in either normal and Salla fibroblasts. This behavior is in agreement with previous findings (3, 4). In both fibroblast lines a substantial part of the cell associated radioactivity was released into the medium during chase. Released radioactivity was carried solely by G_M3, with the exception of a portion of tritiated water. Both the serum- and the trypsin-stable associated radioactivity were markedly higher (40 and 60%, respectively) in Salla than in the normal fibroblasts.

7

The release of radioactivity in the medium, rapid in the first day of chase and decreasing gradually thereafter, became almost zero after 3 days when the total associated radioactivity corresponded to about 3 and 4.8% of the total administered radioactivity in normal and Salla fibroblasts, respectively. Very similar results were obtained when the feeding experiments were performed with [Sph-³H]G_M3.

After fibroblast feeding with [Neu5Ac-³H]G_M3 (15-h pulse plus 72-h chase), cell associated radioactivity was carried, besides G_M3, by free sialic acid, the protein pellet (sialoglycoproteins), and G_D3 ganglioside (Table II and Fig. 2), and in trace amounts by gangliosides chromatographically behaving as G_M2, G_M1, and G_D1a. A very low amount of tritiated water was also found in the medium.

Table II. Distribution of radioactivity in normal and Salla cultured fibroblasts fed with [Neu5Ac-3H]GM3, [Sph-3H]GM3, or [stearoyl-14C]GM3 The pulse time was 15 h and was followed by 72-h chase. The radioactivity linked to individual substances has been expressed as dpm/µg cell protein, and % of the total cell associated radioactivity plus, in the case of the experiments with [Neu5Ac-3H]GM3 and [Sph-3H]GM3, that carried by tritiated water in the medium. Results are referred to fibroblasts from cell lines numbers 1 (normal) and 4 (Salla) of Table I. Compound [Neu5Ac-3H]GM3 [Sph-3H]GM3 [stearoyl-14C]GM3 Normal Salla Normal Salla Normal Salla dpm/µg cell protein dpm/µg cell protein dpm/µg cell protein GM3 277.3 552.2 247.6 501.7 3.77 7.35 GD3 64.5 56.2 22.5 24.0 0.04 0.03 LacCer 29.5 42.0 0.06 0.27 GlcCer 11.9 18.7 0.04 0.09 Cer 13.1 33.0 TR.a 0.28 SM 45.0 68.2 0.01 0.03 Glycoproteins 82.6 79.6 Neu5Ac 2.1 86.6 Sph TR. TR. Water 3.4 5.4 40.2 62.2 Glycerolipids 3.25 7.45 Fatty acids 0.15 0.37 % of total dpm % of total dpm % of total dpm GM3 64.5 70.8 60.4 66.9 52.6 49.1 GD3 15.0 7.2 5.5 3.2 0.5 0.2 LacCer 7.2 5.6 0.8 1.8 GlcCer 2.9 2.5 0.6 0.6 Cer 3.2 4.4 1.8 SM 11.0 9.1 0.2 0.2 Glycoproteins 19.2 10.2 Sph TR. TR. Neu5Ac 0.5 11.1 Water 0.8 0.7 9.8 8.3 Glycerolipids 45.3 48.5 Fatty acids 2.1 2.4 a  TR., traces.

7

In normal fibroblasts radioactive free sialic acid was hardly appreciable, in contrast with the impressive accumulation occurring in Salla fibroblasts (about 11% of total associated radioactivity). Conversely, the radioactivity carried by glycoproteins and G_D3 in normal fibroblasts doubled that present in the Salla at all chase times (Fig. 2).

After fibroblast feeding with [Sph-³H]G_M3 (15-h pulse plus 72-h chase), cell associated radioactivity was carried, besides G_M3, by LacCer, GlcCer, Cer, G_D3, and SM (Table II). Radioactive sphingosine was detected in traces. Substantial amounts of tritiated water, the final catabolite from radioactive sphingosine, were also found in the medium (Table II). As shown in Figs. 3, A and C, and 4, A, C, and E, all the above ³H-metabolites markedly increased (as % of total associated radioactivity) during chase in fibroblasts. Remarkably, SM, G_D3, LacCer, and GlcCer were much more abundant in normal than in Salla fibroblasts, whereas Cer prevailed in Salla fibroblasts.

7

7

It should be noted that radioactive SM can be solely produced by a biosynthetic process which re-cycles radioactive sphingosine liberated during exogenous GM₃ degradation (28, 29). Instead, radioactive LacCer, GlcCer, and Cer derive from two processes: catabolism of taken up and processed [Sph-³H]G_M3 and biosynthesis by re-cycling of released sphingosine. The concomitant use of [stearoyl-¹⁴C]G_M3 provides more precise evidence on this dual process.

After fibroblast feeding with [stearoyl-¹⁴C]G_M3 (15-h pulse plus 72-h chase), cell associated radioactivity was carried, besides G_M3, by glycerolipids, fatty acids, and in very small amounts, G_D3, LacCer, GlcCer, Cer, and SM (Table II). Noteworthy, the proportion of radioactivity linked to glycerolipids was almost as high than that of G_M3, indicating that the fatty acid liberated from G_M3 degradation was massively re-cycled for the biosynthesis of glycerolipids. Instead the amount of radioactivity carried by LacCer, GlcCer, and Cer was very low, and, particularly very much lower than that carried by the same sphingolipids obtained after cell feeding with [Sph-³H]G_M3. This difference indicates that with [stearoyl-¹⁴C]G_M3, LacCer, GlcCer, and Cer result primarily from degradation and, by consequence, the substantially higher formation of these compounds observed with [Sph-³H]G_M3 is substained by a process of biosynthesis by re-cycling of liberated sphingosine.

During chase (Figs. 3 and 4) the formation of radioactive G_D3 and LacCer was higher in normal than in Salla fibroblasts, whereas ceramide, practically undetectable in normal cells, represented as much as 1.8% of total associated radioactivity in the pathological cells. Both SM and GlcCer were present in very small amounts in normal and pathological fibroblast lines.

Radioactive G_D3 was isolated from normal fibroblasts fed simultaneously with equimolar amounts of [Neu5Ac-³H]G_M3 and [stearoyl-¹⁴C]G_M3. After extraction and purification, radioactive G_D3 was subjected to partial hydrolysis with V. cholerae sialidase and the resultant G_M3, LacCer, and sialic acid, together with the remaining G_D3, were analyzed for their [³H] and [¹⁴C] contents and specific radioactivity. Table III shows the results of these experiments. G_D3 and G_M3 carried both [³H] and [¹⁴C]; Neu5Ac carried only [³H] and LacCer only [¹⁴C], as expected. Released sialic acid and G_M3 showed similar values of [³H] specific radioactivity, about half that of G_D3. G_D3, the released G_M3, and LacCer had the same [¹⁴C] specific radioactivity and a low [¹⁴C] content.

Table III. Incorporation of radioactivity into the two sialic acid units and into the fatty acyl moiety of GD3 Radioactive GD3 was extracted from cells fed simultaneously (15-h pulse plus 72-h chase) with [Neu5Ac-3H]GM3 (2.3 Ci/mmol) and [stearoyl-14C]GM3 (52 mCi/mmol), purified, mixed with standard nonlabeled GD3 and subjected to partial enzymatic hydrolysis with Vibrio cholerae sialidase, to obtain a mixture of GD3 (remainder), GM3, LacCer, and Neu5Ac. These compounds were separated by HPTLC and analyzed for mass or radioactivity content ([3H] S.D. ± 2%; [14C] S.D. ± 10%). Picomoles of labeled compound were calculated on the basis of starting ganglioside specific radioactivity. For details see ``Experimental Procedures.'' nmol of nonlabeled carrier GD3 [3H] [14C] [3H] [14C] [3H] [14C] [3H/14C] dpm dpm/nmol nonlabeled carrier GD3 pmol of labeled compound pmol ratio GD3 10.8 39,400 31 3,648 2.87 7.8 0.27 28.89 GM3 16.1 29,900 40 1,857 2.48 5.8 0.35 16.57 LacCer 7.2 0 21 0 2.91 0 0.18 Neu5Ac 30.0 54,850 0 1,828 0 10.8 0

If all the radioactive G_D3 represents the direct sialylation of exogenous radioactive G_M3 brought about by the recycling of catabolic radioactive Neu5Ac, partly diluted with nonradioactive Neu5Ac, then its [³H] content, in terms of starting picomoles of ganglioside, should be only slightly higher than the [¹⁴C] content. But if all the radioactive G_D3 were the result of a neobiosynthetic process, then the amount of incorporated [³H] and [¹⁴C] would be related only to the extent of the [³H]Neu5Ac and [¹⁴C]stearic acid salvage processes and the dilution of the two catabolites in the nonlabeled corresponding compounds. Moreover the two Neu5Ac units must have the same specific radioactivity. Our results clearly show that the radioactive G_D3 is mainly a product of neobiosynthesis. In fact, the [³H] content, in terms of picomoles, was much higher than the [¹⁴C] content, about 30 and 15 times in G_D3 and G_M3, respectively, and the two sialic acid units of G_D3 had the same specific radioactivity. [³H]Neu5Ac is only slightly degraded, as proven by the very low amount of tritiated water found in the medium. This, together with the expected high dilution of [¹⁴C]stearic acid in the nonradioactive stearic acid, explains the wide difference in the incorporation of [³H] and [¹⁴C] in G_D3.

The radioactivity found in the delipidized pellet from normal and Salla fibroblasts fed with [Neu5Ac-³H]G_M3 was carried mainly by glycoproteins showing molecular mass of 70-180 kDa (revealed by SDS-polyacrylamide gel electrophoresis). Radioactivity was carried, almost completely, by the sialic acid of glycoproteins. In fact over 93% of this radioactivity was released from the delipidized pellet by treatment with sialidase and was found to be carried by free sialic acid. In the delipidized pellet from cells fed with [Sph-³H]G_M3 or [stearoyl-¹⁴C]G_M3 no radioactive proteins could be detected.

DISCUSSION

In the present work we explored the metabolic processing of exogenous gangliosides with the aim of verifying: (a) the extent of biosynthetic re-cycling of the individual components produced during ganglioside degradation, and (b) the contribution of these salvage processes to the overall ganglioside turnover. To this purpose, we employed cultured human fibroblasts fed with G_M3 ganglioside, isotopically radiolabeled at the sialic acid acetyl group ([Neu5Ac-³H]G_M3) or at C-3 of sphingosine ([Sph-³H]G_M3) or at C-1 of the stearoyl chain ([stearoyl-¹⁴C]G_M3. In comparison with normal lines we used fibroblast lines from subjects affected with Salla disease, where the egress of sialic acid from lysosomes is genetically impaired.

The molecular aspects of exogenous ganglioside-cell interactions and the modalities of ganglioside insertion of the cell surface and penetration into cells has been studied in detail (3, 4, 30, 31). The general view is that the major portion of exogenous gangliosides which bind to cells is loosely associated to the cell surface, and can be removed by rapid and repeated washings with protein containing solutions (albumin), owing to the easy formation of tight and stable lipoproteic complexes between gangliosides and proteins (32, 33). However, a minor but definite portion of ganglioside interacts strongly with proteins protruding from the membrane surface and can be released by mild trypsin treatment. Finally, a smaller portion of associated gangliosides is constituted by molecules that resist to trypsin treatment and appeared to be inserted into the membrane layer. This portion may be submitted to endocytosis (34). Our results introduce a new feature in this view. In fact, under our experimental conditions, we found that despite the rapid washing treatments with proteins, a portion of the cell associated G_M3 continued to be released into the medium throughout the chase period. The time course of this release followed a hyperbolic trend and reached a constant value after about 72 h chase. This suggests that a portion of the total cell associated gangliosides interacts with the cell surface differently from that easily removed by treatment with protein containing media. Of course, the portion of G_M3 that is released into the medium during chase does not undergo internalization into cells and is not available to metabolic processing.

Considering (a) the specific radioactivity of [³H]G_M3 ([Neu5Ac-³H]G_M3 or [Sph-³H]G_M3 were used to this purpose), (b) the total cell associated radioactivity from which was subtracted the radioactivity released as G_M3 in the medium in the considered period of chase, and (c) the amount of tritiated water found in the medium in the same chase period, we calculated that about 79 pmol/mg protein of exogenous G_M3 available to endocytosis and metabolic processing were taken up by normal cells. This amount constituted only 2.3% of the cell ganglioside content. This small change in the ganglioside membrane content should not prevent exogenous ganglioside to distribute throughout the cell membrane and to follow the routes of internalization into cells and intracellular metabolization of the endogenous gangliosides.

G_M3 taken up by the cells was subjected to degradation and the catabolic fragments were re-cycled for biosynthetic purposes (salvage processes). Particularly, the salvage processes of Neu5Ac, sphingosine, and stearic acid were followed.

The catabolism of [Neu5Ac-³H]G_M3 produced free radioactive sialic acid that was used for the biosynthesis of sialoglycoproteins and gangliosides (namely G_D3). The presence of radioactive sialic acid in cell glycoproteins, contained in the protein pellet, is direct evidence of sialic acid re-cycling. A similar direct conclusion is not necessarily valid for the radioactive sialic acid present in G_D3. In fact, it cannot be excluded that a portion of formed radioactive G_D3 derives from a process of direct G_M3 sialylation assuming that endocytosed G_M3, besides being routed to lysosomes, can directly reach the Golgi apparatus (35). In order to verify this possibility we analyzed the radioactivity distribution in the dual-labeled G_D3 formed after simultaneous cell feeding with [Neu5Ac-³H]G_M3 and [stearoyl-¹⁴C]G_M3 (Table III). The finding that the two sialic acid residues present in G_D3 have the same ³H-specific radioactivity and that G_D3 incorporates much more [³H]Neu5Ac than [¹⁴C]stearic acid strongly supports the notion that the formed radioactive G_D3 is mainly a product of biosynthesis brought about by re-cycling of [³H]Neu5Ac obtained from [Neu5Ac-³H]G_M3 degradation.

The salvage of Neu5Ac is a very important process in fibroblasts. In fact, in normal fibroblasts fed with [Neu5Ac-³H]G_M3, 15 h pulse plus 72 h chase, almost all the radioactive sialic acid produced from catabolism of [Neu5Ac-³H]G_M3 was re-cycled; in all the three cell lines (Table I, Fig. 2) only a very minor portion remained free in the cells or was degraded to radioactive water, the final catabolite of radioactive sialic acid degradation.

Salvage of sphingosine was also high, and after the administration of [Sph-³H]G_M3 [³H]Sph was found in biosynthesized gangliosides and sphingomyelin. Free radioactive sphingosine was hardly detectable in total cell extracts but its catabolism to tritiated water was quite marked.

The catabolism of gangliosides occurs in the lysosomes. It is known that Neu5Ac leaves lysosomes through a membrane protein mediated process. No information on the exit of sphingosine from the lysosomes is available. However, we do know that sphingosine must leave the lysosomes to reach the endoplasmatic reticulum and Golgi apparatus where it is involved in sphingolipid biosynthesis. Thus knowledge of the exit process is of crucial importance also in view of the suggested involvement of sphingosine as a metabolic bioregulator (36).

Radioactive fatty acids from the catabolism of [stearoyl-¹⁴C]G_M3 were also re-used for sphingolipid biosynthesis, but only to a very small extent. In other words, the occurrence of salvage of stearic acid for ganglioside biosynthesis is a very minor phenomenom, if any. This is due to the dilution of the radioactive stearic acid in the cell fatty acid pool, and to the fact that stearic acid is a minor component of fibroblast gangliosides, which mainly contain C-22 and C-24 acyl chains. However, the fatty acid liberated from G_M3 degradation is largely re-cycled for the biosynthesis of glycerolipids. In normal fibroblats after 72 h chase with [stearoyl-¹⁴C]G_M3 close to 45% of the total cell associated radioactivity was carried by glycerophospholipids and about 50% by G_M3. Since neobiosynthesis of G_M3 from stearic acid is virtually absent at the end of the chase period, by the use of [stearoyl-¹⁴C]G_M3 it can be established that after 3 days of chase 50% of radioactive G_M3 was constituted by no yet processed exogenous G_M3.

The administration of radioactive glucose (6) or glucosamine (37), to fibroblast cells, yielded G_M3 and G_D3 with similar specific radioactivity at long chase. Similarly, in a study on rat brain gangliosides (38), intracranial administration of radioactive N-acetylmannosamine yielded a mixture of gangliosides all of which showed the same specific radioactivity. In addition to this, the recovery of sphingosine in cultured neuronal cells (39) was also shown to occur proportionally to the cell ganglioside content. Thus it can be expected that the administration of [Neu5Ac-³H]G_M3 or [Sph-³H]G_M3 to cells will be followed by the incorporation of the catabolic radioactive sialic acid or sphingosine into the neobiosynthesized G_M3 and G_D3, the two main gangliosides of fibroblasts, in proportion to the endogenous ganglioside content (Table I). We calculated that after 72 h chase with [Neu5Ac-³H]G_M3 neobiosynthetic gangliosides, glycoproteins, free sialic acid, and tritiated water, and with [Sph-³H]G_M3 neobiosynthetic gangliosides, glycosphingolipids, ceramide, sphingomyelin, and tritiated water together cover about 50% of the total cell associated radioactivity.

From all these results it follows that at the end of the pulse-chase period with 4 × 10⁷ M [Sph-³H]G_M3, [Neu5Ac-³H]G_M3, or [stearoyl-¹⁴C]G_M3, half of the total associated radioactivity is still carried by the exogenously associated radioactive G_M3. Thus we can predict, in agreement with previous suggestions (6), a ganglioside half-life of about 2-3 days.

The feeding experiments were also performed on fibroblasts from patients with Salla disease. This was done to evaluate the cell ganglioside processing in a system where the metabolic re-cycling of sialic acid is expected to be at least partially impaired.

The metabolic processing of [Neu5Ac-³H]G_M3 taken up by both Salla fibroblast lines quickly produced, as expected, large amounts of free sialic acid. However, there was a reduced incorporation of [³H]Neu5Ac into both gangliosides and sialoglycoproteins, as compared to normal fibroblats. Furthermore, in spite of the large accumulation of free sialic acid only a very small part of it was catabolized to tritiated water. The lower degree of Neu5Ac salvage processes and the accumulation of free sialic acid appear to be an obvious consequence of the impaired exit of Neu5Ac from the pathological fibroblasts.

The activity of lysosomal glycosidases has been reported to be normal in Salla cells, although some results on the activity of sialidase seem controversial (7). In qualitative terms the metabolic processing of G_M3 in normal and Salla cells appeared to be very similar. However, some steps of the lysosomal degradation of gangliosides in Salla cells seem to be slowed down. The small amount of radioactivity carried by neutral lipids following cell administration of [stearoyl-¹⁴C]G_M3, with respect to that found after administration of [Sph-³H]G_M3, suggests that these lipids are mainly intermediates of the ganglioside catabolism. LacCer and Cer were found to accumulate in Salla fibroblasts. This is an indication that some enzymes of glycosphingolipid metabolism are affected by a specific impairment of lysosomal function. Particularly relevant is the accumulation of ceramide which was hardly recognized in control cells. No data are available on the activity of ceramidase in Salla cells. The increase in ceramide, which has been described to be a highly bioactive intermediate of sphingolipid metabolisms (40), could be associated to some clinical features of the Salla pathology. Of course, this hypothesis calls for further experimental evidence.