Lipid Composition of Human Malignant Melanoma Tumors at Various Levels of Malignant Growth

Size: px

Start display at page:

Download "Lipid Composition of Human Malignant Melanoma Tumors at Various Levels of Malignant Growth"

Loraine Garrison
5 years ago
Views:

1 Eur. J. Biochem., 1- (1) Lipid Composition of Human Malignant Melanoma Tumors at Various Levels of Malignant Growth Jacques PORTOUKALIAN, Georges ZWINGELSTEIN, and Jean-Frdnqois DORE Laboratoire de Physiologie GCnCrale et Comparee, Centre National de la Recherche Scientifique, UniversitC Claude Bernard (Lyon-I), Villeurbanne, et Laboratoire d Immunologie et de CancCrologie ExpCrimentale, Institut National de la SantC et de la Recherche Medicale FRA, Centre LCon Btrard, Lyon (Received September, 1) The lipid pattern of thirteen human melanoma tumors from various tissues were investigated. In seven of the tumors, an estimate was given about the proportion of malignant melanocytes to the total cell population, and a reverse correlation was determined between the proportion of malignant cells in these tumors and their neutral lipid content. The phospholipids did not show any modification, nor did the cholesterol in the cancerous tissues. The ganglioside pattern was found to be similar in all analyzed samples, with GMJ, GM and GD as major components, although no correlation was found between the malignant level and the ganglioside content of the tumors. Available information on the lipids of human cancers shows a wide range of variation in their lipid contents and compositions from those of normal tissue and even among the tumors investigated [l]. The large heterogeneity of the cell population in most of the tumors hampers the significance of the data provided since some alteration in the lipid pattern of the surrounding tissue might occur as an early response to the presence of cancerous cells. Moreover, the modifications induced in the cellular lipid metabolism by the malignant transformation are likely to concern specifically different lipid components with respect to the cell type involved. Thus, a systematic study of human tumors is necessary to approach any connection between the lipid metabolism and the malignant growth. Human malignant melanoma seemed a suitable model for such a study because of the high growth rate and occurrence of metastases of this tumor type. In the aim to understand more precisely the evolution in the lipid pattern of Abbreviations. For sialic acid-containing glycosphingolipids, the symbols proposed by Svennerholm [1] are used. GM, N- acetylneuraminylgalactosylglucosylceramide; GM, N-acetylgalactosaminyl-(N-acetylneuraminyl)-galactosylglucosylceramide ; GMI, galactosyl- N- acetylgalactosaminyl- (N- acetylneuramin yl) - galactosylglucosylceramide ; GDs, N-acetylneuraminyl-N-acetylneuraminylgalactosylglucosylceramide ; GD1,, N-acetylneuramin ylgalactosyl- N-acetylgalactosaminyl - (N-acetylneuraminy1)-galactosylglucosylceramide; GD, N-acetylgalactosaminyl-(N-acetylneuraminyl-Nacet ylneuraminy1)-galactosylglucosylceramide. Enzyme. Neuraminidase (EC..1.1). the tumors, an estimate of the level of malignancy in the tumors should be given whenever possible. MATERIALS AND METHODS In order to minimize the possible influence of individual diet, the tumors analyzed in this study were excised from patients who stayed several weeks in the hospitals before undergoing surgery, and thus were given a similar diet. No chemotherapy known to affect the lipid metabolism was used in the treatment of the disease. The melanoma tumors were provided by the Center Leon BCrard and the Departments of Dermatology of the Hospitals de 1 Antiquaille and Edouard Herriot in Lyons (France). Samples 1,,,, and were obtained by surgery and all others by autopsy performed within one hour after death. A sample of each tumor was taken to anatomo-pathological examination and the remaining part was frozen at - 0 C immediately until utilization. Table 1 indicates the location of each tumor, the pigmentation and whenever possible, an estimate of the proportion of malignant cells to the total cell population. The wet weight of the tumors ranged from.-1 g. The tissues were homogenized in chloroform/ methanol (/1, v/v) with a Polytron homogenizer and the lipids were extracted twice with 0ml of chloroform/methanol (/1, v/v), then chloroform/methanol

2 0 Lipids in Human Malignant Melanoma (l/l, v/v), by g of wet tissue. The delipidized tissues were dried and weighed. The lipid extracts were pooled and evaporated under reduced pressure at "C in a rotary evaporator. Purification of the lipids was achieved on Sephadex G- according to Wells and Dittmer []. The neutral lipid and polar lipid fractions were quantitated gravimetrically after chromatography on a column of silicic acid 0 mesh (Mallinckrodt, St Louis, Mo., U.S.A.) kept at 1"C, as described by Fillerup and Mead []. Total cholesterol was assayed on the neutral lipid fraction with the procedure of Zlatkis et al. []. The same procedure was used to quantify free and esterified cholesterol after thin-layer chromatography of the neutral lipids on silica gel G plates (Merck, Darmstadt, Germany) with a solvent cyclohexane/diethyl ether (0/0, v/v) and visualization with iodine vapors. The neutral lipid classes were identified by thinlayer chromatography on silica gel G on a solvent hexane/diethyl ether/acetic acid (O/lO/l, by vol.) along with known standards. After charring of the plate at 10 "C with potassium dichromate 1 % in M sulfuric acid, the proportion of each lipid class was determined by densitometry with a Vernon integrator, with corrections due to the specific absorbance of each compound. A minor component migrating slightly ahead of triacylglycerols was purified and identified as alkyldiacylglycerol by means of mild saponification and mild acid hydrolysis. Alternatively, the neutral lipid fraction was applied in a small amount of pentane to a silicic acid column and separated into individual lipid classes by a gradient of diethyl ether in pentane. The triacylglycerol fraction was taken to dryness and the constitutive fatty acids were recovered as methyl esters by transesterification with HS0 % in anhydrous methanol. Gas chromatography of the fatty acid methyl esters was carried out isothermally at 10 C on a -ft column, % butanediol succinate on HMDS Chromosorb W in a Perkin Elmer F gas chromatograph. Lipid phosphorus was assayed with a slight modification [] of the Bartlett's method, on the polar lipid fraction. The phospholipid distribution was determined as already described after two-dimensional thin-layer chromatography []. The total glycolipid content was evaluated by gravimetric difference between polar lipids and total phospholipids (pg P assayed X ). In some instances, the glycolipids were separated from phospholipids by the acetylation procedure of Saito and Hakomori [] and weighed. The two methods used to evaluate the glycolipid content gave comparable results. The glycolipids were identified by thin-layer chromatography, along with authentic standard extracted from erythrocytes [S], on silica gel G thin-layer plates in two solvent systems: chloroform/methanol/water (//, by vol.) and chloroform/methanol/0. % aqueous calcium chloride (0//, by vol.) []. Visualization was made with orcinol/sulfuric acid and resorcinol/hydrochloric acid [lo]. The ganglioside fraction was found to contain hematoside and some major components migrating above hematoside on thin-layer chromatography. The high-molecular-weight ganglioside fraction was extracted from the polar lipids by aqueous partition [ll], purified on Sephadex G- and weighed. In some instances, the total lipid-bound sialic was assayed by the thiobarbituric acid method [] which gave values consistent with the amount of partitioned gangliosides. Fig. 1 shows that the ganglioside profile was very similar in the tumors. Three major gangliosides were purified by thin-layer chromatography and the carbohydrate content was analysed as trimethylsilyl derivatives with the procedure outlined by Chambers and Clamp [] on a -ft column packedwith%ov-1ongaschromqs0-0mesh (Applied Sciences Lab., U.S.A.) with temperature programming 0- "C,. "C/min, on a Perkin Elmer 00 Gas Chromatograph. Alternatively, they were hydrolyzed by neuraminidase from Clostridium perfringens (type V, Sigma Chemical Co., St Louis, Mo., U.S.A.), and rechromatographed. These gangliosides were tentatively identified as GM, GM and GD using the nomenclature of Svennerholm [1]. GD was found as a minor component. RESULTS The tumors were listed from number 1 to number on a decreasing lipid content basis. As shown in Tables 1 and, the differences noticed in the lipid content were mainly due to the neutral lipids. This concerned especially the triacylglycerol content, since the free fatty acids comprised about %, and the diacylglycerol fraction about 1 % of the dry delipidized mass throughout all the tumors. The triacylglycerol fraction represented up to % of the total lipids extracted from tumor 1, whereas in sample, these compounds were detected in trace amount. The fatty acid pattern of this fraction was not significantly different between all tumors, with palmitic and oleic acids as major components, and the fatty acid pattern of the small amount of alkylglycerols purified from each tumor.was found to be very close to that of the triacylglycerols fraction. The results given in Table 1 suggested a correlation between the increasing proportion of malignant cells in the tumors and their decreasing total lipid content. On the seven samples in which both data were available, a reverse linear correlation was determined statistically (P < O.Ol), and the correlation was also highly significant using the triacylglycerol content instead of total lipid content.

3 J. Portoukalian, G. Zwingelstein, and J.-F. Dore 1 Table 1. Location and characteristics of the human malignant melanoma tumors analyzed, not determined Sample Location Histo-pathological Lipid content of the tumor examination pigmentation malignant cells a I Inguinal Axillary Tumor grown on location after removal of tumor Axillary Mesenteric Inguinal Ax i a r y Mesenteric Cutaneous primary tumor 0 0 % of total cells mg/0 mg of dry delipidized tissue a All values are mean of three different determinations, with range of variation - %. Table. Lipid composition of human malignant melanoma tumors Results are expressed as mg/0 mg of dry delipidized tissue Sample Neutral Phospho- Glyco- Dry delipilipids lipids" lipids" dized mass 1 mg/0 mg %wet tissue a All values are means of three different determination, with range of variations - %. Fig. 1. Thin-layer chrornatogram on silica gel H of polar lipids from melanoma tumors, after aqueous partition of the gangliosides. From left to right, samples,,,,,,. Solvent: chloroform/ methanol/0. "/, CaClz (0// by vol.). Visualization: orcinol/ sulfuric acid. (a) Ceramide monohexoside, (b) ceramide dihexoside, (c) ceramide trihexoside, (d) ceramide tetrahexoside, (e) GM ganglioside, ( GM ganglioside, () GD ganglioside The cholesterol and phospholipid contents were fairly constant among the tumors investigated and the molar ratio of cholesterol to phospholipid was stable. The total glycolipid content showed a peculiar pattern of variation with an elevation from sample 1 to sample, followed by a sharp decrease, and this pattern was even more obvious concerning the levels of gangliosides. The glycolipid distribution appeared of differ markedly in the lipid extracts as can be seen on Fig. 1, but since the tumors were excised from various tissues, this distribution was not determined in all cases. Comparison of the glycolipid profile of the tumorous livers with that of normal liver is given in Table. The differences in the neutral glycolipids were only quantitative, whereas the ganglioside pattern of the cancer tissues was noticeably modified from the normal counterpart, with a four-fold elevation of the lipid-bound sialic acid content found in human liver [1,1]. DISCUSSION The striking variations in the relative contents of neutral lipids and polar lipids of the tumors obtained for this study have already been reported for other types of cancer. Araki et al. showed the neutral lipid content of hepatoma tissues to vary from the level of normal liver [1] up to ten-times this level, whereas the phospholipid content was in many cases below the values given for the normal tissue [1].

4 ~ Lipids in Human Malignant Melanoma Table. Cholesterol content of human malignant melanoma tumors Expressed as mg/0 mg of dry delipidized tissue Sample Total Cholesterol esters Molar ratio cholesterol" free cholesterol to phospholipids Table. Ganglioside fraction extracted by aqueous partition from human malignant melanoma tumors Results are expressed as mg per g dry delipidized tissue and as % of phospholipids Sample Gangliosides 1 mg/0 mg % total cholesterol mg/g % I a All values are mean of three determinations with range of variations - %. Table. Distriburion of neulral glycolipids and gangliosides in 'liver melanoma tumors and rrormal (iver Neutral glycolipids were quantified by densitometry following thin-layer chromatography of the fraction purified by acetylation procedure, and visualization with orcinol/sulfuric acid. Standardization was carried out with known amounts of standard glycolipids. Percentage distribution of sialic acid was determined on the upper phase after aqueous partition, by densitometry following thin-layer chromatography and visualization with resorcinol/hydrochloric acid; tr., traces;, not detected Glycolipid or ganglioside Tumor sample Normal liver mg/g dry delipidized tissue Ceramide nionohexoside Ceramide dihexoside Ceramide trihexoside Ceramide tetrahexoside % ~. _ ~ tr tr.. tr pmol/g dry delipidized tissue Total lipid-bound sialic acid The tremendous amount of glycerides contained in the tumors at the early cancerous level is likely to reflect a primary response of the surrounding tissue to the presence of fast-growing malignant cells, and the apparent correlation, which we determined between the glyceride content of the tumors and the proportion of melanocytes, might show a progressive utilization by the melanocytes of these glycerides stored in the adjacent tissue. This would account for the fact that only trace amount of glycerides were found in samples and. The question remains whether this correlation is a general feature in cancer, and we cannot discard the possibility that the variability of the neutral lipid content could be due to the type of tumor used rather than the proportion of malignant cells. Further experimentation with other types of tumors is needed for a definite answer.

J. Portoukalian, G. Zwingelstein, and J.-F. Dort The other main result of this study concerns the patterns of variation of the ratio of gangliosides to phospholipids (see Table ).

5 J. Portoukalian, G. Zwingelstein, and J.-F. Dort The other main result of this study concerns the patterns of variation of the ratio of gangliosides to phospholipids (see Table ). The occurrence of high amounts of GM and GD gangliosides in all melanoma tumors reflects the presence of these compounds as components of the melanocytes cells since the ganglioside composition of three malignant melanocyte cell lines freshly established from human tumors, has been recently determined, with GM, GMz and GD as major components and GDz as a minor component [1]. In this respect, the amount of water-extractable gangliosides in the tumors analyzed should have been increased as the malignant cells gradually outnumbered the normal cells. Such a correlation was not found. Moreover, preliminary analysis showed that the ganglioside content in the peripheral zone of some tumors was higher than in the central part. This observation has already been reported by Slagel et al. [1] on human brain tumors. Since malignant cells in culture do not show any density-dependent alteration in the glycolipid pattern 0-1 it might be suggested that the ganglioside biosynthesis in malignant melanocytes is sensitive to some serum factors with a further loss of influence once the malignant cell density reaches a sufficient level. Experiments are now in progress in our laboratory in brder to understand this phenomenon which emphasizes our previous finding on the immunoreactions occurring between some patients sera and the high molecular weight ganglioside fraction purified from melanoma tumors []. A significant elevation in the ganglioside content of the patients blood, concerning GM and GD in plasmas and GM in red cells, has been recently shown [] and there might be a correlation between this alteration and the observed modification of the ganglioside content of the tumors during malignant growth. The authors wish to thank Prof. S. Hakomori for helpful discussion and for reading the manuscript. Histopathological examination of the tumors by Dr C. Bailly is gratefully acknowledged. This work was partly supported by a grant number 1--- from the Institut National de la SantP et la Recherche Mkdicale. REFERENCES 1. Bergelson, L. D. (1) Progress in the Chemistry of Fats and other Lipids (Holman, R. T., ed.) vol., part 1, Pergamon Press, Oxford.. Wells, M. A. & Dittmer, J. C. (1) Biochemistry,, -.. Fillerup, D. L. & Mead, J. F. (1) Proc. Soc. Exp. Biol. Med., -.. Zlatkis, A,, Zak, B. & Boyle, A. J. (1) J. Lah. Clin. Med. 1, -.. Zwingelstein, G., Meister, R. & Brichon, G. (1) Biochimie,,0 -.. Portoukalian, J., Meister, R. & Zwingelstein, G. (1) J. Chromatogr. 1, -.. Saito, T. & Hakomori, S. (11) J. Lipid Res., -.. Hakomori, S., Siddiqui, B., Li, Y.-T. & Hellerqvist, C. G. (11) J. Biol. Chem., van den Eijnden, D. H. (11) Hoppe-Seyler s. Physiol. Chem., Svennerholm, L. (1) Biochim. Biophys. Acta,,0-.. Folch, J., Lees, M. & Sloane Stanley, G. H. (1) J. Biol. Chem., Warren, L. (1) J. Biol. Chem., Chambers, R. E. & Clamp, J. R. (11) Biochem. J., Svennerholm, L. (1) J. Neurochem., Brunngraber, E. G., Berra, B. & Zambotti, V. (1) Clin. Chim. Acta,, Araki, E., Phillips, F. Sr Privett, 0. S. (1) Lipids,, Fredrickson, D. S. (1) in The Metabolic Basis of Inherited Disease (Stanbury, J. B., Wyngaarden, J. B. & Fredrickson, D. S., eds) pp. 0-, McGraw-Hill Book Co., New York. 1. Portoukalian, J., Zwingelstein, G. & DorC, J. F. (1) Fourth International Symposium on Malignant Melanoma, Lyon, France, Rev. Inst. Pasteur Lyon, in the press. 1. Slagel, D. E., Dittmer, J. C. & Wilson, C. B. (1) J. Neurochem. 1, Hakomori, S., Kijimoto, S. & Siddiqui, B. (11) Fed. Proc. 0,. 1. Robbins, P. W. & McPherson, I. (11) Proc. R. Soc. Lond.. Biol. Sci Sakiyama, H., Gross, S. K. & Robbins, P. W. (1) Proc. Nut1 Acad. Sci. U.S.A., -.. Portoukalian, J., Zwingelstein, G., Dork, J. F. & Bourgoin, J. J. (1) Biochimie,, -.. Portoukalian, J., Zwingelstein, G., Abdul-Malak, N. s( Dort, J. F. (1) Biochem. Biophys. Res. Commun., 1-0. J. Portoukalian and G. Zwingelstein*, Laboratoire de Physiologie Gkntrale et Comparee, Universitt Claude-Bernard, Boulevard du Novembre 11, F-1 Villeurbanne, France J.-F. DorC, Laboratoire d Immunologie et de Canctrologie Experimentale, Centre Leon Berard, F- Lyon-Cedex-, France * To whom correspondence should be addressed

UMR 8612, Faculty of Pharmacy Chatenay-Malabry. Natura-Brasil. EA Laboratory of Dermatological Research,

Iuliana Popa 1, Noëlle Remoué 2 and Jacques Portoukalian 3 1 UMR 8612, Faculty of Pharmacy Chatenay-Malabry 2 Natura-Brasil 3 EA 41 69 Laboratory of Dermatological Research, University of Lyon I, Faculty