- Email: [email protected]
Tumoricidal action of cis-unsaturated fatty acids and their relationship to free radicals and lipid peroxidation
Cancer Letters, 56 (1991) 235-243 Elsevier Scientific Publishers Ireland
235 Ltd
Tumoricidal action of &-unsaturated fatty acids and their relationship to free radicals and lipid peroxidation U.N.
Das
Department
of Medicine,
The
Nizam’s
Institute
of
Medical Sciences,
Punjagutta,
Hyderabad
(India)
12 December 1990) (Revision received 10 January 1991) (Accepted 11 January 1991)
(Received
Summary &-unsaturated fatty acids (c-UFAs) such as gamma-linolenic acid (GLA), arachidonic acid (AA) and eicosapentaenoic acid (EPA) can kill tumor cells selectively in vitro. As c-UFAs have the ability to augment free radical generation, the effect ofantioxidants, free radical quenchers and augmenters offree radical generation such as iron and copper salts on fatty acid-induced tumor cell death was studied. In addition, the role of lipid peroxidation in the tumoricidal action of c-UFAs was also examined. Results indicate that vitamin E, uric acid, glutathione peroxidase, superoxide dismutase and ATP can block, whereas iron, copper and catalase enhance the tumoricidal action of GLA. The ability of GLA, AA and EPA to kill tumor cells correlated with the amount of lipid peroxidation these fatty acids can induce as measured by thiobarbituric acid test. It was also observed that 14C-labelled linoleic acid uptake was almost the same whereas that of “C-labelled arachidonic acid and eicosapentaenoic acid were substantially less in tumor cells compared to normal cells. Tumor cells incorporated major portions of the fatty acids in the ether lipid and Correspondence to: U.N. Das, Department of Medicine, The N&am’s Institute of Medical Sciences, Punjagutta. Hyderabad-500
482,
India.
0304.3835/91/$03.50 Published and Printed
0 1991 Elsevier Scientific Publishers in Ireland
phospholipid fractions, whereas normal cells incorporated the fatty acids primarily in the phospholipid fraction. These results suggest that c-UFA-induced tumoricidal action is a free radical dependent process and that there are significant differences between normal and tumor cells in fatty acid uptake and distribution.
Keyarords: cis-unsaturated fatty-acids, tumoricidal, free radicals, lipid peroxidation, superoxide anion, anti-oxidants Introduction Studies [1,2,5,22,23] have shown that some &-unsaturated fatty acids (c-UFAs) such as gamma-linolenic acid (GLA, 18:3 n-6)) dihomoGLA (DGLA, 20:3 n-6)) arachidonic acid (AA, 20:4 n-6), and eicosapentaenoic acid (EPA, 20:5 n-3) and to some extent linoleic acid (LA, 18:2 n-6) and alpha-linolenic acid (ALA, 18:3 n-3) can selectively kill tumor cells without harming the normal cells at the concentrations tested. Inhibition of cycle-oxygenase and lipoxygenase enzymes did not prevent the tumoricidal action, but antioxidants did, suggesting that controlled formation of peroxidation products and free radicals may be involved [3,6]. In an extension of these studies, it was observed that GLA, AA and EPA were 1.5-2 times more efIreland Ltd
236
fective than LA in inducing free radical generation in tumor cells but not in normal cells [7,8]. It is well documented that both hydrogen peroxide (H202) and the superoxide (O,? ) radical can oxidize c-UFAs of the cell membrane, producing such cytotoxic metabolites such as malondialdehyde (MDA), 2-alkanals and hydroxy-alkenals [ 111. These substances exert influences on cross-linking with amino groups of DNA through the formation of Schiff bases [17] and thus, free radicals can exert their action on DNA. In addition, it is known that some metals such as copper and iron can augment free radical generation and lipid peroxidation by a non-enzymatic process [ 141. In view of this, the effect of iron and copper metals on GLAinduced tumoricidal action, the possible relationship between lipid peroxidation and c-UFAinduced tumor cell death and the uptake and distribution of fatty acids in normal and tumor cells were studied and the results are reported here. Materials
and Methods
Cells and culture conditions Normal monkey kidney (CV-l), normal human fibroblasts (CCD-41-SK), murine fibroblast transformed by Abelson leukemia virus (BALB), human breast cancer (ZR-75-l) and human promyelocytic leukemia (HL-60) cells were used for the study. The cells were seeded at 5 x lo4 or 1 X lo4 cells per plate or well depending on the experimental protocol and as described earlier [l]. The cells were grown in 0.5 ml or 2.0 ml (in 24-well tissue culture plates or petri dishes, respectively) of bicarbonate buffered Dulbecco’s modified Eagle’s medium (Sigma Chemical Co., U.S.A.) with or without the added fatty acids at 37 “C in a 5% CO, humidified incubator as described earlier [ 11. The fatty acid esters were initially dissolved in 95% ethanol and the final concentration of ethanol was not more that 0.2% in all control and fatty acid supplemented cultures. Cell viability was determined by the trypan blue dye exclusion method. The cells were checked for possible mycoplasma contamination by fluorescent
technique and confirmed that there was no contamination during the period of these studies. Studies with antioxidants and metals These studies were performed in 24 well tissue culture plates where in the cell lines were seeded at 1 x lo4 cells per well. One day after seeding, 10 pg/ml of GLA was added to the cells. To study the effect of possible enhancers or inhibitors of tumor cell death by GLA, the following were added simultaneously to the cultures with GLA: vitamin A, vitamin E, transretinoic acid, uric acid, superoxide dismutase, ATP, glutathione peroxidase, catalase, heatinactivated catalase, FeCI,, FeCI, (ferrous and ferric chloride, respectively) and CuSO, (copper sulfate). All the chemicals were obtained from Sigma Chemical Co., U.S.A. and were of highest grade of purity available. Appropriate controls without GLA were performed. The cultures were observed every day and cell viability was determined every day until day 7 of supplementation of GLA and the various additives. All these studies were done at 10 or 20 pg/ml of GLA concentration depending on the protocol of the experiment(s). These studies were performed with ZR-75-1 cells as a representative of the tumor cells. TBA
reaction
For these studies, cell cultures seeded with 5 x lo4 cells/petri dish (35 mm) were used. ZR-75-1 cells and CV-1 cells were used in this study. Cell cultures were supplemented with 20 pg/ml of ethyl esters of GLA, AA and DHA (docosahexaenoic acid) and methyl ester of EPA. The cells were grown in 2.0 ml medium with or without added fatty acids at 37°C in a 5% CO, humidified incubator. Similar to the experiments with antioxidants and metals, 1 day after seeding, the cells were supplemented with various fatty acids. At the end of 7 days, medium and cells were harvested separately and assayed for TBA reaction as a measure of lipid peroxidation. TBA reaction was performed as described by Gavin0 et al. [13] and as described earlier [21]. The absorbance of the reaction was measured at 532 nm with growth medium
237
or PBS as the controls. The absorbance values obtained were converted to pmol of MDAequivalent (MDA-eq) from a standard curve obtained with 1,1,3,3-tetramethoxypropane.
c-UFA incorporation studies In order to know whether the selective cytotoxicity shown by fatty acids could be due to increased uptake, we studied the fatty acid uptake by normal and tumor cells in vitro [7,8]. Twenty four hours after seeding (1 x lo4 cells/O.5 ml medium), 0.1 &i of 14C-labelled LA, AA or EPA was added (spec. act.: LA, 52.6 mCi/mmol; AA, 54.5 mCi,/mmol; EPA, 55.4 mCi/mmol) to study the uptake of fatty acids by normal and tumor cells. The cells were washed three times in PBS, detached by trypsinisation and the total amount of fatty acid incorporated was counted in a liquid scintillation counter on day 1, 2 and 3 for all the fatty acids tested.
Determination of c-UFA in different lipid fractions To study the incorporation of labelled fatty acids in different lipid fractions, 5 x 104 cells were incubated with 0.5 PCi of LA, AA and EPA. Twenty four hours after incubation, the cells were washed in PBS, lysed and extracted in chloroform/methanol (2: 1, v/v). Different lipid fractions were separated by thin layer chromatography as described by Roos and Choppin [20]. Regions corresponding to phospholipid (PL) , free fatty acid (FFA), ether lipid (EL), cholesterol ester (CE) and triglyceride (TG) fractions were scraped from the developed chromatography plates and extracted in chloroform/methanol (2:1, v/v). The extract was evaporated under nitrogen and counted in a liquid scintillation counter [8]. All the fatty acids except EPA were ethyl esters and EPA is a methyl ester and were obtained from Sigma Chemical Co., U.S.A. All experiments were done in triplicate/quadruplicate and repeated at least 3 or 4 times.
The effect of 20 pg/ml of various c-UFAs on the survival of human breast tumor (ZR-75- l), normal monkey kidney (CV-1) and normal human fibroblast (CCD-41-SK) cells are given in Fig. 1. It is evident from this that GLA and AA are equipotent in their tumoricidal action, where as EPA has less potent action. On the other hand DHA was without any significant tumoricidal action in the test system used. Thus, AA = GLA > EPA > DHA in terms of their action on tumor cells. It is also noted that the maximum tumoricidal action of c-UFAs is seen by day 6 or 7. The results shown in Fig. 2 indicate that both AA and EPA but not GLA are toxic to normal cells (CCD-41-SK) at doses 2 or 3 times the dose effective in killing the tumor cells. This is especially so when normal cells (CCD-41-SK, 1 x 104) were supplemented with 60 pg/ml of AA
% DEAD CELLS
T
T
100 -
80 -
60 -
40 -
20 -
03
3
6
6
7
DAYS Fig. 1.
Effect of 20 pg/ml
ZR-75-1,
CV-1
and CCD-41
of c-UFAs on the survival of SK cells in vitro.
238 % OF
DEAD CELLS
and
ON DAY 7
100 -
0'
I
I
I
I
60 20 40 10 CONCENTRATION OF FATTY ACID IN W/ml
Fig. 2.
Effect of GLA, AA and EPA on the survival of normal (41 SK) and tumor cells in vitro. N = normal cells; T = tumor cells. Table 1. Effect of antioxidants No. 1 2 3 4 5 6
7 8
9 10
and other chemicals
Chemical/Enzyme
Dose(s)
Control GLA Vitamin A Vitamin A Vitamin E Vitamin E Trans-retinoic acid Uric acid Uric acid Uric acid Cis-retinoic acid Superoxide dismutase SOD SOD SOD ATP Glutathione peroxidase Glutathione peroxidase
-
All chemicals were tested on day 7 of the addition of vitamin A, E, retinoic values are expressed as
EPA. This suggests that GLA has more selective tumoricidal action compared to AA and EPA. Thus, GLA > AA > EPA when their capacity to kill tumor cells selectively and least toxicity to normal cells are taken into account. The effect of various inhibitors or enhancers of the tumoricidal action of GLA on human breast tumor cells are given in Tables I and II. For testing the action of inhibitors, a concentration of 20 pg/ml of GLA was used since, earlier studies and the results shown in Fig. 1 showed that this is the optimum concentration to produce maximum cytotoxicity to tumor cells with least cytotoxicity to normal cells [ 1,2,7]. On the other hand, to test the action of possible enhancers of cytotoxicity, a sub-optimal dose of GLA which is 10 pg/ml was used. It is clear from these results, summarised in Table I, that antioxidants and superoxide quencher, superoxide dismutase (SOD) can effectively block the tumoricidal or cytotoxic action of GLA. Similar results were obtained with AA and EPA (data
on the cytotoxicity used
20 j&ml 10 pg/ml 100 pg/ml 10 &ml 100 pg/ml 1 x 10e5 M/ml 1 x 10e4 M/ml 1 x 10m5 M/ml 1 x 10m6 M/ml 1 x 10m5 M/ml 1 x 10m6 M/ml 3 x 10m6 M/ml 3 x 1O-7 M/ml 3 x 10m8 M/ml 200 PM/ml 0.1 pg/ml 0.01 pg/ml
of GLA to human breast cancer cells in vitro. % Dead cells 20.0 95.0 30.0 32.0 10.0 10.0 35.0 40.0 47.0 51.0 65.0 15.0 20.0 22.0 30.0 22.0 21.0 28.0
f f f f f f + * f f f f f f f f f +
5.0 5.0 7.0 9.0 5.0 7.0 8.0 10.0 8.0 7.0 18.0 5.0 7.0 7.0 10.0 9.0 12.0 8.0
along with 20 pg/ml of GLA on ZR-75-1 cells (1 x lo4 cells/ml). All cultures were harvested of GLA with and without the inhibitors. Percentage of dead cells with various concentrations acid, uric acid, SOD, ATP and glutathione peroxidase alone is approximately 14.0 * 7.0. All mean f S.D.
239 Table 11. Effect of iron and copper salts and catalase on the cytotoxicity induced by GLA on human breast tumor cells. No.
Chemical
Dose tested
% Dead cells
1 2 3
Control GLA Catalase Catalase Catalase Heat-inactivated catalase Heat-inactivated catalase Heat-inactivated catalase FeCI, FeCI, FeCI, FeCI, FeCla FeCI,
-
20.0 f 60.0 zt 75.0 f 80.0 f 81.0 f 72.0 zt
4
5
6
FeCI,
7
cuso4 cuso,
10 500 1000 3000 500
&ml U/ml U/ml U/ml U/ml
5.0 15.0’ 10.0’ 12.0’ 8.0” 12.0
1000 U/ml
77.0
f
3000 U/ml
73.0
f
j&ml &ml 20 &ml 40 &ml
61.0 65.0 98.0
ztz 12.0 zt 9.0 f 2.0”
1 &ml 4 pg/ml
65.0 68.0
1 4
20 pg/ml 1 &ml 4 Irg/ml
l
8.0’ 14.0
98.0zt 2.0' l
LIZ10.0 zt 8.0
95.0f 5.0' 57.0f 11.0 78.0
f
14.0”
All chemicals were tested along with 10 pg/ml of GLA on ZR-75-1 cells (1 x lo4 cells/ml). All the cultures were harvested on day 7 of addtion of GLA with and without other chemicals. Percentage of dead cells with catalase, inactivated catalase, FeCI,, FeCI, and CuSO, alone at various concentrations used was approximately 18 * 5. All values are expressed as mean =t S.D. For other details see Materials and Methods. ‘P < 0.001 compared to control. *‘P < 0.05 compared to GLA group.
not shown). The inhibitory action shown by SOD suggests that possibily, superoxide radical has a role in the tumoricidal action of c-UFAs. Both vitmain A and E are believed to have potent antioxidant actions and prevent lipid peroxidation. Similarly, uric acid is an antioxidant. The inhibitory action of ATP is rather surprising. Both catalase and heat-inactivated catalase enhanced the cytotoxic action of GLA at the doses tested (Table II). This suggests that H,O, radical does not participate in the tumoricidal action of c-UFAs. Both FeCI, (when used at 20 and 40 pg/ml) and CuSO, (when used at 4 pg/ml) enhanced the cytotoxicity of GLA (Table II). Since these metals are known to augment lipid peroxidation and free radical generation (especially that of superoxide anion, 14) their enhancing action is understandable. Catalase is
rich in copper and this may explain why heatinactivated catalase enhanced the tumoricidal action of GLA. Table III depicts the MDA-eq detected in the medium and cells on day 7 after supplementation with 20 pg/ml of various c-UFAs. Both AA and GLA supplemented tumor cells (ZR-75-1) produced large amouts of MDA-eq (approx. 4-fold increase) compared with normal cells (CV-1) tested and controls. But surprisingly, EPA-supplemented normal and tumor cells produced significantly increased amounts of MDAeq as compared with controls. Although this result looks paradoxical, this may explain at least in part, why normal cells were more susceptible to the cytotoxic action of EPA compared with that of GLA and AA (Fig. 2). Figure 3 depicts the fatty acid incorporation
240 Table III.
Effect of various c-UFAs on MDA-eq formation
Cell line
Treatment
cv-1 (Normal monkey
kidney cells)
ZR-75- 1 (Human breast cancer cells)
All the values are expressed
as mean
Control GLA AA EPA DHA Control GLA AA EPA DHA f S.D. of three separate
and tumor cells. Uptake of LA was almost the same in normal (41-SK and CV-1) and tumor (ZR-75-1 and HL-60) cells, especially at 48 and 72 h. In contrast, both AA and EPA
in normal
1x1000 35
COUNTS
a
T
in normal and tumor cells on day 7. pmol MDA-eq In the cells
In the medium
0.5 6.2 7.1 15.0 11.4 0.2 26.5 35.9 16.7 11.3
0.3 1.1 2.8 3.3 10.1 0.9 10.6 3.5 10.2 14.4
estimations.
24HRS
lx1000
48HRS
0.3 1.1 0.4 2.7 3.0 0.3 2.7 10.4 1.7 3.9
f 0.1 f 0.3 f 1.0 f 0.5 f 0.3 f 0.2 z!z 2.4 zt 1.6 f 0.6 f 0.4
For other details see Materials and Methods.
incorporation was low in the tumor cells (Fig. 3b and 3~). There were no significant differences between normal and tumor cells in thymidine incorporation (data not shown) suggesting that their growth rates were similar. Hence, the differences in the uptake of fatty acids can not be attributed to changes in cell growth. Table IV shows the distribution of labelled LA, AA and EPA in different lipid fractions of both normal and tumor cells. In normal (41-SK and
1x1000
0
f f f zt f f +z f zt f
COUNTS
72rliiS
COUNTS
25
Fig. 3. Incorporation of (a) 14C-labeled LA, (b) 14Clabeled AA and (c) 14C-labeled EPA in 1 x 104 cells. ( l ) CV-1, (e) HL-60, ( q) 41-SK, ( l) ZR. All values are expressed as mean f S.D.
241 Table IV. Uptake and distribution of different cis-unsaturated ceils in vitro. All values are expressed as mean + S.D. Cell line
Percentage
Total
fatty acids in different lipid fractions of normal and tumor
of distribution
PL
FFA
TG
CL
CE
Normal
cv-1 LA AA EPA
30,253 37,140 58,460
zt 5810 + 6483 zt 8609
59.0 72.6 80.3
5.0 14.9 1.9
1.9 2.1 0.8
1.0 0.4 1.2
22.8 9.9 17.2
41-SK LA AA EPA
29,997 25,362 82,460
f 5406 f 2375 ztz 5481
72.0 78.2 71.7
2.0 9.1 1.5
1.9 2.0 0.7
6.4 4.0 10.7
21.6 6.6 13.8
11,019 zt 2742 13,435 f 672 9377 zt 1582
64.0 83.0 58.0
3.8 14.5 4.7
0.9 1.5 1.1
1.3 1.2 2.1
22.2 3.5 25.3
8589 + 1296 4396 + 142 5010 f 834
44.0 36.6 30.6
7.8 8.3 7.4
5.7 3.7 2.4
2.8 2.9 2.2
39.8 48.5 49.4
Tumor cells
ZR-75- 1 LA AA EPA BALB LA AA EPA
PL = Phospholipids,
FFA = free fatty acids, TG = triglycerides,
cells, LA, AA and EPA were incormainly in the PL and less than 22% in the EL fractions. On the other hand, in the tumor (BALB and ZR-75-1) cells a major portion of the fatty acid was incorporated both in to PL and EL fractions except in ZR-75-1 in which the distribution of fatty acids was similar to that of the normal cells. Incorporation of fatty acids into FFA, TG and CE fractions was similar in normal and tumor cells [7,8]. CV-1)
porated
Discussion The results presented here support and extend the earlier results that GLA, AA and EPA can selectively kill tumor cells and suggest that superoxide anion and lipid peroxidation process have a role in the tumoricidal action of c-UFAs [l-3,5,7,8]. The observation that SOD but not
CE = cholesterol
esters,
EL = ether lipids.
catalase and mannitol can prevent the cytotoxic action of c-UFAs suggests a role for superoxide radical but not for H,O, and hydroxyl radical (results with mannitol were not presented here as it did not have any significant inhibitory effect on c-UFA-induced cytotoxicity) . The ability of vitamin A and E to prevent the cytotoxic action of c-UFAs indicates that lipid peroxidation process has a role. The enhancing effect of copper and iron lends further support to this contention. The demonstration of an increase in the intracellular MDA-eq in tumor cells supplemented with GLA, AA and EPA adds strength to this concept. The increased amounts of MDA-eq observed in normal cells supplemented with EPA suggests that increased lipid peroxidation and/or free radical generation in the cells beyond a limit can induce cell death since EPA is toxic to normal cells at higher con-
242
centrations (Fig. 2). The increased sensitivity of tumor cells to GLA, AA and EPA may be due to their low content of SOD [12]. The inability of DHA (22:6, w-3) to kill tumor cells (Fig. 1) in spite of its high degree of unsaturation and capacity to undergo peroxidation readily may be due to its unstable nature and/or its ineffectiveness to increase specific type of lipid peroxidation in the cells (Table II). The most surprising observation is the ability of ATP to block the cytotoxic action of GLA. In an earlier study, Hilf et al. [16] showed that hematoporphyrin derivative-induced photosensitization, a procedure which also enhances free radical generation in the cells, in R 3230 AC mammary tumors causes a 60% reduction in cellular ATP levels. If, c-UFA-induced cell death is also mediated by interfering with ATP metabolism, it may explain the inhibitory effect of ATP on GLA-induced cytotoxicity. But, it should be mentioned here that ATP can not enter the cells. Hence, this possibility needs to be tested. The other possibility to be considered is the ability of ATP to form complexes with EFAs and thus, block the availability of fatty acids to the cells. It is interesting to note that tumor cells incorporated less AA and EPA than normal cells and that major portions of the fatty acids are incorporated in PL and EL fractions. In spite of this, AA and EPA-treated tumor cells produced higher amounts of lipid peroxides (Table III) which are known to be toxic to cells [15]. This suggests that the low rates of lipid peroxidation seen in tumor cells [12] could be due in large part to low substrate availability. The amounts of free fatty acids both in normal and tumor cells were almost the same and hence, increased free fatty acids can not explain the enhanced lipid peroxidation and/or toxicity in tumor cells. Compared to normal cells, a significant proportion of the incorporated fatty acids are found in the EL fraction in the tumor cells [7,8]. The possibility that the substrates for lipid peroxidation are derived from the EL fraction, which is considered to be metabolically inactive, remains to be tested. It is also likely that the source of fatty acids for lipid peroxidation and free radical generation may be the PL frac-
tion as the activity of phospholipase A2 in the tumor cells is high compared to normal cells
1181. The results of these studies are interesting since it is known that several anti-cancer drugs such as vincristine, Adriamycin etc., enhance free radical generation and lipid peroxidation in the plasma of patients following chemotherapy [21]. Even lymphokines such as interferon and tumor necrosis factor have the capacity to augment free radical generation in the tumor cells and human neutrophils [4,9]. If so, this indicates that free radicals and free radical-dependent lipid peroxidation process may be a common pathway by which tumor cells are eliminated both by natural products of immune cells and drugs.. Thus, methods designed to specifically enhance superoxide radical generation and lipid peroxidation in the tumor cells may form a novel approach to cancer therapy [lo, 191. Acknowledgements This work was supported in part by a grant from the Department of Science and Technology, New Delhi to U.N. Das. Dr. Das was also in receipt of INSA Research Fellowship during the tenure of this study. References Begin, M.E., Das, U.N., Ells, G. and Horrobin. D.F. (1985) Selective killing of human cancer cells by polyunsaturated fatty acids. Prostaglandins Leukotries Med., 19, 177- 186. Begin, M.E., Ells, G., Das, U.N. and Horrobin, D.F. (1986) Differential killing of human carcinoma cells supplemented with n-3 and n-6 polyunsaturated fatty acids. J. Natl. Cancer Inst., 77, 1053-1062. Begin, M.E, Das, U.N. and Ells, G. (1985) Mechanism of essential fatty acid induced cytotoxicity in malignant cells, 2nd Int. Congress. Essential Fatty acids Prostaglandins and Leukotrienes March, London, U.K., p. 7. Berton, G., Zeni. L., Cassatella. M.A. and Rossi. F. (1986) Tumor necrosis factor provokes superoxide anion generation from neutrophils. Biochem. Biophys. Res. Commun., 138, 1276-1282. Booyens. J.. Engelbrecht, P.. LeRoux, S.. et al. (1984) Some effects of the essential fatty acids; linoleic acid and alpha-linolenic acid and of their metabolites gamma-linolenic acid, arachidonic acid eicosapentaenoic acid, docosahexaenoic acid and of prostaglandins A and E on the prolifera-
243
tion of human osteogenic sarcoma cells in culture. Prostagladins Leukotrienes Med., 15, 15-34. Das, U.N., Huang, Y.S., Begin, M.E. and Horrobin, D.F. (1986) Interferons, phospholipid metabolism, immune responses and cancer. IRCS Med. Sci., 14, 1069-1074. Das, U.N., Begin, M.E., Ells, G., Huang, Y.S. and Horrobin, D .F. (1987) Polyunsaturated fatty acids augment free radical generation in tumor cells in vitro. Biochem. Biophys. Res. Commun.,
145,
icity, oxygen radicals, transition metals and disease. Biochem. J., 219, l-14. 15
16
(1986) Relationship of mitochondrial function and cellular adenosine triphosphate levels to hematoporphyrin derivativeinduced photosentitization in R3230 AC mammary tumors. Cancer Res.. 46, 211-217.
15-24.
8 Das, U.N., Haung, Y.S., Begin, M.E., Ells, G. and Hor-
9
10 11
12
13
14
Halliwell, B. and Gutteridge, J.M.C. (1984) lipid peroxidation oxygen radicals and cell damage and antioxidant therapy. Lancet, i: 1396-1397. Hilf, R.. Murant, R.S., Narayanan. U. and Gibson, S.L.
robin, D.F. (1987) Uptake and distribution of cis-unsaturated fatty acids and their effect on free radical generation in normal and tumor cells in vitro. Free Radical Biol. Med., 3, 9-14. Das, U.N., Ells, G.. Begin, M.E. and Horrobin, D.F. (1986) Free radicals as possible mediators of the actions of interferon.
17
Kappus. H. (1985) Lipid peroxidation: Mechanisms, analysis, enzymology and biological relevance. In: Oxidative stress, pp. 273-303. Editor: H. Sies. Academic Press, New York.
18
J. Free Rad. Biol. Med., 2, 183-188. Das. U.N. (1987) Biological significance of arachidonic acid. Med. Sci. Res., 15, 1485-1490. Esterbauer, H. (1982) Aldehydic products of lipid peroxidation, In: Free Radicals, lipid peroxidation and cancer, pp. 101-128. Editors: D.G. McBrien and T.F. Slater. Academic Press, London. Galeotti, T., Bantoli, G.M. and Bartoli, E. (1982) Superoxide radicals and lipid peroxidation in tumor microsomal membranes. In: Biological and Clinical aspects of superoxide and superoxide dismutase, pp. 106-117. Editors: W.H. Bannister and J.V. Bannister. Proc. Fed. Eur. Biochem. Sot.
19
Levine, L. (1981) Arachidonic acid transformation and tumor promotion. Adv. Cancer Res., 35. 49-79. Rotilio. G., Bozzi. A., Mavelli, B. et al. (1982) Tumor cells as models for oxygen-dependent cytotoxicity. In: Biological and Clinical aspects of superoxide and superoxide dismutase. pp. 118-126. Editors: W.H. Bannister and J.V. Bannister. Proc. Fed. Eur. Biochem. Sot. Symposium. No. 62. Roos, D.S. and Choppin. P.W. (1985) Biochemical studies on cell fusion: I. Lipid composition of fusion resistant cells. J. Biol. Chem., 101. 1578-1590. Sangeeta, P., Das, U.N.. Koratkar, R. and Suryaprabha. P. (1990) Increase in free radical generation and lipid perox-
Symposium No. 62. Gavino, V.C., Miller, J.S.. Ikareblha. S.O., Mile. G.E. and Cornwell, D.G. (1981) Effects of polyunsaturated fatty acids and antioxidants on lipid peroxidation in tissue cultures. J. Lipid Res., 22, 763-769. Halliwell, B. and Gutterridge, J.M.C. (1986) Oxygen tox-
20
21
22
23
idation following chemotherapy in patients with cancer. Free. Rad. Biol. Med., 8, 15-20. Seigel, I., Liu, T.L., Yaghoubzadeh. E Kaskey. T.S. and Gleicher, N. (1987) Cytotoxic effects of free fatty acids on ascites tumor cells. J. Natl. Cancer Inst.. 78. 271-277. Tolnai. S. and Morgan, J.F. (1962) Studies on the in vitro anti-tumor activity of fatty acids. V. Unsaturated fatty acids. Can. J. Biochem Physiol , 40, 869-875.
235 Ltd
Tumoricidal action of &-unsaturated fatty acids and their relationship to free radicals and lipid peroxidation U.N.
Das
Department
of Medicine,
The
Nizam’s
Institute
of
Medical Sciences,
Punjagutta,
Hyderabad
(India)
12 December 1990) (Revision received 10 January 1991) (Accepted 11 January 1991)
(Received
Summary &-unsaturated fatty acids (c-UFAs) such as gamma-linolenic acid (GLA), arachidonic acid (AA) and eicosapentaenoic acid (EPA) can kill tumor cells selectively in vitro. As c-UFAs have the ability to augment free radical generation, the effect ofantioxidants, free radical quenchers and augmenters offree radical generation such as iron and copper salts on fatty acid-induced tumor cell death was studied. In addition, the role of lipid peroxidation in the tumoricidal action of c-UFAs was also examined. Results indicate that vitamin E, uric acid, glutathione peroxidase, superoxide dismutase and ATP can block, whereas iron, copper and catalase enhance the tumoricidal action of GLA. The ability of GLA, AA and EPA to kill tumor cells correlated with the amount of lipid peroxidation these fatty acids can induce as measured by thiobarbituric acid test. It was also observed that 14C-labelled linoleic acid uptake was almost the same whereas that of “C-labelled arachidonic acid and eicosapentaenoic acid were substantially less in tumor cells compared to normal cells. Tumor cells incorporated major portions of the fatty acids in the ether lipid and Correspondence to: U.N. Das, Department of Medicine, The N&am’s Institute of Medical Sciences, Punjagutta. Hyderabad-500
482,
India.
0304.3835/91/$03.50 Published and Printed
0 1991 Elsevier Scientific Publishers in Ireland
phospholipid fractions, whereas normal cells incorporated the fatty acids primarily in the phospholipid fraction. These results suggest that c-UFA-induced tumoricidal action is a free radical dependent process and that there are significant differences between normal and tumor cells in fatty acid uptake and distribution.
Keyarords: cis-unsaturated fatty-acids, tumoricidal, free radicals, lipid peroxidation, superoxide anion, anti-oxidants Introduction Studies [1,2,5,22,23] have shown that some &-unsaturated fatty acids (c-UFAs) such as gamma-linolenic acid (GLA, 18:3 n-6)) dihomoGLA (DGLA, 20:3 n-6)) arachidonic acid (AA, 20:4 n-6), and eicosapentaenoic acid (EPA, 20:5 n-3) and to some extent linoleic acid (LA, 18:2 n-6) and alpha-linolenic acid (ALA, 18:3 n-3) can selectively kill tumor cells without harming the normal cells at the concentrations tested. Inhibition of cycle-oxygenase and lipoxygenase enzymes did not prevent the tumoricidal action, but antioxidants did, suggesting that controlled formation of peroxidation products and free radicals may be involved [3,6]. In an extension of these studies, it was observed that GLA, AA and EPA were 1.5-2 times more efIreland Ltd
236
fective than LA in inducing free radical generation in tumor cells but not in normal cells [7,8]. It is well documented that both hydrogen peroxide (H202) and the superoxide (O,? ) radical can oxidize c-UFAs of the cell membrane, producing such cytotoxic metabolites such as malondialdehyde (MDA), 2-alkanals and hydroxy-alkenals [ 111. These substances exert influences on cross-linking with amino groups of DNA through the formation of Schiff bases [17] and thus, free radicals can exert their action on DNA. In addition, it is known that some metals such as copper and iron can augment free radical generation and lipid peroxidation by a non-enzymatic process [ 141. In view of this, the effect of iron and copper metals on GLAinduced tumoricidal action, the possible relationship between lipid peroxidation and c-UFAinduced tumor cell death and the uptake and distribution of fatty acids in normal and tumor cells were studied and the results are reported here. Materials
and Methods
Cells and culture conditions Normal monkey kidney (CV-l), normal human fibroblasts (CCD-41-SK), murine fibroblast transformed by Abelson leukemia virus (BALB), human breast cancer (ZR-75-l) and human promyelocytic leukemia (HL-60) cells were used for the study. The cells were seeded at 5 x lo4 or 1 X lo4 cells per plate or well depending on the experimental protocol and as described earlier [l]. The cells were grown in 0.5 ml or 2.0 ml (in 24-well tissue culture plates or petri dishes, respectively) of bicarbonate buffered Dulbecco’s modified Eagle’s medium (Sigma Chemical Co., U.S.A.) with or without the added fatty acids at 37 “C in a 5% CO, humidified incubator as described earlier [ 11. The fatty acid esters were initially dissolved in 95% ethanol and the final concentration of ethanol was not more that 0.2% in all control and fatty acid supplemented cultures. Cell viability was determined by the trypan blue dye exclusion method. The cells were checked for possible mycoplasma contamination by fluorescent
technique and confirmed that there was no contamination during the period of these studies. Studies with antioxidants and metals These studies were performed in 24 well tissue culture plates where in the cell lines were seeded at 1 x lo4 cells per well. One day after seeding, 10 pg/ml of GLA was added to the cells. To study the effect of possible enhancers or inhibitors of tumor cell death by GLA, the following were added simultaneously to the cultures with GLA: vitamin A, vitamin E, transretinoic acid, uric acid, superoxide dismutase, ATP, glutathione peroxidase, catalase, heatinactivated catalase, FeCI,, FeCI, (ferrous and ferric chloride, respectively) and CuSO, (copper sulfate). All the chemicals were obtained from Sigma Chemical Co., U.S.A. and were of highest grade of purity available. Appropriate controls without GLA were performed. The cultures were observed every day and cell viability was determined every day until day 7 of supplementation of GLA and the various additives. All these studies were done at 10 or 20 pg/ml of GLA concentration depending on the protocol of the experiment(s). These studies were performed with ZR-75-1 cells as a representative of the tumor cells. TBA
reaction
For these studies, cell cultures seeded with 5 x lo4 cells/petri dish (35 mm) were used. ZR-75-1 cells and CV-1 cells were used in this study. Cell cultures were supplemented with 20 pg/ml of ethyl esters of GLA, AA and DHA (docosahexaenoic acid) and methyl ester of EPA. The cells were grown in 2.0 ml medium with or without added fatty acids at 37°C in a 5% CO, humidified incubator. Similar to the experiments with antioxidants and metals, 1 day after seeding, the cells were supplemented with various fatty acids. At the end of 7 days, medium and cells were harvested separately and assayed for TBA reaction as a measure of lipid peroxidation. TBA reaction was performed as described by Gavin0 et al. [13] and as described earlier [21]. The absorbance of the reaction was measured at 532 nm with growth medium
237
or PBS as the controls. The absorbance values obtained were converted to pmol of MDAequivalent (MDA-eq) from a standard curve obtained with 1,1,3,3-tetramethoxypropane.
c-UFA incorporation studies In order to know whether the selective cytotoxicity shown by fatty acids could be due to increased uptake, we studied the fatty acid uptake by normal and tumor cells in vitro [7,8]. Twenty four hours after seeding (1 x lo4 cells/O.5 ml medium), 0.1 &i of 14C-labelled LA, AA or EPA was added (spec. act.: LA, 52.6 mCi/mmol; AA, 54.5 mCi,/mmol; EPA, 55.4 mCi/mmol) to study the uptake of fatty acids by normal and tumor cells. The cells were washed three times in PBS, detached by trypsinisation and the total amount of fatty acid incorporated was counted in a liquid scintillation counter on day 1, 2 and 3 for all the fatty acids tested.
Determination of c-UFA in different lipid fractions To study the incorporation of labelled fatty acids in different lipid fractions, 5 x 104 cells were incubated with 0.5 PCi of LA, AA and EPA. Twenty four hours after incubation, the cells were washed in PBS, lysed and extracted in chloroform/methanol (2: 1, v/v). Different lipid fractions were separated by thin layer chromatography as described by Roos and Choppin [20]. Regions corresponding to phospholipid (PL) , free fatty acid (FFA), ether lipid (EL), cholesterol ester (CE) and triglyceride (TG) fractions were scraped from the developed chromatography plates and extracted in chloroform/methanol (2:1, v/v). The extract was evaporated under nitrogen and counted in a liquid scintillation counter [8]. All the fatty acids except EPA were ethyl esters and EPA is a methyl ester and were obtained from Sigma Chemical Co., U.S.A. All experiments were done in triplicate/quadruplicate and repeated at least 3 or 4 times.
The effect of 20 pg/ml of various c-UFAs on the survival of human breast tumor (ZR-75- l), normal monkey kidney (CV-1) and normal human fibroblast (CCD-41-SK) cells are given in Fig. 1. It is evident from this that GLA and AA are equipotent in their tumoricidal action, where as EPA has less potent action. On the other hand DHA was without any significant tumoricidal action in the test system used. Thus, AA = GLA > EPA > DHA in terms of their action on tumor cells. It is also noted that the maximum tumoricidal action of c-UFAs is seen by day 6 or 7. The results shown in Fig. 2 indicate that both AA and EPA but not GLA are toxic to normal cells (CCD-41-SK) at doses 2 or 3 times the dose effective in killing the tumor cells. This is especially so when normal cells (CCD-41-SK, 1 x 104) were supplemented with 60 pg/ml of AA
% DEAD CELLS
T
T
100 -
80 -
60 -
40 -
20 -
03
3
6
6
7
DAYS Fig. 1.
Effect of 20 pg/ml
ZR-75-1,
CV-1
and CCD-41
of c-UFAs on the survival of SK cells in vitro.
238 % OF
DEAD CELLS
and
ON DAY 7
100 -
0'
I
I
I
I
60 20 40 10 CONCENTRATION OF FATTY ACID IN W/ml
Fig. 2.
Effect of GLA, AA and EPA on the survival of normal (41 SK) and tumor cells in vitro. N = normal cells; T = tumor cells. Table 1. Effect of antioxidants No. 1 2 3 4 5 6
7 8
9 10
and other chemicals
Chemical/Enzyme
Dose(s)
Control GLA Vitamin A Vitamin A Vitamin E Vitamin E Trans-retinoic acid Uric acid Uric acid Uric acid Cis-retinoic acid Superoxide dismutase SOD SOD SOD ATP Glutathione peroxidase Glutathione peroxidase
-
All chemicals were tested on day 7 of the addition of vitamin A, E, retinoic values are expressed as
EPA. This suggests that GLA has more selective tumoricidal action compared to AA and EPA. Thus, GLA > AA > EPA when their capacity to kill tumor cells selectively and least toxicity to normal cells are taken into account. The effect of various inhibitors or enhancers of the tumoricidal action of GLA on human breast tumor cells are given in Tables I and II. For testing the action of inhibitors, a concentration of 20 pg/ml of GLA was used since, earlier studies and the results shown in Fig. 1 showed that this is the optimum concentration to produce maximum cytotoxicity to tumor cells with least cytotoxicity to normal cells [ 1,2,7]. On the other hand, to test the action of possible enhancers of cytotoxicity, a sub-optimal dose of GLA which is 10 pg/ml was used. It is clear from these results, summarised in Table I, that antioxidants and superoxide quencher, superoxide dismutase (SOD) can effectively block the tumoricidal or cytotoxic action of GLA. Similar results were obtained with AA and EPA (data
on the cytotoxicity used
20 j&ml 10 pg/ml 100 pg/ml 10 &ml 100 pg/ml 1 x 10e5 M/ml 1 x 10e4 M/ml 1 x 10m5 M/ml 1 x 10m6 M/ml 1 x 10m5 M/ml 1 x 10m6 M/ml 3 x 10m6 M/ml 3 x 1O-7 M/ml 3 x 10m8 M/ml 200 PM/ml 0.1 pg/ml 0.01 pg/ml
of GLA to human breast cancer cells in vitro. % Dead cells 20.0 95.0 30.0 32.0 10.0 10.0 35.0 40.0 47.0 51.0 65.0 15.0 20.0 22.0 30.0 22.0 21.0 28.0
f f f f f f + * f f f f f f f f f +
5.0 5.0 7.0 9.0 5.0 7.0 8.0 10.0 8.0 7.0 18.0 5.0 7.0 7.0 10.0 9.0 12.0 8.0
along with 20 pg/ml of GLA on ZR-75-1 cells (1 x lo4 cells/ml). All cultures were harvested of GLA with and without the inhibitors. Percentage of dead cells with various concentrations acid, uric acid, SOD, ATP and glutathione peroxidase alone is approximately 14.0 * 7.0. All mean f S.D.
239 Table 11. Effect of iron and copper salts and catalase on the cytotoxicity induced by GLA on human breast tumor cells. No.
Chemical
Dose tested
% Dead cells
1 2 3
Control GLA Catalase Catalase Catalase Heat-inactivated catalase Heat-inactivated catalase Heat-inactivated catalase FeCI, FeCI, FeCI, FeCI, FeCla FeCI,
-
20.0 f 60.0 zt 75.0 f 80.0 f 81.0 f 72.0 zt
4
5
6
FeCI,
7
cuso4 cuso,
10 500 1000 3000 500
&ml U/ml U/ml U/ml U/ml
5.0 15.0’ 10.0’ 12.0’ 8.0” 12.0
1000 U/ml
77.0
f
3000 U/ml
73.0
f
j&ml &ml 20 &ml 40 &ml
61.0 65.0 98.0
ztz 12.0 zt 9.0 f 2.0”
1 &ml 4 pg/ml
65.0 68.0
1 4
20 pg/ml 1 &ml 4 Irg/ml
l
8.0’ 14.0
98.0zt 2.0' l
LIZ10.0 zt 8.0
95.0f 5.0' 57.0f 11.0 78.0
f
14.0”
All chemicals were tested along with 10 pg/ml of GLA on ZR-75-1 cells (1 x lo4 cells/ml). All the cultures were harvested on day 7 of addtion of GLA with and without other chemicals. Percentage of dead cells with catalase, inactivated catalase, FeCI,, FeCI, and CuSO, alone at various concentrations used was approximately 18 * 5. All values are expressed as mean =t S.D. For other details see Materials and Methods. ‘P < 0.001 compared to control. *‘P < 0.05 compared to GLA group.
not shown). The inhibitory action shown by SOD suggests that possibily, superoxide radical has a role in the tumoricidal action of c-UFAs. Both vitmain A and E are believed to have potent antioxidant actions and prevent lipid peroxidation. Similarly, uric acid is an antioxidant. The inhibitory action of ATP is rather surprising. Both catalase and heat-inactivated catalase enhanced the cytotoxic action of GLA at the doses tested (Table II). This suggests that H,O, radical does not participate in the tumoricidal action of c-UFAs. Both FeCI, (when used at 20 and 40 pg/ml) and CuSO, (when used at 4 pg/ml) enhanced the cytotoxicity of GLA (Table II). Since these metals are known to augment lipid peroxidation and free radical generation (especially that of superoxide anion, 14) their enhancing action is understandable. Catalase is
rich in copper and this may explain why heatinactivated catalase enhanced the tumoricidal action of GLA. Table III depicts the MDA-eq detected in the medium and cells on day 7 after supplementation with 20 pg/ml of various c-UFAs. Both AA and GLA supplemented tumor cells (ZR-75-1) produced large amouts of MDA-eq (approx. 4-fold increase) compared with normal cells (CV-1) tested and controls. But surprisingly, EPA-supplemented normal and tumor cells produced significantly increased amounts of MDAeq as compared with controls. Although this result looks paradoxical, this may explain at least in part, why normal cells were more susceptible to the cytotoxic action of EPA compared with that of GLA and AA (Fig. 2). Figure 3 depicts the fatty acid incorporation
240 Table III.
Effect of various c-UFAs on MDA-eq formation
Cell line
Treatment
cv-1 (Normal monkey
kidney cells)
ZR-75- 1 (Human breast cancer cells)
All the values are expressed
as mean
Control GLA AA EPA DHA Control GLA AA EPA DHA f S.D. of three separate
and tumor cells. Uptake of LA was almost the same in normal (41-SK and CV-1) and tumor (ZR-75-1 and HL-60) cells, especially at 48 and 72 h. In contrast, both AA and EPA
in normal
1x1000 35
COUNTS
a
T
in normal and tumor cells on day 7. pmol MDA-eq In the cells
In the medium
0.5 6.2 7.1 15.0 11.4 0.2 26.5 35.9 16.7 11.3
0.3 1.1 2.8 3.3 10.1 0.9 10.6 3.5 10.2 14.4
estimations.
24HRS
lx1000
48HRS
0.3 1.1 0.4 2.7 3.0 0.3 2.7 10.4 1.7 3.9
f 0.1 f 0.3 f 1.0 f 0.5 f 0.3 f 0.2 z!z 2.4 zt 1.6 f 0.6 f 0.4
For other details see Materials and Methods.
incorporation was low in the tumor cells (Fig. 3b and 3~). There were no significant differences between normal and tumor cells in thymidine incorporation (data not shown) suggesting that their growth rates were similar. Hence, the differences in the uptake of fatty acids can not be attributed to changes in cell growth. Table IV shows the distribution of labelled LA, AA and EPA in different lipid fractions of both normal and tumor cells. In normal (41-SK and
1x1000
0
f f f zt f f +z f zt f
COUNTS
72rliiS
COUNTS
25
Fig. 3. Incorporation of (a) 14C-labeled LA, (b) 14Clabeled AA and (c) 14C-labeled EPA in 1 x 104 cells. ( l ) CV-1, (e) HL-60, ( q) 41-SK, ( l) ZR. All values are expressed as mean f S.D.
241 Table IV. Uptake and distribution of different cis-unsaturated ceils in vitro. All values are expressed as mean + S.D. Cell line
Percentage
Total
fatty acids in different lipid fractions of normal and tumor
of distribution
PL
FFA
TG
CL
CE
Normal
cv-1 LA AA EPA
30,253 37,140 58,460
zt 5810 + 6483 zt 8609
59.0 72.6 80.3
5.0 14.9 1.9
1.9 2.1 0.8
1.0 0.4 1.2
22.8 9.9 17.2
41-SK LA AA EPA
29,997 25,362 82,460
f 5406 f 2375 ztz 5481
72.0 78.2 71.7
2.0 9.1 1.5
1.9 2.0 0.7
6.4 4.0 10.7
21.6 6.6 13.8
11,019 zt 2742 13,435 f 672 9377 zt 1582
64.0 83.0 58.0
3.8 14.5 4.7
0.9 1.5 1.1
1.3 1.2 2.1
22.2 3.5 25.3
8589 + 1296 4396 + 142 5010 f 834
44.0 36.6 30.6
7.8 8.3 7.4
5.7 3.7 2.4
2.8 2.9 2.2
39.8 48.5 49.4
Tumor cells
ZR-75- 1 LA AA EPA BALB LA AA EPA
PL = Phospholipids,
FFA = free fatty acids, TG = triglycerides,
cells, LA, AA and EPA were incormainly in the PL and less than 22% in the EL fractions. On the other hand, in the tumor (BALB and ZR-75-1) cells a major portion of the fatty acid was incorporated both in to PL and EL fractions except in ZR-75-1 in which the distribution of fatty acids was similar to that of the normal cells. Incorporation of fatty acids into FFA, TG and CE fractions was similar in normal and tumor cells [7,8]. CV-1)
porated
Discussion The results presented here support and extend the earlier results that GLA, AA and EPA can selectively kill tumor cells and suggest that superoxide anion and lipid peroxidation process have a role in the tumoricidal action of c-UFAs [l-3,5,7,8]. The observation that SOD but not
CE = cholesterol
esters,
EL = ether lipids.
catalase and mannitol can prevent the cytotoxic action of c-UFAs suggests a role for superoxide radical but not for H,O, and hydroxyl radical (results with mannitol were not presented here as it did not have any significant inhibitory effect on c-UFA-induced cytotoxicity) . The ability of vitamin A and E to prevent the cytotoxic action of c-UFAs indicates that lipid peroxidation process has a role. The enhancing effect of copper and iron lends further support to this contention. The demonstration of an increase in the intracellular MDA-eq in tumor cells supplemented with GLA, AA and EPA adds strength to this concept. The increased amounts of MDA-eq observed in normal cells supplemented with EPA suggests that increased lipid peroxidation and/or free radical generation in the cells beyond a limit can induce cell death since EPA is toxic to normal cells at higher con-
242
centrations (Fig. 2). The increased sensitivity of tumor cells to GLA, AA and EPA may be due to their low content of SOD [12]. The inability of DHA (22:6, w-3) to kill tumor cells (Fig. 1) in spite of its high degree of unsaturation and capacity to undergo peroxidation readily may be due to its unstable nature and/or its ineffectiveness to increase specific type of lipid peroxidation in the cells (Table II). The most surprising observation is the ability of ATP to block the cytotoxic action of GLA. In an earlier study, Hilf et al. [16] showed that hematoporphyrin derivative-induced photosensitization, a procedure which also enhances free radical generation in the cells, in R 3230 AC mammary tumors causes a 60% reduction in cellular ATP levels. If, c-UFA-induced cell death is also mediated by interfering with ATP metabolism, it may explain the inhibitory effect of ATP on GLA-induced cytotoxicity. But, it should be mentioned here that ATP can not enter the cells. Hence, this possibility needs to be tested. The other possibility to be considered is the ability of ATP to form complexes with EFAs and thus, block the availability of fatty acids to the cells. It is interesting to note that tumor cells incorporated less AA and EPA than normal cells and that major portions of the fatty acids are incorporated in PL and EL fractions. In spite of this, AA and EPA-treated tumor cells produced higher amounts of lipid peroxides (Table III) which are known to be toxic to cells [15]. This suggests that the low rates of lipid peroxidation seen in tumor cells [12] could be due in large part to low substrate availability. The amounts of free fatty acids both in normal and tumor cells were almost the same and hence, increased free fatty acids can not explain the enhanced lipid peroxidation and/or toxicity in tumor cells. Compared to normal cells, a significant proportion of the incorporated fatty acids are found in the EL fraction in the tumor cells [7,8]. The possibility that the substrates for lipid peroxidation are derived from the EL fraction, which is considered to be metabolically inactive, remains to be tested. It is also likely that the source of fatty acids for lipid peroxidation and free radical generation may be the PL frac-
tion as the activity of phospholipase A2 in the tumor cells is high compared to normal cells
1181. The results of these studies are interesting since it is known that several anti-cancer drugs such as vincristine, Adriamycin etc., enhance free radical generation and lipid peroxidation in the plasma of patients following chemotherapy [21]. Even lymphokines such as interferon and tumor necrosis factor have the capacity to augment free radical generation in the tumor cells and human neutrophils [4,9]. If so, this indicates that free radicals and free radical-dependent lipid peroxidation process may be a common pathway by which tumor cells are eliminated both by natural products of immune cells and drugs.. Thus, methods designed to specifically enhance superoxide radical generation and lipid peroxidation in the tumor cells may form a novel approach to cancer therapy [lo, 191. Acknowledgements This work was supported in part by a grant from the Department of Science and Technology, New Delhi to U.N. Das. Dr. Das was also in receipt of INSA Research Fellowship during the tenure of this study. References Begin, M.E., Das, U.N., Ells, G. and Horrobin. D.F. (1985) Selective killing of human cancer cells by polyunsaturated fatty acids. Prostaglandins Leukotries Med., 19, 177- 186. Begin, M.E., Ells, G., Das, U.N. and Horrobin, D.F. (1986) Differential killing of human carcinoma cells supplemented with n-3 and n-6 polyunsaturated fatty acids. J. Natl. Cancer Inst., 77, 1053-1062. Begin, M.E, Das, U.N. and Ells, G. (1985) Mechanism of essential fatty acid induced cytotoxicity in malignant cells, 2nd Int. Congress. Essential Fatty acids Prostaglandins and Leukotrienes March, London, U.K., p. 7. Berton, G., Zeni. L., Cassatella. M.A. and Rossi. F. (1986) Tumor necrosis factor provokes superoxide anion generation from neutrophils. Biochem. Biophys. Res. Commun., 138, 1276-1282. Booyens. J.. Engelbrecht, P.. LeRoux, S.. et al. (1984) Some effects of the essential fatty acids; linoleic acid and alpha-linolenic acid and of their metabolites gamma-linolenic acid, arachidonic acid eicosapentaenoic acid, docosahexaenoic acid and of prostaglandins A and E on the prolifera-
243
tion of human osteogenic sarcoma cells in culture. Prostagladins Leukotrienes Med., 15, 15-34. Das, U.N., Huang, Y.S., Begin, M.E. and Horrobin, D.F. (1986) Interferons, phospholipid metabolism, immune responses and cancer. IRCS Med. Sci., 14, 1069-1074. Das, U.N., Begin, M.E., Ells, G., Huang, Y.S. and Horrobin, D .F. (1987) Polyunsaturated fatty acids augment free radical generation in tumor cells in vitro. Biochem. Biophys. Res. Commun.,
145,
icity, oxygen radicals, transition metals and disease. Biochem. J., 219, l-14. 15
16
(1986) Relationship of mitochondrial function and cellular adenosine triphosphate levels to hematoporphyrin derivativeinduced photosentitization in R3230 AC mammary tumors. Cancer Res.. 46, 211-217.
15-24.
8 Das, U.N., Haung, Y.S., Begin, M.E., Ells, G. and Hor-
9
10 11
12
13
14
Halliwell, B. and Gutteridge, J.M.C. (1984) lipid peroxidation oxygen radicals and cell damage and antioxidant therapy. Lancet, i: 1396-1397. Hilf, R.. Murant, R.S., Narayanan. U. and Gibson, S.L.
robin, D.F. (1987) Uptake and distribution of cis-unsaturated fatty acids and their effect on free radical generation in normal and tumor cells in vitro. Free Radical Biol. Med., 3, 9-14. Das, U.N., Ells, G.. Begin, M.E. and Horrobin, D.F. (1986) Free radicals as possible mediators of the actions of interferon.
17
Kappus. H. (1985) Lipid peroxidation: Mechanisms, analysis, enzymology and biological relevance. In: Oxidative stress, pp. 273-303. Editor: H. Sies. Academic Press, New York.
18
J. Free Rad. Biol. Med., 2, 183-188. Das. U.N. (1987) Biological significance of arachidonic acid. Med. Sci. Res., 15, 1485-1490. Esterbauer, H. (1982) Aldehydic products of lipid peroxidation, In: Free Radicals, lipid peroxidation and cancer, pp. 101-128. Editors: D.G. McBrien and T.F. Slater. Academic Press, London. Galeotti, T., Bantoli, G.M. and Bartoli, E. (1982) Superoxide radicals and lipid peroxidation in tumor microsomal membranes. In: Biological and Clinical aspects of superoxide and superoxide dismutase, pp. 106-117. Editors: W.H. Bannister and J.V. Bannister. Proc. Fed. Eur. Biochem. Sot.
19
Levine, L. (1981) Arachidonic acid transformation and tumor promotion. Adv. Cancer Res., 35. 49-79. Rotilio. G., Bozzi. A., Mavelli, B. et al. (1982) Tumor cells as models for oxygen-dependent cytotoxicity. In: Biological and Clinical aspects of superoxide and superoxide dismutase. pp. 118-126. Editors: W.H. Bannister and J.V. Bannister. Proc. Fed. Eur. Biochem. Sot. Symposium. No. 62. Roos, D.S. and Choppin. P.W. (1985) Biochemical studies on cell fusion: I. Lipid composition of fusion resistant cells. J. Biol. Chem., 101. 1578-1590. Sangeeta, P., Das, U.N.. Koratkar, R. and Suryaprabha. P. (1990) Increase in free radical generation and lipid perox-
Symposium No. 62. Gavino, V.C., Miller, J.S.. Ikareblha. S.O., Mile. G.E. and Cornwell, D.G. (1981) Effects of polyunsaturated fatty acids and antioxidants on lipid peroxidation in tissue cultures. J. Lipid Res., 22, 763-769. Halliwell, B. and Gutterridge, J.M.C. (1986) Oxygen tox-
20
21
22
23
idation following chemotherapy in patients with cancer. Free. Rad. Biol. Med., 8, 15-20. Seigel, I., Liu, T.L., Yaghoubzadeh. E Kaskey. T.S. and Gleicher, N. (1987) Cytotoxic effects of free fatty acids on ascites tumor cells. J. Natl. Cancer Inst.. 78. 271-277. Tolnai. S. and Morgan, J.F. (1962) Studies on the in vitro anti-tumor activity of fatty acids. V. Unsaturated fatty acids. Can. J. Biochem Physiol , 40, 869-875.