Effects of gamma-linolenic acid, dihomo-gamma-linolenic acid and ethanol on cultured human mammary carcinoma cells

ProstaglandinsLeukotrienesand Medicine20: 209-221,1985

EFFECTS OF GAMMA-LINOLENIC ACID, DIHOMO-GAMMA-LINOLENIC ACID AND ETHANOL ON CULTURED HUMAN MAMMARY CARCINOMA CELLS K.M. Robinson and J.H. Botha, Departments of Physiology and Pharmacology, University of Natal, Box 17039, Congella (Reprint requests to KMR) 4013, South Africa.

ABSTRACT A number of fatty acids have been shown to inhibit the In particular, gammagrowth of malignant cells in vitro. linolenic acid (GLA) has hen proposed to act as a precursor for the production of prostanoids especially prostaglandin Yl To test this hypothesis, the effects of GLA on %l~rE&);* human breast carcinoma cells were compared with acid (DGLA) the metabolite those of dihomo-gamma-linolenic The influence of GLA and the immediate precursor of PGEI. of ethanol (which has been shown to enhance conversion of DGLA to PGEI) on the actions of each of the fatty acids was also investigated. In contrast to the inhibitory effects observed with all concentrations of GLA cell growth was 50 ug DGLA. promoted by the presence of Ethanol reduceId the action of both GLA and DGLA possibly due to some physicochemical reaction between the alcohol and the fatty acids. The fact that the actions of GLA were not mimicked by DGLA which is the next step towards PG production casts doubtuponthe role of PGEl as mediator of the effects which have been observed with GLA in malignant cells. INTRODUCTION It has been suggested that malignant cells may be deficient in the enzyme a-6-desaturase which is necessary for the conversion of l_inoleic acid (LA) to gamma-linolenic acid (GLA) (Fig. 1). Supplementation with exogenous GLA could circumvent this enzyme and thus provide the precursor for

209

LINOLEIC ACID (LA) C1Bz2' we

A'DESATURASE

(-2H)

GAMMA-LINOLENIC ACID (GLA> C

18:

06

3’

ELONGASE (+ZC)

DIHOMO-GAMMA-LINOLENIC ACID (DGLA> 50:3’

w6

\ A'DESATURASE

(-2H) \

ARACHIDONIC AC D ‘20:4’

PG 1 SERIES

FIG, 1

w6

PG 2 SERIES

DERIVATION OF PROSTAGLANDINS OF THE 1 AND 2 SERIES FROM ESSENTIAL FATTY ACIDS

210

the synthesis of prostanoids particularly prostaglandin El (PGEI) (1). Several studies have indeed shown that addition of GLA (2,3,4,5) and a number of other polyunsaturated fatty acids (6) to a range of malignant cells growing in vitro has caused pronounced inhibition of cell growth anTe=cell death. While it is possible that some of the observed effects of GLA and the other fatty acids studied are much is still uncertain mediated by the prostaglandins its effect on about the mechanism whereby GLA exerts cultured cells. If the supposition that GLA is acting via the PGs is then a reasonable sequiter would be that correct, such as dihomo-gamma-linolenic acid metabolites of GLA (DGLA) would induce similar inhibition of cell growth when added to cell cultures. Furthermore, substances such as ethanol which has been reported to facilitate conversion of DGLA to PGEl (7) could be expected to inhibit cell growth when acting alone or enhance the effects of GLA or DGLA when used in combination with these fatty acids.In order to test the premise that the observed effects of exogenous GLA on cultured malignant cells may be attributed to elevated levels of its metabolites the effects of a range of concentrations of GLA and DGLA on human breast carcinoma cells of the continuous line NUB 1 were investigated, as were the effects of ethanol, both alone and in combination with these fatty acids. MATERIALS -AND METHODS The continuous cell line NUB 1 was derived in 1982 from a poorly differentiated mammary carcinoma in a 44 year old Black woman. Following surgical removal, the diced tumour in Eagle's Minimum Essential Medium (MEM) was inoculated subcutaneously into 4 week old BALB/c nu/nu mice (8) of both sexes. Clearly defined encapsulated tumours grew rapidly after a latent period of 9 weeks. The human origin of the resulting tumours was confirmed by lactate dehydrogenase isoenzyme electrophoresis (9). Explant cultures were initiated from tumour xenografts as described previously (IO). Cultures were maintained in MEM supplemented with 10% foetal calf serum (FCS) and antibiotics. Cells were passaged weekly at a ratio of 1:4 and thus the line NUB 1 was established. Under the described conditions NUB 1 cells lack oestrogen and progesterone receptors, are of epithelial morphology and grow in tightly adherent clusters. The fatty acids GLA (6, 9, 12-Octadecatrienoic acid Sigma L 2378) and DGLA (8, 11, 14-Eicosatrienoic acid Sigma E 4504) were diluted aseptically in sterile 0,lM Na2C03 under N2 according to the method of Ingerman-Wojenski et al (11) to

211

give stock solutions of lOmg/ml. Two different batches of each fatty acid were used during the course of the study. Stock solutions were stored at -2OOC until required when final dilutions were rapidly performed in MEM prior to addition to cell cultures. Absolute ethanol was added to MEM as required to give final concentrations of 1% or 2%. In the fi st experiment quadruplicate cultures of both dens (i4 x 10 8 cells/25cm2 flask) and less dense (28 x 10 !? cells/25cm2 flask) NUB 1 cells received 5ml MEM with 10% FCS containing: a) no additives b) 0,lml Na CO3 c) 10, 20, 30, 40 or 50 ug/ml GLA d) 10, 20, 30, 40 or 50 ug/ml DGLA (Total 128 flasks) All flasks were examined daily by phase contrast microscopy and observations recorded. One flask from each set was in methanol fixed and stained with 0,25% May-Grunewald followed by 5% Giemsa in water 24, 48, 72 and 96 hours after the start of the experiment. In the second experiment duplicate dense and less dense cultures as described above were treated with 5ml MEM with 10% FCS containing: no additives a) b) 0,lml Na2C0 c) 50, 100 or 200 ug/ml GLA d) 50, 100 or 200 ug/ml DGLA e) 0,lml or 0,05ml absolute ethanol (50, 100 or 200 ug/ml) of both GLA f) equal concentrations and DGLA together with each concentration of GLA 4) each concentration of ethanol together with each concentration of DGLA h) each concentration of ethanol (Total LOO flasks) All flasks were examined daily and one member of each pair fixed and stained as described above 24 and 72 hr after the experiment commenced. For ultrastructural investigation of the effects of the fatty acids duplicate dense cultures received 5ml MEM with 10% FCS containing: a) no additives b) 50 ug/ml GLA c) 50 ug/ml DGLA Seventy-two hours later these cultures were prepared for electron microscopy as described previously (12).

212

RESULTS Markedly different effects on cell growth and morphology were observed in NUB 1 cells treated with GLA, DGLA and ethanol (Table I and Figs 2-7). In all cultures receiving GLA cells accumulated prominent cytoplasmic granules 24-48hr after initiation of treatment (Fig. 2). In contrast, at low concentrations (lo-50 ug/ml) DGLA administration caused no obvious morphological changes studies confirmed the suspected (Fig. 3). Ultrastructural lipid nature of the cytoplasmic granules in GLA treated cells (Fig 4). Similar lipid accumulation was not observed in cells treated with up to 100 ug/ml DGLA (Fig. 5). Inhibition of cell growth was evident within 24-72hr in cultures receiving doses in excess of 10 ug/ml GLA (Fig. 6) while treatment with high (100 and 200 uglml) concentrations of GLA caused cell death and desquamation within 24-48hr. At concentrations of up to 50 ug/ml DGLA either exerted no effect on cell growth or promoted growth slightly (Fig. 6). Such growth stimulation was most marked at 50 ug/ml DGLA in both dense and sparse cultures. However, further increasing the concentration of DGLA to 100 or 200 ug/ml progressively caused more inhibition of cell growth. Nonetheless, even at 200 ug/ml DGLA was not as effective at suppressing cell growth as was 100 ug/ml GLA. Although observed consistently in all treated cultures growth inhibitory (GLA and high concentrations of DGLA) (low and growth promoting concentrations of DGLA) effects were more pronounced in sparse cultures, presumably because of the higher relative dose per cell. Furthermore a combination of GLA and DGLA at 50 uglml each interfered less with cell growth than did treatment with either 50 or 100 ug GLA alone. However, at higher concentrations (100 or 200 ug/ml) of each fatty acid cell death and desquamation occurred within 24hr. Ethanol at the concentrations used had slightly less effect on cell growth than did GLA alone, but caused more inhibition of cell growth than did DGLA (Fig. 7). When used in combination with the fatty acids under investigation ethanol appeared to diminish the effects of GLA, while cultures exposed to ethanol and DGLA (Fig. 7) were characterised by more growth inhibition than cells receiving DGLA alone. Thus, although GLA and ethanol inhibited growth as anticipated, it is notable that DGLA only suppressed cell growth if present in high doses and actually promoted growth at lower concentrations.

213

Fig.

Phase contrast micrograph GLA.

Paranuclear

showing NUB 1 cells 24hr dfter. the addition of 31

cytoplasmic

granules (both dense and refractile)

in all cells and there is no evidence of cell division.

Fig.

Phase contrast micrograph

Llg/ml

dre promirlent

Bar = 50 urn.

showing an adherent cluster of NUB 1 cells 24hr aft er

the addition of 50 ug/ml OGLA.

Cells are morphologically

indistinguishable

from those in control cultures and mitotic figures (m) are evidence of contin uing growth.

Bar = 50 urn.

214

Fig.

4

Transmission

electron micrograph showing the cytoplasm of a NUB 1 cell 72hr

after the addition of 50 ug/ml GLA.

the

Fig.

5

cytoplasm.

Lipid droplets (1) are prominent

Transmission electron micrograph shouing portions after the addition of 50 ug/ml OGLA. Such cells cultures and show desmosomes reticulum

in

Bar = 1 urn.

(d), mitochondria

of tuo resemble

those

72hr

in control

(m) and abundant endoplasmic

(er), features typical of NUB 1 cells.

215

NUB 1 cells

Bar = 1 urn.

ADDITIONS TO 5ML MEDIUM

GLA (w/ml)

DGLA (w/ml)

PERCENTAGE CELLS IN RELATION TO CONTROL 48HR AFTER START OF EXPERIMENT

EtOH (ml)

DENSE*

10

-

30

-

50 50

50

100 100

100

95

90

70

65

45

40

60

65

6

4
I

105

105

30

110

110

50

130

135

100

90

80

-

35

30

-

0,05

40

50

-

0,05

25

25

100

0,05

50

55

071

30

10

091

35

25

091

40

55

. 100 -

100

10

200

100

100

(1 -

LESS DENSE+

100

* Dense Control Cultures = *8

x lo6 cells/flask

+Less Dense Control Cultures = *2 x lo6 cells/flask

TABLE 1 : PERCENTAGE CELLS IN RELATION TO CONTROL NUB 1 CULTURES 48HR AFTER EXPOSURE TO GLA, DGLA, ETHANOL (EtOH) AND COMBINATIONS THEREOF

216

Fig. 6

Appearance

of dense cultures of NUB 1 cells 48hr after addition of various

doses of GLA and DGLA. May-Grunewald

Giemsa.

DISCUSSION The fact that DGLA produces effects opposite to those of GLA growth when present in (2,3,4,5) and actually stimulates low concentrations casts doubt upon the role of PGEl as mediator of the effects which have been observed with GLA in cultured malignant cells. In all the cell lines which have been studied to date GLA has been shown to consistently inhibit cell growth and cause cell death (2,3,4,5). If PGEl were involved then DGLA which is further down the metabolic pathway (Fig. 1) and is the immediate precursor of PGs of the 1 series could be expected to produce a more marked and/or more rapid response than an equivalent quantity of GLA. Further evidence against PG involvement is the finding that indomethacin (which could be expected to block production of PGsfrom fatty acid precursors) does not prevent or even diminish the actions of GLA but actually enhances them (13). 217

Fig.

7

Appearance or

in

of

cultured

combination

witir

NUB

1 cells

ethanol.

24hr

after

May-Grunewald

exposure

to

GLA

and

DGLA

alone

Giemsa.

In addition docosahexaenoic acid (DHA) has been shown to inhibit cell proliferation even though it is not a PG precursor. Besides the fact that mediation via PGs appears uncertain the possibility that GLA bypasses a deficient a6-desaturase enzyme may also be questioned as LA has been reported to retard cell growth (6). The effect of ethanol was investigated because it has been reported to enhance conversion of DGLA to PGE1 (7). Used alone it did suppress cell growth, possibly by promoting endogenous PGEl p reduction or perhaps by an inherant toxic effect unrelated to the PG pathway. The finding that it did not potentiate the actions of GLA may not be unexpected if, in fact, PGEl is not the mediator of the effects observed with this fatty acid. The reduced activity of GLA in the presence of ethanol may be related to a physicochemical reaction between the alcohol and the fatty acid reducing the net amount of the latter available to the cell. A similar reaction between DGLA and ethanol could likewise decrease availability

and

hence

the

effect

218

of the

fatty

acid.

If in fact the effects of GLA in malignant cells are not attributable to PGsthen it is possible that growth retardation and cell death may be related to a nonspecific Certainly many fatty acids with related fat effect. structures have been reported to produce similar effects acid, AC-linolenic acid, arachidonic i.e., LA, GLA, and even DGLA at high eicosapentaenoic acid and DHA (6) not explain the anomalous doses. This would, however, While GLA and DGLA display finding with DGLA at low doses. only a slight difference in chemical structure the extent to which they are incorporated into the cell membrane or considerably. metabolised within the cell may vary of cells treated with the two fatty acids Microscopy (Figs. 2-5). The revealed markedly different morphologies accumulation of lipid in the cytoplasm of cells exposed to GLA may indicate an inability to metabolise this fatty acid, the resultant overload being responsible for altered cell In contrast DGLA treated function and subsequent death. cells did not accumulate obvious lipid and appeared to benefit from administration of up to 50 uglml of this fatty acid. Since Gaspar et al. have shown that minimal deviation hepatoma 7288 C cells can readily assimilate DGLA (14) it is possible that NUB 1 cells are also able to metabolise it. The contrasting results obtained with GLA and DGLA may possibly imply a deficiency of the elongase enzyme required to convert the former to the latter. The cell death which resulted from combined treatment with high doses of GLA and DGLA is presumably related to the effect of the excessive fat load upon the cells. However combined treatment with the two fatty acids at 50 ug each inhibited cell growth to a lesser extent than did 100 ug GLA Explanations for this could be that either DGLA alone. interfered in some way with the actions of GLA or by promoting cell growth DGLA allowed cells to use GLA more efficiently such that it was relatively less toxic. Whatever the exact explanation may be for the effects observed in malignant cells when GLA and DGLA are combined either with one another or with ethanol the very clear finding that DGLA produces opposite effects to those of GLA casts doubt upon the role of PGEI as mediator of the inhibition of cell growth and cell death produced by GLA. ACKNOWLEDGEMENT We wish to thank N. Ramchurren and S. Bux for technical assistance. This work was partially funded by the National Cancer Association (S.A.3.

219

REFERENCES 1.

Horrobin, D.F. The reversibility of cancer : The relevance of cyclic AMP, calcium, essential fatty acid and prostaglandin El. Med. Hypotheses 6: 469, 1980.

2.

Dippenaar, N., Booyens, J., Fabbri, D. and Katzeff, I.E. The reversibility of cancer : Evidence that malignancy in BL6 mouse melanoma cells is Gammalinolenic acid deficiency dependant. S. Afr. Med. J. 62:

505,

1982.

3.

Dippenaar, N., Booyens, J., Fabbri, D., Engelbrecht, P. and Katzeff, I.E. The reversibility of cancer : Evidence that malignancy in human hepatoma cells is S. Afr. Gamma-linolenic acid deficiency dependent. Med. J. 62: 683, 1982.

4.

Booyens, J., Dippenaar, N., Fabbri, D., Engelbrecht, P. and Katzeff, I.E. The effect of Gamma-linolenic acid on the growth of human osteogenic sarcoma and oesophageal carcinoma cells in culture. S. Afr. Med. J. 65:

5.

6.

240,

1984.

Leary, W.P., Robinson, K.M., Booyens, J. and Dippenaar, N.G. Some effects of Gamma-linolenic acid on cultured human oesophageal cells. S. Afr. Med. J. 62: 681, 1982. Booyens, J., Engelbrecht, P., le Roux, S., Louwrens, I.E. Some C.C., van der Merwe, C.F. and Katzeff, effects of the essential fatty acids linoleic acid, gamma alpha- linolenic acid and their metabolites linolenic acid, arachidonic acid, eicosapentaenoic acid, docosahexaenoic acid and of prostaglandins Al and El on the proliferation of human osteogenic Prostaglandins cells in culture. sarcoma Leukotrienes Med. 15: 15, 1984.

7.

Manku, M.S., Oka, M. and Horrobin, D.F. Differential regulation of the formation of prostaglandins and related substances from arachidonic acid and from dihomogammalinolenic acid. I. Effects of ethanol. Prostaglandins Med. 3: 119, 1979.

8.

Rygaard, J. and Povlsen, C.O. Heterotranplantation of a human malignant tumour to nude mice. Acta Pathologica et Microbiologica Scandinavica A. 77: 758, 1969.

220

9.

Halton, D.M., Peterson, W.D. and Hukku, B. Cell culture quality control by rapid isoenzymatic characterization. In Vitro. 19(l): 16, 1983.

10.

Robinson, K-M., Haffejee, A.A. and Angorn, I.B. Tissue culture and prognosis in carcinoma of the oesophagus. Clin. Oncol. 6: 125, 1980.

11.

Ingerman-Wojenski, C., Silver, M.I., Smith, B., Nissenbaum, M. and Sedar, A.W. Prostacyclin production in rabbit arteries in situ: inhibition of AAinduced endothelial cell damage. Prostaglandins 21, 655, 1981.

Scanning and 12. Robinson, K.M. and Maistry, L. transmission electron microscopy of continuous human esophageal carcinoma cell lines. In: Pfeiffer, C.J., Cancer of the Esophagus. Vol. II, CRC, Boca Raton, 113-118, 1982. 13.

Botha, J.H. and Robinson, K.M. The response of human carcinoma cell lines to gamma-linolenic acid with special reference to the effects of agents which influence prostaglandin and thromboxane synthesis. Prostaglandins Leukotrienes Med. (in press).

14.

Gaspar, G., de Alaniz, M.J.T. and Brenner, R.R. Uptake and metabolism of exogenous eicosa-8, 11, 14, trienoic acid in minimal deviation hepatoma 7288C cells. Lipids 10: 726, 1975.

221