Closely related colon cancer cell lines display different sensitivity to polyunsaturated fatty acids, accumulate different lipid classes and downregulate sterol regulatory element-binding protein 1

Abstract

N-6 polyunsaturated fatty acids (PUFAs) may be associated with increased risk of colon cancer, whereas n-3 PUFAs may have a protective effect. We examined the effects of docosahexaenoic acid (DHA), eicosapentaenoic acid and arachidonic acid on the colon carcinoma cell lines SW480 derived from a primary tumour, and SW620 derived from a metastasis of the same tumour. DHA had the strongest growth-inhibitory effect on both cell lines. SW620 was relatively more growth-inhibited than SW480, but SW620 also had the highest growth rate in the absence of PUFAs. Flow cytometry revealed an increase in the fraction of cells in the G₂/M phase of the cell cycle, particularly for SW620 cells. Growth inhibition was apparently not caused by increased lipid peroxidation, reduced glutathione or low activity of glutathione peroxidase. Transmission electron microscopy revealed formation of cytoplasmic lipid droplets after DHA treatment. In SW620 cells an eightfold increase in total cholesteryl esters and a 190-fold increase in DHA-containing cholesteryl esters were observed after DHA treatment. In contrast, SW480 cells accumulated DHA-enriched triglycerides. Arachidonic acid accumulated in a similar manner, whereas the nontoxic oleic acid was mainly incorporated in triglycerides in both cell lines. Interestingly, nuclear sterol regulatory element-binding protein 1 (nSREBP1), recently associated with cell growth regulation, was downregulated after DHA treatment in both cell lines. Our results demonstrate cell-specific mechanisms for the processing and storage of cytotoxic PUFAs in closely related cell lines, and suggest downregulation of nSREBP1 as a possible contributor to the growth inhibitory effect of DHA.

Abbreviations

AA

arachidonic acid

ACAT

acyl CoA:cholesterol acyltransferase

BHA

butylated hydroxyanisole

BHT

butylated hydroxytoluene

COX

cyclooxygenase

DGAT

diacylglycerol acyltransferase

DHA

docosahexaenoic acid

EPA

eicosapentaenoic acid

FAMEs

fatty acid methyl esters

GSH

glutathione

GSH-Px

glutathione peroxidase

h-ALAS

human 5-aminolevulinate synthase

h-PBGD

human porphobilinogen deaminase

LOX

lipoxygenase

LPPs

lipid peroxidation products

MDA

malondialdehyde

MTT

3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide

NDGA

nordihydroguaiaretic acid

NS-398

N-[2-cyclohexyloxy)-4-nitrophenyl]-methanesulfonamide

OA

oleic acid

PI

propidium iodide

PUFAs

polyunsaturated fatty acids

SOD

superoxide dismutase

SREBP

sterol regulatory element-binding protein

TBA

2-thiobarbituric acid

TBARS

thiobarbituric acid reactive substances

TUNEL

terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end labelling

Epidemiological studies indicate that there is an inverse association between intake of polyunsaturated fatty acids (PUFAs) and incidence of breast, colon and prostate cancers, although these studies are not uniformly conclusive [1–4]. Also, cancer xenografts in immunosuppressed mice as well as cell culture studies demonstrate that PUFAs may slow down cancer cell growth, induce apoptosis and increase the efficiency of chemotherapeutic drugs [5–8]. The mechanisms behind these effects are clearly complex and not well understood. However, factors possibly implicated in the PUFA-mediated effects include modification of tumour cell membranes which can affect cell signalling pathways [9], lipid peroxidation and oxidative stress [10], eicosanoid production [11], fatty acid metabolism [12] and the regulation of gene expression [13]. Also, the activity of different antioxidant defence enzymes seems to vary among different cell lines and between normal and malignant cells, and may be of importance for cancer cells susceptibility towards n-3 PUFAs [14,15]. One of the major mechanisms of PUFA-mediated toxicity on cancer cells is thought to be the lipid peroxidation process, which is initiated by free radical attack on membrane PUFAs, leading to the formation of a wide range of very reactive and genotoxic lipid peroxidation products (LPPs) [16]. Altogether, these data suggest that intake of n-3 PUFAs may modulate cell behavior and growth by a variety of mechanisms.

Previously we have shown that the growth inhibitory effect of docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA) on A-427 cells (human lung adenocarcinoma cell line) may be reversed by vitamin E, sodium selenite and ebselen, a synthetic glutathione peroxidase (GSH-Px) mimic, indicating that lipid peroxidation is responsible for the cytotoxic effect [10]. The high PUFA sensitivity of this cell line seems to be caused by a low level of the antioxidant defence enzyme GSH-Px. We have also shown that this cell line is sensitive to the hydroperoxy derivatives of the PUFAs, indicating that cytotoxicity of n-3 PUFAs may be mediated via the formation of these primary products [17].

In this study we wanted to explore possible mechanisms of cytotoxicity induced by PUFAs. For these studies we used two colon carcinoma cell lines, SW480 and SW620, derived from a primary and a secondary tumour of the same patient, respectively. We find that both cell lines are growth-inhibited by PUFAs, although SW620 cells display a higher degree of PUFA sensitivity than SW480 cells. Our results indicate that lipid peroxidation or deficient defence against oxidants is not involved in sensitivity, contrary to some other cell lines. Rather, our results indicate that differences between cell lines in regard to processing and storage of potentially harmful PUFAs and downregulation of nuclear sterol regulatory element-binding protein 1 (nSREBP1) may be related to effects on cancer cell growth.

Results

Growth inhibitory effects of n-3 and n-6 PUFAs

The effects of DHA, EPA and arachidonic acid (AA) on the growth of SW480 and SW620 are shown in Fig. 1. All three PUFAs inhibited the growth of both cell lines in a time- and concentration-dependent manner. Cell proliferation of SW480 was reduced by 77% compared with control cultures after exposure to 70 µm DHA for 144 h, while EPA and AA reduced cell growth by 50 and 44%, respectively. Cell proliferation of SW620 was somewhat more affected; DHA decreased cell proliferation by 95% after 144 h incubation compared with control cultures, whereas EPA and AA reduced cell proliferation by 75% each. The strongest effect on cell growth was seen with DHA for both cell lines, which is in agreement with previous results using a different cell line [10]. SW620 was relatively more growth-inhibited than SW480, but SW620 also had the highest growth rate in the absence of PUFAs (Fig. 1). Oleic acid (OA) had no effect on cell growth (results not shown).

Figure 1.  Effect of docosahexaenoic acid (DHA), eicosapentaenoic acid (EPA) and arachidonic acid (AA) on the growth of the human colon carcinoma cell lines SW480 and SW620. Cell survival was assessed using the MTT assay as described in Experimental procedures. The values represent the mean of at least four parallels ± SD from one experiment, and are verified through at least two experiments. D, attenuance.

Cell-cycle analysis and apoptosis

Total cellular DNA content and DNA fragmentation were analysed by flow cytometry. Treatment of the cells for 96 h with DHA (70 µm) resulted in a significant increase in the population of cells in the G₂/M phase compared with control cells, 1.7- and 2.5-fold for SW480 and SW620, respectively (Fig. 2). However, the strong inhibitory effect on growth indicates that other phases must also be affected. After 120 h, the population of cells arrested in the G₂/M phase was diminished to 1.3- and 1.7-fold increase relative to controls for SW480 and SW620, respectively (data not shown). The terminal deoxynucleotidyl transferase (TdT)-mediated dUTP-biotin nick end labelling (TUNEL) assay revealed no evidence for DNA-fragmentation and therefore no evidence for apoptosis (data not shown).

Figure 2.  Docosahexaenoic acid (DHA)-induced G₂/M arrest in SW480 and SW620 cells. Cells were treated with DHA (70 µm) for the indicated times, stained with propidium iodide (PI) and examined by flow cytometry as described in Experimental procedures. (A) Untreated SW480 cells after 96 h; (B) SW480 cells treated with DHA (70 µm) for 96 h; (C) untreated SW620 cells after 96 h; (D) SW620 cells treated with DHA (70 µm) for 96 h.

Effect of vitamin E, acyl CoA:cholesterol acyltransferase inhibitor and inhibitors of eicosanoid synthesis on DHA-induced growth inhibition

In order to determine whether vitamin E could counteract the DHA-induced growth inhibition of the two cell lines, SW480 and SW620 were treated with DHA (70 µm) and vitamin E (10 or 50 µm) simultaneously. Vitamin E was not able to prevent growth inhibition in either cell line and even enhanced the antiproliferative effect of DHA in SW480 cells (Fig. 3). Vitamin E alone had no effect on cell survival (data not shown). We also examined whether the lipoxygenase inhibitor nordihydroguaiaretic acid (NDGA) and the cyclooxygenase inhibitor N-[2-cyclohexyloxy)-4-nitrophenyl]-methanesulfonamide (NS-398) could reverse the growth inhibition induced by DHA in these two cell lines. When the cells were coincubated with DHA (70 µm) and NDGA (0.01–1 µm) or NS-398 (0.01–10 µm), no effect was seen on cell survival (data not shown). NDGA and NS-398 had no effect on cell survival alone at the concentrations indicated. Also, coincubation of cells with the acyl CoA:cholesterol acyltransferase (ACAT) inhibitor Sandoz 58-035 (2.1–8.5 µm) and DHA, had no effect on cell survival (data not shown).

Figure 3.  Lack of effect of vitamin E (10 and 50 µm) on DHA-induced growth inhibition in SW480 and SW620 cells at different time points. Cell survival was assessed using the MTT assay as described in Experimental procedures. Control, ▪; Control/EtOH, □; DHA (70 µm), ; DHA(70 µm)/Vit. E (10 µm), ; DHA (70 µm)/Vit.E (50 µm), . The values represent the mean of at least four parallels ± SD from one experiment, and are verified through at least two experiments.

Lipid peroxidation

Because generation of free radicals and subsequent lipid peroxidation are important for PUFA-induced toxicity in several tumour cell lines, we measured the level of malondialdehyde (MDA), a major secondary lipid peroxidation product, after DHA treatment alone or in combination with vitamin E, both in cells and culture medium. A several fold increase in the MDA level was observed in both SW480 and SW620 cells after DHA treatment (70 µm) for 72 h when compared with control (Table 1). A combinatory treatment with DHA (70 µm) and vitamin E (50 µm) reduced the MDA level below the level in control cultures. Only minor amounts of MDA were found in the culture medium and the level did not exceed that in control cultures of the cell lines.

Table 1.   Effect of DHA alone or in combination with vitamin E on lipid peroxidation measured as MDA ^a  in SW480 and SW620 cells.

Cell line/treatment	MDA (nmol·mg protein^−1) ± SD
	Cell extract	Released into media
^a  The value is the mean calculated from measurements performed in triplicate ±  SD in one of two representative experiments.
SW480
Control	3.20 ± 0.21	0.64 ± 0.15
DHA (70 µm)	44.82 ± 0.71	3.00 ± 0.40
DHA (70 µm)/vit.E (50 µm)	2.46 ± 0.29	2.17 ± 0.24
SW620
Control	1.13 ± 0.09	1.06 ± 0.08
DHA (70 µm)	12.62 ± 0.60	0.88 ± 0.07
DHA (70 µm)/vit.E (50 µm)	0.48 ± 0.05	0.79 ± 0.07

Effect of DHA on glutathione levels and GSH-Px activities

The total level of glutathione (GSH) in SW480 and SW620 cells is shown in Fig. 4. The level of GSH is approximately equal in the two cell lines and comparable with the GSH level in other PUFA-sensitive cell lines [10]. An approximately twofold increase in the total amount of glutathione was observed in both cell lines after DHA treatment for 48 h. This increase may represent a rebound effect, which was similar in the two cell lines. However, the sensitivity of the cell lines to PUFAs does not seem to correlate with the GSH level.

Figure 4.  Effect of 48 h treatment with DHA (70 µm) on the GSH level in the human colon carcinoma cell lines SW480 and SW620. Values represent the mean ± SD of triplicate measurements from three separate experiments. *Results analysed by Student's t-test and considered significant different from control (P < 0.05).

To examine whether deficiencies in the antioxidant defence mechanisms could be used as a determinant for PUFA sensitivity, we also measured the GSH-Px activities in the two cell lines (Fig. 5). Compared with the PUFA-senstitive cell line A-427, in which the activity level was very low [10], the activity level in SW480 and SW620 was significantly higher, about two- and sixfold, respectively. However, the fact that the most PUFA-sensitive cell line, SW620, had a 2.5-fold higher level of activity compared with SW480, indicates that there is no relationship between the activity level of GSH-Px and PUFA sensitivity in this case. DHA treatment (70 µm) alone and pretreatment with sodium selenite (250 nm) for 20 h before DHA administration did not lead to any significant change in the GSH-Px activity level in the two cell lines. In accordance with this, pretreating SW480 and SW620 with sodium selenite (250 nm) did not affect cell growth after DHA treatment (data not shown). This gives further support for the finding that lipid peroxidation is not involved in the DHA-mediated cytotoxicity in SW480 and SW620 and that deficiencies in the activity level of GSH-Px can not be used as a determinant for cancer cells susceptibility to PUFA-treatment.

Figure 5.  Effect of 48 h treatment with DHA (70 µm) on the GSH-Px activity in the human colon carcinoma cell lines SW480 and SW620. GSH-Px activity was measured with or without pretreatment with sodim selenite (250 nm) for 20 h before DHA supplementation for 48 h in both cell lines. Results are expressed in nmol NADPH oxidized·min⁻¹·mg⁻¹ protein. Values represent the mean ± SD from one experiment. Each determination was performed in triplicate, and verified through at least two experiments.

Effect of DHA on cell morphology and lipid accumulation

In order to determine whether the antiproliferative effect of DHA was associated with changes in morphology, cells were examined by transmission electron microscopy. After incubation with 70 µm DHA for 72 h, electron microscopy revealed accumulation of lipid droplets in both SW480 and SW620 cells, but no pyknotic nuclei indicative of apoptosis were observed (Fig. 6). Lipid extraction followed by GC-analysis revealed accumulation of DHA in all the lipid fractions. The most striking observation, however, was an ∼190-fold increase in the amount of DHA in the triglyceride- and cholesteryl ester fraction of DHA-treated SW620 cells compared with control cells. In SW480 cells, a 170-fold increase in the amount of DHA was observed in the triglyceride fraction after DHA treatment, whereas there was an increase from nondetectable levels to 16.7 µg·mg⁻¹ protein in the cholesteryl ester fraction. Despite an increase of DHA in both the triglyceride- and cholesteryl ester fraction in both cell lines, DHA accumulated mainly in the form of triglycerides in SW480 cells as opposed to cholesteryl esters in SW620 cells (Tables 2–5). DHA-treatment resulted in an eightfold increase in total cellular cholesteryl ester content in SW620 cells, whereas a 10-fold increase in total cellular triglyceride content was observed in SW480 cells, compared to control cells (Table 6).

Figure 6.  Transmission electron micrographs of SW480 and SW620 cells. (A) Control cells. (B) Cells treated with DHA (70 µm, 72 h). The nucleus and lipid droplet is indicated by N and L, respectively. Bars = 2 µm.

Table 2.   Fatty acid composition of different lipid fractions in SW480 cells.

Fatty acids	Lipid fraction analysed by GC (µg·mg protein^−1)^a
	Free fatty acids	Phospholipids	Monoglycerides	Diglycerides	Triglycerides	Cholesteryl esters
^a  Results are expressed in µg·mg protein^−1 and given as mean ± SD from three separate experiments. ND, not detectable.
16:0	3.06 ± 1.07	49.10 ± 9.46	5.60 ± 0.45	5.21 ± 3.17	4.03 ± 3.76	12.36 ± 7.68
16:1	0.09 ± 0.08	5.98 ± 1.60	0.54 ± 0.02	ND	0.33 ± 0.29	ND
18:0	4.06 ± 1.53	18.86 ± 3.52	1.86 ± 0.25	4.99 ± 2.70	5.09 ± 1.52	14.51 ± 8.90
18:1 (n-9)	1.06 ± 0.37	53.08 ± 9.02	5.08 ± 0.17	2.16 ± 1.06	3.84 ± 1.08	1.01 ± 1.74
18:2 (n-6)	0.11 ± 0.09	4.64 ± 0.73	0.46 ± 0.04	0.26 ± 0.45	0.39 ± 0.68	0.23 ± 0.41
18:3 (n-6)	0.16 ± 0.18	ND	ND	ND	0.24 ± 0.21	0.15 ± 0.26
18:3 (n-3)	ND	ND	0.13 ± 0.11	5.10 ± 1.93	ND	ND
20:0	ND	0.17 ± 0.29	ND	ND	ND	ND
20:3 (n-6)	0.25 ± 0.22	1.60 ± 0.22	ND	ND	ND	ND
20:4 (n-6)	0.10 ± 0.09	20.37 ± 3.06	0.33 ± 0.58	ND	ND	ND
20:3 (n-3)	0.05 ± 0.09	ND	ND	ND	0.74 ± 0.76	ND
20:5 (n-3)	0.43 ± 0.37	3.49 ± 0.38	ND	ND	ND	ND
22:0	ND	0.75 ± 0.23	ND	ND	ND	ND
22:4 (n-6)	ND	1.22 ± 0.10	ND	ND	ND	ND
22:5 (n-3)	ND	8.36 ± 1.47	0.28 ± 0.25	ND	ND	ND
22:6 (n-3)	ND	13.15 ± 1.07	0.19 ± 0.34	ND	0.49 ± 0.86	ND

Table 3.   Fatty acid composition of different lipid fractions in SW480 cells treated with DHA.

Fatty acids	Lipid fraction analysed by GC (µg/mg protein)^a
	Free fatty acids	Phospholipids	Monoglycerides	Diglycerides	Triglycerides	Cholesteryl esters
^a  Results are expressed in µg·mg protein^−1 and given as mean ± SD from three separate experiments. *P < 0.05 compared with control cells, data were analysed by student's paired t-test. ND, not detectable.
16:0	3.48 ± 0.37	54.94 ± 14.79	7.37 ± 1.54	5.44 ± 1.35	15.48 ± 6.60	13.18 ± 8.23
16:1	0.14 ± 0.16	3.87 ± 1.18	0.31 ± 0.30	0.01 ± 0.01	3.10 ± 1.57	0.19 ± 0.33
18:0	4.89 ± 0.61	23.26 ± 5.36	2.97 ± 1.10	6.62 ± 1.04	8.23 ± 3.48	11.90 ± 9.78
18:1 (n-9)	1.25 ± 0.55	35.68 ± 8.66	3.82 ± 0.59*	2.00 ± 0.56	22.94 ± 10.31	5.35 ± 2.72
18:2 (n-6)	0.29 ± 0.04	3.37 ± 0.74	0.45 ± 0.08	0.16 ± 0.27	3.37 ± 1.26	ND
18:3 (n-6)	0.07 ± 0.06	ND	ND	ND	0.28 ± 0.25	2.30 ± 2.03
18:3 (n-3)	ND	ND	0.08 ± 0.14	6.27 ± 1.00	ND	ND
20:0	0.12 ± 0.02*	0.38 ± 0.09	ND	ND	0.13 ± 0.23	ND
20:3 (n-6)	0.20 ± 0.11	2.62 ± 0.47*	0.10 ± 0.18	ND	2.65 ± 1.00*	0.75 ± 1.30
20:4 (n-6)	0.41 ± 0.22	17.12 ± 3.33	1.27 ± 0.34	0.22 ± 0.38	3.44 ± 1.13*	ND
20:3 (n-3)	0.11 ± 0.03	0.07 ± 0.13	ND	0.17 ± 0.29	1.59 ± 1.12	1.33 ± 2.30
20:5 (n-3)	0.15 ± 0.15	3.71 ± 1.38	0.12 ± 0.21	ND	1.05 ± 1.16	ND
22:0	ND	0.72 ± 0.11	0.06 ± 0.10	ND	ND	ND
22:4 (n-6)	0.02 ± 0.04	1.37 ± 0.15	0.04 ± 0.07	ND	1.42 ± 0.49*	ND
22:5 (n-3)	0.23 ± 0.21	5.72 ± 1.00*	0.27 ± 0.24	ND	4.52 ± 1.95*	0.34 ± 0.59
22:6 (n-3)	4.25 ± 2.12*	56.81 ± 15.05*	4.30 ± 0.80*	2.92 ± 2.53	84.26 ± 44.54*	16.70 ± 3.50*

Table 4.   Fatty acid composition of different lipid fractions in SW620 cells.

Fatty acids	Lipid fraction analysed by GC (µg·mg protein^−1)^a
	Free fatty acids	Phospholipids	Monoglycerides	Diglycerides	Triglycerides	Cholesteryl esters
^a  Results are expressed in µg/mg protein and given as mean ±  SD from three separate experiments. ND, not detectable.
16:0	1.58 ± 0.52	19.73 ± 1.77	2.56 ± 1.23	2.17 ± 1.11	5.00 ± 2.72	3.98 ± 1.85
16:1	ND	2.37 ± 0.56	0.20 ± 0.17	ND	0.55 ± 0.57	0.44 ± 0.38
18:0	2.02 ± 0.32	7.34 ± 0.78	0.88 ± 0.27	1.84 ± 1.07	2.17 ± 0.78	1.92 ± 0.69
18:1 (n-9)	0.51 ± 0.56	24.27 ± 3.24	2.61 ± 1.36	0.47 ± 0.82	7.51 ± 5.55	5.54 ± 3.03
18:2 (n-6)	0.05 ± 0.09	1.78 ± 0.24	0.19 ± 0.17	ND	0.52 ± 0.56	0.46 ± 0.43
18:3 (n-6)	ND	ND	ND	ND	ND	ND
18:3 (n-3)	ND	ND	ND	ND	ND	ND
20:0	ND	ND	ND	ND	ND	ND
20:3 (n-6)	ND	0.32 ± 0.29	ND	ND	ND	ND
20:4 (n-6)	ND	8.61 ± 0.91	0.37 ± 0.32	ND	ND	ND
20:3 (n-3)	ND	ND	ND	ND	ND	ND
20:5 (n-3)	ND	0.79 ± 0.69	ND	ND	ND	ND
22:0	ND	ND	ND	ND	ND	ND
22:4 (n-6)	ND	0.30 ± 0.27	ND	ND	ND	ND
22:5 (n-3)	ND	3.62 ± 0.44	0.18 ± 0.16	ND	0.14 ± 0.24	0.36 ± 0.31
22:6 (n-3)	ND	4.69 ± 0.56	0.19 ± 0.17	ND	0.12 ± 0.20	0.35 ± 0.30

Table 5.   Fatty acid composition of different lipid fractions in SW620 cells treated with DHA.

Fatty acids	Lipid fraction analysed by GC (µg·mg protein^−1)^a
	Free fatty acids	Phospholipids	Monoglycerides	Diglycerides	Triglycerides	Cholesteryl esters
^a  Results are expressed in µg·mg protein^−1 and given as mean ± SD from three separate experiments. ND, not detectable. *P < 0.05 compared with control cells, data were analysed by Student's paired t-test. **P = 0.0539 compared with control cells, data were analysed by Student's paired t-test.
16:0	2.05 ± 0.21	27.00 ± 3.95	2.90 ± 1.54	2.64 ± 0.55	5.01 ± 1.40	10.88 ± 5.68*
16:1	ND	1.63 ± 0.63	0.06 ± 0.10	ND	0.78 ± 0.11	1.91 ± 1.65
18:0	3.12 ± 0.20*	13.64 ± 2.28*	1.31 ± 0.73	2.19 ± 1.39	2.41 ± 0.40	4.70 ± 1.23*
18:1 (n-9)	0.52 ± 0.48	18.98 ± 3.95	1.93 ± 0.99	1.28 ± 1.11	7.22 ± 2.29	16.09 ± 9.19*
18:2 (n-6)	ND	1.99 ± 0.28	0.19 ± 0.18	ND	1.10 ± 0.24	2.37 ± 1.52*
18:3 (n-6)	ND	ND	ND	ND	0.23 ± 0.40	ND
18:3 (n-3)	ND	ND	ND	ND	ND	ND
20:0	ND	ND	ND	ND	ND	ND
20:3 (n-6)	ND	1.27 ± 0.13	ND	ND	0.50 ± 0.50	0.60 ± 1.03
20:4 (n-6)	ND	9.17 ± 1.07	0.45 ± 0.49	ND	0.66 ± 0.60	0.89 ± 1.53
20:3 (n-3)	ND	ND	ND	ND	ND	ND
20:5 (n-3)	0.07 ± 0.11	2.24 ± 0.93	ND	ND	ND	0.54 ± 0.94
22:0	0.19 ± 0.34	ND	ND	ND	ND	ND
22:4 (n-6)	ND	ND	ND	ND	ND	0.35 ± 0.61
22:5 (n-3)	ND	2.34 ± 0.54	ND	ND	1.14 ± 0.29*	3.11 ± 2.75
22:6 (n-3)	2.00 ± 0.67*	29.42 ± 0.85*	2.69 ± 0.91*	7.29 ± 5.16	23.55 ± 9.61*	67.23 ± 41.75**

Table 6.   Total amount of fatty acids in different lipid fractions in SW480 and SW620 cells.

Lipid fraction	SW480 Control (µg·mg protein^−1) (% of total)	DHA treated (µg·mg protein^−1) (% of total)	SW620 Control (µg·mg protein^−1) (% of total)	DHA treated (µg·mg protein^−1) (% of total)
Free fatty acids	9.4 (3.5)	15.6 (3.3)	4.2 (3.5)	8.0 (2.7)
Phospholipids	180.8 (68.1)	209.6 (44.2)	73.8 (62.2)	107.7 (37.2)
Monoglycerides	14.5 (5.4)	21.2 (4.5)	7.2 (6.0)	9.5 (3.3)
Diglycerides	17.7 (6.7)	23.8 (5.0)	4.5 (3.8)	13.4 (4.6)
Triglycerides	15.2 (5.7)	152.5 (32.1)	16.0 (13.5)	42.6 (14.7)
Cholesteryl esters	28.3 (10.6)	52.0 (11.0)	13.1 (11.0)	108.7 (37.5)

These observations were accompanied by a slight, but significant increase in the amount of 18:0 (free fatty acids), 18:0 (phospholipids) and 16:0, 18:0, 18:1 (n-9) and 18:2 (n-6) (cholesteryl esters) in SW620 cells. In addition to the accumulation of DHA in different lipid fractions of SW480, a slight, but significant increase could be observed in 20:0 (free fatty acids), 20:3 (n-6) and 22:5 (n-3) (phospholipids), 18:1 (n-9) (monoglycerides) as well as 20:3 (n-6), 20:4 (n-6), 22:4 (n-6) and 22:5 (n-3) (triglycerides). In a similar manner, AA was found to be incorporated preferentially in the triglyceride fraction in SW480 cells, whereas the cholesteryl ester fraction was most affected in SW620 cells (Fig. 7).

Figure 7.  Amount of lipid (% of total) in the phospholipids (▪), triglyceride (□), and cholesteryl ester () fraction in SW480 and SW620 cells treated or not treated with DHA, AA or OA (70 µm, 72 h). Values represent the mean ± SD for SW620 treated with DHA and AA and SW480 treated with DHA, otherwise values represent the mean of two replicates. *P < 0.05 compared with control cells; data were analysed by Student's paired t-test.

In contrast to the cell-type-dependent lipid accumulation pattern of DHA and AA, an increase in the total amount of triglycerides in both cell lines was observed after treatment with OA, whereas the cholesteryl ester fraction was unaffected.

Effect of DHA on mRNA levels of lipid-metabolizing enzymes

The expression of major genes implicated in the synthesis of triglycerides and cholesteryl esters, diacylglycerol acyltransferase 1 and 2 (DGAT1, DGAT2) and ACAT1, were analysed in both cell lines after treatment with DHA at different time periods. The mRNA levels of DGAT1 and DGAT2 were similar and did not change significantly with time in SW480 and SW620 control cells, nor did DHA treatment affect the mRNA expression levels of DGAT1 and DGAT2 (data not shown). The mRNA level of ACAT1 in SW620 control cells was similar to SW480 control cells. After DHA treatment, a 2.5-fold reduction in the mRNA level of ACAT1 was observed in SW620 cells at 72 h, whereas no significant effect was observed at earlier time points in either cell lines (data not shown).

Effect of DHA on protein levels of ACAT1 and SREBP1

Western blotting was used to estimate the relative amounts of ACAT1 and precursor and nuclear forms of SREBP1 (Fig. 8). The protein level of ACAT1 started to decrease after 48 and 24 h treatment with DHA relative to control in SW480 and SW620 cells, respectively. SREBP1 required higher protein input than ACAT1 for acceptable detection and the level was lower in SW480 than in SW620. In addition, the precursor form of SREBP1 was significantly less abundant than the nuclear form and hardly detectable in SW620 cells. The protein level of nSREBP1 decreased with time in both cell lines (Fig. 8). After 6 h treatment with DHA, the protein level of nSREBP1 was reduced by ∼26% in SW480 cells and ∼18% in SW620 cells (Fig. 8).

Figure 8.  ACAT1 and SREBP1 [uncleaved precursor (p) and nuclear (n)] western analysis of whole-cell extracts of DHA treated SW480 and SW620 cells. Aliquots of each sample – SW480: 30 µg (ACAT1), 60 µg (SREBP1); SW620: 20 µg (ACAT1), 50 µg (SREBP1) – were electrophoresed on 10% SDS-polyacylamide gels, and western analysis performed with ACAT1 antiserum at 1:1000 dilution and SREBP1 antiserum at 1:500 dilution. The membranes were reprobed with actin as a lane loading control (*ACAT1, **pSREBP1/nSREBP1). The numbers beneath the bands denote the average fold change relative to control of three independent experiments.

Discussion

Epidemiological and experimental studies conducted over the past few decades suggest a protective role for n-3 PUFAs against the development of colon cancer [3,4,18,19]. N-3 PUFAs have also been shown to reduce the growth of human colon cancer cells in nude mice [5]. In this study we show that the two human colon cancer cell lines SW480 and SW620 are growth-inhibited by both n-3 and n-6 PUFAs, although to different extents. DHA was found to elicit the most pronounced effect on cell growth in both cell lines, whereas EPA and AA induced growth inhibition more equally and to a lesser extent. This is in agreement with previous results with other cell lines [10,20,21]. Both cell lines used here are derived from colon cancer in a single patient, one from the primary tumour and one from a metastatic lesion, and a monoclonal origin is likely [22,23]. The difference between the two cell lines with regard to PUFA sensitivity may therefore be ascribed to genetic changes during cancer progression or in establishing the cell lines. The cell lines also accumulate different lipid classes in cytoplasmic lipid droplets and express some lipid-metabolizing enzymes differently. We do not infer that these differences apply to pairs of primary and metastatic cell lines in general, as we have only studied one pair. Rather, the results point to diverse properties of cancer cells in general, even between closely related cell lines. A correlation between the sensitivity in vitro and in vivo has been observed after implantation of different tumour cell lines in mice [24]. These results and similar studies indicate that in vitro studies are relevant for the in vivo situation. However, this may not always be the case, particularly if high concentrations are required to obtain a significant effect on growth.

Several factors have been implicated in the mechanism of PUFA-induced cytotoxicity in tumour cells, one being the formation of different lipid peroxidation products (LPPs) [16]. PUFAs are highly susceptible to both enzymatic and nonenzymatic peroxidation [25]. The enzymatic peroxidation may either be by cyclo-oxygenase (COX) or lipoxygenase (LOX) enzymes leading to the formation of different eicosanoids, and their activities may be inhibited by NS-398 and NDGA, respectively. When NS-398 (0.01–10 µm) or NDGA (0.01–1 µm) was added to the cells together with DHA (70 µm), no effect was seen on cell growth (data not shown). These results indicate that the DHA-induced cytotoxicity is not mediated through the formation of eicosanoids. This is in agreement with several other reports concluding that prostaglandins and leukotrienes do not regularly participate in the growth inhibitory action of PUFAs [7,26,27].

However, several reports have shown that antioxidants like vitamin E, butylated hydroxytoluene (BHT) and butylated hydroxyanisole (BHA) block the cytotoxic effect of different PUFAs, indicating that nonenzymatic lipid peroxidation is frequently involved [7,17,28]. Previous studies also revealed a good correlation between the MDA concentration and inhibition of colon cancer cell proliferation by PUFAs [29,30]. When measuring the level of MDA induced by DHA in SW480 and SW620, we found that the level was significantly increased in both cell lines and that vitamin E was able to reduce the level below that in control cells. In contrast to previous studies, however, vitamin E was not able to restore cell growth, although it eliminated accumulation of MDA after treating these cells with DHA, indicating that lipid peroxidation is not the cause of cytotoxicity in SW480 and SW620 cells.

Tumour cells seem to frequently have a low level of antioxidant enzymes like superoxide dismutase (SOD), GSH-Px and catalase making them more vulnerable to oxidative stress [31]. GSH-Px catalyses the reduction of hydrogen peroxide and fatty acid hydroperoxides and contributes to cells defence against deleterious lipid peroxidation products [32]. Under normal conditions, inhibition of GSH-Px may affect cell survival and low GSH-Px activity is especially important under conditions of induced oxidative stress such as PUFA treatment [32]. In accordance with this and our previous findings [10], we measured the activity level of selenium-dependent GSH-Px. Compared with the activity level of GSH-Px in PUFA-resistant cell lines [10], the level was about the same in SW480 cells and even twofold higher in SW620 cells. The activity level was not influenced by DHA-treatment. Pretreatment with sodium selenite before DHA supplementation did not affect the activity level of GSH-Px nor cell survival in both cell lines (data not shown). We have previously reported that the effect of DHA in one particular cell line (A-427) is likely to be mediated through the formation of hydroperoxy-DHA [17]. The capacity of PUFAs to arrest growth of breast cancer cells has also been reported to correlate with the accumulation of lipid peroxides and conjugated dienes in cellular lipids [28]. However, it seems more likely that the cytotoxic effect of DHA in SW480 and SW620 is mediated by a mechanism independent of lipid peroxidation and the formation of toxic fatty acid hydroperoxides.

The growth-inhibitory effect of PUFAs on cancer cells is often related to induction of apoptosis characterized by morphological changes such as cell shrinkage, chromatin condensation, membrane blebbing and nuclear fragmentation [33]. Previous studies indicate that a threshold level exists regarding PUFA concentrations and incubation times below which induction of apoptosis is not observed. DHA-induced apoptosis of HT-29 colon cancer cells has been shown to be dose dependent with a threshold at 50 µm with maximal DNA fragmentation at 150 µm[20]. Also, caspase 3 activation in Jurkat cells by DHA has been shown to be dose- and time dependent, the effective DHA concentration being 60 µm[34]. However, studies on the effect of DHA (50 or 100 µm) on the breast cancer cell line MCF-7 revealed no induction of apoptosis or effects on the cell-cycle profile; therefore all phases of the cell cycle must be affected in this cell line [21]. The observations reported by Chamras et al. are in agreement with our results and may indicate a concentration-dependent threshold for induction of apoptosis [21].

Lipid synthesis is critical for cell growth and proliferation. Previous studies have suggested that uptake and storage of PUFAs in the form of triglyceride-containing lipid droplets may be related to PUFA-induced apoptosis [35,36]. Recent findings also suggest that the synthesis and storage of cholesteryl esters are related to proliferative processes in different experimental models and in several types of human neoplasms ensuring the cells need for readily utilized cholesterol during active growth [37,38]. Our results demonstrate that DHA-induced cell-cycle arrest targeted at least in part to the G₂/M phase of cancer cells may be related to the accumulation of cholesteryl esters. Hakumaki et al. report that the accumulation of lipids subsequent to gene therapy may be due to cells undergoing growth arrest prior to apoptosis [39]. Ultrastructural investigations of a human breast cancer cell line (HBL-100) treated with a cationic aldehyde or a cationic acylhydrazine revealed progressive development of lipid droplets and substantial damage to mitochondria, but no morphological changes characteristic of apoptosis [40]. Also, Hirakawa et al. demonstrated that inhibition of cell growth was accompanied by accumulation of cholesteryl esters, triglycerides and glycerol ethers when tsA58/T24H-ras transformants ceased to grow [41]. Lipid production may not merely be related to the apoptotic process, but may be a more general stress indicator of the cells as proposed previously [40].

Sterol regulatory element-binding proteins (SREPBs) are transcription factors playing important regulatory roles in lipid homeostasis. The isoform most commonly found in cultured cells, SREBP1, is involved in both fatty acid and cholesterol biosynthesis [42]. Zhou et al. recently showed that suppression of SREBPs in vascular endothelial cells by either SCAP siRNA or 25-hydroxycholesterol leads to suppression of DNA synthesis and cell cycle arrest in the G₀/G₁ phase [43]. These results indicate that downregulation of SREBP is associated with growth reduction, but the mechanism is not obvious. Field et al. showed suppression of SREBP1 by PUFAs of both n-3 and n-6 type in Caco-2 cells, whereas no change was observed with 18:0 and 18:1, indicating a relationship between chain length, degree of unsaturation and the ability of individual fatty acids to inhibit SREBP1 protein expression [44]. A lower level and a more rapid decrease of nSREBP1 in SW480 cells compared with SW620 cells after DHA treatment may contribute to the observed differences in growth inhibition. Thus, the observed growth arrest in the G₂/M phase and the differences in accumulation of lipids between SW480 and SW620 cells may be mediated through differential regulation of nSREBPs as chain length and degree of unsaturation of fatty acids also are correlated with inhibition of tumour cell growth. Rumsey et al. have observed that incubating macrophages with free fatty acids in the absence of an exogenous source of cholesterol did not produce changes in total cell cholesterol, but increased cholesteryl ester formation presumably by redistributing intracellular free cholesterol [45]. Wang et al. suggested that regulation of SREBP processing is sensitive to alterations in sterol content of endoplasmic reticulum [46]. Depletion of free cholesterol from endoplasmic reticulum or any other intracellular pool could potentially lead to endoplasmic reticulum stress with growth arrest/apoptosis as a consequence. This needs to be further explored in our system.

Changes in cellular membrane fatty acid composition have been correlated with regulation of membrane-bound enzymes like acyl CoA:cholesterol acyltransferase (ACAT) as well several other cellular functions [47]. Both diacylglycerol acyltransferase (DGAT), which catalyses the final step in triglyceride synthesis, and ACAT, which is responsible for the synthesis of cholesteryl esters by transacylating long chain fatty acyl CoA to cholesterol, are located in endoplasmic reticulum [48–50]. Both cell lines showed reduced protein level of ACAT1 after DHA-treatment to almost the same extent after 48 h. However, the ACAT inhibitor did not affect cell survival after DHA treatment, and we therefore consider it unlikely to be involved in the PUFA-mediated growth inhibitory effects observed in SW480 and SW620 cells.

In conclusion, growth inhibition induced by DHA in SW480 and SW620 does not seem to be related to lipid peroxidation or deficiencies in the antioxidant defence mechanisms. DHA did not induce apoptosis under the experimental conditions used in this study, but a relative growth arrest in G₂/M phase was observed. Electron microscopy revealed extensive lipid droplet accumulation mainly in the form of cholesteryl esters in SW620 cells as opposed to triglycerides in SW480 cells. The 190- and 170-fold increase in the amount of DHA in the cholesteryl ester fraction of SW620 and in the triglyceride fraction of SW480, respectively, indicate that these cells exhibit different protective mechanisms for processing potentially damaging molecules like DHA. This is, to our knowledge, the first report demonstrating large and cell-specific increases in cellular cholesteryl esters in tumour cells after treatment with long-chain n-3 PUFAs. Although further studies are needed to reveal the molecular mechanisms by which n-3 PUFAs reduce cell growth in these two cell lines, our results demonstrate that downregulation of nSREBP1 may be a possible contributor to the growth inhibitory effect of DHA.

Experimental procedures

Chemicals

5,8,11,14-Eicosatetraenoic acid (AA), 4,7,10,13,16,19-docosahexaenoic acid (DHA), 5,8,11,14,17-eicosapentaenoic acid (EPA), 9-octadecanoic acid (OA) and N-[2-(cyclohexyloxy)-4-nitrophenyl]-methanesulfonamide (NS-398) were obtained from Cayman Chemical (Ann Arbor, MI) as solutions in ethanol. 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT), 2-thiobarbituric acid (TBA), 1,1,3,3-tetramethoxypropane, (+/–) α-tocopherol, NADPH, glutathione (GSH) reduced form, 5,5′-dithio-bis(2-nitrobenzoic acid), sodium selenite, GSH-reductase (type IV) from Bakers yeast, bovine serum albumin (BSA), nordihydroguaiaretic acid (NDGA), propidium iodide (PI) and Sandoz 58-035 were obtained from Sigma Chemical Co. (St. Louis, MO). Growth media and additives were purchased from Gibco BRL, Life Technologies (Inchinnan, UK). Complete® protease inhibitor cocktail tablets were obtained from Roche Diagnostics (Mannheim, Germany). Molecular mass markers: SeeBlue Plus 2 and Magic Mark were obtained from (Invitrogen™, Life Technologies) and Restore™ Western Blot Stripping Buffer from Pierce Biotechnology (Rockford, IL).

Cell cultures

Human colon adenocarcinoma cell lines, SW480 and SW620, were obtained from The American Type Culture Collection (ATCC) (Rockville, MD). Both cell lines were cultured in Leibovitz's L-15 medium with l-glutamine supplemented with fetal bovine serum (FBS; 10%) and gentamicin (45 mg·L⁻¹). Both cell lines were maintained in a humidified atmosphere of 5% CO₂, 95% air at 37 °C.

Fatty acid supplementation and MTT survival

Stock solutions of fatty acids in ethanol were stored at −20 °C and further diluted in complete growth medium with 10% FBS before experiments, such that the final concentration of ethanol was < 0.025% (v/v). The growth inhibitory effect of AA, EPA, DHA and OA was determined by plating 3 × 10³ cells per well in quadruplicate in 96-well microtitre plates. Cell cultures were incubated at 37 °C in 5% CO₂ atmosphere for 4 h before changing to medium containing the fatty acids. Cultures were supplemented with 35, 50 and 70 µm fatty acids in growth medium and incubated from 1 to 6 days. Medium and additives were changed after 3 days. The growth of the cells was assessed by the MTT-reduction assay essentially as described [10,52]. After the incubation period, MTT (5 mg·mL⁻¹ in NaCl/P_i) was diluted 1:11 in growth medium and added to each well (100 µL) after removing growth media, and the plates were further incubated at 37 °C for 4 h in 5% CO₂. Subsequently, 50 µL of medium were removed from each well and 100 µL of 2-propanol with HCl (0.04 m) were added to solubilize the MTT-formazan. The plate was placed on a mechanical shaker for 20–60 min at room temperature for complete solubilization. The optical density of each well was read on a Titertek Multiscan Plus Reader using a 588 nm wavelength filter.

Analysis of apoptosis and cell cycle by flow cytometry

Cells were seeded at a concentration of 1.5 × 10⁶ in 75 cm² flasks. After 4 h, medium was replaced with medium supplemented with DHA (70 µm) or control medium and further incubated for 96 and 120 h. Medium was changed after 96 h and floating cells were reallocated to their respective culture flasks. Cells were harvested by trypsination and resuspended in NaCl/P_i together with floating cells collected from medium. Cells were then fixed in 1% paraformaldehyde on ice for 15 min, washed twice in NaCl/P_i and then permeabilized and stored in 70% (v/v) ethanol at −20 °C. Detection of apoptosis was performed using the APO-BRDU™ Kit (PharMingen, San Diego, CA), a two-colour staining method for labelling DNA breaks and total cellular DNA (TUNEL). Staining of cells was performed according to the manufacturer's protocol. Briefly, 1 × 10⁶ cells were washed twice and resuspended in DNA-labelling buffer containing TdT-enzyme and Br-dUTP and then incubated 1 h at 37 °C. After washing, cells were stained with fluorescein-conjugated anti-BrdU for 30 min at room temperature and then with PI in a RNase A solution for another 30 min. Samples were analysed on a Coulter Epics Elite ESP flow cytometer (Beckman-Coulter, Hialeah, FL) using a 15 mW Ar laser (488 nm), a 550 nm dichronic longpass filter and a 525/30 nm band pass filter detecting fluorescein (log scale) and a 675/20 nm band pass filter detecting PI (peak and linear scale). According to the dot plots of forward scatter versus side scatter and PI linear versus PI peak, debris and aggregates could be identified and excluded from the analysis. Total counts per sample were 10 000 cells.

Cell-cycle analysis was performed after staining cells with PI alone. 1 × 10⁶ cells stored in 70% ethanol at −20 °C (described above) were washed twice in NaCl/P_i, treated with 100 µL RNase A (100 µg·mL⁻¹) for 30 min at 37 °C and then stained with 400 µL PI (62.5 µg·mL⁻¹) for another 30 min at 37 °C. Samples were analysed by flow cytometry as described above, except using a 640 nm dichronic filter and a 610 nm band pass filter detecting PI peak and PI linear signals. On the basis of the dot plot PI linear versus PI peak, aggregates were identified and excluded from the analysis (‘doublet discrimination’). Total counts were 10 000 cells. Cell-cycle distribution of the generated DNA histograms was analysed using the software multicycle for Windows version 3.0.

Effect of vitamin E, ACAT inhibitor and inhibitors of eicosanoid synthesis on DHA-induced toxicity

Stock solutions of vitamin E and NDGA were prepared in ethanol and Sandoz 58-035 was dissolved in DMSO. Final concentration of ethanol and DMSO never exceeded 0.4% (v/v) and 0.025% (v/v), respectively, which had no effect on cell survival alone. We seeded 3 × 10³ cells per well in quadruplicate in 96-well microtitre plates. Cell cultures were incubated at 37 °C in 5% CO₂ atmosphere for 4 h before changing to medium containing DHA (70 µm) alone or in combination with vitamin E (10 and 50 µm), NS-398 (cyclooxygenase inhibitor, 0.01–10 µm), NDGA (lipoxygenase inhibitor, 0.01–1 µm) or Sandoz 58-035 (2.1–8.5 µm). Cell survival was measured after 72 h or by time points indicated by the MTT assay as described [10,52].

Analysis of MDA by HPLC

We seeded 1.5 × 10⁶ cells in 75 cm² cell culture dishes and after 4 h supplemented with DHA (70 µm) alone or in combination with vitamin E (50 µm). After 72 h of treatment the medium was collected, and the cells were washed twice (NaCl/P_i) before harvesting by scraping. The cells were centrifuged (5 min, 800 g), washed twice in NaCl/P_i and resuspended in 0.8 mL NaCl (0.9%, w/v) at 4 °C. Aliquots were taken for protein analysis by the Bio-Rad assay with bovine serum albumin (BSA) as standard [53]. The cells were lysed and protein precipitated with 200 µL trichloroacetic acid (20% w/v). After 10 min on ice, 250 µL thiobarbituric acid (TBA; 0.67%) was added to the rest of the cell suspension and the mixture heated at 100 °C for 20 min. TBA (0.67%, 4 mL) and trichloroacetic acid (20%, 2 mL) were added to the medium collected (4 mL) and the mixture treated as described for cell suspension. Thiobarbituric acid reactive substances (TBARS) were measured using reduced reagent volumes to increase sensitivity [26,51,54]. Estimation of MDA was performed using a HPLC system consisting of a Hewlett Packard (Avondale, PA) 1050 gradient pump equipped with an automatic injector, a 1050 Hewlett Packard diode-array absorption detector (532 nm) and a computer using chem station from Hewlett Packard. Aliquots of the TBARS samples were injected on a 5 µm Supelcosil LC-18 reversed-phase column (30 cm × 4.6 mm). The mobile phase consisted of 15% methanol in double-distilled water degassed by filtering through a 0.5 µm filter (Millipore, Bedford, MA). The flow rate was 1 mL·min⁻¹. MDA–TBA external standards were prepared using 1,1,3,3,tetramethoxypropane (0–10 µm). The absorption spectra of standards and samples were identical with a characteristic peak at 532 nm. Measurements were expressed as nmol MDA per mg protein.

GSH assay

Cells were seeded at a concentration of 1.5 × 10⁶ in cell culture flasks (75 cm²). After 4 h, medium was replaced with complete control medium or complete medium supplemented with DHA (70 µm). At the end of the incubation period (48 h), the cell layer was rinsed twice with cold NaCl/P_i, harvested by scraping and then sonicated. Total GSH content was assayed using a glutathione reductase assay [55]. The GSH content was quantitated by comparison with a standard curve generated with known amounts of GSH (reduced form) and expressed as nmol per mg protein. Protein concentrations were determined by the Bio-Rad-assay using BSA as standard.

GSH-Px enzyme assay

Cells were seeded at a concentration of 3 × 10⁶ in cell culture flasks (175 cm²). After 4 h, cells were preincubated with fresh medium supplemented with/without Na₂SeO₃ (250 nm) for 20 h. The medium was then replaced with a medium containing DHA (70 µm) or control medium and the cells were further incubated for 48 h. Cells were washed, harvested by scraping and then sonicated. The resulting homogenates were used as source for enzyme measurements. GSH-Px activity was measured by the Colorimetric Assay for Cellular Glutahione Peroxidase (BIOXYTECH GPx-340) according to the manufacturer's protocol (OXIS International, Inc., Portland, OR). The method is based on monitoring the oxidation of NADPH at 340 nm (Hitachi U-2000 spectrophotometer), and enzyme activity was calculated using a molar extinction coefficient of 6220 m⁻¹·cm⁻¹ for NADPH. The rate of decrease in the A₃₄₀ is directly proportional to the GSH-Px activity in the sample.

Electron microscopy

Cells were seeded at a concentration of 1.5 × 10⁶ in cell culture flasks (75 cm²). After 4 h, medium was replaced with medium containing DHA (70 µm) or control medium and the cells were further incubated for 72 h. Cells were washed twice in NaCl/P_i and harvested by trypsination. After centrifugation, the cell pellets were fixed for 12 h at 4 °C in 2% glutaraldehyde in 0.1 m phosphate buffer (pH 7.2) and postfixed for 1 h in 2% OsO₄ in 0.1 m phosphate buffer (pH 7.2). After fixation, the pellets were rinsed with 0.1 m phosphate buffer (pH 7.2), dehydrated through graded ethanol, embedded in Epon and examined in a JEOL 100CX electron microscope after contrasting with uranyl acetate and lead citrate.

Fatty acid analysis

Total lipids were extracted from cells treated or not with DHA (70 µm) for 48 h according to a method modified after Bligh & Dyer [56] and separated into free fatty acids, phospholipids, cholesteryl esters, triglycerides, cholesterol, diglycerides and monoglycerides on Bond Elut aminopropyl columns (Varian SPP., Harbor City, CA) as described by Kaluzny et al. [57] after adding nonadecanoic acid, dinonadecanoin, trinonadecanoin, cholesteryl nonadecanoate (Nu-Chek-Prep, Inc., Elysian, MN), monononadecanoin and dinonadecanoyl glycero-3-phosphatidylcholine (Larodan AB, Malmö, Sweden) as internal standards for quantification. Lipid extracts were transmethylated with BF₃/methanol at 135 °C for 35 min, and fatty acid methyl esters (FAMEs) were extracted into isooctane. GC analysis was performed using a Hewlett Packard 6890 GC (Wilmington, DE) equipped with a flame ionization detector. One microlitre lipid sample in isooctane was injected using a splitless injection technique onto a 25 m × 250.0 µm I.D. capillary column with 0.2 µm film thickness of the polyethylene glycol stationary phase (Chrompack CP-Wax 52CB) (Varian Inc., Walnut Creek, CA) with helium as carrier gas. The FAMEs were separated with a temperature program starting at 90 °C for 5 min, then it was increased 30 °C·min⁻¹ to 165 °C, 1 °C·min⁻¹ to 180 °C and finally 4 °C·min⁻¹ to 225 °C. Initial pressure at 40 psi was held for 4 min, whereafter it was changed at 10 psi·min⁻¹ to 20 psi, which was kept for the rest of the run time. Total run time was 33.75 min. Instrument response was calibrated using a mixture of C16:0, C18:0, C20:0 and C22:0 to relate relative peak areas of these fatty acid methyl esters to each other. Chromatographic purity of eluted lipid fractions was checked by using Silica gel 60F₂₅₄ aluminium sheets (Merck K GaA, Darmstadt, Germany). At least 2 µg lipid was spotted versus known standards to assess contamination of each fraction with other lipids. Plates were developed vertically in a solvent system of hexane/diethylether/acetic acid 40:60:1 (v/v/v). Plates were sprayed with phosphomolybdic acid solution and placed on a heater with gradual heating to 180 °C for visualization. Protein analysis was performed by the Bio-Rad assay as described earlier.

Quantitative real-time RT-PCR analysis

Cells were treated with DHA (70 µm) for different time periods and total RNA was isolated using the High Pure RNA Isolation Kit (Roche, Mannheim, Germany) according to the instruction manual. Total RNA was added the RNase inhibitor rRNasin (40 u·µL⁻¹, 1 µL) (Promega, Madison, WI). The RNA concentration and quality was determined by measuring the absorbance at 260 and 280 nm. RNA was isolated from three independent biological replicates.

Real-time RT-PCR reactions were performed using the LightCycler System 2.0 equipped with lightcycler 4.0 (Roche Diagnostics) in a total volume of 20 µL containing Mn(OAc)₂ (3.25 mm), primers (0.3 µm each for DGAT1, otherwise 0.5 µm each), probes (0.4 µm each for ACAT1, otherwise 0.2 µm each), LightCycler RNA Master Hybridization Probes (7.5 µL) (Roche Diagnostics), RNA (10 ng·µL⁻¹, 5 µL) and dH₂O for volume adjustment. Water instead of RNA was used for negative controls. ACAT1 and DGAT1/DGAT2 were normalized against human 5-aminolevulinate synthase (h-ALAS) and human porphobilinogen deaminase (h-PBGD), respectively. Standard curve construction was performed using lightcycler h-ALAS and h-PBGD Housekeeping Gene Set (Roche Diagnostics). Different concentrations of h-ALAS and h-PBGD standard templates (5 × 10² to 5 × 10⁶ copies) as well as target templates (0.5–250 ng) were used in triplicate to calculate the standard curves. When performing relative quantification, duplicates of the calibrators, duplicates of each sample and one negative control were included. The nucleotide sequences of primers and probes are shown in Table 7.

Table 7.   Sequences of primers and probes used in this study.

Gene	Sequence	Product size (bp)
ACAT1
Primers	(+)ATGGTATGCACGTCGG
	(–)CAGGGAGCTACCCAATC	141
Probes	ATCCCACATTTTTGGATTATGTCCGG-F
	LC Red640-ACGTTCCTGGACTTGTCGT-P
DGAT1
Primers	(+)ATGAAGCCCTTCAAGGACAT
	(–)AGTAGGTGACAGACTCGGAGTT	196
Probes	ATCTGGCTCATCTTCTTCTACTGGCTCT-F
	LC Red640-CCACTCCTGCCTGAATGCCGT-P
DGAT2
Primers	(+)GAAAAGCAGCTACAGGTCATCTC
	(–)TCAGCAGGTTGTGTGTCTTC	262
Probes	TGTTCCAGTCAAACACCAGCCAAGTGAAGTA-F
	LC Red640-AGCACAGCGATGAGCCAGCAATCA-P

The thermal cycling conditions were (a) initial reverse transcription at 61 °C for 20 min, (b) denaturation at 95 °C for 30 s, and (c) amplification (ACAT1): three cycles each of 2 s at 95 °C, 15 s at 60 °C and 18 s at 72 °C; three cycles each of 2 s at 95 °C, 15 s at 58 °C and 18 s at 72 °C, followed by 35 cycles each of 2 s at 95 °C, 15 s at 55 °C and 18 s at 72 °C; (DGAT1): four cycles each of 1 s at 95 °C, 15 s at 65 °C and 15 s at 72 °C; four cycles each of 1 s at 95 °C, 15 s at 62 °C and 15 s at 72 °C, followed by 40 cycles each of 1 s at 95 °C, 15 s at 60 °C and 15 s at 72 °C; (DGAT2): 45 cycles each of 1 s at 95 °C, 15 s at 60 °C and 15 s at 72 °C. The temperature transition rate was 20 °C·s⁻¹ during the reverse transcription, denaturation and amplification steps.

Preparation of whole-cell extracts and western analysis

Cells were treated with DHA (70 µm) for different time periods (3, 6, 12, 24 and 48 h) as described previously. At the end of the incubation period, the cells were rinsed twice in cold NaCl/P_i, scraped and pelleted at 805 g (4 °C). Cell pellets were flash frozen in liquid nitrogen and stored at −80 °C until cell lysis.

Whole-cell extracts were prepared essentially as described by Tanaka et al. [58]. Briefly, cell pellets were resuspended at 1× packed cell volume in buffer I (10 mm Tris/HCl, pH 8.0, 200 mm KCl) and 1× packed cell volume of buffer II (10 mm Tris/HCl pH 8.0, 200 mm KCl, 2 mm EDTA, 40% v/v glycerol, 0.5% NP-40, 2 mm dithiothreitol, Complete® protease inhibitor). The mixture was rocked at 4 °C for 2 h and cell debris was pelleted at 16 100 g at 4 °C for 10 min. The supernatant was recovered and protein concentration measured using the Bio-Rad Protein Assay (Bio-Rad, Hercules, CA). Extracts were snap frozen in liquid nitrogen and stored in small aliquots at −80 °C.

Equal amounts of whole-cell proteins were separated on precast 10% denaturing NuPAGE gels (Invitrogen™, Life Technologies) and transferred to PVDF membranes (Immobilon™, Millipore). Primary antibodies were diluted in 5% fat-free dry milk in NaCl/P_i containing 0.1% Tween®-20. Membranes were incubated for 1 h with antibodies for SREBP1 (sc-367) and ACAT1 (sc-20951) (Santa Cruz Biotechnology, Santa Crux, CA) followed by incubation for 1 h with horseradish peroxidase-conjugated polyclonal swine anti-rabbit immunoglobins (P0399) (DAKO, Denmark). Membranes were reprobed with beta-Actin (ab6276–100) (Abcam, Cambridge, UK) as a lane loading control. For ACAT1, the membrane was stripped before reprobing with actin. Membranes were treated with Super Signal West Femto Maximum Sensitivity Substrate (Pierce) and bands were visualized by Kodak Image Station 2000R. Immunoreactive bands were quantitated with Kodak molecular imaging software v 4.0.0.

Acknowledgements

The project was financed by The Norwegian Women's Public Health Association, The Cancer Research Foundation, Trondheim University Hospital, Peter Möller Department of Orkla ASA, Oslo and the Norwegian Research Council. We especially appreciate the support and help from Bruno Monterotti with cell culture.