Eicosapentaenoic acid (EPA), an n-3 polyunsaturated fatty acid that is abundant in the fish-based diets of populations that exhibit a remarkably low incidence of cancer, exerts anticancer activity in vitro and in animal models of experimental cancer. Here we define the molecular basis for the anticancer effects of EPA. EPA inhibits cell division by inhibiting translation initiation. This is a consequence of the ability of EPA to release Ca2+ from intracellular stores while inhibiting their refilling via capacitative Ca2+ influx that results in partial emptying of intracellular Ca2+ stores and thereby activation of protein kinase R. Protein kinase R phosphorylates and inhibits eukaryotic initiation factor 2α, resulting in inhibition of protein synthesis at the level of translation initiation, preferentially reducing the synthesis and expression of growth-regulatory proteins, including G1 cyclins, and causes cell cycle arrest in G1. In a KLN-205 squamous cell carcinoma mouse model, daily oral administration of EPA resulted in a significant reduction of tumor size and expression of cyclin D1 in the tumor tissues. Furthermore, EPA-treated tumors showed a significant increase in the proportion of diploid cells, indicative of cell cycle arrest in G0-G1, and a significant reduction of malignant hypertetraploid cells. These results characterize EPA as a member of an emerging new class of anticancer compounds that inhibit translation initiation.

The n-3 PUFA3EPA inhibits the proliferation of cancer cells in vitro(1, 2, 3) and the growth and metastatic potential of tumors in animal models of experimental cancer (4, 5, 6). Although alterations of eicosanoid formation and free radical-mediated lipid peroxidation have been mentioned as possible explanations for these observations (7, 8), the molecular basis for the pharmacological anticancer activity of EPA has not been elucidated. Epidemiological evidence suggests that populations consuming diets rich in fish oils containing EPA have a remarkably low incidence of cancer in general and of breast and prostate cancer in particular(9, 10, 11). Defining this mechanism of anticancer activity of EPA is important for its potential impact not only on cancer therapy but also on cancer prevention. In this report, we define the molecular basis for the anticancer activity of EPA.

Recent work from our laboratory has established a distinct connection between partial depletion of intracellular Ca2+stores and inhibition of both cell proliferation and tumor growth(12). Depletion of Ca2+ stores activates IFN-inducible PKR, which phosphorylates and thereby inhibits the translation initiation factor eIF2, resulting in inhibition of translation initiation, reduced synthesis and expression of G1 cyclins, and cell cycle arrest in G1(13). Indeed, regulation of translation initiation plays a critical role in the control of cell growth and division(14) because translation of most growth-regulatory proteins and oncogenes is highly inefficient because of the presence of stable secondary structures in the 5′ UTR of their mRNAs (15, 16). The length and secondary structure of the 5′ UTR of a mRNA molecule is the most critical feature influencing its translation efficiency. A moderately long, unstructured 5′ UTR with low content of G and C bases seems to be optimal to ensure high translational efficiency. In contrast, long 5′ UTRs with a high GC content are a major barrier to translation because G and C bases tend to form bonds that stabilize the secondary structure of the 5′ UTR. Sequence analyses of a large number of vertebrate cDNAs have shown that mRNAs with complex, highly structured 5′ UTRs include a disproportional high number of proto-oncogenes such as the G1 cyclins, transcription and growth factors, cytokines, and other proteins critical in cell growth regulation. In contrast, mRNAs that encode housekeeping proteins rarely have highly structured GC-rich 5′ UTRs (14, 15, 17). These structural differences in the 5′ UTR of their mRNA explain why translation of growth-regulatory proteins including cyclins but not of housekeeping proteins is highly dependent on the activity of translation initiation factors such as eIF2 and eIF4E. The critical role of translation initiation in cell growth regulation is illustrated by the finding that activating mutations of some translation initiation factors cause malignant transformation. For example, malignant transformation of cells can be induced by the expression of a dominant-negative form of PKR, which regulates eIF2 activity(18) by a constitutively active, noninhibitable mutant of eIF2 (eIF2-α51A; Refs. 19 and 20) or by the overexpression of eIF4E (21). In contrast, repression of eIF4E activity by overexpression of eIF4E binding proteins inhibits cell growth and can reverse a malignant phenotype(22, 23, 24). Furthermore, compounds such as clotrimazole that activate PKR and phosphorylate eIF2 by causing partial depletion of intracellular Ca2+ stores are potent inhibitors of cell growth (13) and tumor growth in experimental cancer models (12).

EPA is likely to induce partial depletion of intracellular Ca2+ stores because it releases Ca2+ from IP3-sensitive Ca2+ pools (25) and inhibits store-dependent capacitative Ca2+ influx(26). This suggested to us that EPA might inhibit translation initiation and thereby inhibit synthesis and expression of cell growth-regulatory proteins. We report here that by releasing Ca2+ from intracellular stores while preventing their refilling via capacitative Ca2+ influx, EPA causes PKR-mediated phosphorylation of eIF2α, inhibition of translation initiation, and preferential inhibition of synthesis and expression of G1 cyclins with consequent cell cycle arrest in G1. In a mouse KLN-205 squamous cell carcinoma model, oral EPA administration significantly reduced tumor growth and the expression of cyclin D1 in tumor cells while causing cell cycle arrest in G0-G1. This work defines EPA as an inhibitor of translation initiation, a new emerging class of drugs for cancer therapy.

Materials.

EPA (Sigma Chemical Co., St. Louis, MO), 99% pure, was diluted in DMSO to a concentration of 60 mm, and aliquots were blanketed with nitrogen and stored at −80°C. ω-3 PUFA concentrate containing 300 mg of EPA, 200 mg of docosahexaenoic acid, and 1 IU vitamin E/gelatin capsule.

Cell Culture and Transfection.

NIH 3T3 cells were cultured in DMEM/10% calf serum/penicillin/streptomycin, except during the mitogenic assays and in experiments for synthesis and expression of G1 cyclins with EPA treatment, when the cells were in DMEM/0.1% calf serum/bFGF(5 ng/ml). NIH 3T3 cells were transfected with pBABE (vector alone),cotransfected pBABE either with eIF2α-51A or dominant-negative PKR(PKR-K296), and puromycin-resistant stable colonies were constructed and maintained in DMEM/10% calf serum with 2.5 μg/ml puromycin, as described (13). Human solid tumor-based cell lines obtained from American Type Culture Collection (Rockville, MD)were grown in RPMI 1640 with 5% FBS.

Cell Growth Assay.

Adherent human solid tumor cells plated in 96-well plates were treated for 5 days in the presence and absence of various doses of EPA ranging from 20 to 100 μm in RPMI 1640 containing 5% FBS. At the time of termination, cells were stained with sulforhodamine B and counted as described (27).

Mitogenic Assay.

Synthesis of DNA by incorporation of[3H]thymidine was determined as described(12). Briefly, 4 × 103 NIH 3T3 cells, cells transfected with either vector, PKR-K296, or eIF2α-51A, were made quiescent by serum withdrawal for 36 h (0.1% calf serum/DMEM) and then exposed to a mitogenic stimulus (5 ng/ml bFGF) in the absence or presence of different concentrations of EPA. After 15 h, 1 μCi/ml[3H]thymidine (DuPont NEN, Boston, MA) was added, and the cells were incubated at 37°C in 5%CO2 for 5 h. Cells were harvested and counted in a Packard Top Count (Packard, CT).

Ca2+ Measurements.

Exponentially growing cells were harvested, washed with PBS, and resuspended in Krebs-Ringer medium buffered with 25 mmHEPES (pH 7.4 at 37°C). Cells were loaded in the same buffer with 5μ m Fura-2 AM (Molecular Probes, Eugene, OR), and 4 × 106 were used for cytosolic Ca2+ measurement with a Photon Technology International dual excitation spectrofluorometer (South Brunswick, NJ),exactly as described (28).

Polysome Profile Analysis.

Exponentially growing NIH 3T3 cells were exposed to EPA (30μ m) for 2 h in DMEM/5 ng/ml bFGF/0.1% calf serum,followed by addition of cycloheximide (25 μg/ml), and polysome profiles were determined exactly as described (13, 29).

Protein Synthesis.

Total protein synthesis was measured as described (13). Briefly, NIH 3T3 cells and stable transfected cells were treated with EPA for 1 h and pulse labeled with[35S]Met-Cys for 15 min, and incorporation of label into TCA-precipitated total protein determined.

Western and Northern Blot Analyses.

Protein extraction and Western blot analysis were done as described(30). Total protein extract (40 μg/lane) was analyzed on blots incubated overnight in specific antibodies (Santa Cruz Laboratories, Santa Cruz, CA). Horseradish peroxidase-conjugated secondary antibodies were detected by ECL kit (Pierce, Rockford, IL). RNA (25 μg) was fractionated on a 1% formaldehyde agarose gel,blotted onto nylon membranes, and hybridized with either cyclin D1 or 18S RNA cDNA probe labeled by random-prime labeling with[α-32P]dCTP as described (13).

Phosphorylation of eIf2-α.

Exponentially growing NIH 3T3, PKR-K296, eIF2α-51A, or vector control stable transfected cells were labeled with 32P-orthophosphoric acid (200 μCi/ml) for 3.5 h. One-half of the cells were challenged with EPA (30μ m) for 0.5 h, lysed with IP buffer, and RNase-DNase treated. TCA precipitable counts were determined, and equal number of counts were immunoprecipitated with anti-eIF2α antibody (kindly provided by Dr. N. Sonenberg, McGill University, Montreal, Quebec,Canada) and separated by SDS-PAGE; phosphorylation of eIF2α was quantified by PhosphorImager as described (13).

Cell Cycle Analysis.

NIH 3T3 cells (1 × 105) were treated with 7.5 μm EPA for up to 48 h in DMEM with 2% calf serum and 5 ng/ml bFGF, changed each day. The cells were fixed with ethanol and stained with propidium iodide for cell cycle analysis by flow cytometry.

Animal Studies.

To establish the KLN-205 squamous cell carcinoma model, KLN-205 tumor cells were grown in culture, harvested and injected as a suspension(2.5 × 105 cells in 0.1 ml of PBS) s.c. into 6-week-old female DBA/2J mice (Jackson Laboratories). Four days after tumor implantation, mice received either vehicle or EPA(2.5 g EPA/kg body weight) in the form of a lipid concentrate by gavage daily for 5 weeks. Three-dimensional tumor size was measured by calipers every week, and volumes were calculated using the equation: v (in mm3)= length × width × height. Statistical differences between tumor volumes in control and the EPA-treated group was determined by Student’s t test. The general health of the animals was monitored during the treatment period. At the end of the study, animals were sacrificed by CO2 asphyxiation, and tumors were excised and fixed in 10% buffered formalin.

Expression of Cyclin D1 in Tumor Tissue.

To analyze for cyclin D1 expression, 4-μm sections of formalin-fixed,paraffin-embedded tumor tissues from all of the animals in each group were cut and fixed on slides. Sections were deparaffinized in graded alcohol, immunostained with anti-cyclin D1 monoclonal antibody (Santa Cruz Biotechnology, Santa Cruz, CA), and counterstained with hematoxylin. A minimum of three different areas from each slide per animal were quantitated for cyclin D1 using BioQuant software. The percent labeling index for cyclin D1 was calculated by the formula:(cyclin D1 positive nuclei/total nuclei) × 100.

DNA Analysis for Ploidy Levels and Tumor Cell Kinetics.

Formalin-fixed, paraffin-embedded tumor sections were stained by the Feulgen method and microscopically analyzed for DNA content using the BioQuant system. The DNA content was measured in 350 tumor cell nuclei selected randomly from distinct areas of all slides from each treatment group. The proportion of cell populations with differing DNA content were estimated by peak fitting using Jandel software that identifies normally distributed populations in the data. The area percent under the fitted curves represents cells with differing ploidy content. A Feulgen-stained mouse liver sample was used to verify the 2C DNA content. A 2C content characterizes cells in G0-G1, a 2C-4C content characterizes cells in S phase, and a 4C content, cells in G2-M phase. Malignant tumors characteristically show a cell population containing hypertetraploid (≫4C) DNA content,considered an index of tumor malignancy (31, 32).

EPA Inhibits the Growth of Cancer Cells and DNA Synthesis in NIH 3T3 Cells in Vitro.

The effect of EPA on cell growth was assessed in a diverse panel of human cancer cells using the sulforhodamine B assay. Exposure of cancer cells to EPA for 5 days in a medium containing 5% FCS inhibited cell growth with a half maximal inhibition in the range of 20–100μ m (Table 1). In quiescent NIH 3T3 cells stimulated in DMEM/bFGF (5 ng/ml)/0.1%calf serum, EPA inhibited DNA synthesis, as measured by[3H]thymidine incorporation, with half-maximal inhibition at 7.5 μm (Fig. 1).

EPA Releases Intracellular Ca2+ and Blocks Capacitative Ca2+ Influx.

There is an established relationship between the filling state of the intracellular Ca2+ pools and cell proliferation(33, 34, 35). Indeed, we have demonstrated that compounds such as clotrimazole inhibit proliferation of normal and cancer cells in vitro and tumor growth in vivo because they induce sustained partial depletion of intracellular Ca2+ stores (12). Because EPA releases Ca2+ from internal stores(25) and at the same time inhibits their refilling via capacitative Ca2+ influx (26), we predicted that EPA would likely induce partial Ca2+ store depletion. This prediction was confirmed by the finding that in quiescent 3T3 cells, in which EPA inhibited bFGF-induced DNA synthesis (Fig. 1), EPA abolished the typical cytosolic Ca2+ transient elicited by bFGF(Fig. 2,a). This is because of EPA-mediated depletion of IP3-sensitive intracellular Ca2+ pools rather than blocking of bFGF signaling upstream of the IP3 receptor, because EPA:(a) potently released Ca2+ from intracellular stores (Fig. 2,b); (b) inhibited thapsigargin-induced capacitative Ca2+ influx(Fig. 2,c); and (c) did not affect bFGF-induced transcription of cyclin D1 (see below). To investigate further the relationship between the inhibitory effect of EPA on cancer cell growth and its effect on intracellular Ca2+ pools, human cancer cells were loaded with Fura-2 AM and challenged with EPA to assess cytosolic Ca2+ transients. Fig. 2,d shows that at a concentration of 30μ m, EPA induced rapid Ca2+ release in a representative EPA-sensitive cancer cell line (U118-MG; IC50 for growth inhibition, 36 μm) but not in a EPA-resistant cancer cell line (HTB-174; IC50 for growth inhibition, 80 μm). The inset to Fig. 2 d shows that in the HTB-174 cell line, 120μ m EPA induced Ca2+release in a manner comparable with that obtained in U118-MG cells with 30 μm. These results provide evidence for a direct link between the inhibitory effect of EPA on cancer cell growth and Ca2+ depletion.

Vitamin E reverses the antiproliferative activity of EPA, suggesting that the anticancer effect of EPA is mediated through lipid peroxidation (36, 37). To evaluate whether peroxidation products of EPA are responsible for EPA-induced Ca2+ release, experiments were performed in NIH 3T3 cells cultured in the presence of vitamin E (100 μm)prior to their loading with Fura-2 AM and exposure to EPA. EPA did not cause Ca2+ release in cells cultured with vitamin E (Fig. 2 b, lower curve). In contrast, pretreatment with vitamin E did not affect the Ca2+ release induced by clotrimazole or thapsigargin (data not shown) indicating that vitamin E does not affect the ability of the cellular Ca2+ stores to release Ca2+, and that peroxidation products of EPA induce Ca2+ release. Together, these results provide a direct evidence for a link between the effect of EPA on Ca2+ pools and its growth-inhibitory activity and indicate that these effects are related to a peroxidation product(s) of EPA on the filling state of the cellular Ca2+pools.

EPA Inhibits Protein Synthesis at the Level of Translation Initiation.

The filling state of the Ca2+ pools influences the rate of cellular protein synthesis (38). To investigate whether EPA inhibits protein synthesis, NIH 3T3 cells were treated with EPA and pulse labeled with[35S]Met-Cys, and we determined the incorporation of label into total protein. EPA inhibited protein synthesis (Fig. 3,a) in a dose-dependent manner (Fig. 3,a, inset). The rate of total protein synthesis as a function of EPA concentration reaches a minimum plateau at ∼33% of control (Fig. 3,a);this is most likely because EPA only inhibits synthesis of a subset of proteins whereas other proteins are not affected, as we will show below. Fig. 3 shows that the concentration of EPA required for half-maximal inhibition of protein synthesis is ≈15μ m, a value comparable with the concentration required for half-maximal inhibition of cell growth (Fig. 1). Similar protein synthesis inhibition was obtained with our cancer cell panel(data not shown). To investigate whether EPA inhibits protein synthesis as a direct result of Ca2+ store depletion,exponentially growing NIH 3T3 cells were transiently incubated with or without EPA for 1 h. The cells were washed with Ca2+-free medium to remove EPA and also to render the media Ca2+ free. Cells were then pulse-labeled with [35S]Met-Cys in the presence or absence of external Ca2+, and incorporation of label was determined as above. Cells transiently exposed to EPA reinitiated protein synthesis only when external Ca2+ was supplied; pulse labeling in the absence of external Ca2+ resulted in marked inhibition of protein synthesis. In contrast, removal of external Ca2+ during the pulse labeling period in control cells not exposed to EPA had negligible effect on protein synthesis. This experiment indicates that inhibition of protein synthesis by EPA is the direct result of Ca2+ store depletion(Fig. 3 b).

Depletion of intracellular Ca2+ stores activates PKR that in turn phosphorylates eIF2α and inhibits translation initiation, the rate-limiting step of protein synthesis(39, 40, 41). To investigate whether EPA inhibits protein synthesis at the level of translation initiation, we determined the polyribosome profiles in sucrose density fractions derived from exponentially growing NIH 3T3 cells incubated with EPA for 2 h. EPA shifted the ribosomal profile from heavy polyribosomes to lighter polysomes, monosomes, and free ribosomal subunits (Fig. 4). This shift toward lighter fractions clearly demonstrates that EPA inhibits translation initiation.

Translation in eukaryotic cells starts with the assembly of the eIF2.GTP-Met-tRNA ternary complex, recruitment of the 40S ribosomal subunit, followed by binding to mRNA at its 5′end. The ribosomal complex then migrates along the 5′ UTR of the mRNA in a process facilitated by an array of initiation factors including eIF2 and eIF4(42). At the end of each initiation event, GTP is hydrolyzed, and the eIF2.GDP complex is released. Regeneration of the eIF2.GTP complex by a GDP-GTP exchange reaction catalyzed by eIF2B is required to start a new round of translation initiation. Phosphorylation of serine 51 in the α subunit of eIF2 by PKR inhibits GDP-GTP exchange, suppressing the initiation of translation. To investigate whether the inhibitory effect of EPA on translation initiation is attributable to phosphorylation of eIF2α by PKR, cells expressing either a nonphosphorylatable mutant of eIF2α (eIF2α-51A;Ref. 19) or a dominant-negative mutant of PKR [PKR-K296;which has no enzymatic activity and acts as a dominant-negative mutant because of its ability to heterodimerize with endogenous protein(43)] were tested for their sensitivity to EPA. Compared with vector-transfected cells, eIF2α-51A and PKR-K296 transfected cells were resistant to the inhibitory action of EPA on both cell growth and protein synthesis (Fig. 5 and b). This observation does not reflect a general resistance of the eIF2α-51A and PKR-K296 transfected cell lines to antiproliferative agents because their sensitivity to etoposide was similar to that of vector-transfected control cells (data not shown). These results indicate that PKR-mediated phosphorylation of eIF2α is responsible for the inhibitory effect of EPA on protein synthesis and cell growth.

To confirm this interpretation, we measured phosphorylation of eIF2α in response to EPA treatment. Exponentially growing NIH 3T3 cells transfected with vector only, eIF2α-51A, or PKR K296 were incubated with [32P]Pi to label the intracellular ATP pool and challenged with EPA. Phosphorylation of eIF2α was determined by the amount of 32P incorporated into immunoprecipitated eIF2α. EPA treatment markedly increased eIF2α phosphorylation in both parental NIH 3T3 (not shown) and vector-transfected cells; in contrast,phosphorylation of eIF2α was minimal in eIF2α-51A and PKR-K296 transfected cell lines (Fig. 5 c). It has been reported recently that mouse embryonic fibroblasts that are nullizygous for PKR are not impaired in their ability to inhibit protein synthesis in response to endoplasmic reticulum stress (44). This is because the endoplasmic reticular-resident kinase PERK that is also activated by Ca2+ depletion phosphorylates eIF2α on serine 51 and thereby inhibits mRNA translation. Our results showing abrogation of eIF2α phosphorylation in PKR-K296 transfected cells treated with EPA indicate that PERK is not involved in the inhibition of protein synthesis that we observe in response to EPA. Alternatively, PKR-K296 may also block PERK activity and/or compete for a rate-limiting cofactor required for phosphorylation of eIF2α.

EPA Preferentially Abrogates G1 Cyclin Synthesis and Expression and Causes Cell Cycle Arrest in G1.

Reducing the rate of translation initiation preferentially inhibits the synthesis of proteins that are inefficiently translated (45, 46), mostly because they are coded for by mRNAs with highly structured 5′ UTR. Because the mRNAs of G1 cyclins contain highly structured 5′ UTRs, their synthesis and expression requires optimal activity of translation initiation factors (16). Thus,pharmacological agents that decrease the rate of translation initiation are expected to inhibit preferentially synthesis and expression of G1 cyclins and other cell growth-regulatory proteins (15). Exponentially growing NIH 3T3 cells were treated with EPA and[35S]Met-Cys for 1 h and assayed for incorporation of label into immunoprecipitated cyclin D1, cyclin E, and Ras, as well as into β-actin and ubiquitin. EPA inhibited synthesis of cyclin D1, cyclin E, and Ras, all coded for by mRNAs with highly structured 5′ UTRs, but not of β-actin or ubiquitin, proteins coded for by mRNAs with simple 5′ UTRs (Fig. 6,a). EPA treatment also resulted in a parallel reduction in the expression of cyclin D1 and cyclin E, whereas expression ofβ-actin and ubiquitin was not affected (Fig. 6,a). Because cyclin D1 has a short half-life of ∼30 min (47), the effect of EPA on incorporation of [35S]Met-Cys into cyclin D1 and on cyclin D1 expression could be attributable to reduced synthesis or increased degradation. In contrast, the effect of EPA on Ras is unlikely attributable to an increased degradation of this protein because it has a longer half-life that is comparable with that of ubiquitin or β-actin (>16 h). These results support the conclusion that EPA exerts a differential inhibitory effect on the synthesis and expression of a subset of proteins including G1 cyclins. This is further confirmed by the finding that in PKR-K296 and eIF2α-51A transfected cells that are resistant to its growth-inhibitory effect, EPA did not affect cyclin D1 or cyclin E expression (Fig. 6 c). EPA-induced inhibition of translation initiation may explain the reduced expression of farnesyl protein transferase, an enzyme essential for the transforming activity of ras oncoproteins in EPA-treated mice with colon cancer (48).

To determine whether the effect of EPA on synthesis and expression of cyclin D1 is primarily mediated through inhibition of translation initiation rather than of transcription, quiescent NIH 3T3 cells were stimulated with bFGF for 8 h and then treated with or without EPA for an additional hour. Cells were tested simultaneously for the synthesis of cyclin D1 protein and expression of cyclin D1 mRNA. EPA treatment strongly inhibited cyclin D1 synthesis without affecting the expression of its mRNA (Fig. 6,b). Furthermore, Northern blot analysis with a cyclin D1-specific probe of mRNAs isolated from polysome fractions (described in Fig. 4) showed a significant shift of cyclin D1 mRNA from heavier to lighter polysomal fractions (Fig. 6 d). These experiments confirm that EPA inhibits cyclin D1 synthesis at the level of translation initiation. Importantly, they also demonstrate that EPA does not affect cyclin D1 transcription,indicating that EPA inhibits cell growth without interfering with growth factor signaling upstream from cyclin D1 transcription.

In eukaryotic cells, successive expression of cyclin D1 and E during the G1 phase of the cell cycle is critical for progression beyond the restriction point and into S phase. To test whether EPA blocks cell cycle progression before the restriction point,quiescent NIH 3T3 cells were stimulated with bFGF, and the effect of adding EPA at hourly intervals on entry into S phase was monitored. Quiescent cells enter S phase 12–13 h after bFGF stimulation(13). EPA addition to these cells until late G1 (10–11 h after bFGF stimulation) prevented G1-S progression, even when cells were exposed to EPA for only 1 h before beginning of S phase. In contrast, the addition of EPA at later times failed to inhibit DNA synthesis (data not shown), indicating that EPA blocks cell proliferation specifically in G1, most probably before the restriction point. These results were further confirmed by FACS cell cycle analysis of propidium iodide-stained cells. Treatment with EPA (7.5μ m) for 48 h resulted in a significant accumulation of cells in G1: Control cells:G0-G1, 65%; S, 31%;G2-M, 4%; EPA-treated cells:G0-G1, 87%; S, 11%;G2-M, 2%. These results indicate that by reducing synthesis and expression of G1 cyclins, EPA blocks cell cycle progression in G1.

EPA Reduces Tumor Growth and Induces G0-G1Arrest in Vivo.

The in vivo consequences of EPA-mediated inhibition of translation initiation were elucidated in the KLN mouse squamous cell carcinoma model. Starting 4 days after inoculation, either vehicle or marine lipid concentrate containing EPA was administered daily by gavage at a dose of 2.5 g/kg for 5 weeks (n = 10 animals/group). Tumors were measured on a weekly basis for 5 weeks and then analyzed by immunohistochemistry for expression of cyclin D1 and by Feulgen stain for nuclear DNA content. EPA treatment significantly reduced tumor growth (Fig. 7,a) while showing no signs of toxicity, as indicated by the absence of weight loss or any behavioral abnormalities with treatment. Consistent with the effect of EPA on G1 cyclin expression in cultured cells, EPA-treated tumors showed a significant decrease in the number of cells expressing cyclin D1 (Fig. 7,b). EPA reduced the cyclin D1 labeling index from 29 ± 5 in vehicle-treated animals to 4 ± 1 (mean ± SE, P < 0.001). Nuclear DNA content analysis of tumors documented that EPA significantly increased the number of cells with a diploid (2C) DNA content, indicating cell cycle arrest in G0-G1. In addition, it significantly decreased the number of hypertetraploid cells (>4C DNA content), representative of the rapidly dividing malignant cell population (Fig. 8).

The results reported in this paper demonstrate that EPA and/or its peroxidation product(s) inhibits cell proliferation by causing partial but sustained depletion of intracellular Ca2+stores. As a result of Ca2+ store depletion, EPA causes cell cycle arrest in G1 because it inhibits translation initiation and thereby preferentially abrogates the synthesis and expression of G1 cyclins both in vitro and in vivo. This mechanism of the antiproliferative action of EPA closely resembles that of clotrimazole, another translation initiation inhibitor that shares with EPA potent antiproliferative effects in tumor cells in vitro and in tumor-bearing animals with very low toxicity, even after prolonged daily administration. Furthermore, the inhibitory effect of EPA on translation initiation and its preferential inhibitory action on the synthesis of growth-regulatory proteins provides a possible explanation for the epidemiological evidence that populations consuming food rich in EPA and other n-3 PUFAs have a low incidence of cancer.

We are grateful to Drs. Daniel C. Tosteson, Magdalena T. Tosteson, and Victor J. Dzau for continuous support. We are also grateful to Dr. M. Davis for kindly providing the PKR and eIF2 plasmids and to Dr. N. Sonenberg for providing the anti-eIF2 antibodies.

Fig. 1.

EPA inhibits DNA synthesis in NIH 3T3 cells. DNA synthesis was measured by [3H]thymidine incorporation in confluent,quiescent NIH 3T3 cells stimulated with DMEM/0.1% calf serum/bFGF (5 ng/ml) in the presence and absence of EPA. Bars,SE.

Fig. 1.

EPA inhibits DNA synthesis in NIH 3T3 cells. DNA synthesis was measured by [3H]thymidine incorporation in confluent,quiescent NIH 3T3 cells stimulated with DMEM/0.1% calf serum/bFGF (5 ng/ml) in the presence and absence of EPA. Bars,SE.

Close modal
Fig. 2.

EPA releases intracellular Ca2+, inhibits capacitative Ca2+ influx, and abrogates mitogen-induced Ca2+ transients. Fura-2 AM-loaded NIH 3T3(a–c) or cancer (d) cells were suspended in Ca2+-free medium or in medium containing 2 mm CaCl2 (c), and the cytosolic Ca2+ was monitored continuously. Calcium release was stimulated with 500 ng/ml bFGF after 5 min preincubation with or without 30 μm EPA (a),with 30 μm EPA (b and d),or with 300 nm thapsigargin (c). For the experiment shown in b, cells were incubated with or without 100 μm vitamin E for 24 h before testing the effect of EPA. Note that in a, the concentration of bFGF required to elicit cytosolic Ca2+ changes is 100-fold higher than that used for the mitogenic assay in Fig. 1because 100 times more cells were used for the Ca2+ measurements. d, inset,Ca2+ release after 120 μmEPA addition to HTB-174 cells.

Fig. 2.

EPA releases intracellular Ca2+, inhibits capacitative Ca2+ influx, and abrogates mitogen-induced Ca2+ transients. Fura-2 AM-loaded NIH 3T3(a–c) or cancer (d) cells were suspended in Ca2+-free medium or in medium containing 2 mm CaCl2 (c), and the cytosolic Ca2+ was monitored continuously. Calcium release was stimulated with 500 ng/ml bFGF after 5 min preincubation with or without 30 μm EPA (a),with 30 μm EPA (b and d),or with 300 nm thapsigargin (c). For the experiment shown in b, cells were incubated with or without 100 μm vitamin E for 24 h before testing the effect of EPA. Note that in a, the concentration of bFGF required to elicit cytosolic Ca2+ changes is 100-fold higher than that used for the mitogenic assay in Fig. 1because 100 times more cells were used for the Ca2+ measurements. d, inset,Ca2+ release after 120 μmEPA addition to HTB-174 cells.

Close modal
Fig. 3.

EPA inhibits protein synthesis. a,exponentially growing NIH 3T3 cells were incubated for 1 h at 37°C with increasing concentrations of EPA and pulse-labeled for 15 min with [35S]Met-Cys in the presence or absence of EPA. Gray area of the bars, minimum plateau of protein synthesis that is unaffected by EPA. Bars, SE. b,cells were incubated for 1 h without (columns 1 and 2) or with (columns 3 and 4) 30 μm EPA and washed with Met-Cys and Ca2+-free medium containing 20 μm EGTA to remove both EPA and Ca2+. Cells were then pulse-labeled with [35S]Met-Cys for 15 min in the absence(columns 2 and 4) or presence(columns 1 and 3) of 2 mmCa2+. TCA precipitable counts were measured in aliquots of cell lysates containing equal amounts of protein. Bars,SE.

Fig. 3.

EPA inhibits protein synthesis. a,exponentially growing NIH 3T3 cells were incubated for 1 h at 37°C with increasing concentrations of EPA and pulse-labeled for 15 min with [35S]Met-Cys in the presence or absence of EPA. Gray area of the bars, minimum plateau of protein synthesis that is unaffected by EPA. Bars, SE. b,cells were incubated for 1 h without (columns 1 and 2) or with (columns 3 and 4) 30 μm EPA and washed with Met-Cys and Ca2+-free medium containing 20 μm EGTA to remove both EPA and Ca2+. Cells were then pulse-labeled with [35S]Met-Cys for 15 min in the absence(columns 2 and 4) or presence(columns 1 and 3) of 2 mmCa2+. TCA precipitable counts were measured in aliquots of cell lysates containing equal amounts of protein. Bars,SE.

Close modal
Fig. 4.

EPA inhibits translation initiation. Exponentially growing 3T3 cells were exposed to 30 μm EPA (solid line) or vehicle (dotted line) for 2 h. Extracts were prepared, equal A260 nm units were separated by sucrose density gradient centrifugation, and the polysome profiles of the gradients were obtained.

Fig. 4.

EPA inhibits translation initiation. Exponentially growing 3T3 cells were exposed to 30 μm EPA (solid line) or vehicle (dotted line) for 2 h. Extracts were prepared, equal A260 nm units were separated by sucrose density gradient centrifugation, and the polysome profiles of the gradients were obtained.

Close modal
Fig. 5.

EPA causes phosphorylation of eIF2α, and dominant-negative PKR- and eIF2α-51A-expressing cell lines are resistant to EPA. a, DNA synthesis([3H]thymidine incorporation); and b, protein synthesis ([35S]Met-Cys incorporation) were measured in pBABE vector, dominant-negative PKR, or eIF2α51A stable transfected NIH3T3 cell lines in the presence of increasing concentrations of EPA. The apparent higher concentration of EPA required for inhibition of total protein than of DNA synthesis is the consequence of the preferential inhibitory effect of EPA on the synthesis of a subset of proteins. Bars, SE. c, effect of EPA on phosphorylation of eIF2α in transfected cell lines; pBABE vector,dominant-negative PKR, or eIF2α51A stable transfected NIH 3T3 cell lines labeled with [32P]Pi (200μCi/ml) for 3.5 h were treated with (+) or without (−) 30μ m EPA for 0.5 h, and an equal number of TCA precipitable counts were immunoprecipitated with anti-eIF2α antibody,separated by SDS-PAGE, and phosphorylation of eIF2α was quantified by PhosphorImager.

Fig. 5.

EPA causes phosphorylation of eIF2α, and dominant-negative PKR- and eIF2α-51A-expressing cell lines are resistant to EPA. a, DNA synthesis([3H]thymidine incorporation); and b, protein synthesis ([35S]Met-Cys incorporation) were measured in pBABE vector, dominant-negative PKR, or eIF2α51A stable transfected NIH3T3 cell lines in the presence of increasing concentrations of EPA. The apparent higher concentration of EPA required for inhibition of total protein than of DNA synthesis is the consequence of the preferential inhibitory effect of EPA on the synthesis of a subset of proteins. Bars, SE. c, effect of EPA on phosphorylation of eIF2α in transfected cell lines; pBABE vector,dominant-negative PKR, or eIF2α51A stable transfected NIH 3T3 cell lines labeled with [32P]Pi (200μCi/ml) for 3.5 h were treated with (+) or without (−) 30μ m EPA for 0.5 h, and an equal number of TCA precipitable counts were immunoprecipitated with anti-eIF2α antibody,separated by SDS-PAGE, and phosphorylation of eIF2α was quantified by PhosphorImager.

Close modal
Fig. 6.

EPA inhibits synthesis and expression of growth-regulatory proteins at the level of translation initiation. a, exponentially growing NIH 3T3 cells were pulsed with[35S]Met-Cys with or without 30 μm EPA for 1 h, and equal protein containing cell lysates immunoprecipitated with anti-cyclin D1, cyclin E, Ras, β-actin or ubiquitin antibodies. Immunocomplexes were separated by SDS-PAGE and visualized by PhosphorImager (upper panel) or by Western Blot(bottom panel). b, quiescent NIH 3T3 cells were stimulated with bFGF (5 ng/ml)/0.1% calf serum for 8 h and pulsed for 1 h with [35S]Met-Cys with or without 30 μm EPA, and samples were processed as above. Simultaneously, mRNAs from cells with similar treatments were also analyzed by Northern blot using a cyclin D1-specific probe using 18S mRNA as loading control. c, a similar protocol as in a was used to determine expression of cyclin D1 and cyclin E in pBABE vector and eIF2A and PKR-K296 transfected cells by Western blot analysis. d, Northern blot analysis of sucrose gradient fractions, as shown in Fig. 4.

Fig. 6.

EPA inhibits synthesis and expression of growth-regulatory proteins at the level of translation initiation. a, exponentially growing NIH 3T3 cells were pulsed with[35S]Met-Cys with or without 30 μm EPA for 1 h, and equal protein containing cell lysates immunoprecipitated with anti-cyclin D1, cyclin E, Ras, β-actin or ubiquitin antibodies. Immunocomplexes were separated by SDS-PAGE and visualized by PhosphorImager (upper panel) or by Western Blot(bottom panel). b, quiescent NIH 3T3 cells were stimulated with bFGF (5 ng/ml)/0.1% calf serum for 8 h and pulsed for 1 h with [35S]Met-Cys with or without 30 μm EPA, and samples were processed as above. Simultaneously, mRNAs from cells with similar treatments were also analyzed by Northern blot using a cyclin D1-specific probe using 18S mRNA as loading control. c, a similar protocol as in a was used to determine expression of cyclin D1 and cyclin E in pBABE vector and eIF2A and PKR-K296 transfected cells by Western blot analysis. d, Northern blot analysis of sucrose gradient fractions, as shown in Fig. 4.

Close modal
Fig. 7.

EPA inhibits tumor growth and cyclin D1 expression in vivo. a, KLN-205 squamous cell carcinoma cells were implanted s.c. in DB/2J mice. Treatment with either EPA (□, 2.5 g/kg body weight) or vehicle (▪) administered daily by gavage was started 4 days after tumor implant. Each point represents mean of tumor volume measured at the end of each week (n = 10; bars, SE; when not shown, SE is smaller than the size of the point); Ps were calculated using Student’s t test. b, tumor sections immunostained with monoclonal anti-cyclin D1 antibody and horseradish peroxidase secondary antibodies.

Fig. 7.

EPA inhibits tumor growth and cyclin D1 expression in vivo. a, KLN-205 squamous cell carcinoma cells were implanted s.c. in DB/2J mice. Treatment with either EPA (□, 2.5 g/kg body weight) or vehicle (▪) administered daily by gavage was started 4 days after tumor implant. Each point represents mean of tumor volume measured at the end of each week (n = 10; bars, SE; when not shown, SE is smaller than the size of the point); Ps were calculated using Student’s t test. b, tumor sections immunostained with monoclonal anti-cyclin D1 antibody and horseradish peroxidase secondary antibodies.

Close modal
Fig. 8.

EPA blocks cell cycle progression in G1 and reduces the population of hypertetraploid cells in tumor tissue. a, tumor sections from vehicle and EPA-treated mice were stained with Feulgen, and the absorbance of the nuclei was measured using BioQuant software. The normal ploidy distributions calculated after pooling of values obtained from at least one field from each animal from both groups are shown. The table indicates the area percent of 2C cells, representative of G0-G1 phase of the cell cycle, and of ≫4C cells representative of malignant hypertetraploid cells, in vehicle- or EPA-treated animals.

Fig. 8.

EPA blocks cell cycle progression in G1 and reduces the population of hypertetraploid cells in tumor tissue. a, tumor sections from vehicle and EPA-treated mice were stained with Feulgen, and the absorbance of the nuclei was measured using BioQuant software. The normal ploidy distributions calculated after pooling of values obtained from at least one field from each animal from both groups are shown. The table indicates the area percent of 2C cells, representative of G0-G1 phase of the cell cycle, and of ≫4C cells representative of malignant hypertetraploid cells, in vehicle- or EPA-treated animals.

Close modal

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported in part by NIH Grant CA 78411.

3

The abbreviations used are: PUFA,polyunsaturated fatty acid; EPA, eicosapentaenoic acid; PKR, protein kinase R; eIF2α, eukaryotic initiation factor 2α; UTR, untranslated region; IP3, inositol triphosphate; bFGF, basic fibroblast growth factor; TCA, trichloroacetic acid.

Table 1

Effect of EPA on human cancer cell growth

Cancer cellsTypeIC50m)
HeLa Cervix 23.5 ± 1.9 
Hep-G2 Liver 24.1 ± 1.1 
U-118 MG Glioblastoma 36.0 ± 3.2 
DU-145 Prostate 36.4 ± 2.3 
MMRU Skin 36.4 ± 1.8 
MDAMB231 Breast 37.7 ± 2.8 
MCF-7 Breast 43.3 ± 3.6 
SK-Mel-28 Skin 46.4 ± 2.9 
HCT-15 Colon 47.4 ± 1.8 
A549 Lung 54.0 ± 3.8 
ACHN Kidney 62.9 ± 3.9 
HT-29 Colon 72.8 ± 2.8 
Caski-1 Cervix 76.0 ± 1.7 
HTB-174 Lung 79.5 ± 4.1 
SK-OV3 Ovary 96.0 ± 2.1 
Cancer cellsTypeIC50m)
HeLa Cervix 23.5 ± 1.9 
Hep-G2 Liver 24.1 ± 1.1 
U-118 MG Glioblastoma 36.0 ± 3.2 
DU-145 Prostate 36.4 ± 2.3 
MMRU Skin 36.4 ± 1.8 
MDAMB231 Breast 37.7 ± 2.8 
MCF-7 Breast 43.3 ± 3.6 
SK-Mel-28 Skin 46.4 ± 2.9 
HCT-15 Colon 47.4 ± 1.8 
A549 Lung 54.0 ± 3.8 
ACHN Kidney 62.9 ± 3.9 
HT-29 Colon 72.8 ± 2.8 
Caski-1 Cervix 76.0 ± 1.7 
HTB-174 Lung 79.5 ± 4.1 
SK-OV3 Ovary 96.0 ± 2.1 
1
Falconer J. S., Ross J. A., Fearon K. C., Hawkins R. A., O’Riordain M. G., Carter D. C. Effect of eicosapentaenoic acid and other fatty acids on the growth in vitro of human pancreatic cancer cell lines.
Br. J. Cancer
,
69
:
826
-832,  
1994
.
2
Lai P. B., Ross J. A., Fearon K. C., Anderson J. D., Carter D. C. Cell cycle arrest and induction of apoptosis in pancreatic cancer cells exposed to eicosapentaenoic acid in vitro.
Br. J. Cancer
,
74
:
1375
-1383,  
1996
.
3
Finstad H. S., Drevon C. A., Kulseth M. A., Synstad A. V., Knudsen E., Kolset S. O. Cell proliferation, apoptosis and accumulation of lipid droplets in U937-1 cells incubated with eicosapentaenoic acid.
Biochem. J.
,
336
:
451
-459,  
1998
.
4
Ramesh G., Das U. N. Effect of free fatty acids on two-stage skin carcinogenesis in mice.
Cancer Lett.
,
100
:
199
-209,  
1996
.
5
Hudson E. A., Beck S. A., Tisdale M. J. Kinetics of the inhibition of tumour growth in mice by eicosapentaenoic acid-reversal by linoleic acid.
Biochem. Pharmacol.
,
45
:
2189
-2194,  
1993
.
6
Noguchi M., Minami M., Yagasaki R., Kinoshita K., Earashi M., Kitagawa H., Taniya T., Miyazaki I. Chemoprevention of DMBA-induced mammary carcinogenesis in rats by low-dose EPA and DHA.
Br. J. Cancer
,
75
:
348
-353,  
1997
.
7
Das U. N. Essential fatty acids and their metabolites and cancer.
Nutrition
,
15
:
239
-241,  
1999
.
8
Jiang W. G., Bryce R. P., Horrobin D. F. Essential fatty acids: molecular and cellular basis of their anti-cancer action and clinical implications.
Crit. Rev. Oncol. Hematol.
,
27
:
179
-209,  
1998
.
9
Eskimo diets and diseases. Lancet, 21: 1139–1141, 1983.
10
Kaizer L., Boyd N. F., Kriukov V., Tritchler D. Fish consumption and breast cancer risk: an ecological study.
Nutr. Cancer
,
12
:
61
-68,  
1989
.
11
Caygill C. P., Charlett A., Hill M. J. Fat, fish, fish oil and cancer.
Br. J. Cancer
,
74
:
159
-164,  
1996
.
12
Benzaquen L. R., Brugnara C., Byers H. R., Gattoni-Celli S., Halperin J. A. Clotrimazole inhibits cell proliferation in vitro and in vivo.
Nat. Med.
,
1
:
534
-540,  
1995
.
13
Aktas H., Flückiger R., Acosta J. A., Savage J., Palakurthi S. S., Halperin J. A. Depletion of intracellular Ca++ stores, phosphorylation of eIF2α and sustained inhibition of translation initiation mediate the anti-cancer effects of clotrimazole.
Proc. Natl. Acad. Sci. USA
,
95
:
8280
-8285,  
1998
.
14
Sonenberg N. mRNA translation: influence of the 5′ and 3′ untranslated regions.
Curr. Opin. Genet. Dev.
,
4
:
310
-315,  
1994
.
15
Kozak M. An analysis of vertebrate mRNA sequences: intimations of translational control.
J. Cell Biol.
,
115
:
887
-903,  
1991
.
16
Pelletier T., Sonenberg N. The involvement of mRNA secondary structure in protein synthesis.
Biochem. Cell Biol.
,
65
:
576
-581,  
1987
.
17
Rastinejad F., Conboy M. J., Rando T. A., Blau H. M. Tumor suppression by RNA from the 3′ untranslated region of α-tropomyosin.
Cell
,
75
:
1107
-1117,  
1993
.
18
Koromilas A. E., Roy S., Barber G. N., Katze M. G., Sonenberg N. Malignant transformation by a mutant of the IFN-inducible dsRNA-dependent protein kinase.
Science (Washington DC)
,
257
:
1685
-1689,  
1992
.
19
Donze O., Jagus R., Koromilas A. E., Hershey J. W., Sonenberg N. Abrogation of translation initiation factor eIF-2 phosphorylation causes malignant transformation of NIH 3T3 cells.
EMBO J.
,
14
:
3828
-3834,  
1995
.
20
Raught B., Gingras A-C., James A., Medina D., Sonenberg N., Rosen J. M. Expression of a translationally regulated, dominant-negative CCAAT/enhancer-binding protein β isoform and up-regulation of the eukaryotic translation initiation factor 2α are correlated with neoplastic transformation of mammary epithelial cells.
Cancer Res.
,
56
:
4382
-4386,  
1996
.
21
Lazaris-Karatzas A., Montine K. S., Sonenberg N. Malignant transformation by a eukaryotic initiation factor subunit that binds to mRNA 5′ ca.
p. Nature (Lond.)
,
345
:
544
-547,  
1990
.
22
Rousseau D., Gingras A. C., Pause A., Sonenberg N. The eIF4E-binding proteins 1 and 2 are negative regulators of cell growth.
Oncogene
,
13
:
2415
-2420,  
1996
.
23
Graff J. R., Boghaert E. R., De Benedetti A., Tudor D. L., Zimmer C. C., Chan S. K., Zimmer S. G. Reduction of translation initiation factor 4E decreases the malignancy of ras-transformed cloned rat embryo fibroblasts.
Int. J. Cancer
,
60
:
255
-263,  
1995
.
24
Sonenberg N. Regulation of translation and cell growth by eIF-4E.
Biochimie
,
76
:
839
-846,  
1994
.
25
Chow S. C., Jondal M. Polyunsaturated free fatty acids stimulate an increase in cytosolic Ca2+ by mobilizing the inositol 1,4,5-trisphosphate-sensitive Ca2+ pool in T cells through a mechanism independent of phosphoinositide turnover.
J. Biol. Chem.
,
265
:
902
-907,  
1990
.
26
Gamberucci A., Fulceri R., Benedetti A. Inhibition of store-dependent capacitative Ca2+ influx by unsaturated fatty acids.
Cell Calcium
,
21
:
375
-385,  
1997
.
27
Monks A., Scudiero D., Skehan P., Shoemaker R., Paull K., Vistica D., Hose C., Langley J., Cronise P., Vaigro-Wolff A., Gray-Goodrich M., Campbell H., Mayo J., Boyd M. Feasibility of a high-flux anticancer drug screen using a diverse panel of cultured human tumor cell lines.
J. Natl. Cancer Inst.
,
83
:
757
-766,  
1991
.
28
Clementi E., Scheer H., Zacchetti D., Fasolato C., Pozzan T., Meldolesi J. Receptor-activated Ca2+ influx.
J. Biol. Chem.
,
267
:
2164
-2172,  
1992
.
29
Rousseau D., Kaspar R., Rosenwald I., Gehrke L., Sonenberg N. Translation initiation of ornithine decarboxylase and nucleocytoplasmic transport of cyclin D1 mRNA are increased in cells overexpressing eukaryotic initiation factor 4E.
Proc. Natl. Acad. Sci. USA
,
93
:
1065
-1070,  
1996
.
30
Aktas H., Hong C., Cooper G. M. Ras links growth factor signaling to the cell cycle machinery via regulation of cyclin D1 and the Cdk inhibitor p27KIPI.
Mol. Cell. Biol.
,
17
:
3850
-3857,  
1997
.
31
Kiss R., Salmon I., Camby I., Gras S., Pasteels J-L. Characterization of factors in routine laboratory protocols that significantly influence the Feulgen reaction.
J. Histochem. Cytochem.
,
41
:
935
-945,  
1993
.
32
Salmon I., Kiss R. Relationship between proliferative activity and ploidy level in a series of 530 human brain tumors, including astrocytomas, meningiomas, schwannomas, and metastases.
Hum. Pathol.
,
24
:
329
-335,  
1993
.
33
Meldolesi J., Pozzan T. The endoplasmic reticulum Ca++ store: a view from the lumen.
Trends Biochem. Sci.
,
23
:
10
-14,  
1998
.
34
Berridge M. J. Capacitative calcium entry.
Biochem. J.
,
312
:
1
-11,  
1995
.
35
Gill D. L., Waldron R. T., Rys-Sikora K. E., Ufret-Vincenty C. A., Graber M. N., Favre C. J., Alfonso A. Calcium pools, calcium entry, and cell growth.
Biosci. Rep.
,
16
:
139
-157,  
1996
.
36
Sagar P. S., Das U. N. Cytotoxic action of cis unsaturated fatty acids on human cervical carcinoma (HeLa) cells in vitro.
Prostaglandins Leukotrienes Essent. Fatty Acids
,
53
:
287
-299,  
1995
.
37
Chajès V., Sattler W., Stranzl A., Kostner G. M. Influence of n-3 Fatty Acids on the Growth of Human Breast Cancer Cells in Vitro: Relationship to Peroxides and Vitamin E
200
-212, Kluwer Academic Publishers Dordrecht, the Netherlands  
1995
.
38
Brostrom C. O., Brostrom M. A. Calcium-dependent regulation of protein synthesis in intact mammalian cells.
Annu. Rev. Physiol.
,
52
:
577
-590,  
1990
.
39
Prostko C. R., Brostrom M. A., Malara E. M., Brostrom C. O. Phosphorylation of eukaryotic initiation factor (eIF) 2α and inhibition of eIF-2B in GH3 pituitary cells by perturbants of early protein processing that induce GRP78.
J. Biol. Chem.
,
267
:
16751
-16754,  
1992
.
40
Srivastava S. P., Davies M. V., Kaufman R. J. Calcium depletion from the endoplasmic reticulum activates the double-stranded RNA-dependent protein kinase (PKR) to inhibit protein synthesis.
J. Biol. Chem.
,
270
:
16619
-16624,  
1995
.
41
Brostrom C. O., Brostrom M. A. Regulation of translational initiation during cellular responses to stress.
Prog. Nucleic Acid Res. Mol. Biol.
,
58
:
79
-125,  
1998
.
42
Pain V. M. Initiation of protein synthesis in eukaryotic cells.
Eur. J. Biochem.
,
236
:
747
-771,  
1996
.
43
Sharp T. V., Xiao Q., Justesen J., Gewert D. R., Clemens M. J. Regulation of the interferon-inducible protein kinase PKR and (2′-5′) oligo (adenylate) synthetase by a catalytically inactive PKR mutant through competition for double-stranded RNA binding.
Eur. J. Biochem.
,
230
:
97
-103,  
1995
.
44
Harding H. P., Zhang Y., Ron D. Protein translation and folding are coupled by an endoplasmic-reticulum-resident kinase.
Nature (Lond.)
,
397
:
271
-174,  
1999
.
45
Lodish H. F. Model for the regulation of mRNA translation applied to haemoglobin synthesis.
Nature (Lond.)
,
251
:
385
-388,  
1974
.
46
Heinrich R., Rapoport T. A. Mathematical modelling of translation of mRNA in eucaryotes; steady states, time-dependent processes and application to reticulocytes.
J. Theor. Biol.
,
86
:
279
-313,  
1980
.
47
Lebwohl D. E., Muise-Helmericks R., Sepp-Lorenzino L., Serve S., Timaul M., Bol R., Borgen P., Rosen N. A truncated cyclin D1 gene encodes a stable mRNA in a human breast cell line.
Oncogene
,
9
:
1925
-1929,  
1994
.
48
Singh J., Hamid R., Reddy B. S. Dietary fish oil inhibits the expression of farnesyl protein transferase and colon tumor development in rodents.
Carcinogenesis (Lond.)
,
19
:
985
-989,  
1998
.