Although epidemiological data support a strong positive correlation between certain tumours and high-fat diets(Reference Byers and Graham1, Reference Willett and MacMahon2), it is still debated how a high consumption of dietary lipids affects carcinogenesis. Animal studies provide evidence that a high content of linoleic acid (18 : 2n-6) in the diet influences the development of carcinomas of the breast, colon and prostate, by acting at the promotional stage of carcinogenesis (for a review, see Carroll & Khor(Reference Carroll and Khor3)). The reproduction of experimental metastasis has also been shown to be favoured by 18 : 2n-6-rich diets(Reference Hubbard and Erickson4–Reference Mannini, Calzolari, Calorini, Mugnai and Ruggieri7). These experimental data prompted interest in 18 : 2n-6 as a possible factor promoting the development of human tumours, in view of the fact that 18 : 2n-6 is the most abundant fatty acid of the n-6 series in vegetable oils which are the major lipid component in the diets of the Western world. The promoting activity of 18 : 2n-6 on carcinogenesis and tumour progression appears to be mediated by the bioactive metabolites of arachidonic acid (AA), the product of elongation and desaturation of 18 : 2n-6(Reference Hubbard, Chapkin and Erickson8, Reference Marks, Muller-Decker and Furstenberger9).
More recently, epidemiological studies have reported that the risk of breast(Reference Sasaki, Horacsek and Kesteloot10), colon(Reference Caygill, Charlett and Hill11) and prostate(Reference Leitzmann, Stampfer, Michaud, Augustsson, Colditz, Willett and Giovannucci12) cancers is lower in populations whose dietary habits are characterised by a high intake of fish oil, which is particularly rich in EPA and DHA, the higher homologues of the α-linolenic acid (18 : 3n-3) family(Reference Terry, Rohan and Wolk13). The protective effect of n-3 PUFA may also be derived from the observation of an enhancement of breast cancer rate(Reference Lands, Hamazaki, Yamazaki, Okuyama, Sakai, Goto and Hubbard14) among Japanese women during the past decades associated with a replacement of fish oil with maize oil in their diet. The recognition of an inhibitory activity of n-3 PUFA on tumour development was further supported by animal studies showing that feeding diets enriched with n-3 PUFA inhibited spontaneous and chemically induced carcinogenesis(Reference Kamano, Okuyama, Konishi and Nagasawa15–Reference Cave17), growth of mammary(Reference Karmali18–Reference Gabor and Abraham20), colon(Reference Reddy and Maruyama21) and liver carcinomas(Reference Calviello, Palozza, Piccioni, Maggiano, Frattucci, Franceschelli and Bartoli22) in rats and mice, and growth of human lung carcinoma in nude mice(Reference de Bravo, de Antueno, Toledo, De Tomás, Mercuri and Quintans23).
Metastatic diffusion of melanoma(Reference Reich, Royce and Martin24), colon(Reference Iigo, Nakagawa, Ishikawa, Iwahori, Asamoto, Yazawa, Araki and Tsuda25), mammary(Reference Rose and Connolly26) and Lewis lung(Reference Yam, Peled, Huszar and Shinitzky27) carcinomas was also inhibited when they were transplanted into animals fed n-3 PUFA-rich diets. In these studies, however, a few histotypes of malignant cells have been considered, and the effects of n-3 PUFA on metastatic diffusion have been examined by using various experimental approaches, for example, by feeding diets containing fish oil(Reference Rose and Connolly26) or purified EPA and DHA in variable proportions(Reference Iigo, Nakagawa, Ishikawa, Iwahori, Asamoto, Yazawa, Araki and Tsuda25), and by subcutaneous administration of fish oil(Reference Salem, Kishihara, Abe, Matsuzaki and Nomoto28). Moreover, there are also a few studies that demonstrated a promoting activity of dietary n-3 PUFA on tumour growth and experimental metastasis(Reference Salem, Kishihara, Abe, Matsuzaki and Nomoto28, Reference Griffini, Fehres, Klieverik, Vogels, Tigchelaar, Smorenburg and Van Noorden29). Thus, more information should be available before introducing the use of appropriate dietary formulations for the chemoprevention and chemosuppression of human tumours.
To expand our knowledge about the inhibitory effects of dietary n-3 PUFA on metastatic diffusion, in the present study we investigated to what extent a diet containing 5 % fish oil, compared with a diet containing 5 % maize oil, affects the reproduction of experimental metastases developed from F10-SR cells, a highly metastatic variant isolated from the murine B16 melanoma F10 line. In the present study, we also investigated whether an alteration of the metastatic potential of F10-SR cells induced by our dietary protocol involves some biological activities crucial in tumour progression, such as cellular proliferation(Reference Cifone30), apoptosis(Reference Larsson, Kumlin, Ingelman-Sundberg and Wolk31, Reference Malaguarnera32) and angiogenesis(Reference Larsson, Kumlin, Ingelman-Sundberg and Wolk31, Reference Li, Shan, Cao and Dewhirst33). Cellular proliferation, apoptosis and angiogenesis were examined by immunolocalisation of their respective markers in metastatic pulmonary colonies developed from F10-SR cells transplanted into animals fed diets containing maize oil or fish oil. These markers were: proliferating cell nuclear antigen (PCNA)(Reference Hall, Levison and Woods34), bax-bcl2(Reference Granville, Carthy, Hunt and McManus35), vascular endothelial growth factor (VEGF)(Reference Ferrara and Gerber36) and endothelial-associated von Willebrand factor (vWF)(Reference Müller, Hermanns, Skrzynski, Nesslinger, Müller and Kirkpatrick37).
Materials and methods
Cells and culture conditions
Cells used in the present study were a highly metastatic variant (F10-SR cells) isolated in our laboratory, from a metastasis reproduced by the murine B16 melanoma F10 line in adrenal gland. F10-SR cells (5 × 105) were seeded in 100 mm tissue culture dishes (Sarstedt), and grown in Dulbecco's modified Eagle's medium containing glucose (4500 mg/l) supplemented with 10 % fetal calf serum (Gibco) at 37°C in a 10 % CO2 humidified atmosphere. Cell cultures were propagated every 3 d by trypsinisation.
Diets
Diets used in the present study were: (a) a diet composed of 70 % sucrose, 16 % casein, 3 % α-cellulose, 1 % choline chloride, 1 % vitamins, 4 % salt mixture and 5 % maize oil (maize oil diet); (b) an isoenergetic diet in which 5 % maize oil was replaced with 5 % fish oil (fish oil diet). The major fatty acids of the maize oil diet were: 14 : 0, 5·3 %; 16 : 0, 1·6 %; 16 : 1, 16·7 %; 18 : 0, 3·3 %; 18 : 1, 22·1 %; 18 : 2, 50·0 %. The fish oil diet contained: 14 : 0, 21·1 %; 16 : 0, 31·6 %; 16 : 1, 10·5 %; 18 : 0, 5·0 %; 18 : 1, 18·6 %; 18 : 2, 3·0 %; 20 : 5n-3, 5·5 %; 22 : 6n-3, 4·5 %. The maize oil and fish oil diets, supplied by Laboratorio Dottori Piccioni (Milan, Italy), were stored at 4°C in sealed plastic bags in vacuo.
Animals and dietary treatments
Newly weaned female mice (C57Bl/6 strain; Charles River, Frederick, MD, USA) were placed on the maize oil diet for 1 week, and then divided in two groups, each of six animals: one group continued to be fed the maize oil diet, while the other group was switched to the fish oil diet. The degree of n-3 PUFA incorporation into the liver phosphatidylcholine was experimentally evaluated in one representative animal by weekly measuring the EPA+DHA:AA ratio. The highest value of this ratio was reached in animals fed the fish oil diet after 5 weeks. At this moment, EPA+DHA:AA ratios in animals fed the maize oil or fish oil diet were 0·2 and 3·7, respectively. The food consumption (4·0 (sd 1·7) g/d per animal) and body weight of the experimental animals were assessed twice per week from the beginning of the dietary treatment to the end of the experiment. At the end of the experiments, animals showed a similar increase (30·4 %) of body weight, regardless of whether they were fed the maize oil or fish oil diet (data not shown).
Determination of lung colonisation of F10-SR cells transplanted in animals fed maize oil or fish oil diets
A F10-SR cell suspension (0·2 ml; 5 × 105 cells/ml Dulbecco's modified Eagle's medium) was injected into the lateral tail vein of experimental animals of the two groups fed their respective diets for 5 weeks. The animals were maintained at their respective diets until the end of experiment, i.e. 3 weeks after tumour cell injection. Animals were killed by cervical translocation under diethyl ether anaesthesia. Lung surface was inspected with the aid of a dissection microscope, and surface pulmonary colonies were counted, exsected from lung parenchyma and submitted to immunohistochemical analyses.
Gas-chromatographic analyses of fatty acids
Total lipids were extracted by the method of Folch et al. (Reference Folch, Lees and Sloane Stanley38) from liver homogenates. Phosphatidylcholine was separated from the other lipid classes by HPLC following the conditions reported in a previous paper(Reference Mannini, Calzolari, Calorini, Mugnai and Ruggieri7). Phosphatidylcholine, solubilised in 0·2 ml benzene, was submitted to transmethylation with 5 % H2SO4 in methanol, at 80°C, overnight. The methyl-ester derivatives of fatty acids were analysed on a 30 m × 0·53 mm internal diameter wide-bore capillary column (Megabore W DB-23, 50 % cyanopropyl polysiloxan; J & W Scientific, Folsom, CA, USA), mounted in a PerkinElmer gas chromatograph equipped with a hydrogen flame detector (PerkinElmer, Waltham, MA, USA). Fatty acid composition was determined from the total integrated area in the chromatogram. Percentages of EPA, DHA and AA were used to calculate EPA+DHA:AA ratios in liver phosphatidylcholine.
Immunohistochemistry
Immunoreactivities of PCNA, bax/bcl-2, VEGF and vWF were determined in 5 μm thick sections of formalin-fixed paraffin-embedded metastatic colonies, according to a procedure described in a previous paper(Reference Mannini, Calzolari, Calorini, Mugnai and Ruggieri7). Monoclonal antibodies against PCNA, bax/bcl-2 and VEGF were supplied by Santa Cruz Biotechnology (Santa Cruz, CA, USA), while the polyclonal antibody against mouse vWF was supplied by Dako (Glostrup, Denmark). Negative controls were represented by consecutive tissue sections submitted to the immunohistochemical procedure in which the primary antibodies were omitted.
The immunoreactivities of PCNA, bax, bcl-2 and VEGF were scored following a semi-quantitative procedure that graded the percentages of reactive cells as follows: +, focal (less than 10 %); ++, moderate (10 to 50 %); +++, strong staining (>50 % stained cells); − , absent. Microvessel densities in the pulmonary colonies grown in animals fed the maize oil or fish oil diets were derived from the number of vWF-positive cells/microscopic field.
Detection of apoptotic nuclei
In order to confirm apoptotic activity in the pulmonary colonies developed in animals fed the maize oil or fish oil diets, DNA fragments in apoptotic nuclei were visualised according to the method of Gavrieli et al. (Reference Gavrieli, Sherman and Ben-Sasson39). This method is based on the reaction, catalysed by terminal deoxynucleotidyl transferase, of 3′-OH ends of DNA fragments with biotin-labelled deoxynucleotides (Oncogene Research Products, Cambridge, MA, USA). The biotinylated nucleotides were complexed with a streptavidin–horseradish peroxidase conjugate, followed by visualisation with diaminobenzidine chromogen. Sections were counterstained with a 0·3 % methyl green solution and mounted on slides in a xylene mounting medium.
Statistical analysis
The Mann–Whitney test was used to determine the statistical significance of the differences between animals fed the maize oil diet and animals fed the fish oil diet in lung colonisation and microvessel densities.
Results
As shown in Fig. 1, F10-SR cells reproduced a lower number of metastatic colonies in the lungs of mice fed the fish oil diet, as compared with mice fed the maize oil diet. Regardless of the dietary regimen, PCNA (Figs. 2 (a) and (b)) was highly expressed in the pulmonary colonies developed in animals transplanted with F10-SR cells. Bax was moderately immunoreactive in the pulmonary colonies grown in animals fed the maize oil diet (Fig. 2 (c)), but highly expressed in the pulmonary colonies grown in animals fed the fish oil diet (Fig. 2 (d)). Regardless of the dietary regimen, pulmonary colonies did not express any reactivity for bcl-2 (see insert). Apoptotic nuclei, visualised by specific labelling of DNA fragments, were detectable only in metastatic colonies grown in animals fed the fish oil diet, as indicated by arrows in Fig. 2 (f).
Fig. 1 Lung colonisation of F10-SR cells in animals fed maize oil (
) or fish oil (□) diets. Values are the means of number of lung colonies per animal (n 5), with standard errors represented by vertical bars. * Mean value was significantly different from that of the animals fed the maize oil diet (P = 0·004).
Fig. 2 Immunoreactivities of proliferating cell nuclear antigen (PCNA), bax and bcl-2 (see inserts in c and d) and visualisation of apoptotic nuclei ( ← ) in pulmonary colonies developed from F10-SR cells in animals fed maize oil (a, c, e) or fish oil (b, d, f) diets. The immunoreactivity of each marker was evaluated as reported in Materials and methods. +, Focal staining (less than 10 %); +++, strong staining (>50 % stained cells).
As shown in Fig. 3, VEGF immunoreactivity was rather weak in the pulmonary colonies developed in animals fed the maize oil diet (Fig. 3 (a)), but fairly strong in colonies grown in animals fed the fish oil diet (Fig. 3 (b)). In these latter, however, there was a reduction of microvessel density, as determined by the visualisation of endothelial-associated vWF (Fig. 3 (d)).
Fig. 3 Immunoreactivities of vascular endothelial growth factor (VEGF) and endothelial von Willebrand factor (vWF) in pulmonary colonies developed from F10-SR cells in animals fed maize oil (a, c) or fish oil (b, d) diets. The immunoreactivity of each marker was evaluated as reported in Materials and methods. +, Focal staining (less than 10 %); +++, strong staining (>50 % stained cells). The mean number of vWF-positive cells/microscopic field was (n 10) 41·6 (sem 3·6) in (c) and 22·0 (sem 1·9) in (d) (P < 0·05).
Discussion
In the present study, we demonstrated that the development of experimental metastases from a highly metastatic murine melanoma line, the F10-SR cells, was inhibited in animals fed a diet containing 5 % fish oil, a concentration much lower than that used in other studies(Reference Rose and Connolly40). In contrast with the present results, Salem et al. (Reference Salem, Kishihara, Abe, Matsuzaki and Nomoto28) found an accentuated growth of primary tumour and an increased development of lung metastases from B16 melanoma cells in mice treated with subcutaneous administration of fish oil before and after subcutaneous transplantation of tumour cells. The discrepancy between our and Salem's results might be related to the different experimental protocols used in the respective studies. An antimetastatic activity of n-3 PUFA was also found by Reich et al. (Reference Reich, Royce and Martin24) who obtained a reduction of lung colonisation of B16-F10 melanoma cells which had been previously enriched in vitro with EPA, an observation also made in our laboratory (A Mannini and S Ruggieri, unpublished results). Moreover, Iigo et al. (Reference Iigo, Nakagawa, Ishikawa, Iwahori, Asamoto, Yazawa, Araki and Tsuda25) found an inhibition of lung colonisation in animals injected with colon carcinoma cells which had been previously submitted to a serial transplantation into mice fed an n-3 PUFA-rich diet. Reich's and Iigo's observations would suggest that the inhibitory activity of dietary n-3 PUFA on metastatic growth is mediated by their incorporation into tumour cells, although a contribution of lipid changes in host cells may not be ruled out.
In order to investigate which biological activities mediate the inhibition of lung colonisation of F10-SR cells in animals fed a fish oil diet and injected with F10-SR cells, we examined immunohistochemically PCNA, bax/bcl-2 and VEGF/vWF, as markers of growth, apoptosis and vascularisation, respectively. This investigation revealed that the reduced lung colonisation of melanoma cells in animals fed a fish oil diet is associated with an increased apoptotic activity, as shown by the enhancement of immunoreactivity of bax combined with a visualisation of nuclear DNA fragments in pulmonary colonies. This finding is in agreement with the positive correlation between the consumption of fish oil diets and apoptotic activity reported in several studies(Reference Rose and Connolly40–Reference Latham, Lund and Johnson42). The unchanged PCNA immunoreactivity in the lung colonies developed in animals fed a fish oil diet suggests that the antimetastatic effect of dietary n-3 PUFA is not mediated by an inhibition of tumour cell proliferation, a finding in contrast with the selective antiproliferative effects of n-3 PUFA observed in different systems of tumour cells grown in tissue culture(Reference Mund, Pizato, Bonatto, Nunes, Vicenzi, Tanhoffer, de Oliveira, Curi, Calder and Fernandes43–Reference Bégin, Ells, Das and Horrobin47). However, the artificial conditions produced in in vitro studies, not comparable with the complex in vivo tumour microenvironment, limit the significance of the data derived from in vitro studies regarding the influence of specific lipid components, including PUFA, on tumour growth (for a review, see Diggle(Reference Diggle48)).
The diminution of microvessel density found in pulmonary colonies developed from F10-SR cells in animals fed a fish oil diet might contribute to the antimetastatic effect of dietary n-3 PUFA. This observation is supported by the finding that the reduction of murine mammary tumour growth by dietary n-3 PUFA is mediated by an inhibition of angiogenesis(Reference Rose and Connolly49, Reference Mukutmoni-Norris, Hubbard and Erickson50). The association of a reduced microvessel density with a high VEGF immunoreactivity in pulmonary colonies developed in animals fed a fish oil diet is quite an unexpected finding. However, it is possible that n-3 PUFA inhibit vascularisation by counteracting the stimulatory activity of VEGF on vascular endothelia. Indeed, the competition between dietary n-3 PUFA and n-6 PUFA inhibits the synthesis of PGE2(Reference Horia and Watkins51, Reference Henderson, Black and Wolf52), an essential factor in angiogenesis(Reference Ziche, Jones and Gullino53, Reference Salcedo, Zhang, Young, Michael, Wasserman, Ma, Martins-Green, Murphy and Oppenheim54). Moreover, Salcedo et al. (Reference Salcedo, Zhang, Young, Michael, Wasserman, Ma, Martins-Green, Murphy and Oppenheim54) showed that the inhibition of PGE2 synthesis abrogates the VEGF-mediated expression of CXCR4, a key receptor in angiogenesis.
In conclusion, the data derived from the present study point out the possibility that even limited amounts of n-3 PUFA in the diet are able to inhibit the reproduction of experimental metastases from F10-SR cells. The reduced formation of metastases from F10-SR cells transplanted into animals fed an n-3 PUFA-rich diet might be the final step of a series of phenomena, consisting of: a decreased generation of PGE2– reduced microvessel density – hypoxia – induction of apoptosis. This sequence, however, does not rule out the possibility that promotion of apoptosis and inhibition of angiogenesis represent two independent responses to the accumulation of n-3 PUFA.
Acknowledgements
The present study was supported by Ente Fondazione Cassa di Risparmio di Firenze (Italy) and by Ente Fondazione Cassa di Risparmio di Lucca (Italy). None of the authors has conflicts of interest with respect to the study. S. R. and A. M. designed the study and prepared the paper. A. M. and N. K. contributed to the successful execution of experimental work. L. C. and G. M. contributed to the lipid analysis.
