PUFAs and cancer .1 Lipid content of the diet

Epidemiological studies, often based on the comparison of populations living for a long time in a country and recently migrating people, suggest that the incidence of different types of cancer is more linked to environmental fac-tors than to genetic facfac-tors (RIBOLIand DECLOÎTRE, 1996). Among the environ-mental factors, diet seems to play a major role and, among constituents of the diet, lipids could have a prominent role in cancer development. From a quantita-tive point of view, it has been demonstrated that animals fed lipid-rich diets

develop tumors more rapidly than those fed lipid-poor diets, independent of the nature of the fatty acids used in the experiments (CARROLL, 1991).

3.1.4.2 Nature of fatty acids

Epidemiological and experimental studies have shown that one must consi-der the nature of fatty acid intake, especially the n-3 or n-6 family. Epidemiolo-gical studies have compared a population consuming a large amount of marine products (n-3 LC-PUFAs) and a population from the same origin having a diffe-rent diet. The most frequent conclusion from these studies is that an abundant consumption of n-3 LC-PUFAs decreases the frequency of different types of cancer (breast, colon, prostate…) (NIELSENand HANSEN, 1980; SCHLOSS et al., 1977; NORRISHet al., 1999). N-3 PUFAs are in competition with n-6 and increa-sing one decreases the other in cellular phospholipids; these results suggest that, a contrario, n-6 PUFAs could promote tumor development. However, a controversy remains concerning the results of other epidemiological studies:

some authors do not observe any significant correlation between tissular amounts (serum, adipose tissue) of n-6 PUFAs in healthy or cancer-bearing people in the same population (ZOCKand KATAN, 1998), although a large Euro-pean study (SIMONSENet al., 1998) indicates a strong inverse correlation bet-ween the ratio n-3 LC-PUFAs/total n-6 PUFAs and breast cancer frequency (for review, BARTSCHet al., 1999).

The results of experimental studies, either on cell cultures from mouse or man, or on animal models, are clearer. They indicate that, with the same lipid content in the diet, linoleic acid from vegetal oils promotes tumor growth (JUR -KOWSKIand CAVE, 1985; CANNIZO and BROITMAN, 1989; ROSE et al., 1995) and increases carcinogenesis (YETIV, 1988) and metastasis formation in animal models (CARROL, 1997). Inversely, oils rich in n-3 fatty acids, α-linolenic acid (HIROSEet al., 1990; KAMANO et al., 1989; NAKAYAMA et al., 1993) or EPA and DHA of fish oils (IP et al., 1986; CAVE, 1994; WELSCH, 1992; ROSEet al., 1995), can slow down tumor growth.

Among possible mechanisms, the most frequently suggested concerns the high level of prostaglandin synthesis of the series 2, especially PGE2, arising from a high tissular level of arachidonic acid in animals fed a linoleic acid-rich diet (HORROBIN, 1986). Many studies demonstrated that PGE2 stimulates cellu-lar proliferation (OKUYAMA et al., 1997), by promoting c.myc oncogene expres-sion (BANDYOPADHYAYet al., 1988; HANDLERet al., 1990). Recent data confirm that n-6 PUFAs increase cyclooxigenase expression in healthy mammary gland (BADAWI et al., 1998) and that this enzyme is over-expressed in tumor cells (HWANG et al., 1998). Besides, protracted inflammation, associated with over-production of other derivatives of C20:4 n-6 by lipoxigenases, frequently results in tumoral transformation. Thus, molecules presenting anti-inflammatory acti-vity, by inhibition of lipoxigenase activity on arachidonic acid, appear to exhibit carcinogenesis inhibition (MARNETT, 1992; RAO et al., 1995). In this context, a recent study shows that 5-hydroxieicosatetraenoic acid (5-HETE), issued from arachidonic acid by lipoxigenase activity, stimulates prostate tumor cell prolife-ration and is a powerful factor for tumor cell survival (GHOSHand MYERS, 1998).

In brief, arachidonic acid could promote tumorigenesis via its derivatives:

from cyclooxigenase activity (especially PGE2) and from lipoxigenase activity (especially leucotriens). Thus linoleic acid, precursor of C20:4 n-6, would be

involved in the process. Moreover, its oxidation products have been directly suspected (KELERet al., 1992).

Conversely, n-3 LC-PUFAs (EPA and DHA), essentially present in fish oils, are generally considered as inhibitors of tumor growth (CAVE, 1994; WELSCH, 1992; ROSE et al., 1995). First, they are precursors of prostanglandins of the series 3, which do not exert mitogenic effect (MARNETTand HONN, 1994; NOGU -CHIet al., 1995). Secondly, they inhibit arachidonic acid synthesis from linoleic acid, decreasing the production of prostaglandins of series 2. Some research works suggest that α-linolenic acid is also able to decrease the growth of some cancers (OKUYAMAet al., 1997 for review); this is explained by the fact that this fatty acid limits by competition arachidonic acid synthesis from linoleic acid.

Another possibility lies with the antitumoral effect of polyunsaturated fatty acid lipoperoxides (not only n-3) derived from radical peroxidation. Indeed, tumor cells are very sensitive to these lipoperoxides (WELSCH, 1987; BEGIN et al., 1988; HORROBIN, 1990; GONZALES et al., 1993; CHAJES et al., 1995), which exert a cytotoxic effect on them (PADMAand DAS, 1996). In other words, PUFAs incorporation into tumor cells, and especially incorporation of the most readily oxidizable n-3 PUFAs, leads to an increase in the sensitivity of these cells to oxidative stress (BEGINet al., 1988; VARTAK et al., 1997), increasing in parallel their sensitivity to radical producing treatments: hyperthermia (SPECTOR and BURNS, 1987), pro-oxidant drugs (BURNS and NORTH, 1986) and irradiation.

These data are confirmed by studies showing that PUFAs toxicity against tumor cells, in vitro and in vivo, is suppressed when culture medium or diet contains excessive amounts of vitamin E (CHAJES et al., 1995; LHUILLERY et al., 1996, 1997; FINSTADet al., 1998), whereas pro-oxidant drugs, like quinones, are cyto-toxic for tumor cells (WORKMAN, 1994; CORNWELLet al., 1998).

In addition, PUFAs can modulate hormone secretion (WIDMAIER, 1997), thus having a possible effect on hormono-dependent tumors (breast, prostate). For example, it has been shown that n-3 PUFAs deficiency decreases melatonin synthesis, whose role in sexual hormone synthesis is well known (GAZZAHet al., 1993). However, few specific research works have been devoted to the study of individual effects of fatty acids on hormonal secretion.

In summary, from research works published during the last 15 years, there is a true convergence concerning the effects of PUFAs on tumor development:

– n-6 PUFAs, producing arachidonic acid and thus prostaglandins of the series 2, are considered to promote tumor growth.

– n-3 LC-PUFAs, by limiting the synthesis of arachidonic acid, fights against tumor growth.

– All PUFAs could be considered as inhibitors of tumor growth, being sources of cytotoxic lipoperoxides, but n-3 PUFAs are major sources due to their high instauration level, insofar as lipoperoxides can be produced. Conse-quently, dietary intake of anti-oxidant molecules, and particularly vitamin E, must remain reasonable. Thus, n-3 PUFAs could be used as preventive agents in tumorigenesis and as help in cancer treatment using radiotherapy or chemotherapy with prooxidant drugs.

– The value of 5 for the ratio linoleic acid/α-linolenic acid seems also to be appropriate for cancer prevention, as suggested by the follow-up after the Lyon Study (DE LORGERILet al., 1998).