Relationship between amount and type of dietary fat in promotion of mammary carcinogenesis induced by 7,12-dimethylbenz[a]anthracene.

Female Sprague-Dawley rats were fed semipurified diets containing various fats, either alone or in combination, to provide different amounts of dietary fat and linoleic acid. One week before commencing the diets, each rat received an intra-gastric dose of the carcinogen 7,12-dimethylbenz[a]anthracene. Rats fed diets containing mixtures of 3% sunflower seed oil and 17% of either tallow or coconut oil developed twice as many tumors as those fed 3% sunflower seed oil or 20% of either saturated fat alone. Tumor yields in the rats fed these mixed-fat diets were comparable to those in rats fed a 20% lard diet, which provided about the same amount of linoleic acid. No further increase in tumor yield was observed in rats fed a 20% sunflower seed oil diet that contained more than five times as much linoleic acid. These results show that a certain amount of polyunsaturated fat, as well as a high level of dietary fat, is required to promote mammary carcinogenesis.

Effect of dietary polyunsaturated fat on the growth of a transplantable adenocarcinoma in C3HAvyfB mice.

Weaning male and female C3HAvyfB mice were fed a low-fat (4.5%) diet until they were 60-70 days of age when they were fed high-fat (18.6%) diets containing either sunflower-seed oil (polyunsaturated fat diet) or tallow (saturated fat diet). After receiving either of the high-fat diets for 4 weeks, each mouse received an inoculum of approximately 1,700 single cells from a transplantable mammary adenocarcinoma. The cumulative incidence of tumor-bearing mice was significantly greater among both males and females fed the polyunsaturated fat diet than among males and females fed the saturated fat diet. The mean times elapsed before palpable tumors developed were less when mice were fed the polyunsaturated fat diet than when mice were fed the saturated fat diet, but these differences were not statistically significant. The cumulative incidence of tumor-bearing mice was also significantly greater among females than males. The results supported the suggestion from previous work in this laboratory that the polyunsaturated fat diet exerts its effect on the promotional stage of carcinogenesis rather than on the initial event of neoplastic transformation.

Polyunsaturated fatty acids as promoters of mammary carcinogenesis induced in Sprague-Dawley rats by 7,12-dimethylbenz[a]anthracene.

The development of mammary tumors was examined in female noninbred Sprague-Dawley rats fed either a low-fat diet or high-fat diets containing different fats and fatty acid esters. Each rat was given 5 mg 7,12-dimethylbenz[a]anthracene by stomach tube 1 week before diets were introduced. Addition of 3% ethyl oleate (an ethyl ester of an unsaturated fatty acid) to a diet high in saturated fat (coconut oil) had no significant effect on tumor development, but the addition of 3% ethyl linoleate (an ethyl ester of a polyunsaturated fatty acid) increased the tumor yield to about twice that in rats fed either the high-saturated fat diet or a low-fat diet. Animals fed the high-saturated fat diet containing 3% ethyl linoleate developed as many tumors as those fed a 20% sunflower seed oil diet, though the sunflower seed oil diet contained about four times as much linoleate. Rats fed a high coconut oil diet containing 3% menhaden fish oil, which contains polyunsaturated fatty acids of the linolenate family (but having little linoleic acid), also developed as many tumors as those fed the 20% sunflower seed oil diet. These differences in mammary tumor yield could not be explained by alterations in the serum levels of prolactin, estrogen, or progesterone. However, the higher tumor yields were associated with increased unsaturation of mammary tissue phospholipids.

Carcinogenesis induced by 7,12-dimethylbenz[a]anthracene in c3H-A vyfB mice: influence of different dietary fats.

The effect of a diet containing either sunflower-seed oil (polyunsaturated fat diet) or tallow (saturated fat diet) on 7,12-dimethylbenz[a]anthracene (DMBA)-induced carcinogenesis in C3H-A vyfB mice was examined. After receiving either diet for 28 days, some of the mice were given an intragastric dose of 5 mg DMBA. To identify the stage of carcinogenesis that might be influenced by dietary fat, the diets of half of the mice were then interchanged so that those previously fed the saturated fat diet were fed the polyunsaturated fat diet and vice versa. The cumulative incidence of tumor-bearing mice was significantly greater among the females fed the polyunsaturated fat diet compared to those fed the saturated fat diet. This enhancement of carcinogenesis was observed only when the mice were fed the polyunsaturated fat diet after DMBA administration. Similar trends were observed in the male mice, but these mice developed fewer tumors and none of the differences between the tumor incidences were statistically significant. The most common sites for tumors in the male mice were the liver, lungs, and skin, whereas those for tumors in the females were the mammary glands and ovaries. The differences in tumor incidence suggest that carcinogenesis was enhanced by the polyunsaturated fat diet during the promotion stage of carcinogenesis.

Relative rates of incorporation of esterified cholesterol into human very low density lipoproteins and low density lipoproteins. In vitro studies of two separate pathways.

It has been shown previously that there are two pathways by which the esterified cholesterol formed in human plasma in the reaction catalysed by lecithin: cholesterol acyltransferase may be delivered to very low density lipoproteins (VLDL) and low density lipoproteins (LDL): (a) an indirect pathway in which esterified cholesterol which was incorporated initially into high density lipoproteins (HDL) is transferred subsequently to VLDL and LDL in a process mediated by an esterified cholesterol transfer/exchange protein and (b) a direct pathway in which a small proportion of the esterified cholesterol formed in the lecithin:cholesterol acyltransferase reaction is delivered to VLDL and LDL directly from its site of synthesis via a pathway which bypasses the bulk HDL fraction. These present studies have been designed to examine the incorporation of esterified cholesterol into VLDL relative to that into LDL via each of these two pathways. It has been found that a delivery of esterified cholesterol from HDL to VLDL and LDL via the indirect pathway has a marked preference for VLDL over LDL; equating the concentrations of esterified cholesterol in the two fractions revealed an incorporation into VLDL which was 7-11 times greater than that into LDL. By contrast, delivery via the direct pathway showed a marginal preference for LDL over VLDL.

An effect of very low density lipoproteins on the rate of cholesterol esterification in human plasma.

Rates of esterification of plasma cholesterol have been compared in two groups of human subjects with widely differing concentrations of very low density lipoproteins (VLDL). Group 1 consisted of subjects with fasting plasma triacylglycerol concentrations over 2.0 mM and group 2 consisted of subjects with concentrations under 1.6 mM. The rate of esterified cholesterol production in incubations of whole plasma from group 1 subjects was much greater than that from group 2 subjects. A clear difference between the rates of esterification was also evident when either total lipoproteins, total high density lipoproteins (HDL) or the HDL subfraction, HDL3, from the two types of subjects were incubated in the presence of a common source of lecithin:cholesterol acyltransferase. Since these findings appeared to reflect fundamental differences within the HDL3 subfractions, which may have been modified by prior exposure to the high concentrations of VLDL in group 1 subjects, VLDL-deficient plasma and the plasma fraction of d greater than 1.125 g/ml (containing HDL3) from hypotriglyceridaemic subjects were preincubated at 37 degrees C with an excess of added VLDL in the presence of a reversal of the inhibition of lecithin:cholesterol acyltransferase, the capacity of the original fractions to esterify cholesterol had been markedly increased. These studies, therefore, show that the lecithin:cholesterol acyltransferase substrate capacity of particles within the HDL3 subfraction is enhanced by exposure to VLDL and that this enhancement is not dependent on the continued presence of VLDL during the actual esterification reaction.

Capacity of lipoproteins to act as substrates for lecithin:cholesterol acyltransferase. Enhancement by pre-incubation with an artificial triacylglycerol emulsion.

Previous studies have shown that high concentrations of very low density lipoproteins (VLDL) stimulate the formation of cholesteryl esters in human plasma, possibly by acting as recipients of cholesteryl esters transferred from high density lipoproteins (HDL). To gain further insight into this phenomenon, experiments were performed to determine whether the triacylglycerol emulsion Intralipid, which acts as an artificial recipient of HDL cholesteryl esters, has an effect on cholesterol esterification similar to that of VLDL. Intralipid, in contrast to VLDL, is devoid of apolipoproteins which may stimulate cholesterol esterification. Human plasma, which had previously been depleted of VLDL, was pre-incubated in the presence or absence of Intralipid. After pre-incubation, the Intralipid was removed and rates of cholesterol esterification were measured in subsequent incubations of the Intralipid-depleted fractions. The presence of Intralipid during the pre-incubation had a marked stimulatory effect on cholesterol esterification, comparable to that of VLDL in earlier studies. The pre-incubation with Intralipid also markedly reduced the cholesteryl ester content, and increased the triacylglycerol content, of the HDL. These changes in composition coincided with two changes in the elution profile of HDL after density-gradient ultracentrifugation which were (i) a reduction in the density of particles in the HDL3 subfraction, which virtually disappeared as an identifiable peak, and (ii) the emergence of a discrete population of very dense lipoproteins consisting primarily of protein and phospholipid. Since all these changes were related to redistributions of lipids, the results highlight the importance of lipid transfers in the regulation of plasma cholesterol esterification.

Comparison of different methods of isotopically labelling the esterified cholesterol in human high density lipoproteins.

Different methods of incorporating esterified [3H]cholesterol into human high density lipoproteins (HDL) have been assessed in terms of their influence on the subsequent rate at which esterified [3H]cholesterol was transferred from HDL to other plasma lipoproteins in 37 degrees C incubations containing tracer amounts of the labelled preparations added to aliquots of a common pool of unlabelled human plasma. Two basic methods were used to label the HDL esterified cholesterol: (a) by the action of lecithin: cholesterol acyltransferase in incubations of plasma containing exogenous [3H]cholesterol and (b) by exchange in incubations of plasma containing preparations of prelabelled LDL. The exchange labelling approach was also used to examine the effects of various HDL compositional changes on the subsequent behaviour of its esterified [3H]cholesterol. It was found that the source of the label, per se, whether from the esterification of exogenous [3H]cholesterol or from exchange with labeled LDL, had no influence on the subsequent behaviour of HDL esterified [3H]cholesterol. However, modification of the HDL composition during the labelling process had profound effects. For example, in preparations in which the esterified cholesterol content of the labelled HDL was increased by the action of lecithin: cholesterol acyltransferase, there was a reduced rate of transfer of esterified [3H]cholesterol out of the HDL fraction in subsequent incubations of the labelled HDL with fresh, unlabelled plasma. Conversely, when the esterified cholesterol content of the labelled HDL had been decreased by pre-incubation with Intralipid or very low density lipoproteins (VLDL), the subsequent rate of transfer of esterified [3H]cholesterol out of the HDL fraction was increased markedly. It has been concluded that to obtain a biologically valid, labelled tracer of human HDL esterified cholesterol, the labelling should be achieved with minimal modification to the HDL composition. This means labelling by exchange in incubations in which lecithin: cholesterol acyltransferase is inhibited in plasma from which VLDL has been removed.

Comparison of human plasma low- and high-density lipoproteins as substrates for lecithin: cholesterol acyltransferase.

A recent observation that lecithin: cholesterol acyltransferase (EC 2.3.1.43) interacts with both low-density lipoproteins (LDL) and high-density lipoproteins (HDL) in human plasma is in apparent conflict with an earlier finding that the purified enzyme, while highly reactive with isolated HDL, was only minimally reactive with LDL. There is evidence, however, that lecithin: cholesterol acyltransferase may exist physiologically as a component of a complex with other proteins and that studies with the isolated enzyme may therefore provide misleading results. Consequently, interactions of the enzyme with isolated human lipoproteins have been re-examined in incubations containing lecithin: cholesterol acyltransferase as a component of human lipoprotein-free plasma in which a physiologically active complex of the enzyme with other proteins may have been preserved. In this system there was a ready esterification of the free cholesterol associated with both LDL and HDL-subfraction 3 (HDL3) in reactions that obeyed typical enzyme-saturation kinetics. For a given preparation of lipoprotein-free plasma the Vmax values with LDL and with HDL3 were virtually identical. The apparent Km for free cholesterol associated with HDL3 was 5.6 X 10(-5) M, while for that associated with LDL it was 4.1 X 10(-4) M. This implied that, in terms of free cholesterol concentration, the affinity of HDL3 for lecithin: cholesterol acyltransferase was about 7-times greater than that of LDL. When expressed in terms of lipoprotein particle concentration, however, it was apparent that the affinity of LDL for the enzyme was considerably greater than that of HDL3. When the lipoprotein fractions were equated in terms of lipoprotein surface area, the apparent affinities of the two fractions for the enzyme were found to be comparable.

The rabbit as an animal model of hepatic lipase deficiency.

A natural deficiency of hepatic lipase in rabbits has been exploited to gain insights into the physiological role of this enzyme in the metabolism of plasma lipoproteins. A comparison of human and rabbit lipoproteins revealed obvious species differences in both low-density lipoproteins (LDL) and high-density lipoproteins (HDL), with the rabbit lipoproteins being relatively enlarged, enriched in triacylglycerol and depleted of cholesteryl ester. To test whether these differences related to the low level of hepatic lipase in rabbits, whole plasma or the total lipoprotein fraction from rabbits was either kept at 4 degrees C or incubated at 37 degrees C for 7 h in (i) the absence of lipase, (ii) the presence of hepatic lipase and (iii) the presence of lipoprotein lipase. Following incubation, the lipoproteins were recovered and subjected to gel permeation chromatography to determine the distribution of lipoprotein components across the entire lipoprotein spectrum. An aliquot of the lipoproteins was subjected also to gradient gel electrophoresis to determine the particle size distribution of the LDL and HDL. Both hepatic lipase and lipoprotein lipase hydrolysed lipoprotein triacylglycerol and to a much lesser extent, also phospholipid. There were, however, obvious differences between the enzymes in terms of substrate specificity. In incubations containing hepatic lipase, there was a preferential hydrolysis of HDL triacylglycerol and a lesser hydrolysis of VLDL triacylglycerol. By contrast, lipoprotein lipase acted primarily on VLDL triacylglycerol. When more enzyme was added, both lipases also acted on LDL triacylglycerol, but in no experiment did lipoprotein lipase hydrolyse the triacylglycerol in HDL. Coincident with the hepatic lipase-induced hydrolysis of LDL and HDL triacylglycerol, there were marked reductions in the particle size of both lipoprotein fractions, which were now comparable to those of human LDL and HDL3, respectively.

Competitive inhibition of plasma cholesterol esterification by human high-density lipoprotein-subfraction 2.

Mixtures containing subfractions of human plasma high-density lipoproteins (HDL) and human lipoprotein-free plasma were incubated in vitro at 37 degrees C. Esterification of cholesterol was observed both in incubations containing HDL-subfraction 3 (HDL3) and in those containing HDL-subfraction 2 (HDL2). The implication that the lecithin: cholesterol acyltransferase in lipoprotein-free plasma may therefore interact with lipoproteins in both HDL subfractions was developed further by proposing a simple model in which the two HDL subfractions may compete for interactions with the enzyme. This model was described mathematically and tested in experiments in which a constant amount of the enzyme was incubated with a wide range of concentrations of HDL2 and HDL3 present either alone or in combination. The model was able to predict experimentally observed rates of cholesterol esterification with great accuracy. The best fit was obtained with a Vmax for HDL3 that was 2.4-4-times greater than that for HDL2 and values of the apparent Km for HDL3 free cholesterol and HDL2 free cholesterol of 43-60 nmol/ml and 167-391 nmol/ml, respectively. The model thus predicts that, at physiological concentrations of lipoproteins, HDL2 will function as a competitive inhibitor of the cholesterol esterification reaction by displacing lecithin: cholesterol acyltransferase from a more effective substrate, HDL3, to a less effective substrate, HDL2.

Pathways for the incorporation of esterified cholesterol into very low density and low density lipoproteins in plasma incubated in vitro.

In vitro incubations of human or pig plasma containing a tracer amount of [3H]cholesterol have been performed to determine which lipoprotein fractions are the immediate recipients of the esterified cholesterol formed in the reaction catalysed by lecithin: cholesterol acyltransferase. In pig plasma, which is deficient in activity of the protein which promotes transfer of esterified cholesterol between different lipoprotein fractions, 87-90% of the lecithin: cholesterol acyltransferase-derived esterified cholesterol was incorporated into the high density lipoprotein (HDL) fraction. In human plasma there was an initial recovery of more than 80% in HDL, although the proportion recovered in very low density lipoproteins (VLDL) and low density lipoproteins (LDL) became progressively greater with increasing duration of incubation, consistent with a transfer from an HDL -esterified cholesterol pool of increasing specific activity. Nevertheless, as in the pig plasma incubations, there was evidence that some 10-15% of the esterified cholesterol formed in the lecithin: cholesterol acyltransferase reaction was incorporated directly into human VLDL and LDL. In quantitative terms, however, it was found that most of the esterified cholesterol delivered to human VLDL and LDL was the result of transfers from HDL rather than as a direct incorporation from its site of synthesis.

Apolipoprotein A-I inhibits transformation of high density lipoprotein subpopulations during incubation of human plasma.

We have examined the effect of added apolipoprotein A-I (apoA-I) on the changes in high density lipoprotein (HDL) particle size that occur when human plasma is incubated in vitro. In the absence of added apoA-I, incubation of plasma at 37 degrees C resulted in a dramatic increase in HDL particle size. When these incubations contained an inhibitor of LCAT, an additional population of smaller HDL particles was formed. These changes in particle size were even more pronounced when the incubations were supplemented with an artificial triglyceride emulsion, Intralipid. All these changes in HDL particle size were markedly inhibited when incubations were supplemented with apoA-I. Even when the amount of added apoA-I was as little as 4.5% of the endogenous apolipoprotein there was an obvious inhibition of the changes in HDL particle size. The presence of added apoA-I sufficient to increase the plasma concentration by 18% virtually abolished the changes in HDL particle size. This effect did not relate to an inhibition of cholesterol esterification, nor did it appear to depend on an incorporation of the added apoA-I into the HDL.

Role of esterified cholesterol transfers in the regulation of plasma cholesterol esterification.

The role of esterified cholesterol transfers in mediating the known effect of very low density lipoproteins (VLDL) on plasma cholesterol esterification has been examined. Advantage has been taken of the fact that pig plasma, in contrast to human plasma, is deficient in activity of the protein that transfers esterified cholesterol between plasma lipoproteins. The experimental protocol consisted of a pre-incubation in which VLDL-deficient pig plasma or isolated pig lipoproteins were incubated with added VLDL, in the presence or absence of an exogenous source of the esterified cholesterol transfer protein. After pre-incubation, any VLDL that may have been present was removed and rates of cholesterol esterification measured in subsequent incubations of the recovered VLDL-depleted fractions. It was, thus, possible to assess the effect of VLDL and/or transfer protein activity on the capacity of other lipoproteins to act as substrates for cholesterol esterification. In the absence of transfer protein activity, the presence of VLDL in pre-incubations had no effect on the subsequently measured rate of cholesterol esterification. However, when pig plasma was supplemented with human lipoprotein-free plasma (a source of esterified cholesterol transfer protein activity) the presence of VLDL in pre-incubations resulted in both a reduction in the esterified cholesterol content of the re-isolated VLDL-deficient fraction and a markedly enhanced rate of cholesterol esterification in subsequent incubations of this fraction.(ABSTRACT TRUNCATED AT 250 WORDS)

Lipoprotein substrates for plasma cholesterol esterification. Influence of particle size and composition of the high density lipoprotein subfraction 3.

Subpopulations of lipoproteins within the high density lipoprotein subfraction 3 (HDL3) have been isolated from human plasma and characterized in terms of their chemical composition and particle size. Since HDL3 are the major substrates for the esterification of plasma cholesterol, and thus play a central role in the transport of cholesterol through the plasma, these studies also examined the relative capacities of different HDL3 subpopulations to interact with lecithin: cholesterol acyltransferase. Substrate reactivity of a given preparation of lipoproteins was defined in terms of the Vmax of the cholesterol esterification reaction in incubations containing a fixed amount of human lipoprotein-free plasma as a source of the enzyme. Substrate reactivity was found to correlate inversely with the radius of HDL3 particles. This inverse relationship between particle size and substrate reactivity was independent of the particle content of cholesteryl ester.

Particle size distribution of high density lipoproteins as a function of plasma triglyceride concentration in human subjects.

Polyacrylamide gradient gel electrophoresis has been used to examine the particle size distribution of high density lipoproteins (HDL) in human subjects with a wide range of plasma triglyceride concentrations. In studies of groups of both male and female subjects, it was confirmed that the concentration of HDL cholesterol decreases with increasing plasma triglyceride concentration. The HDL fraction from subjects with elevated concentrations of plasma triglyceride was depleted of cholesteryl ester and enriched in triglyceride. It was also confirmed that the proportion of HDL subfraction 2 (HDL2) declines as the plasma triglyceride increases. A new finding was that there were also significant changes in the size of particles in HDL subfraction 3 (HDL3). At low concentrations of plasma triglyceride the predominant subpopulation of HDL3 comprised particles of mean radius 4.3 nm. As the triglyceride concentration increased, however, there was a progressive appearance of HDL3 particles of radius 3.9 nm; in plasma samples with the highest concentrations of triglyceride there was an almost complete disappearance of the 4.3-nm particles, with the population of 3.9-nm particles now predominant.

Lipoprotein lipase prevents the hepatic lipase-induced reduction in particle size of high density lipoproteins during incubation of human plasma.

Human plasma lipoproteins or human whole plasma have been incubated in vitro with canine hepatic lipase (HL) and bovine milk lipoprotein lipase (LPL) to determine the effects of lipases on the particle size distribution of HDL. Confirming previous reports, HL preferentially hydrolysed high density lipoprotein (HDL) triacylglycerol while LPL hydrolysed predominantly very low density lipoprotein (VLDL) triacylglycerol; however, neither lipase altered HDL particle size unless both VLDL and cholesteryl ester transfer protein (CETP) were present. Under these conditions HL promoted marked reduction in HDL particle size in a process dependent on the concentration of VLDL triacylglycerol while LPL was virtually without effect. When both LPL and HL were included in the same incubation, however, LPL prevented the effects of HL. These results are consistent with a proposition that HL has a direct effect on HDL particle size in a process which is dependent on concurrent lipid transfers between HDL and VLDL and that LPL reduces the effect of HL by reducing the concentration of VLDL triacylglycerol.

A unified model of esterified cholesterol exchanges between human plasma lipoproteins.

Exchanges of esterified cholesterol between human high density lipoproteins (HDL) and very low density lipoproteins (VLDL) and between VLDL and low density lipoproteins (LDL) have been fitted to a mathematical model previously developed to describe exchanges between human HDL and LDL. In all cases the fit of the model predicted exchanges to those measured experimentally was extremely good, thus greatly increasing confidence in the validity of the model. The model assumes that esterified cholesterol exchanges are achieved by means of a transfer protein which interacts with lipoprotein particles from which it picks up and deposits esterified cholesterol. The values generated for the model constants indicated that, given equal concentrations of esterified cholesterol in each fraction, the relative probability that the transfer protein will pick up a molecule of esterified cholesterol in HDL vs. VLDL vs. LDL is in the ratio 28.9:4.65:1. According to the model the transfer protein may 'bind' to lipoproteins. The model predicts that, at physiological lipoprotein concentrations, the proportion of transfer protein bound to HDL will be more than double that which is unbound to lipoprotein and that bound to VLDL will be about one tenth that unbound. The model was unable to detect evidence of the transfer protein binding to LDL.

Dissociation of the in vitro transfers of esterified cholesterol and triglyceride between human lipoproteins.

This report describes the detailed time-course of in vitro net mass transfers of esterified cholesterol (EC) and triglyceride (TG) between human plasma lipoproteins incubated in the presence of EC and TG transfer proteins. In incubations containing high density lipoproteins (HDL), very low density lipoproteins (VLDL) and the thiol-group blocker parachloromercuriphenyl sulfonate (PCMPS) (which was added to inhibit EC production), the rate of EC transfer to VLDL was considerably more rapid than that of TG to HDL and equilibrium between the pools of EC was achieved much sooner than was the case for the pools of TG. When VLDL and low density lipoproteins (LDL) were incubated, the presence of PCMPS in the incubation mixture resulted in a marked reduction in the TG transfer to LDL, even though the EC transfer to VLDL was apparently unaffected. The molar ratio of the transfers of EC and TG was also examined in incubations of lipoproteins from a variety of human subjects. In incubations of HDL and VLDL (performed in the absence of PCMPS) there was a large individual variation (sixfold difference) in the molar ratio of the transfers of TG:EC. In incubations of LDL and VLDL from different subjects there was a twofold individual variation in the molar ratio of the transfers of TG:EC. It has been concluded that the net mass transfers of EC and TG between human lipoproteins do not involve a simple molecular exchange of one moiety for the other, as has been suggested previously, but are achieved by processes which are at least partially independent.

Dietary polyunsaturated fat versus saturated fat in relation to mammary carcinogenesis.

High levels of dietary fat have been shown to promote the development of mammary tumors induced in rats by 7,12-dimethylbenz(alpha)anthracene, and polyunsaturated fats were found to be more effective than saturated fats. In further studies it was found that diets containing 3% sunflowerseed oil (polyunsaturated fat) and 17% beef tallow or coconut oil (saturated fats) enhance tumorigenesis as much as a diet containing 20% sunflowerseed oil. Rats on these diets developed at least twice as many tumors as those fed diets containing either 3% sunflowerseed oil or 20% of the saturated fats alone. These results are in accord with human epidemiological data which show that breast cancer mortality in different countries is positively correlated with total fat intake but not with intake of polyunsaturated fat. Total fat intake varies greatly in different countries, but most human diets probably contain levels of polyunsaturated fat at least equivalent to 3% sunflowerseed oil.