Extra-Virgin Olive Oil and Its Minor Compounds Influence Apoptosis in Experimental Mammary Tumors and Human Breast Cancer Cell Lines

Simple Summary Breast cancer is a disease influenced by dietetic factors, such as the type and amount of lipids in a diet. In this work, we aimed to elucidate the different effects of two high-fat diets on the histopathological and molecular characteristics of mammary tumors in an experimental model. Animals fed with a diet high in extra-virgin olive oil (EVOO), compared to those fed with a diet high in seed oil, developed tumors with less aggressiveness and proliferation. Tumor molecular analyses of several cell death pathways also suggested an effect of EVOO in this process. In vitro experiments indicated the role of EVOO minor compounds on the effects of this oil. Obtaining insights into the influence and the mechanisms of action of dietary compounds are necessary to understand the relevance that dietetic habits from childhood may have on health and the risk of disease. Abstract Breast cancer is the most common malignancy among women worldwide. Modifiable factors such as nutrition have a role in its etiology. In experimental tumors, we have observed the differential influence of high-fat diets in metabolic pathways, suggesting a different balance in proliferation/apoptosis. In this work, we analyzed the effects of a diet high in n-6 polyunsaturated fatty acids (PUFA) and a diet high in extra-virgin olive oil (EVOO) on the histopathological features and different cell death pathways in the dimethylbenz(a)anthracene-induced breast cancer model. The diet high in n-6 PUFA had a stimulating effect on the morphological aggressiveness of tumors and their proliferation, while no significant differences were found in groups fed the EVOO-enriched diet in comparison to a low-fat control group. The high-EVOO diet induced modifications in proteins involved in several cell death pathways. In vitro analysis in different human breast cancer cell lines showed an effect of EVOO minor compounds (especially hydroxytyrosol), but not of fatty acids, decreasing viability while increasing apoptosis. The results suggest an effect of dietary lipids on tumor molecular contexts that result in the modulation of different pathways, highlighting the importance of apoptosis in the interplay of survival processes and how dietary habits may have an impact on breast cancer risk.


Introduction
Breast cancer is the most common cancer in women worldwide, with high incidence, prevalence, and mortality rates [1]. This neoplasia has a multifactorial etiology, in which diet, as an environmental factor, has an important role. Epidemiological and especially (hydroxytyrosol, oleuropein, and luteolin) were also carried out to acquire insight into the molecular basis of the influence of EVOO on breast cancer.

Animals and Experimental Design
All animals received humane care under an institutionally approved experimental animal protocol, following the legislation applicable in this country. Female Sprague-Dawley rats (stain Crl: OFA (SD), n = 100) were obtained from Charles River Lab (L'Arbresle Cedex, France) at 23 days of age. Three semisynthetic diets were designed: the low-fat diet (LF; 3% corn oil -w/w-), the high corn oil diet (HCO; 20% corn oil) and the high extra-virgin olive oil diet (HEVOO; 3% corn oil + 17% EVOO). The definition, preparation, and suitability of the experimental diets were previously described in studies [8,21,22]. Animals were distributed in five groups depending on the diet and the timing of dietary intervention (n = 20 each group): the LF diet from weaning (LF group), the HCO diet from weaning (HCO group) or after induction (LF-HCO group), and the HEVOO diet from weaning (HEVOO group) or after induction (LF-HEVOO group). At 53 days of age, mammary tumors were induced by oral gavage with one single dose of 5 mg of DMBA (Sigma-Aldrich, St. Louis, MO, USA) dissolved in corn oil (25 mg DMBA/kg body weight). From day 74 onwards, animals were monitored weekly for the appearance of mammary tumors. At 236-256 days of age, animals were euthanized by decapitation, selecting the rats that were in the diestrus phase of the estrous cycle, and between 10:00 h and 13:00 h to avoid circadian rhythm variability ( Figure S2). Tumors were removed, a portion of each was fixed in 4% formalin for histopathological analysis, and the rest was flash-frozen and stored at −80 • C for molecular analysis.
Tumor histopathology characterization was determined by applying the Scarff-Bloom-Richardson grading method (SBR3; scoring 1-3), and also the method adapted to rat mammary carcinomas that we have previously described (SBR11; scoring 3-11) [23], with the highest categories indicating the most aggressive tumors. SBR scores include differentiation pattern, nuclear pleomorphism, and mitotic index. Number of mitoses were determined in 10 high power fields (HPF) and classified into five categories [23]. Only confirmed mammary adenocarcinomas were included in this study.

Protein Extraction
Tumor samples were homogenized in an extraction buffer (50 mM Tris-HCl pH 7.2, 250 mM sucrose, 2 mM EDTA, 1 mM EGTA, 5 mM MgCl 2 , 50 µM NaF, 100 µM Na 3 VO 4 , 10 µL/mL protease inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA), 10 mM β-mercaptoethanol, and 1% Triton X-100) for total protein extraction. For fractioned extracts, samples were cold centrifugated at 105,000× g for one hour to separate the soluble fraction (corresponding to cytosol) from the particulate fraction, which was further resuspended in extraction buffer with 0.1% Triton X-100 and centrifuged at 105,000× g (supernatant corresponding to the membranous fraction containing all membranes). Mitochondrial extract was obtained from homogenized samples by centrifuging at 1500× g to separate nuclei, followed by supernatant centrifuging at 10,000× g to obtain precipitated mitochondria. Protein quantification was determined by Lowry's method using the commercial DC Protein Assay Kit (Bio-Rad, Hercules, CA, USA) (cytosolic, mitochondrial, and membrane).

Western Blot
The different protein extracts (10-20 µg) were subjected to SDS-PAGE electrophoresis on an acrylamide gel (7.5% or 12% Mini-Protean TGX Stain-Free Gels, Bio-Rad, Hercules, CA, USA) and transferred to a PVDF membrane with Trans-Blot Turbo Transfer System (Bio-Rad). Membranes were blocked with 5% BSA or 5% skimmed milk in TBS-0.1% Tween for 1 h at room temperature and incubated with primary antibody overnight at +4 Membranes were incubated with the secondary antibody (anti-mouse or anti-rabbit IgG peroxidase-conjugated) for 1 h at room temperature and developed with Luminata Forte Western HRP Substrate (Merk Millipore, Burlington, MA, USA) luminogen for 3-5 min. Proteins bands were visualized using the Chemidoc XRS+ hardware associated with Image Lab Software 5.1-Beta (Bio-Rad). Densitometric values were normalized with the total protein loaded [24] and relativized to an internal control sample loaded in duplicates in all blots.

Cell Culture Treatment
For in vitro analyses, we used the MCF-7 cell line, representing the molecular subtype of human breast cancer Luminal A, and MDA-MB-231 cells, representing the triple-negative subtype. Cell lines were obtained from the American Type Culture Collection (ATCC; LGC Standards, Middlesex, UK). MCF-7 cells were grown in EMEM (Gibco, Thermo Fisher Scientific, Waltham, MA, USA) and supplemented with 10% fetal bovine serum (FBS; Gibco) and 0.01 mg/mL of insulin. MDA-MB-231 cells were grown in DMEM (Gibco) and supplemented with 10% FBS. Cells were maintained at 37 • C in a humidified atmosphere of 95% air and 5% CO 2 . The cell cultures were subjected to Mycoplasma screen tests periodically.

Viability Assays
For analysis of cellular metabolic activity as an indicator of cell viability and proliferation, we used MTT colorimetric assay based on the reduction of thiazolyl blue tetrazolium bromide (Sigma-Aldrich) to formazan. Briefly, MCF-7 cells were seeded in 96-multiwell plates at 6 × 10 3 cells/well in 100 µL of complete EMEM (Gibco) medium with 5% FBS. MDA-MB-231 cells were seeded in 96-multiwell plates at 5 × 10 3 cells/well in 100 µL of complete DMEM (Gibco) medium with 5% FBS. After 24 h, the medium was replaced with fresh one containing the specific treatment and incubated for 24 to 72 h. MTT reagent was added to each well (10 µL/well of MTT solution at 5 mg/mL in PBS) and incubated 2 h at +37 • C. Then, formazan crystals were solubilized in 100 µL of 100% DMSO and absorbances were read at 570 nm. Relative cell viability was calculated as the ratio of absorbance of treated samples and control samples.

Determination of Apoptotic Cells by Flow Cytometry
MCF-7 and MDA-MB-231 cells were seeded 24 h before treatment in 12-multiwell plates at 2 × 10 5 cells/well in 1 mL of complete EMEM and DMEM (Gibco) medium, respectively, with 5% FBS. Cells were treated with the specific concentration of oleic acid, linoleic acid, hydroxytyrosol, oleuropein, or lutein for 24, 48, and 72 h. Cells were trypsinized, washed in PBS, and labelled with Annexin-V-Fluorescein and Propidium iodide (PI) using an Annexin-VFLUOS staining kit (Inmunostep, Salamanca, Spain) according to the manufacturer instructions. Then, data were acquired on a FACSCalibur flow cytometer at SCAC Facility, UAB.

Mitochondrial Membrane Potential Assay by JC-1 Staining
MCF-7 cells were seeded 48 h before treatment in 6-multiwell plates at 4 × 10 5 cells/well and treated for 24, 48, and 72 h with 0 or 400 µM of hydroxytyrosol (Sigma Aldrich). Briefly, cells were collected by trypsinization, stained with JC-1 dye (15 µg/mL), and incubated at room temperature for 15 min at 37 • C in a humidified atmosphere of 95% air and 5% CO 2 . Data were acquired on a FACSCalibur flow cytometer at SCAC Facility, UAB.

Statistical Analysis
Statistical analyses were performed using R Deducer. The statistical test to be used was determined depending on the distribution of each variable (Shapiro-Wilk test) and the equality of variances among groups (Levene's test). Data from tumors usually do not follow a normal distribution, and the non-parametrical Kruskal-Wallis test-Wilcoxon method was used. Quantitative parametrical data were analyzed with one-way ANOVA test (t-test equal variance) and paired t-test. For qualitative data (distributions and percentages) the Pearson's Chi-squared test was used. The level of significance was established at p < 0.05. For in vivo results, distribution of data is represented using box plots. Box indicates interquartiles (IQR; first quartile Q1, second quartile = median, and third quartile Q3); whiskers extend from maximum (Q3 + 1.5 × IQR) to minimum (Q1 − 1.5 × IQR); dots represent outliers; and cross indicates the mean value.

The High-Corn Oil Diet Influenced the Manifestation of the Disease
Carcinogenesis parameters are shown in Table 1. The tumor latency period (days from induction to first palpable malignant tumor) was shorter in the HCO and LF-HCO groups, although differences did not reach statistical significance. On the other hand, the percentage of tumor-bearing animals and total number of tumors were higher in the groups fed the HCO diet in comparison to the LF and HEVOO groups. Tumor multiplicity (mean number of tumors per animal) was also increased in the HCO group in comparison to the LF group. Results for the morphological degree of tumor aggressiveness and proliferation are shown in Figure 1. Histopathological analyses indicated a higher degree of malignancy (both in SBR3 and SBR11 score) in the groups fed the HCO diet in comparison to the groups fed the control and high-EVOO diets. A histological quantitation of proliferation indicated a higher mitosis score in the high-corn oil diet groups versus control, especially in LF-HCO, and a great number of mitoses in both HCO-fed groups in relation to the control. LF-HCO also showed a great number of mitoses in comparison to the high-EVOO diet groups.

The High-Corn Oil Diet, but Not the High-EVOO Diet, Had a Clear Effect on Histopathological Malignancy and Proliferation of Tumors
Results for the morphological degree of tumor aggressiveness and proliferation shown in Figure 1. Histopathological analyses indicated a higher degree of maligna (both in SBR3 and SBR11 score) in the groups fed the HCO diet in comparison to groups fed the control and high-EVOO diets. A histological quantitation of prolifera indicated a higher mitosis score in the high-corn oil diet groups versus control, espec in LF-HCO, and a great number of mitoses in both HCO-fed groups in relation to control. LF-HCO also showed a great number of mitoses in comparison to the high-EV diet groups.

The EVOO-Enriched Diet Promoted the Expression of Extrinsic Apoptosis Pathway Proteins
The protein expression levels of extrinsic pathway-related proteins (Fas, FADD total and membrane fraction, TNFR1, TRADD membrane fraction, TRAF2, pro-Caspase-8, and cleaved-Caspase-8) were studied by Western blot and the results are shown in Figure 2 (Original Western blots in Supplementary material). Changes in protein expression levels were observed in FADD, TNFR1, and pro-Caspase-8 with higher levels in the LF-HEVOO group when compared to LF group. When comparing between high-fat diets, FADD (total and membrane fraction), TNFR1, and pro-Caspase-8 showed higher levels in the LF-HEVOO group versus LF-HCO group, while the HEVOO group showed higher FADD and TRADD in comparison to the HCO group. Higher expression levels of TNFR1 and pro-Caspase-8 were also found in LF-HEVOO versus HEVOO.
representative determination of mitoses in tumors with different mitosis score. Stacked bar chart depicts the distribution of mammary adenocarcinomas of each group according to the mitosis score. Box and whisker plot depicts data of the number of mitoses in 10 high power field (HPF, 400× magnification) for each tumor. Arrows connecting groups indicate differences statistically significant (p < 0.05), Chi-squared test (distribution data: SBR3 and mitosis score), non-parametric Mann-Whitney U test (quantitative data: SBR11 and number of mitoses). LF, n = 46; HCO, n = 100; LF-HCO, n= 87; HEVOO, n = 57; LF-HEVOO, n = 82. Scale bar.

The EVOO-Enriched Diet Promoted the Expression of Extrinsic Apoptosis Pathway Proteins
The protein expression levels of extrinsic pathway-related proteins (Fas, FADD total and membrane fraction, TNFR1, TRADD membrane fraction, TRAF2, pro-Caspase-8, and cleaved-Caspase-8) were studied by Western blot and the results are shown in Figure 2 (Original Western blots in Supplementary material). Changes in protein expression levels were observed in FADD, TNFR1, and pro-Caspase-8 with higher levels in the LF-HEVOO group when compared to LF group. When comparing between high-fat diets, FADD (total and membrane fraction), TNFR1, and pro-Caspase-8 showed higher levels in the LF-HEVOO group versus LF-HCO group, while the HEVOO group showed higher FADD and TRADD in comparison to the HCO group. Higher expression levels of TNFR1 and pro-Caspase-8 were also found in LF-HEVOO versus HEVOO. . FADD (membrane) and TRADD (membrane) were detected in the membrane fraction, the other proteins were detected in total protein extract. Box: 25th percentile, median, 75th percentile; whiskers: maximum to minimum; dots: outliers; cross: mean value. Solid arrows connecting groups indicate statistically significant differences (p < 0.05), dashed lines indicate differences close to significance (p < 0.1), Kruskal-Wallis test.

The EVOO-Enriched Diet Increased the Expression of Intrinsic Apoptosis Pathway Proteins
We have also studied proteins involved in the apoptotic intrinsic pathway: the proapoptotic Bid and the active tBid, Bak, Bax, Cytochrome C in mitochondrial and cytoplasm fraction, APAF, pro-Caspase-9, cleaved-Caspase-9, and the anti-apoptotic Bcl2 (Figure 3 and Original Western blots in Supplementary material). Relative protein expression levels of the total Bid, tBid, and Cytochrome C (membrane fraction) were higher in the LF-HEVOO group than in the LF group. This LF-HEVOO group also showed higher levels of total Bid, tBid, Bak, and Bax when compared to the HCO-fed groups, and those of Bid and Bak when compared to the HEVOO group.

The EVOO-Enriched Diet Increased the Expression of Intrinsic Apoptosis Pathway Proteins
We have also studied proteins involved in the apoptotic intrinsic pathway: the proapoptotic Bid and the active tBid, Bak, Bax, Cytochrome C in mitochondrial and cytoplasm fraction, APAF, pro-Caspase-9, cleaved-Caspase-9, and the anti-apoptotic Bcl2 (Figure 3 and Original Western blots in Supplementary material). Relative protein expression levels of the total Bid, tBid, and Cytochrome C (membrane fraction) were higher in the LF-HEVOO group than in the LF group. This LF-HEVOO group also showed higher levels of total Bid, tBid, Bak, and Bax when compared to the HCO-fed groups, and those of Bid and Bak when compared to the HEVOO group.  Bcl2, Cytochrome C (cyt.C) in membrane and cytosol, APAF in cytosol, and Caspase-9 (pro-and cleaved-Casp9). Cyt.c (membrane) was detected in membrane fraction, Cyt.C (cytosol) and APAF (cytosol) were detected in cytosol fraction, the other proteins were detected in total protein extract. Box: 25th percentile, median, 75th percentile; whiskers: maximum to minimum; dots: outliers; cross: mean value. Solid arrows connecting groups indicate statistically significant differences (p < 0.05), dashed lines indicate differences close to significance (p < 0.1), Kruskal-Wallis test.

Effect of High-Fat Diets on Endoplasmic Reticulum (ER) Stress-Induced Cell Death and in Apoptosis Pathways Convergence
Results for the Caspase-12 protein, which is involved in ER stress-induced cell death, are shown in Figure 4A and in Original Western blots in Supplementary material. There was a trend in the HCO group to have lower levels of pro-Caspase-12, while a trend to have higher levels of the active cleaved form was observed in the LF-HCO group. was a trend in the HCO group to have lower levels of pro-Caspase-12, while a trend to have higher levels of the active cleaved form was observed in the LF-HCO group.
On the other hand, the protein expression levels of pro-Caspase-3 and active Caspase-3 are shown in Figure 4B and in Supplementary material. There was a trend to have increased levels of 17 kDa cleaved-Caspase-3 in the LF-HEVOO group compared to the LF-HCO and HEVOO groups. The Caspase-3 inhibitor, XIAP, was increased in the LF-HEVOO group in comparison to LF-HCO group. The key protein in confluence of apoptotic pathways Caspase-3 (pro-Casp3 and the 19 KDa and 17 Da cleaved forms), and its inhibitor XIAP. (D) Caspaseindependent cell death proteins AIF (total in mitochondria and active cytosolic AIF levels) and EndoG (mitochondria and cytosolic). (E) Total protein levels of p53. AIF and EndoG were detected in mitochondria fraction and total protein extract, the other proteins were detected in total protein extract. Box: 25th percentile, median, 75th percentile; whiskers: maximum to minimum; dots: outliers; cross: mean value. Solid arrows connecting groups indicate statistically significant differences (p < 0.05), dashed lines indicate differences close to significance (p < 0.1), Kruskal-Wallis test.
On the other hand, the protein expression levels of pro-Caspase-3 and active Caspase-3 are shown in Figure 4B and in Supplementary material. There was a trend to have increased levels of 17 kDa cleaved-Caspase-3 in the LF-HEVOO group compared to the LF-HCO and HEVOO groups. The Caspase-3 inhibitor, XIAP, was increased in the LF-HEVOO group in comparison to LF-HCO group.

Effect of High-Fat Diets on Caspase Independent Cell Death and on p53
We studied AIF and EndoG as proteins related to caspase-independent cell death ( Figure 4C and Original Western blots in Supplementary material). The total AIF (67 KDa) in mitochondria was decreased in the HCO group in comparison to the LF group. Active AIF (57 KDa) in cytosol was detected in a few samples, with a higher frequency (Pearson's Chi-squared test, p < 0.05) in the LF-HEVOO group. EndoG levels in mitochondria were significantly higher in the LF-HEVOO and HEVOO groups compared to the LF-HCO group, while there was a trend toward higher levels of cytosolic EndoG in the LF-HEVOO group.
In relation to p53, total levels were significantly higher in the LF-HEVOO group in comparison to all other high-fat diet groups ( Figure 4D and Supplementary material).

Hydroxytyrosol Promoted Apoptosis in Breast Cancer Cell Lines
Annexin V/PI staining and flow cytometry were used to evaluate the effect of EVOO polyphenols in cell apoptosis. MDA-MB-231 and MCF-7 cells were treated with increasing doses of HT, OLE, and LUT for 24 h and 48 h.  compared to all other doses. Few differences were found with OLE and LUT treatments, with no clear trend

Hydroxytyrosol Did Not Significantly Affect Cell Cycle or Mitochondrial Membrane Potential in MCF-7
Further analysis on the effect of HT was carried out on MCF-7 cells. We studied by flow cytometry the cycle of cells treated with HT at 400 µM for 24, 48, and 72 h ( Figures 8A and S4A). On the other hand, mitochondrial membrane potential was analyzed by staining with JC-1, a cationic dye indicator of potential-dependent accumulation in the mitochondria (Figures 8B and S4B). No statistically significant differences were observed in the HT treatment.
Further analysis on the effect of HT was carried out on MCF-7 cells. We studied by flow cytometry the cycle of cells treated with HT at 400 μM for 24, 48, and 72 h ( Figures  8A and S4A). On the other hand, mitochondrial membrane potential was analyzed by staining with JC-1, a cationic dye indicator of potential-dependent accumulation in the mitochondria (Figures 8B and S4B). No statistically significant differences were observed in the HT treatment.

Discussion
In this work, we aimed to analyze the effects of two high-fat diets, rich in n-6 PUFA or rich in EVOO, on the characteristics of the aggressiveness and proliferation of DMBAinduced mammary tumors, and to obtain insight into the molecular mechanisms of such effects. We have previously demonstrated a different influence of these diets on the manifestation of experimental tumors, clearly stimulating in the case of the high n-6 PUFA diet, and with the high-EVOO diet having a weak influence [4,[9][10][11][12]. Here, we have observed that tumors from both groups fed with the high n-6 PUFA diet, from weaning (HCO group) or after induction (LF-HCO group), displayed the highest tumor histopathological grade, indicating more aggressiveness [23]. The number of mitoses and the mitosis score were also higher in such groups. On the other hand, tumors from groups fed with the high-EVOO diet were not significantly different to those from the control low-fat group. Considering that high-fat intake is associated with an increase in breast cancer risk [28,29], and that both experimental high-fat diets were isocaloric, our results suggest a specific beneficial effect of the EVOO that can partially counteract the detrimental impact of an excessive intake of fat.

Discussion
In this work, we aimed to analyze the effects of two high-fat diets, rich in n-6 PUFA or rich in EVOO, on the characteristics of the aggressiveness and proliferation of DMBAinduced mammary tumors, and to obtain insight into the molecular mechanisms of such effects. We have previously demonstrated a different influence of these diets on the manifestation of experimental tumors, clearly stimulating in the case of the high n-6 PUFA diet, and with the high-EVOO diet having a weak influence [4,[9][10][11][12]. Here, we have observed that tumors from both groups fed with the high n-6 PUFA diet, from weaning (HCO group) or after induction (LF-HCO group), displayed the highest tumor histopathological grade, indicating more aggressiveness [23]. The number of mitoses and the mitosis score were also higher in such groups. On the other hand, tumors from groups fed with the high-EVOO diet were not significantly different to those from the control low-fat group. Considering that high-fat intake is associated with an increase in breast cancer risk [28,29], and that both experimental high-fat diets were isocaloric, our results suggest a specific beneficial effect of the EVOO that can partially counteract the detrimental impact of an excessive intake of fat.
The effects that dietary lipids may exert on mammary tumorigenesis are likely driven by multiple and complex molecular mechanisms acting simultaneously. We have recently demonstrated that these high-fat diets induced metabolic changes depending on the type and timing of dietary intervention [13]. The high-EVOO diet, in comparison to the high n-6 PUFA diet, induced modifications suggesting an increase in the glucose uptake and glycolysis, pentose phosphate pathway (PPP), tricarboxylic acid cycle, and energy dissipation by UCP2. Although these pathways have been reported as deregulated in cancer [14][15][16], the clinical and morphological characterization of tumors showed that such changes were not associated to aggressiveness. Our data also pointed out that the relevance of metabolic changes depend on the interplay with other signaling pathways, such as apoptosis [30]. In this sense, in a previous transcriptomic analysis, we have observed different effects of these high-fat diets on the expression profiles of genes related to metabolism, proliferation, and apoptosis pathways [9]. Thus, we have analyzed in the experimental tumors the effect of diets on the protein levels of several pathways leading to cell death: intrinsic, extrinsic, ER stress-induced, and caspase-independent ( Figure S1). The results showed that several pro-apoptotic proteins increased in the tumors from animals fed the EVOO diet, especially in the LF-HEVOO group. In this group, we found higher levels of the proteins of the extrinsic apoptosis pathway TNFR1, FADD, and pro-Caspase-8, although there were no differences in active-Caspase-8 levels. On the other hand, several mitochondrial proteins with a role in the intrinsic apoptosis pathway were also increased (tBid, Bax, Bak, and Cytochrome C), and no differences were found in cytoplasmic Cytochrome C or active-Caspase-9. Caspase-12, involved in ER stress-induced apoptosis, showed a tendency to be increased in the LF-HEVOO group, and the active-Caspase-12 increased in the LF-HCO group. Caspase-3, the key protein in the confluence of the apoptotic pathways, showed no marked changes, which could be related to the increase in its inhibitor XIAP. Despite this, a statistical trend towards higher levels of 17 kDa cleaved-Caspase-3 was found in the LF-HEVOO group. Regarding the caspase-independent pathway, membrane EndoG levels were increased in such EVOO diet groups, also showing a trend towards higher levels of cytosolic EndoG. Although few samples showed active AIF, the percentage of tumors with detectable cytoplasmatic levels of active AIF was significantly higher in the LF-HEVOO group. Finally, we also observed significantly higher levels of the p53 protein in this group. Taken together, all these results raise several hypotheses. One is that, given the complexity and heterogeneity of the experimental tumors, the set of results is much more informative than the specific proteins, so these data suggest greater cell death due to the effect of the high-EVOO diet through several pathways. Another possibility is that despite observing increases in various pro-apoptotic proteins, anti-apoptotic proteins are also upmodulated, thus some effectors are not increased. Additionally, little changes in these proteins are enough to contribute to the aggressive phenotype in the context of other pathways, and methodological issues cannot be ruled out. In any case, the analysis of all these proteins and pathways suggests that cell death could be increased in the LF-HEVOO group (which displays a tendency towards greater levels of cleaved-Caspase-3, active AIF, EndoG, and p53), which is consistent with the histopathological characteristics of this group. The fact that the same results have not been found in the group fed this diet from weaning (HEVOO) is consistent with the greater volume of their tumors [9,12]. It is also concordant with metabolic changes observed in the HEVOO group, i.e., an increase in glycolysis and PPP, which has been associated with a decrease in apoptosis [13]. This metabolic profile favors NADPH generation, which may lead to caspase inhibition [31,32].
Further analyses were performed with the components of the oils in different cell lines characteristic of different molecular subtypes of human breast cancer: MDA-MB-231, representing the triple-negative molecular subtype, and MCF-7, representing the Luminal A subtype. MCF-7 is a non-aggressive, hormone receptor-positive, and non-invasive cell line, while the MDA-MD-231 cell line is an estrogen receptor-, progesterone receptor-, and HER2negative, highly proliferative and invasive cell line [33]. In accordance, the proliferation rate was higher in MDA-MD-231 cells, while the percentage of cells in apoptosis was lower in comparison to MCF-7 cells. We analyzed cell viability and apoptosis in cells treated with the main fatty acids of the two oils used in vivo: oleic acid (main fatty acid of EVOO) and linoleic acid (main fatty acid of corn oil). In MDA-MB-231, but not in MCF-7, both fatty acids had a slight effect on viability, while the higher doses after 72 h decreased apoptosis. It is likely that the effect of oleic acid depends on time, dose, and the type of cell. It has been reported in other cell lines that oleic acid treatment induced apoptosis [34,35], but it has also been shown to increase proliferation, migration, or invasion in cancer cells [36,37]. It has been suggested that the heterogeneity in fatty acid metabolism may be at the basis of the different response of cell lines to fatty acid treatments [38]. On the other hand, we analyzed the effect of some of the main minor compounds of EVOO: the secoiridoid oleuropein (OLE), its metabolite hydroxytyrosol (HT), and the main flavonoid luteolin (LUT) [39]. HT and OLE had a dose-and time-dependent effect, decreasing cell viability on both MDA-MB-231 and MCF-7 cells. HT also stimulated apoptosis in both cell lines, while little effect was observed with OLE and LUT. The preliminary results with HT have not shown changes by this polyphenol in cell cycle or membrane potential, the latter suggesting that apoptosis would not be led from the intrinsic pathway, although further analyses are needed to elucidate the involvement of the different cell death pathways. The results with hydroxytyrosol and oleuropein agree with those described by other authors in relation to their anti-proliferative and apoptotic properties [40,41]. On the contrary, with the doses used, and in these cell lines, we did not find a significant effect of luteolin on viability and apoptosis, although other authors reported anti-inflammatory and antioxidant effects [42], and tumor suppressor activity in DMBA-induced tumors [43,44]. Recently, a mitogenic effect of oleic acid has also been observed in colon cancer cells, while the effect of several minor compounds of EVOO was the opposite [45], which highlights the importance of the specific composition of the oils consumed and the complexity of the interaction of nutritional factors in vivo.
All this data suggest that dietary fats have a molecular effect on tumors that influences their histopathological degree and proliferation through various mechanisms, specific and nonspecific, that interact with each other, creating a myriad of responses. The results are consistent, for example, with changes in the membrane composition of tumors as a function of the type of lipid consumed [46,47], concomitantly with a specific effect of different minor compounds, which would have effects in different signaling pathways heterogeneously. Thus, we have observed increases in the EVOO groups in the proteins involved in the extrinsic apoptosis pathway, in which changes in microdomains can have a role, and in intrinsic and caspase-independent pathways, which can be also related to the effect of EVOO minor compounds. In this sense, we have previously observed changes in p21Ras location and activity, which has been related to an oleic acid-enrichment of membranes that can modify microdomains and thus several signaling pathways [12,46]. In addition, our in vitro results, both in hormone receptor-positive (representing Luminal A human breast cancer) and hormone receptor-negative (representing the triple-negative subtype) cells, suggest that HT and OLE are able to influence different breast cancer cell lines through estrogen receptor-independent pathways. In the case of the rodent DMBA model, the mammary tumors developed reflect the most common less-aggressive human breast cancer, being hormone-sensitive, low proliferative, and with no overexpression of HER2 [23,48]. In any case, data in the literature suggest several mechanisms of action of EVOO minor compounds, including hormone receptor-dependent and -independent effects [49]. Many EVOO components have shown anticancer activities by modulating gene expression, the direct activation/inhibition of proteins and enzymes, or antioxidant activity, which may in turn regulate apoptosis [49,50].
The results also showed the effect of the timing of dietary intervention. High-fat diets were administered from weaning (groups HCO and HEVOO), or after induction (groups LF-HCO and LF-EVOO). Both groups fed the high-EVOO diet had lower clinical and histopathological levels of malignancy than high corn oil diet groups. However, the LF-HCO group had the highest degree of clinical and morphological malignancy, while LF-EVOO is the group in which more differences in cell death proteins have been found. In previous works, we have observed that the high-fat diets, especially the high corn oil diet, advanced the puberty onset [10]. Thus, with an accelerated puberty maturation, the observed difference in carcinogenesis yield can be related to the different degree of differentiation that the mammary gland was at during the time of the carcinogenic insult. If the mammary gland of groups LF-HCO and LF-HEVOO were less differentiated at the time of induction, the gland could be more sensitive to the influence of the high-fat diets, clearly carcinogenesis-promoting in the case of the high corn oil diet (resulting in the highest degree of malignancy in the LF-HCO group), and with an effect of EVOO on cell death pathways (more evident in the LF-EVOO group). Moreover, we also previously observed an increase in body weight and mass if the high corn oil diet is administered from weaning [10]. Considering that advanced puberty onset and obesity are known risk factors of breast cancer, our results highlight the importance of acquiring healthy choices from an early age.
Despite the variability of some results, the group of tumors from animals fed with the high-EVOO diet have different characteristics, such as low proliferation, modifications in several cell death pathways, and the reported metabolic changes [13]. However, neither of these pathways, without the interrelation of all networks, characterized by itself the clinical manifestation and the degree of tumor aggressiveness in vivo. In vitro analyses support an effect of EVOO on proliferation and apoptosis, with minor compounds such as hydroxytyrosol and oleuropein playing a role. In vivo, it is likely that the fatty acids also have a role as carriers of minor compounds and/or in the absorption of such compounds, in addition to the reported importance of the ratio between fatty acids on breast cancer risk [51,52]. Taking into account the tumor-promoting influence of a high intake of total fat [28,29], our investigations support the beneficial effect of the moderate intake of extra-virgin olive oil (which contains all the bioactive elements) on the risk of breast cancer. The complexity of this disease suggests that the effects of nutritional factors are subtle, heterogeneous, and probably in the long-term, but potentially relevant, especially considering that exposure, depending on dietary habits, can occur chronically throughout life. Thus, more investigation on the effects and mechanisms of these modifiable factors is needed, which can provide important scientific bases for dietary recommendations in health and disease.

Conclusions
The results reported show the specific effect of the type of dietary lipid in breast carcinogenesis with the diet high in n-6 PUFA diet but not the high-EVOO diet, promoting mammary tumors with a high degree of histopathological malignancy and proliferation. This is the first study reporting that EVOO increased proteins leading to different cell death pathways in experimental mammary tumors. In vitro results support that the potential beneficial effect of EVOO is mainly elicited by its minor components, such as hydroxytyrosol (decreased tumor cell variability and increased cell death), and highlight the importance of specific food compositions (e.g., different effects of olive oil, extra-virgin olive oil, and oleic acid-enriched seed oil) in breast cancer prevention and treatment strategies.

Institutional Review Board Statement:
The study was conducted according to the guidelines of the Declaration of Helsinki, and approved by the Ethical Committee of Animal and Human Experimentation of Universitat Autonoma de Barcelona (CEEAH 566/3616).

Informed Consent Statement: Not applicable.
Data Availability Statement: Molecular data from Western blots generated in this study are included in Supplementary Materials.