Modulation of prostate cancer genetic risk by omega-3 and omega-6 fatty acids

Abstract

Although a causal role of genetic alterations in human cancer is well established, it is still unclear whether dietary fat can modulate cancer risk in a predisposed population. Epidemiological studies suggest that diets rich in omega-3 polyunsaturated fatty acids reduce cancer incidence. To determine the influence of fatty acids on prostate cancer risk in animals with a defined genetic lesion, we used prostate-specific Pten-knockout mice, an immune-competent, orthotopic prostate cancer model, and diets with defined polyunsaturated fatty acid levels. We found that omega-3 fatty acids reduced prostate tumor growth, slowed histopathological progression, and increased survival, whereas omega-6 fatty acids had opposite effects. Introducing an omega-3 desaturase, which converts omega-6 to omega-3 fatty acids, into the Pten-knockout mice reduced tumor growth similarly to the omega-3 diet. Tumors from mice on the omega-3 diet had lower proportions of phosphorylated Bad and higher apoptotic indexes compared with those from mice on omega-6 diet. Knockdown of Bad eliminated omega-3-induced cell death, and introduction of exogenous Bad restored the sensitivity to omega-3 fatty acids. Our data suggest that modulation of prostate cancer development by polyunsaturated fatty acids is mediated in part through Bad-dependent apoptosis. This study highlights the importance of gene-diet interactions in prostate cancer.

Figure 1. Suppression of prostate tumor proliferation by omega-3 PUFAs in vivo.

Pten^P+/+, Pten^P+/–, and Pten^P–/– mice were fed the high–omega-3, low–omega-3, and high–omega-6 diet for a period of up to 24 weeks. Mouse AP, DL, and VP lobes were weighed, and the sums were expressed as milligrams per 25 gram body weight. Five mice were used per data point in a cohort of 180 mice. Open circles represent mice on the high–omega-3 diet; shaded squares, mice on the low–omega-3 diet; filled triangles, mice on the high–omega-6 diet. Horizontal bars represent averages. SDs are shown for Pten^P–/– mice fed with the high–omega-3 and high–omega-6 diet.

Figure 2. Pathological evaluation of prostate.

(A) Gross appearance of 8-week-old prostates. (B) Histological evaluation of AP, DL, and VP lobes. Histology was evaluated for the cohort of 180 mice used in Figure 1, but qualitative differences were observed for 8-week-old Pten^P–/– mice when the high–omega-3, low–omega-3, and high–omega-6 diets were compared. Two sections of AP, DL, and VP from each mouse (5 mice per group) were sampled from separate areas of each tissue block and evaluated by 2 veterinary pathologists. When complex histology was found, the most advanced type was indicated. Hyper, hyperplasia; cis, carcinoma in situ; Inv ca, invasive carcinoma. (C) Representative H&E-stained sections from VPs of Pten^P–/– mice. Scale bar: 100 μm. Additional sections are shown in Supplemental Figure 3.

Figure 3. Pten deletion rate in mice on high–omega-3 and high–omega-6 diets.

(A) Schematic representation of Pten and lacZ alleles before and after Cre-loxP–mediated recombination. In the ROSA26^LoxZ allele, recombination removes a sequence that interrupts the coding frame of β-galactosidase, resulting in activation of β-galactosidase enzyme. (B) Real-time quantification of Pten Δ5 allele. The inactive (exon 5 deletion) Pten allele and a control gene (wild-type Il-2) were quantified in 6-, 7-, and 8-week-old AP, DL, and VP lobes from mice on the high–omega-3 and high–omega-6 diet. Levels of the Pten Δ5 allele normalized to Il-2 are shown. Three mice were used for each data point, with bars representing SEM. Pten was deleted to similar extents in mice on either diet. (C) Prostates from 6-week-old Pten^loxP/loxPROSA26^Z/ZPB-cre4^T/– mice on the high–omega-3 and high–omega-6 diet were dissected, snap-frozen, and used for β-galactosidase staining as well as activity measurement. Three mice were used for each data point. Bars represent SEM. Pictures were taken at ×20 magnification. In 4 of the 6 mice that had sufficient protein remaining, the ratio of phosphorylated to total Akt, which is indicative of Pten inactivation, was quantified by Western blotting.

Figure 4. Effect of the fat-1 omega-3 desaturase on prostate tumor growth.

Pten^P–/–fat1^T/– and Pten^P–/–fat1^–/– mice were fed the high–omega-6 diet. Prostate tumor weight was compared in 12-week-old mice, with 5 mice per group. Bars represent SD. P < 0.006, Student’s t test.

Figure 5. Survival rate of mice on different diets.

A cohort of 90 mice (10 mice per group) on the high–omega-3, low–omega-3, or high–omega-6 diet was monitored for a period of 1 year. The Kaplan-Meier cumulative survival plot shows the cumulative probability of survival of each group of mice. Pten^P+/+ and Pten^P+/– mice had 100% survival regardless of diet. Twelve-month survival of Pten^P–/– mice was 60%, 10%, and 0% on the high–omega-3, low–omega-3, and high–omega-6 diets, respectively.

Figure 6. Alterations in Bad phosphorylation in vivo.

(A) Protein extracts from the prostates of 8-week-old Pten^P+/+ and Pten^P–/– mice fed the high–omega-3 and high–omega-6 diets were used for Western blotting analysis of Akt, pSer473 Akt, Bad, pSer112 Bad, and β-actin. Ratios of phospho-Akt to total Akt, phospho-Bad to total Bad, and total Bad to β-actin were calculated. Samples with the lowest ratio in each comparison were arbitrarily set as 1, and relative ratios are indicated at the bottom of the corresponding panels. Group averages and SEM for total Bad/β-actin and phospho-Bad/total Bad ratios are shown in the graph. (B) Immunohistochemistry for apoptotic cells was performed with an anti–cleaved caspase-3 antibody in prostate tissues of 8-week-old Pten^P–/– mice. Arrowheads indicate apoptotic cells. Scale bar: 50 μm.

Figure 7. Dependence on Bad for omega-3 PUFA–induced cell death.

(A) PC3 cells were infected with lentivirus expressing scrambled shRNA (C shRNA) or Bad-specific shRNA (Bad shRNA) or mock infected (no shRNA) for 2 days. Cells were seeded and incubated with FAs for 6 days, then photographed using a fluorescence microscope. The presence of lentivirus is evidenced by GFP expression. Western blotting was performed to confirm the successful knockdown of Bad in both PC3 and LNCaP cells. Day 0, prior to FA treatment; day 6, after incubation with media containing FA for 6 days. C, control. (B) Live and dead cells were enumerated by the trypan blue exclusion method using a hemocytometer. P values were determined by Student’s t test. LNCaP cells have higher background apoptosis compared with PC3 cells, but omega-3 FA treatment increased apoptosis approximately 4-fold in both cell lines.(C) LNCaP cells were transfected with control vector or mouse Bad expression vector. G418-resistant colonies were isolated, verified for successful expression of the HA-tagged mouse Bad by Western blotting, and pooled. Vector control– (V) and mouse Bad–expressing (mBad) LNCaP cells were infected with shRNA lentivirus as described above. Cells were incubated with FAs for 6 days, and live and dead cells were enumerated as described above. Cell lysates were used for confirmation of successful knockdown of the endogenous Bad but not the exogenous mouse Bad (whose sequence varies from human Bad in the region targeted by the shRNA).