Dietary Energy Balance Modulates Prostate Cancer Progression in Hi-Myc Mice

Obesity is a long-term consequence of energy imbalance that occurs when energy intake exceeds expenditure. Although data associating obesity and prostate cancer risk have been inconclusive (1, 2), studies have shown increased risk of biochemical failure and metastasis, as well as poorer survival among obese patients with prostate cancer with androgen-dependent tumors, especially those who experienced rapid weight gain (3). In addition, obesity has been shown to be associated with more aggressive tumors and adverse outcome including mortality (1, 4–7). Waist-to-hip ratio (an indicator of body fat distribution) has also been positively correlated with other hormone-dependent cancers (8, 9). Men with low-volume prostate cancer have a lower body mass index (BMI), less body fat, and a smaller waist-to-hip ratio than men with high-volume prostate cancer, which agrees with other reported findings (10, 11). Several biologic mechanisms have been postulated to explain the association between obesity and aggressive disease, including increases in circulating levels of growth factors [i.e., insulin-like growth factor-1 (IGF-1) and leptin], hyperinsulinemia, and increased systemic and tissue inflammation (12–14)

Epidemiologic evaluations of the independent role of energy intake in prostate cancer incidence, disease progression, or mortality are more limited. High total energy intake, independent of BMI, was associated with a statistically significant increased risk of fatal prostate cancer in the Health Professionals Follow-Up Study Cohort (15). Similarly, energy intake was significantly higher among cases than controls in the U.K. multicenter, population-based case–control study of incident prostate cancer; however, risk analyses were not computed (16). Fat, the most calorically dense component of total energy intake, has been the focus of most prostate cancer dietary epidemiologic investigations. High consumption of fat, especially saturated fat (17–19), is associated with advanced-stage prostate cancer and mortality (17, 20). Several biologic mechanisms, similar to those of obesity per se, have been postulated to explain the role of energy intake and/or fat consumption in prostate cancer progression.

In the current study, we have used the Hi-Myc transgenic mouse model (21) of prostate cancer to explore the effects of dietary energy balance manipulation, including a standard diet-induced obesity (DIO) regimen, a control diet regimen that results in an overweight phenotype, and a calorie restriction (CR) regimen that leads to a lean phenotype, on prostate cancer progression. Given that approximately 68% of the adult population in the United States is currently overweight (22), a comparison of the effects and underlying mechanisms of a lean, overweight, and obese phenotype on prostate cancer progression was therefore evaluated. In the Hi-Myc transgenic mouse model, overexpression of c-Myc in the prostate is directed via the ARR₂Pb probasin promoter (21). Prostatic epithelial expression of c-Myc in the dorsolateral, ventral, and anterior prostate lobes reportedly results in complete penetrance of prostatic intraepithelial neoplasia (PIN) as early as 2 to 4 weeks of age, which progressed to locally invasive adenocarcinomas within 6 to 12 months of age (21). Recently, Kobayashi and colleagues (23) showed that if Hi-Myc mice were maintained on a low-fat (12 kcal%) versus a high-fat (42 kcal% fat) isocaloric diet, the transition from PIN to prostate cancer was delayed. This delay was associated with a decrease in prostatic Akt activity as well as a decrease in phosphorylation of the downstream targets p70S6K and GSK3β. More recently, increased NFκB signaling also was reported to play a possible role in prostate cancer progression in the Hi-Myc model (24). The current study shows that, relative to a control diet regimen that results in overweight mice, a DIO regimen significantly enhances the progression of prostate cancer in Hi-Myc mice whereas a 30% CR regimen significantly delays prostate cancer progression. Further studies indicated that the modulating effects of both DIO and CR (relative to the overweight control diet regimen) on prostate cancer progression in Hi-Myc mice were associated with alterations in Akt/mTORC1, STAT3, and NFκB signaling pathways as well as downstream inflammation–associated signaling pathways. The current data in Hi-Myc mice support the hypothesis that dietary energy balance modulates signaling through multiple pathways leading to altered prostate cancer progression.

Hi-Myc mice were obtained from the NIH MMRRC on an FVB/N genetic background. All mice for the current experiments were bred in-house. All diets were purchased from Research Diets, Inc. Mice were placed on a modified AIN76A semipurified diet (catalogue no. D12450B, fed ad libitum) at 5 to 7 weeks of age for a 1-week equilibration period and then randomized into the following 3 dietary groups for the duration of the study: (a) 30% CR diet (D03020702); (b) overweight control diet (continued on the modified AIN76A diet, # D12450B, 10 kcal% fat), fed ad libitum; and (c) DIO diet (60 kcal% fat; #D12492), fed ad libitum as previously described (25). The diet composition is shown in Supplementary Table S1. When fed in aliquots equivalent to 70% of the overweight control group's average daily intake, the CR diet (which was 14 kcal% fat, 29 kcal% protein, and 57 kcal% carbohydrate and supplemented with additional vitamins and minerals) provided 100% of the vitamins, minerals, fatty acids, and amino acids but only 70% of the carbohydrate calories relative to the overweight control group (which is 10 kcal% fat and 70 kcal% carbohydrate). The overweight control, 30% CR, and DIO diets are commonly used to induce a range of body size phenotypes in rodents (26–28).

Average body mass and food consumption were determined bimonthly for each dietary group. Groups of mice were terminated by CO₂ asphyxiation, and the complete urinary tract was taken at 3 and 6 months of age for histopathologic diagnosis and immunohistochemical (IHC) analyses. An additional set of mice from each group was used for protein and RNA analyses (6-month time point only). For all the studies, mice were housed in suspended polycarbonate cages or individually ventilated cages (Lab Products) on autoclaved hardwood bedding at room temperatures of 20°C to 22°C, relative humidity of 60% to 70%, and 14/10-hour light/dark cycle.

For histologic analyses, the male reproductive tract was removed intact from Hi-Myc mice at 3 and 6 months of age, fixed in 10% formalin, embedded in paraffin, and transversely sectioned. Sections of 4 μm were stained with hematoxylin and eosin (H&E) for histopathologic diagnosis. All histopathologic diagnoses of prostate lesions were based on published criteria (29–31). Hi-Myc transgenic mice used for the current study develop prostate lesions starting from PIN at 3 months of age or younger which progresses to invasive tumors by 6 months of age (21). Under our experimental conditions, Hi-Myc mice developed lesions primarily in the ventral and the dorsolateral prostate, with fewer lesions in the anterior prostate. Furthermore, at 6 months of age, more than 90% of the mice developed invasive tumors primarily in the ventral prostate. In contrast, at the same time point, the number of invasive tumors in the dorsolateral and anterior prostate was significantly lower. Lesions in the dorsolateral prostate showed a more variable consistency in the number of hyperplasia and PINs as well. Thus, we focused our analyses of diet effects on the ventral prostate gland which displayed a more homogeneous and consistent development of the lesions from hyperplasia to invasive adenocarcinoma within the 6-month time frame of our experiments.

IHC analyses were conducted on paraffin-embedded prostate tissue sections. Primary antibodies (all from Cell Signaling Technology) were used to detect phospho-Akt (Ser⁴⁷³; 1:50), phospho-mTOR (Ser²⁴⁴⁸; 1:50), phospho-S6 ribosomal protein (Ser^235/236), phospho-STAT3 (Tyr⁷⁰⁵), phospho-NFκB (p65, Ser²⁷⁶; 1:100), and cyclin D1 (DCS6; 1:100). Additional antibodies were used to detect the following markers: CD31 (1:100; Pharmingen Clone MEC 13.3), Ki67 (1:50; Santa Cruz), macrophages (RM0029-11H3, 1:200; Santa Cruz), and T-Lymphocytes, (CD3, 1:100; Serotec). Antibodies were detected with biotinylated secondary antibodies, followed by peroxidase-conjugated avidin/biotin (Vectastain ABC kit; Vector Laboratories) and 3,3′-diaminobenzidine (DAB) substrate (Dako). All IHC slides were scanned and digitalized using a scanscope system (Scanscope XT, Aperio Technologies), and quantitative analyses of IHC staining were conducted using the image analysis software provided (ImageScope).

Individual prostate lobes were excised and homogenized in lysis buffer [20 nmol/L Tris, pH = 7.5, 150 nmol/L NaCl, 1 mmol/L EDTA, 1 nmol/L EGTA, 1% Triton X-100, 2.5 nmol/L sodium pyrophosphate, 1 mmol/L β-glycerophosphate, protease inhibitor cocktail (Sigma-Aldrich)]. Fifty micrograms of cell lysate was electrophoretically separated on a 7% to 10% SDS-PAGE gel and transferred onto nitrocellulose membrane (BioRad Laboratories). Transfer was conducted at 20V overnight. After blocking with 5% skim milk in PBS containing 0.1% Tween, the membranes were incubated with antibodies as follows: Akt (1:1,000), phospho-Akt (Thr³⁰⁸ and Ser⁴⁷³; 1:500), mTOR (1:1,000), phospho-mTOR (Ser²⁴⁴⁸; 1:1,000), phospho-S6 ribosomal protein (Ser^235/236 and Ser^240/244; 1:1,000), 4EB-P1 (1:1,000), phospho-4EB-P1 (Thr^37/46; 1:1,000), phospho-GSK3β (Ser⁹; 1:1,000), phospho-FOXO3a (Ser²³⁵; 1:1,000), STAT3 (1:1,000), and phospho-STAT3 (Tyr⁷⁰⁵; 1:1,000). All of these antibodies for Western blot analyses were obtained from Cell Signaling Technology. β-Actin was used as a loading control (mouse monoclonal, 1:4,000; Sigma-Aldrich). Protein bands were detected using enhanced chemiluminescence (Pierce Biotechnology, Inc.). Quantitation was conducted with a G-Box system (Syngene) using the Genetool quantitation software.

The relative density of each protein band was normalized to the density of the corresponding β-actin band and where possible, phosphorylated proteins were presented as the ratio of phosphorylated/total protein.

Blood was collected by cardiac puncture immediately following CO₂ asphyxiation (12 mice per diet) and spun at 7,500 rpm for 7 minutes. Serum was then collected, flash frozen in liquid nitrogen, and stored at −80°C until analysis. Total mouse serum IGF-1 concentration was then measured using a 25 μL sample with an RIA kit (Diagnostic Systems Laboratories, Inc.). Levels of insulin, leptin, and resistin were measured using a 10 μL sample with a Milliplex MAP Mouse Serum Adipokine panel multiplexing Luminex assay (Millipore). Adiponectin levels were determined using a 10 μL sample with a Milliplex MAP Mouse Adiponectin single plex Luminex assay (Millipore).

Total RNA from the ventral prostate was extracted using a Qiagen RNeasy Protect Mini kit according to the manufacturer's protocol. The RNA was eluted in 100 μL of nuclease-free water and stored at −80°C until further use. The RNA concentration was determined from the absorbance at a wavelength of 260 nm (by using an OD₂₆₀ unit equivalent to 40 μg/mL of RNA).

RNA was analyzed for integrity using the Agilent 2100 Bioanalyzer (Agilent Technologies, Inc.). Total RNA (1 μg) was then used as template to synthesize cDNA with the High Capacity cDNA Archive Kit (Applied Biosystems) The resultant cDNA was stored at −80°C until further use. Quantitative PCR was subsequently carried out on the ABI 7900HT Fast Real Time PCR System (ABI) with assays on demand (AOD; specific to the relevant genes of interest). RNA levels were normalized to the endogenous control gene, glyceraldehyde-3-phosphate dehydrogenase (GAPDH), by using the ABI Mouse GAPDH Endogenous Control Kit (ABI; catalogue no. 4352339E). Data analysis were conducted using Sequence Detection System software from ABI, version 2.2.2. The experimental C_t (cycle threshold) was calibrated against the GAPDH control product. All amplifications were conducted in duplicate. The ΔΔC_t method was used to determine the amount of product relative to that expressed by calibrator sample–derived RNA (1-fold, 100%).

Differences in levels of serum markers (Table 1) were analyzed using the Wilcoxon rank-sum test. Difference in the incidence of prostate lesions (Supplementary Table S2 and Fig. 2) were analyzed using the Fisher exact test. Differences in subclassification of the adenocarcinomas (Supplementary Table S3 and Fig. 2) were also analyzed using the Fisher exact test. Quantitative differences in phosphorylation and mRNA levels were also evaluated using the Wilcoxon rank-sum test. Statistical analyses on the quantification of the immunohistochemistry (Supplementary Fig. S1A and S1B) were conducted using Wilcoxon rank-sum test. Significance was set in all cases at P ≤ 0.05.

Table 1 shows the average body weights, percentage of body fat, and serum profile of Hi-myc mice maintained on a 30% CR, overweight control, or DIO diet for 6 months. Weight distribution did not differ at the start of the study; however, the differences in body weight (i.e., body mass) among the 3 groups at the end of the study were statistically significant (P < 0.05). These values directly correlate with our previously published data obtained from FVB/N mice maintained on the same dietary regimens (25). In addition, previous studies using the same diet composition and feeding regimen used here have classified mice with body fat content values that correspond to those reported in Table 1 as lean, overweight, and obese (32, 33). Thus, Hi-Myc (FVB/N genetic background) mice effectively respond to these dietary manipulations, producing similar body phenotypes as has been observed in previous studies.

Feed and calorie consumption was significantly reduced in the 30% CR group. In addition, caloric consumption was significantly higher in the DIO group than in the overweight control diet group (147.2 vs. 114.7 kcal/wk, respectively). The levels of circulating insulin, IGF-1, and resistin were significantly different among all 3 diet groups (P < 0.05). The levels of leptin were similar between the overweight control and DIO diet groups but significantly reduced (P < 0.05) in the 30% CR group. Finally, a significant increase in adiponectin level was observed in the 30% CR group relative to both the overweight control and DIO diet groups (P < 0.05).

The effects of the 3 diets varying in caloric density on prostate cancer progression were assessed in Hi-Myc mice at 3 and 6 months of age. For these experiments, male mice (approximately 5–7 weeks of age) were maintained on the previously described dietary regimens for either 8 or 25 weeks. As noted in the Materials and Methods section, dietary energy balance effects were evaluated in the ventral prostate because of the greater variability and lower penetrance of lesions in the dorsolateral and anterior prostate lobes of Hi-Myc under the current experimental conditions. Histopathologic evaluations revealed that a variety of lesions were present at both 3 and 6 months of age including hyperplasia, PIN, and adenocarcinomas (Fig. 1). In this regard, all diet groups had approximately similar incidence of hyperplasia and low-grade PIN (lgPIN) at both 3 and 6 months of age (Fig. 2A and B and Supplementary Table S2). Dietary energy balance, however, affected the progression of premalignant lesions to malignant lesions in the ventral prostate. In this regard, 30% CR significantly reduced the incidence of in situ adenocarcinomas at 3 months of age compared with Hi-Myc mice maintained on the DIO diet. Furthermore, the incidence of in situ adenocarcinomas at 6 months of age was significantly reduced by 30% CR compared with both the overweight control and DIO groups. In addition, the incidence of adenocarcinomas with aggressive stromal invasion was increased in the DIO group relative to the overweight control group at both the 3- and 6-month time points; however, statistically significant increases were only observed at the 6-month time point (96% vs. 65%, respectively; P = 0.02; Fig. 2B). Notably, 30% CR completely suppressed the formation of invasive adenocarcinomas at both 3 and 6 months (see Figs. 2A and B and Supplementary Table S2).

The severity of invasive adenocarcinomas was also evaluated on the basis of differentiation characteristics at both 3 and 6 months using established criteria (31). These results are summarized in Fig. 2 and Supplementary Table S3. At 3 months of age, mice on the overweight control diet developed only well-differentiated adenocarcinomas. In contrast, mice on the DIO diet developed a significant number of moderately differentiated adenocarcinomas. Furthermore, at 6 months of age, both the overweight control and DIO diet groups had poorly differentiated adenocarcinomas, with a significantly greater (P = 0.02) incidence observed in the DIO diet group than in the overweight control diet group. Notably, there were no invasive adenocarcinomas in the 30% CR diet group. Thus, dietary energy balance significantly affected the incidence and severity of the adenocarcinomas that developed at both the 3- and 6-month time points examined in the current study.

IHC analyses were conducted on sections of the ventral prostate from 6-month-old Hi-Myc mice to evaluate potential diet-dependent differences in staining of phosphorylated levels of Akt, mTOR, S6 ribosomal protein, STAT3, and NFκB (p65). Proliferation (Ki67 and cyclin D1 staining), angiogenesis (CD31 staining), and inflammation (staining for macrophages and T-lymphocytes) were also assessed by immunohistochemistry. Relative to the overweight control group, DIO increased whereas 30% CR reduced staining for phospho-Akt (Ser⁴⁷³), phospho-mTOR (Ser²⁴⁴⁸), and phospho-S6 ribosomal protein (Ser^235/236; Fig. 3). Staining for both phospho-mTOR and phospho-S6 ribosomal protein was primarily cytoplasmic in all types of lesions. In contrast, staining for phospho-Akt was seen in cytoplasm, membrane, and nucleus with some areas exhibiting more intense nuclear staining as shown in Fig. 3. Similar diet-induced effects were observed for levels of phosho-STAT3, Ki67, and cyclin D1 (Figs. 3 and 4). CD31 staining was significantly increased in ventral prostate of Hi-Myc mice receiving the DIO regimen, indicating a substantial increase in vascularization (see Fig. 4). Changes in vascularization were further quantified as changes in microvessel density and mean vessel area. Significant differences were seen in both microvessel density and mean vessel area across the range of dietary energy balance (see Supplementary Fig. S1). Furthermore, DIO increased whereas 30% CR reduced the level of nuclear phospho-NFκB (p65). Staining for macrophages and T-lymphocytes also revealed a significant increase with DIO compared with both the overweight control and 30% CR groups. Macrophages were observed primarily in the stromal compartment surrounding adenocarcinomas, whereas T-lymphocytes were found in the stroma as well as infiltrating the adenocarcinomas. Quantitative analyses of the results from these IHC analyses are shown in Supplementary Fig. S1.

Western blot analyses were conducted on protein lysates from ventral prostate of Hi-Myc mice (6 months of age) maintained on the different diets to further analyze diet-induced changes in phosphorylation status of Akt, mTOR, and several Akt and mTORC1 downstream targets. In addition, the status of STAT3 tyrosine phosphorylation was evaluated. As shown in Fig. 5, phosphorylation of Akt (Thr³⁰⁸), mTOR (Ser²⁴⁴⁸), and S6 ribosomal protein (Ser^235/236) was significantly different (P < 0.05) across the 3 diet groups. Specifically, 30% CR reduced phosphorylation of these signaling molecules, whereas both the overweight control and DIO diets led to significant increases in phosphorylation of Akt, mTOR, and S6 ribosomal protein. Furthermore, 30% CR significantly reduced phosphorylation of Akt (Ser⁴⁷³), GSK3β (Ser⁹), FOXO3a (Ser²⁵³), 4EB-P1 (Thr³⁷), S6 ribosomal protein (Ser^240/244), and STAT3 (Tyr⁷⁰⁵) compared with both the overweight control and DIO diet groups (P < 0.05); however, no significant differences were observed between the overweight control and DIO diet groups. These Western blot data are consistent with the IHC analyses presented in Fig. 3 showing that the greatest differences in levels of phosphorylation occurred between the dietary energy balance extremes in this study.

In light of the data in Fig. 4 showing diet-dependent differences in CD31 staining and staining for inflammatory cells (macrophages and T-lymphocytes), we isolated RNA from the ventral prostate of Hi-Myc mice on the different diets at 6 months of age. Expression of a panel of selected genes associated with inflammation was analyzed via quantitative PCR. As shown in Fig. 6A, mRNA levels of interleukin (IL) 1α, IL7, IL23, IL27, NFκB1 (p50), and TNFα were significantly increased in the ventral prostate of mice maintained on the DIO diet relative to both the overweight control and 30% CR groups. IL1β mRNA levels were also increased in the DIO diet group but slightly lower than the overweight control group. Regardless, IL1β mRNA levels were elevated in both of these diet groups relative to the 30% CR diet group. The differences in mRNA levels in the DIO diet group for all the other genes evaluated were significantly different (P < 0.05) compared with both the overweight control and 30% CR diet groups.

In light of the dramatic diet-dependent differences seen in angiogenesis in ventral prostate of Hi-Myc mice, we also analyzed mRNA levels of VEGF family members. For these analyses, we quantified mRNA levels for VEGFA, B, C, D, and (placental growth factor) PLGF. As shown in Fig. 6B, mRNA levels for all VEGF family members were significantly (P < 0.05) upregulated in ventral prostate of Hi-Myc mice on the DIO diet compared with mice on both the overweight control and 30% CR diets. In addition, differences seen in the mRNA levels between the overweight control and 30% CR diet groups for all VEGF family members were also significantly different (P < 0.05). As shown in the figure, the difference between the DIO and 30% CR diet groups were particularly striking.

The present study was designed to investigate the impact and to identify potential mechanisms for dietary energy balance effects on prostate cancer progression in an established mouse model of prostate cancer (i.e., Hi-Myc mice; ref. 21). Diets of varying caloric density, which have been used in numerous studies to evaluate the effects of dietary energy balance on chronic disease (including cancer), were used and the resulting body weights and adiposity levels corresponded with previously published data using the same diets (32, 33). These previous reports classified mouse adiposity levels and provide a corresponding body phenotype as a function of both human body fat content and BMI classification. As such, the mice in the current study correspond to lean, overweight, and obese phenotypes. The DIO regimen used has consistently been shown to effectively induce an obese state in mice, enabling evaluation of the direct effects of obesity on prostate cancer as well as comparisons among obese, overweight, and lean phenotypes (26). While 30% CR significantly reduced body weight, adiposity, and levels of globally active circulatory proteins, differences in body composition and serum profiles between the overweight control and DIO diet groups were not as dramatic (although still statistically significant). Thus, the current study evaluated a range of body weight and adiposity phenotypes with the biggest differences occurring between the 30% CR and DIO diet groups. The data clearly show an effect of dietary energy balance on prostate cancer progression through changes in both growth factor and inflammation signaling pathways.

In the current study, a spectrum of prostate lesions was observed, most consistently in the ventral prostate of Hi-Myc mice across the 30% CR, overweight control, and DIO diet groups. The incidence of hyperplasia and lgPIN was similar across this spectrum of dietary energy balance in Hi-Myc mice at both 3 and 6 months of age. In contrast, dietary energy balance manipulation significantly affected the progression of premalignant lesions to malignant lesions. In this regard, although in situ carcinomas were observed in all diet groups, regardless of caloric consumption, the incidence of these lesions was significantly reduced (P < 0.05) in the 30% CR group relative to the overweight control at both 3 and 6 months of age. Furthermore, 30% CR completely suppressed the formation of invasive adenocarcinomas at both time points. DIO also significantly increased the incidence of adenocarcinomas with aggressive stromal invasion compared with the overweight control diet group at 6 months (96% vs. 65%, respectively; P = 0.02). The severity of invasive adenocarcinomas was also significantly affected by the different diets with more severe lesions (i.e., less differentiated lesions) observed in the DIO diet group than observed in the overweight control diet group at both the 3- and 6-month time points. Both IHC and Western blot analyses revealed that, relative to overweight control, DIO increased whereas 30% CR reduced activation of signaling through both Akt and mTORC1, as well as STAT3. Markers for proliferation (cyclin D1 and Ki67) and angiogenesis (CD3) also showed increased staining in tumors from the DIO group compared with the overweight control and 30% CR diet groups. Further analyses revealed a significant difference in inflammatory cell (macrophages and T-lymphocytes) infiltration in the ventral prostate of Hi-Myc mice maintained on the DIO diet compared with both overweight and 30% CR groups. Finally, significant elevations in inflammatory cytokine and VEGF family member gene expression were observed in the ventral prostate of Hi-Myc mice maintained on the DIO diet compared with both overweight and 30% CR diets. These changes were correlated with changes in expression of NFκB1 and increased expression of nuclear phospho-NFκB (p65). Collectively, the current data show that DIO enhances whereas 30% CR suppresses prostate cancer progression in Hi-Myc mice. Potential mechanisms for these effects involve altered growth factor signaling (Akt/mTOR and STAT3), accompanied by significant changes in inflammatory signaling, tissue inflammation, and angiogenesis.

A number of potential cellular and molecular mechanisms are known to contribute to prostate cancer progression (reviewed in ref. 34). One mechanism important to the current study involves activation of Akt signaling. Tumor suppressor gene phosphatase and tensin homolog (PTEN), a lipid and protein phosphatase, acts as a negative regulator of Akt and loss of PTEN function leads to constitutive activation of Akt (reviewed in ref. 35). Loss of one PTEN allele is a high-frequency event in prostate cancer, observed in up to 70% to 80% of primary tumors (36–39). Homozygous inactivation of PTEN is associated with advanced disease and metastasis (40, 41). PTEN heterozygous null mice (PTEN^+/− mice) develop PIN with a variable penetrance varying from 40% to 50% (42–44) to 90% (45), depending on genetic background. Lesions develop mainly in the anterior and dorsolateral prostate but are also observed in the ventral prostate. Progression to adenocarcinoma is normally not observed in PTEN^+/− mice, possibly related to age-dependent morbidity because of the high incidence of thymic lymphomas that occur in these mice. In contrast, conditional PTEN knockout mice with complete loss of PTEN in the prostate develop invasive prostatic carcinoma but with an extended latent period (45, 46). Inactivation of PTEN in combination with other mutations can promote prostate cancer progression in transgenic mice (47–50). As shown in Figs. 3 and 5 and Supplementary Fig. S1 of the current study, dietary energy balance modulated Akt and mTOR activity (i.e., phosphorylation) and downstream mTORC1 signaling in ventral prostate of the Hi-Myc mice that directly correlated with the diet-induced changes seen in prostate cancer progression in the Hi-Myc mice (highest with DIO and lowest with 30% CR). Furthermore, recent studies have shown that Akt activation can lead to activation of NFκB signaling via stimulation of IKKα (51). Studies in both mouse models and human prostate cancer have implicated deregulated NFκB signaling in mediating responsiveness to androgens, metastasis, and disease outcome (52–56). Jin and colleagues (24) also recently reported that upregulation of NFκB activity in Hi-Myc mice (via crosses with IKBα^+/− mice) led to androgen-independent growth suggesting that NFκB signaling may regulate (at least in part) prostate cancer progression. Thus, differential activation of the Akt/mTOR signaling pathways to downstream effectors, including NFκB, may have contributed to the effect of dietary energy balance on prostate cancer progression seen in Hi-Myc mice in our current study.

Recently, Kobayashi and colleagues (23) showed that the Hi-Myc prostate phenotype was responsive to dietary manipulation. They examined the effect of an increase in dietary fat on the development of prostate lesions. A modified pair-feeding regimen was used to ensure that the caloric intake was equivalent for the 2 groups. It is important to note that there were no differences in body weight between the 2 diet groups in this study. These investigators showed that if the Hi-Myc mice were maintained on a low-fat (12 kcal% fat) diet versus a HF (42 kcal% fat) diet, the transition from PIN to prostate cancer was delayed and there was a decrease in prostatic Akt activity as well as a decrease in phosphorylation of the downstream targets p70S6K and GSK3β. Although we cannot rule out a contribution of increased fat intake in the DIO diet (60 kcal% fat), the fact that the 30% CR mice had such a striking reduction in prostate cancer phenotype despite consuming a diet with a higher fat composition (14 kcal% fat) than the overweight control group (10 kcal% fat) argues against fat alone as the major contributor in our studies. Furthermore, previous work by us and others show that genetically obese mice or rats consuming the same AIN76A-based diet (i.e., greater consumption of the diet due to hyperphagia) have much higher tumor development than nonobese mice. In addition, studies of energy restriction accomplished by carbohydrate restriction (as is done here), protein or fat restriction, or total diet restriction all have similar effects in reducing tumor development in multiple models, suggesting that total energy, independent of the macronutrient manipulated, is the major determining factor in the anticancer effects of such restricted diets (57–60). For example, we previously compared the same 30% CR regimen used here with a low carbohydrate/high fat diet in a colon cancer model in C57BL/6 mice. Only the 30% CR diet, and not the low carbohydrate/high fat diet (which mimicked the induction phase of the Atkins diet), reduced adiposity and impacted colon tumor progression in that model (60). Taken together, these findings suggest that energy balance may be a stronger determinant of tumor progression than fat or carbohydrate intake.

Despite no changes in body weight in the study by Kobayashi and colleagues (23), their findings further illustrate the importance of the Akt/mTOR signaling pathway in prostate tumor formation and progression in the Hi-Myc model. They also support the role of fat, per se, independent of energy balance status, in modulating prostate cancer progression in this model. We previously showed that dietary energy balance manipulation modulated the activity of the Akt and mTORC1 signaling pathways in several normal tissues, including the prostate, in wild-type FVB/N and C57BL/6 mice (25). In these studies, we found that, relative to same overweight control diet used here, 30% CR resulted in a reduction, whereas DIO increased steady-state signaling through these 2 important signaling pathways in the prostate. These effects appeared to be due primarily to alterations in upstream signaling through the IGF-1 receptor (IGF-1R) and not through alterations in nutrient-sensing pathways such as AMPK (25). Thus, changes in Akt/mTOR signaling, as a result of dietary energy balance manipulation, would be expected to be present in ventral prostate of Hi-Myc mice both before and during prostate tumor development and tumor progression.

Evidence also points to activation of STAT3 in prostate cancer progression. In prostate cancer cell lines, STAT3 activation correlated with malignancy (61, 62). Inhibition of STAT3 in DU145 cells by antisense STAT3 oligonucleotides resulted in growth inhibition and apoptosis (61). Analysis of prostate adenocarcinoma specimens also revealed elevated levels of activated STAT3 that were positively correlated with a more advanced stage of tumors exhibiting higher Gleason scores (61). Thus, STAT3 seems to be involved in both proliferation and survival of prostate cancer cells as well as prostate cancer progression. We recently reported the development of a new mouse model of prostate cancer, based on overexpression of a constitutively active form of STAT3 (STAT3C) coupled with PTEN loss in prostate epithelial cells (63). The combination of constitutive STAT3 activation and elevated Akt activity led to progression of prostate tumors and further elevations in endogenous STAT3 activity in these mice. Notably, tumor progression in these mice was also accompanied by a dramatic increase in nuclear staining for phospho-NFκB (p65). The data with this new mouse model suggest that Akt/mTOR signaling cooperates with STAT3 signaling in prostate cancer progression. Thus, the diet-induced changes in STAT3 activation in ventral prostate of Hi-Myc mice seen in our current study, in concert with the changes seen in Akt/mTOR signaling, may cooperate in a similar manner in the effects of dietary energy balance on prostate cancer progression.

In this study, we also found significant diet-associated changes in tissue inflammation both at the cellular level (altered inflammatory cell infiltration and angiogenesis) and at the level of altered inflammatory gene expression that correlated with the progression of prostate cancer in the Hi-Myc mice. The links between obesity and inflammation and between chronic inflammation and cancer have been well described (64, 65). Chronic, low-grade systemic inflammation typically accompanying obesity is associated with 2- to 3-fold increases in the circulating levels of the cytokines TNFα, soluble TNF receptor, IL1β, IL6, IL-1 receptor antagonist, and C-reactive protein (64). The source of these elevated cytokines may be adipocytes or immune cells such as macrophages (64). The diet-induced changes in tissue inflammation seen in Hi-Myc mice could be due to changes in both circulating levels of inflammatory mediators and increased local production by infiltrating macrophages and T-lymphocytes as well as prostate epithelial cells and prostate tumor cells. The fact that both STAT3 and NFκB were activated to a greater extent in ventral prostate of Hi-Myc mice maintained on the DIO diet compared with the overweight control and 30% CR diet groups suggests that local production of inflammatory cytokines is likely an important driving force for obesity-induced tumor progression. Both STAT3 and NFκB regulate expression of a number of inflammatory cytokines as well as VEGF family members (66). The inflammatory genes that we analyzed in this study include a subset of genes that are regulated by both STAT3 and NFκB (66). Because both activated STAT3 and activated NFκB were observed in the epithelial cells of the ventral prostate, including adenocarcinoma cells (Figs. 3 and 4), it is likely that some of the tissue inflammation seen in ventral prostate is due to increased production of inflammatory cytokines from these cells. Recruitment and activation of macrophages and subsets of T-lymphocytes as well as increased numbers of adipocytes in the prostate gland that also produce a battery of proinflammatory cytokines would create a positive proinflammatory loop that may have facilitated prostate cancer progression. Overall, the data we have generated in this study suggest that the significant tissue inflammatory response seen in the DIO group relative to the overweight control and 30% CR diet groups may play an important role in the effects of dietary energy balance on prostate cancer progression in the Hi-Myc mouse model. Further ongoing work in this area is aimed at identifying which inflammatory mediators might be most important in this regard.

In conclusion, we have shown that dietary energy balance modulates prostate tumor progression in an established mouse model of prostate cancer based on overexpression of c-myc in prostate luminal cells. DIO enhanced whereas 30% CR reduced prostate cancer progression. While the exact mechanism(s) for these effects of dietary energy balance on prostate cancer progression in Hi-Myc mice remain to be fully elucidated, several possibilities were discovered including altered Akt/mTOR, STAT3, and NFκB signaling. Some or all of these changes may have contributed to the dramatic changes in tissue inflammation in the ventral prostate of Hi-Myc mice. The ultimate goal of our studies is to interrogate the impact of dietary energy balance manipulation in the Hi-Myc mouse model (and possibly other mouse models of prostate cancer) at different stages of carcinogenesis to illuminate key signaling pathways that may ultimately represent targets for reversing the effects of obesity on prostate cancer progression in humans. Ultimately, it is hoped that studies in mouse models of prostate cancer will accelerate our knowledge of the effectiveness of lifestyle and/or pharmacologic interventions that may offset the effects of obesity on prostate cancer progression. In this way, we hope to rapidly translate findings from the bench to bedside.

No potential conflicts of interest were disclosed.

The authors thank Nancy Otto and John Repass for their outstanding technical assistance in the IHC analyses and quantitative PCR analyses, respectively.

This study was supported by NIH grant P50 CA140388, Start-up funds from the University of Texas at Austin, NIEHS Center grant ES-07784, and University of Texas MD Anderson Cancer Center Support grant CA16672.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.