Can Dietary Phytoestrogens Slow Down Prostate Tumor Proliferation?

Data from ecologic and experimental studies clearly show a protective effect of dietary phytoestrogens against prostate cancer. Genetic factors are also of etiologic importance in prostate cancer and there is growing evidence for the importance of a gene-diet interaction in prostate cancer progression. The investigators have recently found a putative genetic interaction for this protective role of phytoestrogens. The overall decreased risk of prostate cancer in men with a high intake of phytoestrogens was strongly modified by a nucleotide sequence variant in the estrogen receptor-beta (ERß) gene. An American study found a similar interactive effect between intake of phytoestrogens, as well as body mass index (BMI), and another nucleotide sequence variant in ERß.The phytoestrogens isoflavonoids and coumestans bind tightly to the estrogen receptor-beta (ERß) and are mainly found in soy and other beans. ERß expression has been found to be involved in progression of prostate cancer, suggesting that phytoestrogens may interact with ERß in the development of prostate cancer. With regard to potential carcinogenic mechanisms, phytoestrogens may be involved in the endocrine control of prostate cell growth by influencing the balance between AR and ERß.Testosterone and its metabolite 5α-dihydrotestosterone (DHT) cause proliferation of prostate epithelial cells through binding to the AR. In contrast, by binding to ERß, 5α androstane-3ß,17ß-diol (3ßAdiol), a metabolite of DHT, represses the expression of AR and thereby inhibits androgen-driven proliferation while promoting cell differentiation. In terms of proliferation, current data suggests a combined stimulatory role of ERα and AR in the prostate whereas ERß inhibits proliferation and stimulates differentiation. Both estrogen ERα and ß have affinity for estradiol whereas phytoestrogens and 3ßAdiol selectively activate ERß. Phytoestrogens should, as a result, be able to restrict cancer growth by acting as a substitute for 3ßAdiol. This is confirmed by experimental studies.

Results from previous intervention studies have given promising results indicating that certain dietary or lifestyle modifications can influence the progression of prostate cancer, although most of these studies have been of fairly short of duration and based on relatively small numbers of patients. For example, a randomised study showed that men who were given a dietary addition of flaxseeds had a lower percentage of positive cells containing the proliferation marker Ki-67 than men who had not received this addition to their diet. Taken together, the present evidence points towards the possibility for prostate cancer patients to complement with lifestyle-related treatment alternatives to prevent or delay tumor progression.

Specific aim: To perform a controlled randomized dietary intervention study in men with prostate cancer. The investigators will pursue the following specific aims/hypothesis:

1. "Among men with prostate cancer of intermediate risk-level, a diet high in phytoestrogens will reduce tumor progression compared to men with a diet low in phytoestrogens".
2. "The effects of a diet high in phytoestrogens on prostate cancer tumor progression differ between men with different genotypes of polymorphisms in the ERß gene". The men will be divided into two subgroups, those bearing TT or TC/CC alleles, of the single-nucleotide polymorphism (SNP) rs2987983-13950 T/C in the ERß-gene, in which the hypothesis will be tested.
3. To determine whether RNA expression of the AR, ERα and ERβ in prostate tumor tissue, as well as levels of steroid hormones, differs between men with a diet high and low in phytoestrogens, respectively, or between different genotypes of the ERβ gene.
4. To determine whether mRNA expression of genes involved in proliferation and ER signal pathways analyzed by whole transcriptomic profiling in prostatectomy specimens differs between men with a diet high and low in phytoestrogens, respectively, or between different genotypes of the ERβ gene.

Study protocol: Trough treating physicians, the investigators will identify 240 men in the Västra Götaland healthcare region who have been diagnosed with prostate cancer of intermediate risk-level (T1-T2, Gleason score <8, PSA<20)scheduled for radical prostatectomy. The clinic will be informed of which patients have been enrolled in the study, and the surgery for these patients will be scheduled six weeks ahead of time from the date of inclusion. Once a patient has agreed to participate, he will be randomized to one of the two intervention arms. The patient will then receive additional information depending on which group he belongs to and will be asked to donate a blood sample. Under normal clinical routine, data on all prostate cancer cases is collected in the National Prostate Cancer Register (NPCR), for example information on Gleason score, tumor volume, Prostate-specific antigen (PSA) levels. The investigators will collect information from the registry in connection with the analysis of data.

Intervention: Patients will be introduced to the intervention diet by a dietitian and will receive general information about healthy food choices (according to recommendations for the general public) and not to eat any dietary supplements, no other dietary restrictions will be given. The patients will be given a package containing food with a high amount of phytoestrogens (≥100 mg isoflavonoids and ≥100 mg lignans) that are to be consumed daily. The intervention continues until the day before the surgery (at least 6 weeks). The patients in the control group will be informed in the same way as the intervention group except that they will not receive any food-package.

Endpoint: The effect of the diet on tumor proliferation rate will be measured by making comparisons of 1) percentage of positive cells containing the marker Ki-67, 2) PSA-levels, 3) expression of genes involved in proliferation and ER signal pathways and cell-cycle progression (CCP) gene expression. In immediate adjacency to the surgery, the surgeon collects four median-needle biopsies from the fresh prostate specimen, two from the tumor site and two from healthy prostate tissue, based on previous results from the diagnostic biopsies and magnetic resonance imaging.Also at the surgery, the radical prostatectomy tissues are collected and formalin fixed paraffin embedded. Blood samples will be taken at the beginning of the study, at the day before the surgery, and send to analysis of PSA (free and total). The absolute differences in PSA levels and the PSA doubling time will be calculated. In addition the investigators will measure the expression of AR, ERα and ERß in tumor tissue and the blood levels of hormones, testosterone, DHT, estradiol and 3ß-Adiol.

Questionnaire: All patients will answer a password-protected questionnaire at baseline, surgery, weight and height will be measured. Once during the 6-weeks period a 24-h recall interview will be performed, in which participants food intake during the preceding 24 h will be registered. The individual intake of energy and specific nutrients will be estimated by linking information from the questionnaire and diet record to the National Food Agency's nutritional database and our previously developed phytoestrogen database. Information on the possible use of Finasteride, a 5α-reductase inhibitor, will be collected from medical records.

Concentrations of phytoestrogens in blood: In a proportion of participants, plasma concentrations of different phytoestrogens (daidzein, enterodiol, enterolactone, equol, genistein, glycetein, lariciresinol, and secoisolariciresinol) will be analyzed, using LC-MS/MS measurements performed on microLC 200 series (Eksigent/AB Sciex, Redwood City, CA, USA) and QTrap 5500 mass spectrometer (AB Sciex, Framingham, MA, USA).

Analysis of polymorphs:SNP in the ERß-gene will be analysed in blood samples, using the PCR-based method competitive allele specific PCR (KASP™). The allele-constitution will be identified and participants will be assigned to group TT, TC or CC. Analysis of Ki-67: The biopsy sample is fixed by immersion with 4 % paraformaldehyde in 0,1 M phosphate buffer during 24hours at +4°C. Then it is dehydrated with graded ethanol followed by xylene and paraffin embedding. Paraffin sections are heated for 30 minutes at 60ºC, pre-treated with citric acid buffer. Unspecific binding is blocked by donkey serum in phosphate buffered saline (PBS) containing Triton X-100 for 30 minutes. Incubating with Ki-67 antibody over night (4ºC) followed by incubation with the appropriate secondary biotinylated antibody. To enhance the signal, the sections are treated with avidin-biotin-complex solution, and to visualize the immunoreactivity Diaminobenzidine (DAB) is used. Stained tumor cells (Ki-67 positivity) will be counted and the result will be reported as the ratio of positive nuclei divided by the total number evaluated × 100.

Analysis of receptor-expression, CCP score, hormone levels and PSA: Tissue samples taken at surgery is placed in RNA-later before being frozen. The samples will be analysed using RNeasy plus Universal Mini Kits (QIAGEN), and Real-time polymerase chain reaction (PCR). ERα, ERβ and AR mRNA expression will be determined using TaqMan assay. Total RNA will be extracted from formalin-fixed paraffin-embedded sections of tumors and benign tissues using the global transcriptome-wide expression array Clariom ™ D (Thermos Fischer). For analysis of the CCP score, gene expression of the genes involved in CCP based on the Prolaris® gene panel and the Decipher Score, respectively.. PSA-blood levels will be analysed according to standard clinical protocol.

Power calculation: In the Västra Götaland healthcare region in 2012, 321 men had intermediate risk-level prostate cancer for which they received curative treatment. Calculation using the primary outcome (Ki-67) gives for a study group consisting of 118 patients an 80 % power for a two-sided test with a level of significance at 0,05 and effect size of 0,5. The investigators found in earlier studies that approximately 42 % of the male population are heterozygous or homozygous for the variant allele (TC/CC) of the ERβ promoter region SNP (rs2987983-13950 T/C) and 58 % homozygous for the wild type allele (TT). Since the investigators expect the effect of the intervention to be greater among subjects with the variant allele, while smaller among subjects with the wild type allele, this latter group ought to be large enough to be able to find a difference if one exists. Thus a total sample size of 118/58%=203 patients is needed.

Statistical Considerations: The investigators will utilize the intention-to-treat method. In the design and data collection testing of drugs as the template have been made (randomization, placebo, blinding, no attrition, no differential measuring errors, correct analysis); for the deviations from the perfect situation the investigators use epidemiological theory for guiding the analyses and interpreting the results. Preliminary analyses, based on means, medians, standard deviations and interquantile range will be provided to properly describe the outcomes of interest and stratification will be performed to assess the performance of the randomization of the assigned diet. Box-plots will therefore be produced to graphically evaluate the temporal trends in the response. The distribution of the outcomes will be studied and appropriate transformations will be applied, if needed, to improve symmetry and normality. Due to the longitudinal structure of the data, correlation among observations will be incorporated and linear mixed effects models will be carried out to study if (i) there is a temporal trend (ii) if the linear trend differs between treated and untreated individuals (iii) if the treatment effect between the intervention group and the control group is modified according to ERß genotypes, both for the common and the variant. A key step in analysis of longitudinal data is to identify the appropriate covariance structure that describes the correlations among the data points: independent covariance (no correlation), exchangeable covariance (equal correlation) or autoregressive covariance with correlation between responses decreasing with time between measurements. The investigators will identify the most appropriate structure, verified empirically with help of an information theoretic approach. Both statistical tests and confidence intervals will be produced to assess the significance of the effects, partial F test procedures will therefore be used to study the joined effect of main and interaction variables and appropriate statistical methods will be applied to study model fitting. The association between phytoestrogens intake and cancer proliferation will be evaluated by generalized linear models, which will provide estimates of the risk difference (RDs) and corresponding 95% CIs, stratified by ERβ genotypes. Interactions between phytoestrogen intake and ERβ SNPs on proliferation will be evaluated considering additive effect scales. Group of phytoestrogen intake are included as a continuous variable, and each SNP are represented by an indicator variable (variant or not). Interaction will be assessed in a linear odds model by the product term between the covariates representing phytoestrogen intake and SNP genotypes.