Chemoprevention of Human Prostate Cancer by Oral Administration of Green Tea Catechins in Volunteers with High-Grade Prostate Intraepithelial Neoplasia: A Preliminary Report from a One-Year Proof-of-Principle Study

The incidence of prostate cancer (CaP) is steadily increasing in the U.S. and Europe. Actually, it has become the second leading cause of cancer-related deaths among men in western countries, thus representing a major (and growing) health and social problem. When truly organ-confined, radical prostatectomy or radiation therapy are the therapeutic approaches of choice, but after it has spread to local and distant sites, hormonal therapy remains the most generally used chemotherapy for this disease. However, in nearly all men, advanced CaP eventually becomes refractory to hormonal therapy, which results in cancer recurrence. Because of the unfavorable prognosis of high-grade organ-confined and extraprostatic CaP, early detection at potentially curable stages makes sense, however, screening has never been shown to decrease CaP mortality. On the other hand, the fact that CaP onset and progression takes considerable time to occur can be considered as an important opportunity for treating premalignant lesions. Thus, at present, prevention may be the best approach to fight this frequent disease.

Lifestyle-related factors, particularly the diet, are considered to be the major contributors to CaP promotion. Although clinical evidences are still rather sparse and not definitive, several epidemiologic studies have focused on the lower incidence of CaP in Asian countries where green tea is consumed regularly (1), as compared with western populations. Moreover, the risk for CaP returns in Asian immigrants to the U.S. if original dietary habits are abandoned (1). Recently, a case-control study conducted in China showed that green tea consumption is etiologically associated with CaP, suggesting the protective effect of green tea against this disease (2). This body of evidence has suggested that administration of biologically active compounds from green tea might be effective in lowering the incidence of CaP.

It is well known that the active compounds released and found in highest amounts in the dry matter of green tea infusion are catechins, the most common of which are (−)-epigallocatechin-3-gallate (EGCG), (−)-epigallocathechin (EGC), (−)-epicatechin-3-gallate (ECG), and (−)-epicatechin (EC). Green tea catechins (GTCs), and especially EGCG, have been shown to be potent chemopreventive agents in vitro and in many in vivo animal models of induced carcinogenesis (3, 4). The systematic study of the biological and biochemical properties of GTCs only started quite recently, searching for possible molecular explanations for their effect on cancer cells. Endocrine changes occurring upon GTCs or EGCG administration (5), and/or inhibition of 5-α-reductase (the prostatic enzyme transforming testosterone into the more active androgen, 5-α-dihydrotestosterone) have been often suggested as key events capable of inhibiting CaP burden (6, 7). We previously showed that apoptotic cell death was specifically induced by EGCG in both SV40-immortalized (PNT1A) and tumorigenic (PC-3) human prostate cells by the activation of caspase-cascade without any significant effect on benign controls (normal human prostate epithelial cells in primary culture from cystectomy; ref. 8).

The transgenic adenocarcinoma of the mouse prostate (TRAMP) mouse model is a well known in vivo animal model for CaP displaying in situ and invasive carcinoma of the prostate that mimics the whole spectrum of human CaP progression from prostate intraepithelial neoplasia (PIN) to androgen-independent disease (9, 10). Recent studies have shown that GTCs administration can actually prevent CaP development in this model (11). We confirmed this finding, showing that p.o. administration of GTCs to TRAMP mice reduced CaP incidence at 24 weeks from 100% to 20% without any side effects (8), and suggested that GTC's action might be mediated by the induction of the expression of Clusterin (CLU). The CLU gene is potently up-regulated during prostate gland involution (12), but is down-regulated in human CaP specimens (13, 14) and, as previously reported by us, in PNT1a and PC-3 cells, it exerts antiproliferative (15) and proapoptotic activities (16–19). In the prostate glands of TRAMP mice responding to GTCs treatment, CLU expression was maintained at high levels, and, shortly before the induction of casp-9 expression/activation, and concomitantly to decreased expression of histone H3 mRNA (a specific marker of cell proliferation; ref. 8), localized to the nuclei, in which it is known to exert a proapoptotic role (17–19). These data, altogether, confirmed that GTCs exert potent and selective in vitro and in vivo proapoptotic activity on prostate cancer cells.

Nevertheless, no definitive clinical data demonstrating the efficacy of GTCs as chemopreventive agents in humans have thus far been produced. Thus, clinical studies for the evaluation of safety and effectiveness of these agents in cancer chemoprevention, both individually and in combination, are needed (20). Our preclinical results prompted us to do a proof-of-principle clinical trial to assess the possible efficacy of GTCs for the chemoprevention of CaP.

In order to obtain this information as quickly as possible, the study was done in a selected population of 60 human volunteers with high-grade PIN (HG-PIN), the main premalignant lesion of CaP (21), known to result in a substantial number of cancers in a 1-year period (22, 23). Because, at present, no treatment is given to these patients until CaP is diagnosed, our study can be envisaged as an attempt to fill this therapeutic void. The primary end point of this study was to determine a possible difference in the prevalence of CaP in the GTCs-treated arm in comparison to placebo. Possible changes in total serum prostate-specific antigen (PSA) values, as well as possible variations in lower urinary tract symptoms (LUTS), as assessed by International Prostate Symptom Score (IPSS; ref. 24) and quality of life score (QoL; ref. 25) in men with coexistent benign prostate hyperplasia (BPH) were also pursued as secondary observations.

Study design. The study was conducted on volunteers with HG-PIN. At present, in Italy, this condition is clinically managed simply by performing saturation biopsies of the prostate every 3 to 6 months, in search of possible coexisting CaP (21, 26). Recent studies have quantified the risk for invasive prostate cancer in men with HG-PIN, and it was suggested that the prevalence of CaP was as high as 30% within 1 year after repeated biopsy (22, 23). These data suggest that in the near future, therapy of HG-PIN may become a useful approach for inhibiting the development of CaP. Clinical benefits in case of effective treatment would include reduced morbidity, enhanced quality of life, delayed surgery or radiation, and increase in the interval for surveillance requiring invasive procedures (26).

To investigate whether the administration of GTCs could prevent malignancy in men at high-risk, 60 volunteers (Caucasian men) bearing HG-PIN lesions, to whom no therapy is commonly given, were enrolled in this double-blind, placebo-controlled study. Because pure HG-PIN is a rather rare finding, recruitment time extended up to 18 months to reach the appropriate number of subjects for each arm. Volunteers were randomly assessed to a placebo- or GTCs-arm by simple randomization. More precisely, subjects were called for an informative interview within 2 weeks from the time of diagnosis and asked to join the study by signing the informed consent. That same day, they were alternatively assigned to the placebo- or GTCs-arm and given the appropriate treatment. To all subjects, capsules were given by the urologist according to the double blind method. Compliance with study medication was assessed by pill count/returned blister packs. Due to the enthusiastic adherence to the study by highly motivated subjects at high risk for CaP, all volunteers took the complete medication assigned in all cases, thus the compliance was very good in both arms and did not differ between the two groups. The GTCs-arm volunteers received three capsules per day containing 200 mg each of GTCs, for a total of 600 mg of GTCs per day. GTCs were given as a high-quality preparation whose content was determined by high-performance liquid chromatography as previously published (8). Content was as follows: EGC, 5.5%; EC, 12.24%; EGCG, 51.8%; ECG, 6.12%; total GTCs, 75.7%; virtually caffeine-free (caffeine <1%). In the second arm, men received placebo (three identical capsules per day). The primary end point was the prevalence of CaP during the 1-year study in the two arms. Although the study was specifically designed for assessing the possible chemopreventive action of GTCs, we also recorded possible changes in total serum PSA values during the whole study, and possible variations in LUTS, as assessed by IPSS and QoL scores before GTCs administration, and after 3 months of treatment in a subset of men with coexistent BPH, not receiving any other therapy. In case of cancer diagnosis, subjects were excluded from the chemoprevention trial (failure of chemoprevention) and recommended to clinical management.

Eligibility criteria. Men ages 45 to 75 years old with HG-PIN found after the collection of 8 to 18 needle biopsies according to prostate volume. All subjects were volunteers, who were made aware of the study details, and agreed to sign an informed consent form.

Exclusion criteria. Men aged >75 years, consuming green tea or derived products, vegetarians, taking antioxidants, and following antiandrogenic therapy. Patients diagnosed with cancer were excluded from the chemoprevention trial (failure of chemoprevention) and recommended to clinical management.

Clinical trial results and follow-up studies. Medical history, physical examination, and evaluation of total PSA were carried out every 3 months following the recruitment date. Six and 12 months after the beginning of the study, each subject had prostate mapping by 8 to 14 core needle biopsy examination, the number chosen according to prostate volume. Subjects exhibiting a sudden increase of total PSA levels, alterations of other clinical variables or symptoms of prostate disease (referred during physical examination) underwent needle biopsy earlier, at the physician's discretion. During the whole study, the urologists were continuously in contact with volunteers to detect any possible adverse or side effects.

Total serum PSA determination. All total serum PSA determinations were carried out in the same central laboratory of the hospital by using a two-site immunoluminometric in vitro commercially available assay kit for quantitative determination of human prostate–specific antigen in human serum and plasma.

Statistical analysis. For PSA, data are expressed as mean values ± SD. For IPSS and QoL, data are expressed as mean values. Statistical significance was calculated by the Student's t test and P values are indicated with 95% confidence. Multivariate analyses of variables in Table 1 were done with the one-way MANOVA test.

Primary end-point: prevalence of prostate cancer. Table 1A and B shows the complete list of individual variables (a, age of subject at the time of diagnosis; b, total serum PSA at the time of enrollment; c, prostate volume at the time of enrollment; d, prostate volume at the end of study; e, total number of HG-PIN cores versus total cores taken at the time of enrollment; f, total number of HG-PIN cores taken at the end of study; total number of monofocal or plurifocal HG-PIN lesions) of 30 volunteers of placebo-arm (A) and 30 volunteers of GTCs-arm (B). The summary of such data (mean values ± SD), are also provided (C). Statistical analysis (t test analysis) showed that none of the variables considered were significantly different in the two arms of the study with 95% confidence (age, P = 0.670; PSA, P = 0.819; prostate volume at the time of enrollment, P = 0.862; prostate volume at the end of study, P = 0.756). Also, prostate volumes at the enrollment were not significantly different from those at the end of the study in the placebo-arm (P = 0.813) or the GTCs-arm (P = 0.427). In addition, multivariate analysis using all the above variables at the time of initial diagnosis showed no significant differences in the two arms with a 99.99% confidence. Thus, none of these variables may account for the difference in prevalence of CaP in the two arms of the study, which are reported in Table 2. After 1 year of treatment, only one cancer was diagnosed among the 30 men that received GTCs daily, with a final incidence of about 3%; instead, nine cancers were found among the 30 men treated with placebo, with a final incidence of 30%, a figure close to that expected (21, 22). In particular, in the placebo-arm, six cancers were found 6 months after the recruitment (at the 6-month biopsy check), and three more cancers were found at the 12-month biopsy check (end of the study). The only cancer found in the GTCs-arm was detected at the 12-month biopsy check. Altogether, this suggests a 90% chemoprevention efficacy of GTCs in men subjected to high risk for developing CaP (Table 2). Statistical analysis showed that this result was highly significant (P < 0.01; Table 2).

Secondary observation: total serum PSA values. It is known that PIN lesions do not significantly elevate serum PSA per se. In the cohort of 60 men enrolled in the study, total serum PSA ranged from 0.70 to 35.70 ng/mL (mean, 7.7 ng/mL) at the recruitment time. In the 30 men of the placebo-arm, PSA varied from 1.15 to 35.70 ng/mL (mean, 7.97 ng/mL), whereas in those of the GTCs-arm, it ranged from 0.70 to 15.85 ng/mL (mean, 7.57 ng/mL). Although total serum PSA value was not taken into consideration for assigning volunteers to the placebo- or GTCs-arm, the mean PSA values was very similar in the two arms at the beginning of the study (Table 1C). GTCs treatment did not significantly affect PSA values throughout the study, probably because of high individual differences reflected by different total range between placebo- and GTCs-arm (Fig. 1). In any case, it may be worth noticing that the mean value of total PSA was always lower in the GTCs-arm at any time point with respect to control, and a trend toward a more stable total PSA value was clearly evident in GTCs-treated men.

Secondary observation: changes in LUTS as assessed by IPSS and QoL scores. In the cohort of volunteers, 18 out of 30 in the placebo-arm and 17 out of 30 in the GTCs-arm had BPH and were suffering LUTS. We quantified LUTS at the beginning of the study and after 3 months of treatment by assessing both IPSS (27) and QoL (25) scores. Volunteers agreed not to undergo any therapy for LUTS during the 3-month subtrial. Changes in LUTS are shown in Table 3. A decrease, small but significant (P < 0.05), in IPSS score was found in GTCs-treated arm for 3 months as compared with placebo. Improvement of IPSS was found in 65% of GTCs-treated men. Also, QoL score decreased in 35% of the men in the study following GTCs treatment, reaching close to statistical significance (P = 0.08), whereas no changes were found in placebo-treated men.

Medical events and side effects. No significant side or adverse effects were documented throughout the whole study as reported during physical examinations that were done every 3 months. In the study, only two cases of diarrhea on each arm were reported and rated as very mild disorders. These rare events have been considered to be unrelated to GTCs administration because the incidence reported in the GTCs-treated arm was identical to that reported in the placebo arm (two versus two), and they disappeared spontaneously within 3 days, thus allowing study completion for these subjects.

To our knowledge, this is the first study showing that GTCs have potent in vivo chemoprevention activity for human CaP. Altogether, our data suggest that up to 90% of chemoprevention efficacy can be obtained by GTCs administration in men prone to develop CaP. Thus, administration of GTCs could be an effective therapy for treating premalignant lesions of high-risk subjects, thus filling a therapeutic void (21) by taking advantage of an important window of opportunity for treatment before CaP develops. If confirmed, our finding suggests a new scenario in which the incidence of this disease could be greatly reduced by simply making GTCs available to the elderly or men at high-risk, resulting in a tremendous social and clinical impact, especially in the Western countries. The fact that no side or adverse effects have been reported confirm that GTCs, at least at the dosage used here, are safe in humans. This was also shown previously (28). In fact, in preparation for future trials, a study was conducted to determine the safety and pharmacokinetics of green tea polyphenol administration by using pure EGCG or Polyphenon E, a defined, decaffeinated green tea polyphenols mixture with a composition very similar to our GTCs preparation (28). Preliminary clinical trials showed that Polyphenon E is effective in the prevention of human papilloma virus–infected cervical lesions (29).

It is worth noticing that, as a secondary observation, we found a small but statistically significant improvement in LUTS in 65% of GTCs-treated volunteers with coexistent BPH as assessed by IPSS. Improvements were also found with regard to QoL score in the same men. This suggests that GTCs administration might also be of help to relieve the symptoms of BPH, although the possible mechanisms of action of GTCs on benign prostate diseases is still unknown. Nevertheless, because this result was not accompanied by a significant reduction of prostate volume (Table 1C), it seems unlikely to be related to a possible antiandrogenic action. Unfortunately, due to the fact that our study was not specifically designed to investigate the possible effects of GTCs on BPH, no further data are available on this cohort of men. Thus, the positive effect on LUTS recorded here should just be considered a rather promising observation for future research suggesting that other potential benefits may accompany GTCs administration.

It seems rather obvious that the chemopreventive effect exerted by GTCs on CaP development must be quickly confirmed by a larger study. Considering that all volunteers enrolled in this study were Caucasians, it would be particularly important to check whether GTCs treatment is also effective in high-risk men with different genetic backgrounds. Although follow-ups will continue for up to 5 years in the cohort of subjects studied here, a larger confirmatory study extending GTCs administration for up to 5 years would allow us to understand whether CaP onset could be definitively prevented or simply delayed by the treatment, and to exclude possible negative effects caused by long-term GTCs treatment. In addition, comparison of the histologic features of the tumors diagnosed in the two arms will also permit us to verify the possible effects of GTCs on cell differentiation, clinical stage and aggressiveness of CaPs, goals that could not be achieved in the present contribution. The importance of this issue is supported by a recent report showing that Finasteride was effective in reducing the prevalence of CaP by 28.8% over a 7-year period, but apparently, cancers with Gleason grade from 7 to 10 were more common in the Finasteride arm than in the placebo arm (30). The number of cancers found at the end of the trial were too small to evaluate the effect of GTCs on cancer grade. We still believe that knowing whether or not long-term GTCs administration affects the grading and staging of CaPs is crucial information that is needed and could possibly be obtained with a second confirmatory study employing a much larger number of volunteers.

Grant support: In part by PRIN 2004 (Miur, Italy); Dr. Rizzi was supported by Genprofiler Srl (Bolzano, Italy).

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.

We are indebted to the 60 volunteers who participated in this study. We thank Dr. Daniel Remondini, Dipartimento di Morfofisiologia Veterinaria e Produzioni Animali, Università di Bologna, Italy, for statistical and multivariate analysis.