- Genitourinary cancers: prostate
In this issue of Clinical Cancer Research, Khor et al. (1) investigate Bcl-2 and Bak as potential predictors of clinical outcome in prostate cancer patients enrolled in the Radiation Oncology Group Protocol 92-02.
Prostate cancer is the most commonly diagnosed cancer in men in the United States with more than 218,000 cases detected each year (2). Although widespread serum prostate-specific antigen (PSA) screening has enabled detection of the disease at earlier stages (3–6), prostate cancer still accounts for more than 27,000 deaths each year (2). Radiation therapy or surgery is the standard treatment and is effective for most, but not all (7, 8). Men who have locally advanced prostate cancer at diagnosis, defined as stage T2c-4 but with PSA <150 ng/mL, are particularly at risk for treatment failure (9). Consequently, this is a subset of patients for whom hormonal manipulation is often combined with radiation therapy to try to achieve the best chance of cure. Hormonal ablation is most commonly achieved by administering an antiandrogen (e.g., flutamide) and a gonadotropin (e.g., goserelin), which together achieve androgen deprivation and apoptosis of the cancer cells. This strategy is not without cost, as androgen deprivation is associated with erectile dysfunction, osteopenia (10, 11), and hot flashes (12), as well as potentially increased risk of diabetes (13, 14) and hypertension (15, 16). Accurately predicting which patients require that the radiation therapy be supplemented by long-term androgen deprivation or may avoid treatment without such side effects would be of great help (17).
RTOG Protocol 92-02 was a clinical trial that attempted to help define the optimal duration of androgen deprivation when combined with radiation therapy (18, 19). Patients with stage T2c-4 prostate cancer and PSA <150 were given androgen ablation 2 months before and for 2 months during radiation therapy for a total of 4 months of androgen ablation. Patients were then randomly assigned to either the group that received no additional androgen deprivation (short-term AD-RT arm) or 24 months of androgen deprivation (long-term AD-RT). The LTAD-RT arm showed improved disease-free survival, an improvement that was particularly marked in patients with histologically aggressive disease (i.e., those with Gleason score of 8-10). RTOG 92-02 has therefore influenced the current standard of care for treating high-grade prostate cancer, which now consists of androgen deprivation in conjunction with radiation therapy for treating prostate cancer of high grade, stage, or PSA.
Bcl-2 and Bax: Balancing Life and Death?
Given the toxicities associated with androgen deprivation, finding markers that accurately predict outcome could substantially improve our ability to tailor androgen deprivation to derive the best efficacy and least morbidity. The apoptosis induced by androgen deprivation naturally suggested Bcl-2 and Bax as candidate markers, given well-established opposing roles in mediating apoptosis in cells in culture, including prostate cancer cells (Fig. 1 ; refs. 20, 21). Bcl-2 stabilizes the mitochondrial membrane and prevents the release of cytochrome c into the cytosol (22). Bax, in contrast, has the opposite activity, as it permeabilizes the outer membrane of the mitochondria and allows the release of cytochrome c into the cytosol, which precipitates the activation of caspases to bring about apoptosis. Whether or not apoptosis is induced can therefore depend on the balance between these two respectively antiapoptotic and proapoptotic factors (23–25). Cellular pathways or stimuli may disrupt the balance to favor one factor. For example, blockade of the androgen receptor in prostate cancer cells leads to up-regulation of Bax, which then accentuates apoptosis (26).
Reports suggest that this opposing relationship between Bcl-2 and Bax in cells established in cells in culture may also be detectable in tumor cells within tumors afflicting patients in the clinic. For example, immunohistochemical analyses done on prostate cancer tissue in a cohort of patients treated with external radiation therapy showed higher levels of Bcl-2 relative to Bax in patients with persistent or recurrent disease. The correlation between the Bcl-2/Bax ratio and poor therapeutic outcome after treatment remained significant even after correcting for age, PSA, and Gleason score (27). A different study assessed pretreatment Bcl-2 and Bax proteins in patients who received brachytherapy and whose disease status was then verified in biopsies done after treatment. High Bcl-2 levels were noted in the tumor tissues of patients in whom the brachytherapy was unsuccessful. Interestingly, patients who were treated successfully were also noted to have high numbers of apoptotic cancer cells (28).
The relative levels of Bcl-2 and Bax have also been found to be potentially useful prognostic markers for cancers other than prostate. Bcl-2/Bax expression correlated with poor prognoses in patients with bladder cancer, independent of other predictors such as stage or treatment. Patients with tumors expressing higher ratios of Bcl-2/Bax sustained poorer relapse-free survival (29). In an analysis of patients with stage I non–small cell lung cancer, those with tumors classified as expressing Bax but not Bcl-2 had poorer survival rates than those with tumors expressing both (30). It should be noted that although higher Bcl-2 protein levels correlated with poorer treatment response in these reports, the opposite correlation seems to have been reported for breast cancer recurrent after first-line chemotherapy. In patients with breast cancer who failed initial anthracycline-based chemotherapy and were therefore being treated with either docetaxel or methotrexate and 5-fluorouracil, higher Bcl-2 expression levels in the tumor were conversely associated with better overall survival (31).
Bcl-2 and Bax Protein Expression in RTOG 92-02
RTOG 92-02 has shown that a longer duration of androgen deprivation was superior, especially for poor-risk prostate cancer patients. Using archived patient tissues from the trial, Khor et al. (1) set out to determine whether Bcl-2/Bax might correlate with patient outcome in either treatment arm. The immunoreactivity of at least a thousand tumor cells was graded by observers who did not know each patient's clinical outcome. An interesting but potentially vital component of their method was that the levels of Bcl-2 and Bax expression in the prostate cancer cells were compared with those in the background normal prostate tissues, thereby enabling an internal control for each patient and which bolstered the validity of each individual analysis. Alterations of Bax expression were therefore defined as expression of Bax either increased or decreased in at least 20% of the cancer cells as compared with adjacent normal epithelium (in which Bax was also commonly detected). Further bolstering the strengths of the study, the relatively large number of patients enrolled in RTOG 92-02 allowed multivariate analyses and statistically robust comparison between patients that had received short-term versus long-term androgen deprivation.
Khor et al. found that for about half of patients, Bcl-2 was overexpressed or Bax levels were altered (more than 90% of these were overexpression). Bcl-2 expression or Bax alterations did not correlate with clinical variables such as age, PSA, or T-stage, but more tumors with alterations of Bcl-2 or Bax were of Gleason score 7-10 disease. On multivariate analysis, either Bcl-2 expression or Bax alteration correlated with an increased risk of treatment failure in patients assigned to short-term androgen deprivation. Because of the opposing biological roles these two proteins play, these factors were then considered together for each patient. Patients with either Bcl-2 expression or Bax alteration were found to have significantly higher failure rates, but only in the short-term androgen deprivation arm. In the long-term androgen deprivation arm, neither factor correlated with failure. Interestingly, in patients with initial PSA >30, on the multivariate analysis Bcl-2 expression or Bax alteration also correlated with a higher risk of biochemical (PSA) failure.
These results suggest a number of tentative conclusions or hypotheses worthy of further testing. For example, it seems that a longer course of androgen deprivation might overcome the adverse effects of either Bcl-2 expression or alterations of Bax, possibly accounting for the lack of significance of these factors in that treatment arm. These results suggest that patients with tumors expressing Bcl-2 or Bax require longer than a mere 4 months of androgen deprivation. Patients presenting with PSA >30 and expressing Bcl-2 or showing alterations of Bax should not receive a short course of androgen deprivation. In all the analyses, why should tumors expressing Bax—after all, a proapoptotic factor—do so poorly? Is the presence of increased Bax expression reflective of defects in caspase activation or overexpression of as yet unstudied antiapoptotic factors?
The Next Point of Attack?
Where do we go from here? These findings represent tangible advances in a long campaign against prostate cancer. The methods and results of Khor et al. will help guide future efforts. For example, many practitioners now often dichotomize patients between intermediate-course androgen deprivation (e.g. 6 months) versus long-course androgen deprivation. It would be interesting to investigate whether an intermediate course of androgen deprivation is sufficient to overcome the adverse effects of Bcl-2/Bax expression. Future studies should also incorporate percent positive biopsy as a measure of extent of tumor (32). Perhaps most relevant to the future development of novel treatment strategies, the biology behind how perturbed Bcl-2 or Bax expression leads to increased treatment resistance needs to be better clarified. Direct comparisons of Bcl-2/Bax against the expression of other genes (mRNA) or proteins, or against microarray or proteomic approaches, may be warranted. Prospective trials verifying in prostate cancer patients that assessment of Bcl-2 and Bax holds predictive value and complements or is even superior to standard variables (i.e., risk categorization) should be completed.
These and many other battles need to be waged in the future. It will require considerable persistence, cooperation, and resources. Scientists and clinicians need funding restored to build on the advances already achieved. Perhaps the most important question of all should therefore be “Do we still have enough resolve—economic, emotional, intellectual, and political will—to continue to fight for victory in this war on cancer?” At least 27,000 souls each year are desperately hoping that we do.
- Received April 6, 2007.
- Accepted April 9, 2007.