Role of the Bcl-2 gene family in prostate cancer progression and its implications for therapeutic intervention.

Prostate cancer (PC) is an escalating health burden in the western world. A large number of patients still present with extraprostatic (i.e., T3/T4, N0, M0/M1 or any T category and M1 disease or involved lymph nodes) and therefore incurable disease. Since the work of Huggins in 1940, there have been no major therapeutic advances and androgen ablation remains the best treatment option for extraprostatic androgen-responsive PC. Eighty to ninety percent of PC patients respond well to this form of treatment initially. After a median time of approximately 2 years, however, relapse to an androgen-independent (AI) state occurs, followed by death after a further median 6 months. Androgen ablation is rarely curative. The major molecular defect in extraprostatic and AI PC is the inability of PC cells to initiate apoptosis in response to a variety of stimuli, including different forms of androgen ablation and cytotoxic agents. The balance between cellular proliferation and cell death is regulated by multiple genes or families of genes through the cell cycle. The exact mechanisms governing this intricate and complex process are as yet not fully understood. One family of genes involved in cell survival/death control is the Bcl-2 gene family, which consists of homologous proteins that function to regulate distal and crucial commitment steps of the apoptotic pathway. The Bcl-2 family constitutes both agonists and antagonists of apoptosis that function at least in part through protein-protein interactions between various members of the family. The final outcome depends on the relative ratio of death agonists and antagonists. Bcl-2 expression has been closely associated with the AI phenotype of PC. Cytotoxic chemotherapy may be used as palliative therapy in AI PC but has not been found effective. Most chemotherapeutic cytotoxic agents induce apoptosis in cancer cells by direct and indirect action on the cell cycle. In vitro and in vivo studies have established that Bcl-2 expression confers an antiapoptotic activity against androgen withdrawal and cytotoxic chemotherapy. It thus offers a tempting potential target for therapeutic manipulations of PC.

dramatically, such that an individual suffering from metastatic PC may lose approximately 9 to 10 years of life expectancy (2). More than 90% of PCs are diagnosed between 45 and 89 years of age (average, 72 years of age). Because the risk of PC bears a strong relationship to age, the incidence [1 in 10,000 at <39 years of age to 1 in 6 for men > 80 years of age (1)] can be expected to rise because men are living longer than ever before (3). Other risk factors include family history (4), higher intake of dietary fat (5,6), high blood levels of male sex hormones, and Americans of African origin. Vitamins D and E and selenium are suggested to have a protective effect (7,8), whereas the role of vitamin A is controversial.

Pathology of Prostate Cancer
Histopathologically, high-grade prostatic intraepithelial neoplasia (HGPIN) is considered the most likely precursor of invasive PC because of its peripheral zone location and immunohistochemical expression of biomarkers similar to that of PC (9). Prostatic intraepithelial neoplasia is present in 85% of PC samples (10). Histologically, almost all PCs are adenocarcinomas originating from the prostatic glandular epithelium. The most widely used system of grading PC is the Gleason system (11,12), which takes into account the degree of glandular differentiation of epithelial cells in relation to stroma at relatively low magnification. Grades range from 1 to 5 depending on tumor architecture, which varies from well differentiated (showing a preserved glandular pattern) to poorly differentiated (showing no definite glandular architecture). A combined Gleason's score or Gleason's sum (2)(3)(4)(5)(6)(7)(8)(9)(10) is determined by adding the most frequent primary and secondary pattern of cellular architecture in a heterogeneous tumor.
PC spreads to local as well as distant sites. Local spread can involve adjacent viscera including seminal vesicles, ejaculatory ducts, and rectum. The most frequent sites of metastatic spread are lymph nodes, bones, and lungs. Less common sites include bladder, liver, and adrenal glands (13).
Because of the gravity of extraprostatic disease, a screening program designed to detect PC at a curable stage appears logical. However, there are limitations with present screening methods; even after histologic diagnosis has been made, the biologic behavior of individual PC cannot be predicted. Identification of those cases of localized PC that require definitive treatment from those in which no treatment is the appropriate option remains an issue.
At present no definite biologic or molecular marker exists that can predict the future clinical behavior of PC while it is still localized to the prostate [reviewed by Isaacs (14) and Lalani et al. (15)]. New genotypic and phenotypic markers are therefore urgently needed to classify the disease accurately and to guide its management (16). The

Localized Clinically Significant and Latent Cancer
Part of the problem in managing PC lies in the fact that there is a latent form of the disease that is equally prevalent in communities throughout the world. Latent PC is generally an incidental finding, usually identified in postmortem studies or in prostates removed surgically for apparently benign disease or following prostate-specific antigen (PSA) testing and imaging-directed stereotactic biopsy. Latent PC is by definition small, generally well differentiated, and the clinical behavior is usually indolent. This cancer may be found in men as young as in their 30s and has been reported in approximately 30% of 50-year-olds and up to 80% of 80-year-olds at postmortem (18). This is in sharp contrast to clinically important cancer, which is rare in Asia (1:100,000) but reaches epidemic proportions in the West; the highest rates reported are in Americans ofAfrican origin (19).
For the clinician the challenge is to distinguish between patients with localized disease (T1,2NoM0) that is a) dinically significant and that will metastasize if untreated, usually justifying radical therapy; b) dinically significant but is an apparently localized disease with unidentifiable occult metastases using current imaging techniques in whom there is no curative option (but these patients may be subjected to radical therapy in the hope of cure); and c) latent cancer that will not affect them during their lifetime. There are currently no clinical or molecular tests to distinguish between these three groups ofpatients.  (20) asked, "Is cure possible in those for whom it is necessary, and is cure necessary in those for whom it is possible?" There is no doubt that a large number of patients require effective treatment for their disease, whether localized or extraprostatic. Treatment depends primarily on stage but also on the patient's age, general fitness, and comorbidity. Localized PC can be managed by watchful waiting, radiotherapy, or radical prostatectomy. During watchful waiting patients may be assessed regularly by monitoring serum PSA levels or by digital rectal examination to determine disease progression. By following patients conservatively, the morbidity of active treatment can be avoided. The mean age of PC at presentation is 72 years, at which age death due to intercurrent diseases is greater than that from PC itself. However, patients in younger age groups and especially those with high-grade disease may be considered for active treatment. With improvement in surgical techniques and increased knowledge of surgical anatomy of the prostate, radical prostatectomy has become a safe treatment for localized PC with apparent positive long-term results (21). However, it remains to be proven more effective than either radiotherapy or watchful waiting. Radiotherapy is especially used in older patients and in those presenting with intercurrent illnesses and who are therefore unsuitable for radical surgery. Patients best suited for radiotherapy have a low serum PSA level (< 20 ng/ml) and low-grade lowstage tumors (22). Endocrine (hormone) therapy is the mainstay of treatment in patients with locally advanced and/or disseminated metastatic PC. Hormone therapy may take the form of surgical castration, leutinizing-hormone releasing hormone analogues (medical castration), antiandrogens, or a combination of castration and antiandrogens that block both testicular and adrenal androgens-maximum androgen blockade. Cytotoxic chemotherapy in androgen-independent (AI) PCs has limited success. Because of the lack of definite curative therapy for extraprostatic PC, patients require palliative treatment such as hormone downregulation and cytotoxic chemotherapy to alleviate their diseaserelated symptoms and improve their general well being. These treatments aim to kill tumor cells in situ by inducing apoptosis, or if this cannot be achieved, by at least keeping them confined to their current boundaries by cytostasis. The exact underlying mechanisms responsible for the development of AI and chemoresistance in PC is unknown.
The normal prostatic epithelium requires androgens for mitogenesis, as is also true for prostatic adenocarcinomas. This is why most PCs respond initially to androgen ablation therapy, which causes tumor regression by inhibiting cellular proliferation and inducing apoptosis (23). The response is usually short-lived and results in an AI tumor. AI tumors are usually more aggressive and show little or no response to androgen-ablative measures. As the effects of androgen/antiandrogens are mediated through the androgen receptor (AR), much research effort is focused on the role of AR in tumor recurrence and progression (24), which may be a consequence of multiple genetic and epigenetic events, including mutations in and altered expression of the AR (15). Mechanisms by which PC cells circumvent endocrine therapy have thus far focused on loss of AR, amplification of wild-type AR and clonal expression of these cells, and mutations of AR resulting in either increased transactivation and/or loss of specificity for steroid hormones and their antagonists (25)(26)(27)(28)(29)(30)(31). Alternatively, the defects could occur in the AR down stream-signaling pathways, leading to AI in the presence of a normal AR. Whatever the cause of AI tumors, the final clinical outcome will be the uninhibited proliferation of PC cells even after androgen ablation. AI PCs also bear a close association with the expression of cell death inhibitors of the Bcl-2 family (32-34).
Several cytotoxic agents, alone or in combination, have been used in the treatment of AI extraprostatic PC but with little benefit (35). Possible  Several diseases are associated with either impaired or excessive apoptosis including cancer, autoimmunity, HIV-associated immunodeficiency, and some neurodegenerative disorders (36)(37)(38). Apoptosis is an evolutionary conserved and genetically controlled mechanism through which cells are eliminated in both health and disease. Apoptosis constitutes distinct morphological and biochemical changes in the cell. Morphologically, it involves rapid condensation of chromatin, shrinkage of cells, and ultimately formation of membraneenclosed apoptotic bodies that are engulfed by neighboring scavengers (39)(40)(41)(42). Biochemically, there is relocation of phosphatidyl serine from the inner to the outer aspect of the plasma membrane and degradation of double-stranded nuclear DNA, leading to production of oligonucleosomal DNA fragments ( Figure 1). The evolutionary conserved nature of genetic regulation of this process has become apparent from studies of C. elegans, a primitive nematode worm, through Drosophila melanogaster to mammals. Two important gene families, the Bcl-2 gene family and the caspase family of cysteine proteases (previously known as the interleukin-l-,-converting enzyme family) play key roles in the regulation and execution of this program. The Bcl-2 fa.mily, which consists of both inhibitors and promoters'of apoptosis, constitutes mammalian homologs of ced9, which is an inhibitor of apoptosis in C. elegans. ced9 functions upstream of two cell-death promoters, ced4 and ced3. During the development of C. elegans, 131 of 1090 somatic cells normally die. However, these cells fail to die in response to gain of function mutations in ced9 (43). Mammalian counterparts of ced3 are caspase proteases, which can cleave a number of specific death substrates and execute the final stages of apoptotic pathways ( Figure 2). Recently Zou et al. (44) identified another deathpromoting protein, apoptotic proteaseactivating factor (Apafl), which is considered a human homolog of ced4.

Bcl-2 Protein Family and Regulation of Apoptosis
In vitro studies have highlighted the role of Bcl-2 proteins as important regulators of the apoptotic pathway in several cell types (38,(45)(46)(47)(48)(49). Bcl-2, the prototype of this family, was discovered by studies of t(14: 18) chromosomal translocations, which are frequent in non-Hodgkin lymphomas and follicular lymphomas (50-52). The name Bcl-2 (B cell lymphoma/leukemia gene 2) signifies the close association of this gene to these malignancies in which enhanced expression was initially believed to arise solely as a result of these translocations, resulting in the juxtaposition of the Bcl-2 gene to a potent enhancer element sequence of the IgH gene. Several genes have been identified and designated as the Bcl-2 family based on their sequence homology to Bcl-2. These genes include both positive and negative regulators of apoptosis (Table 2). In contrast there is only a single homolog, ced9, in C. elegans. Sequence analyses of Bcl-2 family proteins have identified up to four evolutionary conserved domains, recently named BHI to BH4 (BH stands for Bcl-2 homology domain), as proposed by Oltavi et al.   (53). These domains may have functional significance, allowing Bcl-2 members to act as either death promoters or death inhibitors. Most Bcl-2 family members have a stretch of hydrophobic amino acids at their carboxy terminal, which is responsible for localization (anchorage) to membranes of cellular organelles. Bcl-2 proteins interact with each other through homo-and heterodimerization to regulate apoptosis. For example, Bcl-2 can form homodimers by interaction with another molecule of Bcl-2 or a heterodimer with a molecule of Bax (53). These dimerizations are of functional significance. For example, a Bcl-2:Bax heterodimer will favor cell survival, whereas a Bax:Bax homodimer will favor cell death. Mutations in BH domains can effect dimerization potential and abrogate normal function (54-57). There is great complexity and cell specificity of homoand heterodimerizations [reviewed by Reed (58)].

Biochemical Functions of the Bcl-2 Family Proteins
The Bcl-2 family proteins have been studied in relation to their localization to intracellular organelles, primarily the outer mitochondrial membrane (Figure 3A Bax \, nuclear envelope, and endoplasmic reticulum (59,60). A number of theories based on experimental observations have tried to explain how Bcl-2 functions as a death regulator. Initial observations suggested its role as an antioxidant (61-64), a regulator of intracellular calcium (65)(66)(67)(68)(69), and in transport of proteins across cellular membranes e.g., across the nuclear membrane (68,70,71). Recently it has been proposed that Bcl-2 regulates activation of caspase proteases (72)(73)(74)(75)(76)(77)(78)(79), which are responsible for the final executionary steps of apoptosis ( Figure 3A, B). There is as yet no evidence that these functions are a result of the direct action of Bcl-2. These may, however, be consequences of Bcl-2 action upstream of these events. In the last 2 years, Bcl-2 function has been extensively studied in relation to its regulation of permeability transition (PT) in mitochondria. The induction of PT can result in loss of membrane potential (80,81), inhibition of oxidative phosphorylation, generation of reactive oxygen species, and release of proteins such as cytochrome c. Cytochrome c normally resides in the intermembrane space of the mitochondria and on its translocation to cytosol can induce activation of caspases (81,82) ( Figure 3A  the mitochondrial space and stabilizes the mitochondrial potential (AT) in response to apoptosis-inducing agents, for example, stauroporine (83). Furthermore, a proposed pore-forming capability of Bcl-2, BCI-XL, and Bax in lipid membranes (84)(85)(86)(87), similar to those of the bacterial toxins (e.g., diphtheria) and colicins suggests a novel function for these proteins as channels for ions, proteins, or both.
The function of Bcl-2 is unique because it does not affect the pharmacokinetics of the anticancer drugs. It blocks the transmission of signals originating from the damage to the effectors of apoptosis, allowing cells to survive in the presence of otherwise lethal damage, enhancing the chances of acquiring genetic alterations and favoring the development of a more malignant phenotype.

Role of Apoptos in the Development and Progression of Prostate Cancer
Normal prostatic glandular cells have a low proliferation rate (< 0.3%/day) that is efficiently balanced by the same rate of apoptosis so that there is no net growth (104). During transformation to HGPIN, a net growth of malignant cells occurs because of the disturbance in equilibrium between cell proliferation and apoptosis. The growth kinetics of localized PC Environmental Health Perspectives * Vol 107, Supplement 1 * February 1999 Bcd-2 GENE FAMILY PROTEINS AND PROSTATE CANCER indicate that there is a decrease in cell death (and no further increase in proliferation rate as compared to HGPIN) that is responsible for its continuous growth (104). A similar situation occurs in metastatic PCs in which the proliferation rate is also low (< 3%) (104). These cell kinetic data provide an explanation as to why conventional cytotoxic chemotherapy (which usually targets actively proliferating cancer cells) fails to kill PC cells and suggests that proliferation-independent therapeutic strategies may be more beneficial for the treatment of PC.
Androgen ablative therapies are the mainstay of treatment in extraprostatic PC. Androgen ablation has a dual effect i.e., inhibiting epithelial cellular proliferation and promoting apoptosis (105). Kyprianou et al. (106) demonstrated that castration of mice bearing PC82 prostatic epithelial cell xenografts resulted in tumor regression (106). Histologic analyses revealed a decrease in the number of cells undergoing mitosis and a dramatic increase in the percentage of cells undergoing apoptosis. In vivo, PC cells possessing the androgensensitive phenotype also show inhibited proliferation and apoptosis in response to androgen withdrawal (23). However, as AI develops, this response to androgen withdrawal is abolished. Several studies have demonstrated that androgen ablation does not induce apoptosis in AI PC cells. A close association of the Bcl-2 family antiapoptotic protein expression and the development of AI in PC has been found (32,34). This means that in Al PC, resistance to apoptotic-inducing cytotoxic chemotherapeutics is contributed both by slow proliferation and acquisition of the antiapoptotic potential owing to the dysregulated expression of antiapoptotic Bcl-2 homologs. Therefore, potential measures to overcome cytotoxic chemoresistance in AI PC can be achieved either by introducing proliferation-independent treatment strategies and/or restoring sensitivity to induce apoptosis in cancer cells. Thapsigargin, for example, could induce apoptotic cell death independent of cell proliferation (107,108). Thapsigargin is a specific inhibitor of endoplasmic reticulum (ER) Ca-ATPase and results in the depletion of the ER pool and the subsequent signaling of a sustained influx of extracellular calcium (109). Agents that could target Bcl-2 or related antiapoptotic proteins and therefore sensitize cancer cells to undergo apoptosis could also be useful as new therapeutic modalities in the management of PC.
In vitro studies on PC cell lines suggest a role for Bcl-2 in the inhibition of apoptosis and progression to AI (33). Raffo et al. (33) have shown that Bcl-2 transfection into LNCaP cells (a human androgen-sensitive PC cell line) resulted in the inhibition of apoptosis in response to serum (growth factors) withdrawal in contrast to apoptotic cell death in controls. Moreover, Bcl-2 expression provided a growth advantage in the absence of androgens i.e., LNCaPlBcl-2 transfectant xenografts in nude mice continued their growth after castration, whereas tumors formed by the control transfectant regressed after castration and showed evidence of apoptotic cell death (33). McConkey et al. (95) reported close association of de novo Bcl-2 expression with acquisition of the metastatic phenotype by LNCaP sublines derived by orthotopic implantation into rat prostates.
These Bcl-2-expressing sublines were significantly resistant to adriamycin and thapsigargin-induced apoptosis (95). We have observed a similar chemoresistance against adriamycin and thapsigargin in the Bcl-2 transfectants of the DU145 prostatic epithelial cell line expressing high levels of Bcl-2 protein (128). Enforced Bcl-2 expression in these cells altered the expression of other cell-cycle and cell-death regulators including p53 and proliferating cell nuclear antigen.
Studies utilizing antisense oligonucleotides directed against BcI-2 mRNA demonstrated the development of chemosensitivity to chemical agents in otherwise chemoresistant cells (97)(98)(99). A similar approach using ribozymes (enzyme coupled antisense nucleotides) against Bcl-2 and BCI-XL has also been reported to favor apoptosis (129,130). Taken together, these observations suggest that ectopic BcI-21Bcl-XL expression confers a death-resistant phenotype and imparts an ability to inhibit the initiation of the apoptotic cascade in response to several stimuli including chemical agents and radiation (131).

Potential for Therapeutic Intervention
Several studies, some of which have been cited in this review, demonstrated that inhibition of some of the death antagonists of the Bcd-2 family can significantly potentiate apoptosis. This offers hope in developing potential therapeutic modalities to combat cancers in general and PC in particular. There are several other potential mechanistic targets involving Bcl-2 that may offer additional hope, including antisense strategies utilizing oligonucleotides or ribozymes against Bcl-21Bcl-XL mRNA. Other potential mechanistic targets include therapies to alter Bcl-2/Bcl-XL protein configuration inhibiting homoand heterodimerization, for example, phosphorylating Bcl-2, rendering it nonfunctional (132,133), and upregulating the expression of death promoters e.g., Bax, Bak, etc.
Antisense strategies using anti-Bcl-2 oligonucleotides or ribozymes may abolish Bcl-2 expression by targeting Bcl-2 mRNA and are effective at least in vitro (129,130).
Another mechanism to prevent Bcl-2-mediated inhibition of cell death is to convert Bcl-2/Bd-XL protein configuration to nonfunctional forms. Phosphorylation of Bcl-2 renders it nonfunctional as a death inhibitor (132,133) and abolishes heterodimerization of Bcl-2 with death promoter partners, allowing death promoters to function uninhibited (134). Cytotoxic agents, for example, taxol, vinblastine, and vincristine result in phosphorylation and inactivation of Bcl-2 and can lead to apoptotic cell death in Bcl-2-expressing cells, including those of the prostate (132,133 The protein product of the oncogene bcl-2 is a component of the nuclear envelope, the endoplasmic reticulum, and the outer mitochondrial membrane. Cell Growth Differ 5:411-417