The mammalian prostate gland is a male exocrine sexual accessory organ composed of secretory and non- secretory epithelial cells, fibroblasts, smooth muscle ...
Regulation of Apoptosis in the Prostate Gland by Androgenic Steroids Ralph Buttyan, Ahmad Shabsigh, Harris Perlman and Marc Colombel
androgens (or other steroid hormones) regulate cellular apoptosis. Indeed, as a number of recent studies have established, the concept of apoptosis is becoming increasingly more pertinent to our understanding of prostate disease states and their potential treatments2,3. •
The prostate gland requires androgenic steroids for its appropriate embryological formation and postpubertal growth and, once at adult size, remains dependent on a continuous supply of androgens for its vitality and function. A reduction of the levels of circulating androgens will rapidly induce apoptosis of the cells of the prostate, leading to extensive glandular regression. Studies of rodent models of prostate response to castration have shown that there are some remarkable changes in the gene activity of prostate epithelial cells leading up to apoptosis. There is now evidence for a critical cell signaling pathway, regulated by c-fos expression, necessary for castration-induced apoptosis, as well as evidence that this signaling initiates an abrupt and transient alteration in the synthesis of fas antigen, p53, bax and bcl-2 proteins in the androgen receptor-expressing prostate epithelial cells, the cellular compartment that appears to be the most affected by castration. However, more recent studies suggest that these castration-induced effects on the prostate epithelial cells might be, at least in part, an indirect response to a critical reduction in blood flow to the prostate gland that precedes the onset of epithelial cell apoptosis. The castration effects on blood flow to the prostate gland seem to be related to vascular degeneration associated with apoptosis of a subset of prostate endothelial cells. The mammalian prostate gland is a male exocrine sexual accessory organ composed of secretory and nonsecretory epithelial cells, fibroblasts, smooth muscle, nerves and endothelial cell types interacting to form a branching ductal network (reviewed in Ref. 1). The main glands of the prostate are located at the distal tips of the ductal network and secrete proteinaceous substance and other materials into the ejaculate. Proposed functions of prostatic secretions include nourishment R. Buttyan, A. Shabsigh and H. Perlman are at the College of Physicians and Surgeons of Columbia University, Department of Urology, Atchley Pavilion 11th Floor, 161 Fort Washington Blvd, New York, NY 10032, USA; and M. Colombel is at the Service d’Urologie et Chirurgie de la Transplantation, Centre Hospitalier Universitaire de Lyon 1, Hôpital Edouard Herriot, 69437 Lyon, France.
TEM Vol. 10, No. 2, 1999
and protection of sperm subsequent to copulation as well as potential antimicrobial action. These features are shared by other male sexual accessory tissues. In humans, the prostate is associated with the development of a number of debilitating disease states. Chronic prostatitis, benign prostatic hypertrophy (BPH) and prostate cancer are common during aging. Because androgenic steroids are required for prostate development and adult maintenance (see below), the prostate is also studied frequently as a model for androgen action. Finally, one of the most direct effects of androgenic stimulation of the prostate is the suppression of prostate cell apoptosis. Therefore, the prostate gland is now widely utilized as an in vivo model system to study the potential cellular and molecular mechanisms by which
Overview of Androgens and the Regulation of Prostate Gland Development, Growth and Apoptosis
Just like other male sexual accessory tissues, the prostate gland requires an early but transient stimulation by androgenic steroids for its appropriate embryological development as well as continuous, postpubertal stimulation for its growth and maintenance at the adult size. The primary steroids involved in this stimulation are testosterone and dihydrotestosterone, a metabolite of testosterone produced by the action of the enzyme, 5-α reductase, within the prostate and other target tissues4. Androgen action also requires intracellular androgen receptor (AR) protein, in humans a single 110-kDa species with characteristic conserved domains (steroid-binding domain, DNA-binding domain) typical of a member of the steroid receptor family5. Immunohistochemical studies of human and rat prostate tissues have shown that ARs are most abundant in the nucleus of the epithelial cells of the adult prostate6,7. However, many of the prostatic stromal cells (fibroblasts and smooth muscle) also contain nuclear ARs (Ref. 7), and these cells are therefore likely targets for androgen action as well. Androgenic steroids have three principal influences on prostate cells in vivo that might explain the stimulatory attributes of androgens on prostate gland development, growth and maintenance8, namely: (1) androgens can stimulate proliferation, especially of prostatic epithelial cells; (2) androgens encourage differentiation of the prostate secretory cell phenotype and; (3) androgens suppress prostate cell apoptosis (again, an effect usually noted in the epithelial cells). In general, it is presumed that this multiaction potential of androgens on the
1043-2760/99/$ – see front matter © 1999 Elsevier Science. All rights reserved. PII: S1043-2760(98)00104-0
47
with androgen-regulated prostate cell apoptosis. •
Figure 1. Photomicrographs identify extensive apoptosis in the mature rat ventral prostate gland after castration. (A) Thin sections (5 µm) of a rat (Sprague-Dawley) ventral prostate showing an extensive epithelial cell layer containing glands. (B) Section of ventral prostate gland at 72 h after castration. Arrows identify apoptotic bodies present in vacuoles throughout the ducts and glands of the regressing prostate. (C) Section through a ventral prostate gland at 72 h after castration demonstrates apoptotic bodies within the epithelium as well as extensive shedding of apoptotic bodies into the lumen of the prostate gland. Arrows identify some representative apoptotic bodies within the glandular lumen. (D) High magnification (31000) of a section through a prostate gland at 72 h after castration demonstrates important morphological features of prostate cell apoptosis, including nuclear condensation and pyknosis, cell shrinkage and separation from neighboring cells (thick arrows).
prostate is the result of some direct action of androgens on prostate cells. This presumption implies that androgens might be regulating the synthesis of key molecular components involved in the signaling pathway for each of these three basic cellular processes in prostate cells, or that the three distinct processes might share some common molecular component(s) that is regulated directly by androgens. When prostate epithelial cells maintained in culture are studied, androgens show the ability to affect both the in vitro proliferation rate9 and the differentiation characteristics of these cells10, even though they do not require androgenic steroids for their survival in vitro. In view of the effects of androgens on prostate cells in culture, it is also possible that the cellular and molecular mechanisms through which
48
androgens regulate prostate cell proliferation and differentiation might be distinct from those by which androgens regulate prostate cell apoptosis. There has been much interest in the molecular basis by which androgenic stimulation or withdrawal affects prostate cells (particularly their apoptosis) at the genetic level. Because androgen-regulated prostate cell apoptosis is only observed in vivo, such studies depend on the analysis of prostate tissues obtained under varying conditions of hormonal stimulation. It has been difficult to do these types of studies on human prostates, and experimentation on the prostate glands of large laboratory animals is often limited by the expense. Thus, in vivo studies of the rodent prostate, especially that of the rat, have been used primarily in defining some of the genetic variables associated
The Rat Ventral Prostate Gland as a Model for the Study of Androgen-regulated Apoptosis: Morphology and Biochemistry of Apoptotic Ventral Prostate Epithelial Cells
A relationship between androgenic steroids and prostate cell survival was established conceptually by Huggins and Hutchins in their early studies on the effects of castration on human prostate cancer11. However, a real understanding of this relationship has only come from observations on experimental animal models. Rodent prostates have been most useful in this regard12–14. The rat ventral prostate gland was one of the first sites where the morphological attributes of cells undergoing apoptosis were well described15. The ventral prostate gland is induced to regress by surgical castration of a mature male rat. Within days, the wet weight and size of the ventral prostate glands are reduced significantly, and it has been estimated that up to 85% of the cells of the mature ventral prostate will be deleted by apoptosis by two weeks after castration16. Quantification of apoptotic cells after castration (by direct counting of apoptotic bodies in tissue sections) showed that they start to appear in the rat ventral prostate after 24 h and increase dramatically at 48 and 72 h after castration, declining thereafter17. At the peak of apoptotic activity, it has been estimated that ~20% of the cells are deleted from this tissue by apoptosis per day18. Without testosterone supplementation, these apoptotic cells are not replaced. The loss of this large proportion of cells can explain the reduction in ventral prostate weight and size following castration. In contrast to the human prostate, which contains a much larger proportion of fibroblast and smooth muscle cells, the rat ventral prostate is composed mainly of columnar epithelial cells that form the secretory component of the glands and more cuboidal epithelial cells that line the prostatic ducts leading to the urethra. These are TEM Vol. 10, No. 2, 1999
the prostate cell components which, to date, have received the most attention in the study of the androgen-regulated apoptotic effect. This effect is not homogeneous; there is a relative gradient of sensitivity to androgen withdrawal that depends on the epithelial cells’ location along the prostate duct8,19,20. In an androgenically intact male rat, there is a normal low but detectable basal rate of apoptosis of epithelial cells in the areas of the prostatic ducts most proximal to the urethra. However, after castration the epithelial cells in the glandular area are the first to demonstrate apoptosis, and it is these distal glandular regions that incur the most extensive cell losses associated with prostate regression. In contrast, the basal apoptotic rate of the epithelial cells in the proximal ducts declines. Thus, castration has the paradoxical effect of drastically decreasing the lifespan of the glandular ventral prostate epithelial cells, while seemingly increasing the lifespan of the ductular cells. The regressing rat ventral prostate gland is a model tissue for examining cells undergoing apoptosis21. Morphologically, the columnar epithelial cells of the ventral prostate in a castrated rat demonstrate early blebbing of the apical membrane, cell shrinkage and a distinct shift in the nuclear position away from the basolateral region of the cell (Fig. 1). One of the more remarkable early changes in regressing prostate tissue is the chaotic variation of nuclear positioning in the epithelial cell layer, which contrasts greatly with the regular alignment of the epithelial nuclei adjacent to the basement membrane region of an androgenically intact rat prostate (Fig. 1). Subsequently, apoptotic epithelial cells detach from their neighboring epithelial cells as well as from the basement membrane. This detachment is remarkable when one considers the extensive amount of manipulation and enzymatic digestion that is required to produce single-cell suspensions of epithelial cells from a mature rat ventral prostate gland22. As the epithelial cells detach, the cytoplasm and nuclei continue to shrink as apoptotic bodies form. Fully formed TEM Vol. 10, No. 2, 1999
apoptotic bodies are often found in vacuoles in the epithelial lining directly adjacent to normal appearing epithelial cells or to epithelial cells involved in the earlier stages of apoptosis. Thus, although androgen-regulated prostate cell apoptosis is a relatively synchronous event when quantified in the total population of prostate cells, it is a heterogeneous process, affecting individual cells with identical preapoptotic morphology at different rates. The characteristic pyknosis and fragmentation of the cell nucleus gives the final characteristic to the small apoptotic bodies, clusters of small, dark-staining globular chromatin bodies (Fig. 1). The apoptotic bodies are ultimately removed from the regressing prostate gland epithelium by one of two means: phagocytosis and digestion by a neighboring epithelial cell or release into the lumen of the prostatic duct. Although occasional clumps containing numerous apoptotic bodies can be found within the prostatic lumen (Fig. 1), it is difficult to determine whether either of these two removal mechanisms is more prevalent or important for prostatic regression. The most predominant ‘biochemical marker’ for apoptosis in tissues is the characteristic ‘ladder’ pattern degradation of the nuclear DNA associated with the induced expression or activation of an endonuclease that preferentially degrades DNA within the internucleosomal regions. The regressing rat ventral prostate gland readily demonstrates this 180 base pair ladder pattern when DNA is extracted and electrophoresed on an agarose gel14,21,23,24. Likewise, this DNA degradation characteristic is useful in performing quantitative studies of apoptosis in the rodent (and human) prostate by the use of the modern immunohistochemical staining techniques referred to as ISEL- or TUNELlabeling25. There has been little work to define the endonuclease(s) responsible for this activity in the rodent or other prostate systems. Recently, however, an immunohistochemical study of the regressing rat ventral prostate has demonstrated a significant
upregulation of the gene encoding rat DNase I and intranuclear accumulation of this nuclease in pre-apoptotic prostate epithelial cells, thus suggesting that this particular nuclease is associated with the nucleolytic activity during prostate regression26. Because it is not clear whether the intranucleosomal DNA degradation associated with apoptosis is due to any particular nuclease, it is possible that certain endonucleases, already present in the prostate cells, might be ‘activated’ during apoptosis by the significant increase in intracellular free Ca2+ ion levels that has been shown to accompany prostate regression following castration23. There is also evidence that the characteristic internucleosomal DNA degradation might not be critical for androgen-regulated prostate epithelial cell apoptosis. Treatment of castrated rats with a DNA-synthesis inhibitor drug, mimosine, suppressed the appearance of the DNA ladder as well as the ability of the apoptotic epithelial cell nuclei to be labeled by ISEL techniques, yet did not suppress the appearance of apoptotic bodies27. This strongly implies, as has been reported for other cell systems, that the internucleosomal DNA degradation activity associated with prostate cell apoptosis is probably an end stage of the cell death process, and certainly a dispensable aspect of it. •
Molecular Regulation of Apoptosis by Androgen Deprivation in the Rat Ventral Prostate Model
Most exciting in the study of apoptosis has been the idea that apoptosis is a genetically regulated process, one that requires the expression and subsequent action of discrete gene products and coregulatory molecules to proceed. Many of the molecules that have been characterized in the last decade for their role in apoptosis in other cell systems have also been found to be relevant for apoptosis in the regressing rodent ventral prostate gland model. For example, the gene products c-Fos, c-Myc, p53 and TRPM-2/SGP-2 were shown to be induced in the regressing rat ventral prostate gland before or at the time that these proteins were 49
c-Fos Fas TGF-β
Caspase DNase I
Cathepsin B Cathepsin D Matrilysin
Initiation
Execution
Disposal
Regulation Bax/Bcl-2 p53
Gene products involved in androgen-regulated prostate cell apoptosis Figure 2. Functional stages of apoptosis and their association with gene products that appear to be mediating these individual stages during post-castration regression of the rat ventral prostate gland. TGF-β, transforming growth factor β.
shown to have a general role in other apoptotic cell systems. Other apoptosis-regulating gene products, such as Bcl-2 family members, Fas antigen and caspase family gene members, are also proving to be involved in the apoptosis of the prostate gland due to androgen withdrawal. In general, molecular effectors of apoptosis can be segregated into four categories that reflect one or more of the stages of the cell death process in which they participate. These are: (1) initiation and signaling of apoptosis; (2) regulation of the response to the apoptotic initiator; (3) execution of the apoptotic process and; (4) removal (disposal) of the apoptotic body. Most of the prostate-effector gene products associated with rodent prostate regression can be sorted into the general apoptotic-effector categories described above (Fig. 2). Apoptotic cellular signals have been found in the prostate gland following androgen withdrawal. First, castration significantly alters the local expression and concentration of certain prostatederived growth factors that are known to affect prostate cell proliferation and survival. In particular, transforming growth factor β (TGF-β) gene expression is upregulated in the rat prostate after castration28. TGF-β can induce apoptosis in cultured cells and, in abundance, can directly elicit apoptosis in rat prostate glands maintained in organ culture29. A reduction in the
50
prostatic expression of the epidermal growth factor (EGF) gene occurs after castration30, in conjunction with an upregulation of prostatic TGF-β gene expression. EGF is believed to be a stimulatory factor for the prostatic epithelial cell, and its reduced synthesis after castration might by itself, or in synergy with increased TGF-β synthesis, participate in the signaling process by which apoptosis is activated in prostate epithelial cells. A similar phenomenon is also associated with the onset of apoptosis in the rat kidney31. Although the effects of castration on the growth factor milieu of the prostate remain to be defined completely, there is evidence that many cellular signaling pathways that might be engaged by an altered growth factor environment are activated. This is exemplified by the upregulation of the gene encoding the Fas antigen, a phenomenon that has been linked previously to the initiation of apoptotic signaling in other cell systems and is also reported to accompany castration-induced regression in the mouse and the rat prostate gland32. Induction of one of the functional forms of the signal transduction AP-1 transcription complex is also required for prostate cell apoptosis following castration. c-Fos, one of the early serum-response proteins associated with a cellular proliferative response, is induced in the rat ventral prostate soon after castration33,34. The early induction of c-Fos appears to be absolutely necessary for rat ventral prostate apoptosis after castration. Genetically altered mice that lack a functional c-Fos protein (cfos knockouts) fail to undergo prostate cell apoptosis after castration35. Intriguingly, the ventral prostate glands of these mice show increased amounts of AP-1 binding complexes, similar to the prostate glands of castrated normal mice. However, unlike the regressing prostates of a normal mouse, these AP-1 binding complexes are deficient in the c-Fos component. Within the second category of apoptotic effectors (apoptosis regulators), there is increasing evidence that members of the Bcl-2 family are important for the regulation of castration-
induced prostate cell apoptosis. This family of gene products is characterized by homologous conserved domains that enable them to form functional dimers with themselves (homoduplexes) as well as with other members of the bcl-2 protein family (heteroduplexes)36. The Bcl-2-related protein family is also functionally unique in that individual members can participate in one of two opposing functions: enhancement of the apoptosis response (Bax, Bad or BcXs) or suppression of the apoptosis response (Bcl-2, MC-1 or BclXL) when a cell is challenged by an apoptotic agent37,38. Two of the bcl-2 family genes (bax and bcl-2) have been characterized recently for expression patterns in the regressing rat ventral prostate gland. The prostate epithelial cells’ response to castration appears to include a significant modulation of bax and bcl-2 gene expression. Epithelial cells of the androgenically intact rat prostate gland synthesize both Bax and Bcl-2 at relatively low levels. However, after castration, both of these genes are upregulated in the rat ventral prostate gland, but in strikingly different temporal patterns, which can be correlated with both the onset and the subsequent cessation of apoptosis during regression39. Castration results in a rapid induction of products of the bax gene (mRNAs and protein) in the prostatic epithelium, reaching a peak level of expression that is ~13 times higher in the three-day castrated prostate epithelium than it is in control rats (Fig. 3C). Subsequently, this induced bax gene expression declines back to control levels in the regressed tissue that remains after long-term castration. In contrast, bcl-2 gene expression is induced rather slowly over the first few days after castration but, with longer periods of castration, this gene is very highly expressed in all surviving prostatic epithelial cells (Fig. 3A). When the castration expression patterns of these two gene products are superimposed, the cellular Bax to Bcl-2 ratio is highest on those days in which the prostate epithelial cells undergo the greatest amounts of apoptosis (days 2 and 3 after castration)39. This finding TEM Vol. 10, No. 2, 1999
is consistent with the opposing functional activity of these two gene products and with Korsmeyer and colleagues’ proposition that the cellular Bax to Bcl-2 ratio is analogous to a cellular rheostat that regulates a cell’s ability to respond to apoptotic agents40,41. Their hypothesis that a raised Bax to Bcl-2 ratio is associated with enhanced apoptosis seems to be well illustrated in this in vivo system of the regressing rat ventral prostate. Conversely, those epithelial cells that survive castration in this tissue have extremely high levels of Bcl-2 (and a low Bax to Bcl-2 ratio) when compared with the epithelial cells of an androgenically intact rat39,42, suggesting that the residual prostatic epithelial cells of the long-term castrated rodent prostate survive in the hormone-deficient environment because they have such high levels of Bcl-2. This principle appears to have relevance for human prostate cancers, where androgen deprivation therapy is one of the mainstays of treatment of the advanced form of the disease. Several surveys have shown that prostate cancer cells that survive hormone treatment can have highly raised levels of the Bcl-2 protein42–44. In addition, experimental studies have shown that Bcl-2 can change a hormone-sensitive prostate cancer cell line (LNCaP) into a hormone-insensitive line, when tested in nude mouse xenografting45. Thus, the Bcl-2 protein might be one of the preferred targets for the development of gene therapy or other therapeutic agents against human prostate cancer46. The tumor suppressor protein p53 also appears to play an important role in the regulation of some forms of apoptosis47. This protein, a transcription factor with multifunctional action, is best known for its role in suppressing cell cycle transition. Studies of cancer cells lacking p53 and the analysis of p53 ‘knockout’ mice have shown that certain apoptotic pathways are also lost to cells when p53 function is absent. Whether androgen-induced apoptosis of prostate cells is one of the apoptotic pathways influenced by p53 remains a highly debated issue. This TEM Vol. 10, No. 2, 1999
Figure 3. Photomicrographs of immunohistochemically stained thin sections (6 µm) of rat ventral prostate glands demonstrate induced synthesis of apoptotic regulatory proteins, Bcl-2 and Bax, during prostate regression. Sections of formalin-fixed, ventral prostate tissue from a mature male rat (A) or a seven-day castrated rat (B) were immunostained with a monoclonal antibody against Bcl-2. Red chromogenic substrate demonstrates a low level of the Bcl-2 protein throughout the intact prostate epithelium and much higher levels in epithelial cells that survive one week after castration. Sections of fixed ventral prostate tissue from a mature male rat (C) or a three-day castrated rat (D) were immunostained with a monoclonal antibody against the Bax protein. Red chromogenic substrate demonstrates a low level of the Bax protein throughout the intact prostate epithelium and greatly increased levels in the epithelial cells on the third day after castration, the peak period for epithelial cell apoptosis during castrationinduced prostate regression.
influence was proposed initially in studies in which it was found that p53 gene products (mRNA and protein) are transiently induced in prostate epithelial cells after castration, corresponding with the onset of apoptosis of these cells48,49. However, the prostatic epithelial cells in p53 knockout mice do undergo apoptosis after castration, albeit with a significant temporal delay as compared to p53 normal mice50. The results of these latter experiments suggest that the wild-type p53 protein is acting to promote apoptosis in prostate epithelial cells after castration. However, p53 does not appear to be an obligatory molecule in this pathway. There is a need for a third category to describe the influence of wildtype p53 protein on cellular apoptotic
pathways; ‘p53 sensitivity’ as opposed to only p53 dependence or independence. Finally, a number of proteolytic activities and enzymes are induced during castration-induced prostate regression. It can be anticipated that some of these proteolytic molecules will belong to the caspase family of gene products (and thus be important for the execution phase of apoptosis), but there is little information about which members of the growing caspase family might be most important for prostate cell apoptosis after androgen withdrawal. Other proteolytic enzymes induced after castration (for example, plasminogen activator, cathepsin B, cathepsin D and matrilysin)51–54 might be more involved in the degradation of the extracellular matrix that 51
RBF (ml min–1 gm–1)
0.5 0.4 0.3 0.2 0.1 0
Shamoperated
Castrated
Castrated, testosterone replenished
Figure 4. Relative blood flow (RBF; ml blood gm–1 tissue min–1) measurements in the ventral prostate glands of operated rats. A fluorescent microsphere perfusion method61 was used to measure relative blood flow to various tissues in groups of sham-castrated rats, surgically castrated rats, or surgically castrated rats that were replenished immediately with testosterone by a subcutaneous implant (timerelease testosterone-propionate pellet). There was a 50% reduction in relative blood flow to the ventral prostate glands of the 24-h castrated group (p
