New strategies for the medical treatment of prostate cancer


John T. Isaacs, Sidney Kimmel Comprehensive Cancer Center, Johns Hopkins University, 1650 Orleans St., CRB 1M44, Baltimore, Maryland 21231–1001, USA.


Androgen is a major growth factor in the normal prostate and determines the overall number of prostate cells. Metastatic prostate cancer, while initially responsive to androgen ablation, eventually becomes hormone-refractory and resistant to many treatments. Unfortunately, there are very few agents in the preclinical stage with a seemingly promising future for hormone-refractory prostate cancer (HRPC) that are actually taken through the complete drug development process, including US Food and Drug Administration approval. Many novel strategies under investigation for treating HRPC target metastatic prostate cancer cells that are neither androgen-dependent nor in the proliferative state. Examples of therapies that target this so-called ‘Achilles’ heel’ of HRPC include immune therapy, gene therapy, angiogenesis inhibition, and activation of programmed cell death. Unique properties of HRPC allow for the development of novel treatments that target prostate-specific antigen (PSA), human glandular kallikrein-2, or prostate-specific membrane antigen. An inactive prodrug with a thapsigargin analogue, a sesquiterpene lactone from the plant Thapsia garganica, is currently under investigation specifically for the targeted therapy of HRPC. Preclinical data suggest the PSA-targeting abilities of this novel therapy are associated with a nearly complete cessation of tumour growth with minimal toxicity.


hormone-refractory prostate cancer


Food and Drug Administration


prostate-specific membrane antigen


human glandular kallikrein-2


granulocyte-macrophage colony-stimulating factor


sarcoplasmic/endoplasmic reticulum calcium ATPase (pump).


An appreciation of the general biological principles that govern cancer growth is necessary to understand the logic behind novel approaches to the treatment of hormone-refractory prostate cancer (HRPC). For all cancers, the progression from normal cells to malignancy, and particularly to metastasis, involves a series of genetic changes. In normal tissue, the rates of cell proliferation and cell death are balanced so that there is neither overgrowth nor regression of tissues. In contrast, in cancer or neoproliferative disease, the rate of cell production exceeds that of cell death. This imbalance is targeted via therapeutic interventions that attempt to reduce the cell production rate to less than that of cellular death, resulting in regression of disease. If there is no such alteration death is the eventual result.


Various cancers, including embryonic tumours, malignant lymphomas, sarcomas, squamous cell carcinomas and adenocarcinomas, differ in doubling time (the amount of time required for the tumour to have twice as many cells) (Table 1). Embryonal tumours have a doubling time of ≈ 1 month, while typical adenocarcinomas take several months to double. Similarly, proliferation rates vary dramatically in accordance with the histological origin of the cancer. The proportion of daily cell growth for embryonal tumours is about half, compared with only 3% for adenocarcinomas. Although in this instance the proliferation rates differ by a log order, the actual overall growth rates differ by only two- to three-fold. This reflects the impact of the spontaneous death rates.

Table 1.  Mean kinetic parameters of various histological types of human tumours
Histological typeDoubling time, days% daily cell
  1. The numbers in parentheses indicate the number of patients for whom data were available. Adapted from Tubiana et al.[1]. Scientific Foundation of Oncology. Chicago: Year Book Medical Publishers Inc. 1975: 126–35.

Embryonal tumours27 (76)4946
Malignant lymphomas29 (41)4845.5
Mesenchymal sarcomas41 (87) 6 4.1
Squamous cell carcinomas58 (51)13.612.1
Adenocarcinomas83 (134) 3.3 2.3

The importance of understanding the relationship between the proliferation rate and cell-death rate is best shown with concrete examples. For instance, malignant lymphoma is a relatively fast-growing cancer, with 48% of cells proliferating per day and 45.5% of cells dying per day [1]. The net result is a positive 2.6% difference that allows the tumour to double every 29 days. If a cytostatic drug is given that slows the rate of cellular proliferation by 10%, the rate of cell proliferation changes to 44.8%. Because the rate of cells dying per day is unchanged (45.5%), the net difference is negative, producing a 0.7% daily loss in cancer cells, meaning that the cancer is potentially curable.

If the same cytostatic agent were used in adenocarcinomas, the result would be quite different. Untreated typical adenocarcinoma has a proliferation rate of 3.3% and a death rate of 2.3%, resulting in a doubling time of 83 days [1]. The same chemotherapy that successfully treated malignant lymphoma would reduce the proliferation rate in typical adenocarcinoma by 10% as well, from 3.3% to 3.0%. However, given the rate of cellular death per day (2.3%), the net change is an increase of 0.7% in the number of cancerous cells produced per day, resulting in a prolongation of the doubling time to 116 days. Thus, the same chemotherapeutic agent that effectively treated malignant lymphoma would not produce the same outcome in typical adenocarcinoma. These examples serve to demonstrate the important role that death rates of various histological types of human cancers play in the progression or regression of disease (Table 1).


The positive overall accumulation of cancerous cells in prostate cancer is similar to that in the typical adenocarcinoma example given above. The differential between the rates of cell proliferation and death is such that cancer cells are accumulated. The actual magnitude of this imbalance is problematic in curing HRPC. Prostate cancer is a disease of the epithelial cells of the glandular component of the prostate, with high-grade prostatic intraepithelial neoplasia as the precursor for most peripheral-zone prostatic carcinomas [2,3]. Using archival preparations from pathology and autopsy studies, we measured the rates of cell proliferation and death based on immunocytochemical staining [4]. Remarkably, 99.6% of cells in the normal prostate are neither proliferating nor dying; they are in a proliferatively quiescent state of Go. The low rate of proliferation is balanced with an equally low rate of cell death. By contrast, in metastatic prostate cancer occurring after androgen ablation, the rate of cellular proliferation increases from the 0.2% seen in normal prostate cells to up to 3.1%, but there is no concomitant increase in the rate of cellular death [5]. This 1–2% difference in survival is associated with a 3-month doubling time, which is sufficient to cause death.

The growth fractions in normal tissues other than the prostate have quite different rates of proliferation. For bone marrow, ≈ 25% of cells are in cycle each day, while in the colon the proportion is 10–15%; in the skin, 7–8% of cells are proliferating [6]. Chemotherapeutic agents with no specificity for malignant vs normal cells, and that induce cell death as the cell goes through its cycle, have a smaller target in the prostate than in other tissues. Contrary to popular belief, men with advanced HRPC develop metastatic disease in many sites, including the bone and soft tissue [7]. Thus, any therapeutic agent for advanced disease must be effective in these areas. Similar to proliferation rates found in early-stage prostate cancer, the median growth fraction for metastatic prostatic cancer cells remains low, although its small increase is sufficient to eventually cause death.

The number of cell lines and model systems available for the study of prostate cancer are limited. LNCaP, the human prostatic epithelial cell line, is one of the most commonly used models. When grown in culture, 95% of the cells of LNCaP are in cycle [7], rendering the application of in vitro data to the clinical situation problematic.


In the normal prostate, androgen is the major growth factor through two mechanisms: (i) it is an agonist that stimulates cellular proliferation of prostatic epithelial cells; (ii) it is also an antagonist that blocks the potentially high apoptotic rate of epithelial cells. The net balance determines the overall number of prostate cells [8,9].

In the 1940s, Charles Huggins found that androgen ablation of metastatic prostate cancer was associated with beneficial palliative effects [10]. However, despite decades of use, this treatment is not curative for metastatic disease because it only affects cancer cells that are dependent upon androgen for growth and survival [11]. Indeed, death rates due to metastatic prostate cancer continued to increase even after the introduction of androgen ablation [12]. Androgen-independent, also termed hormone-refractory prostate cancer (HRPC), cells derived through genetic changes from hormonally dependent cells remain able to proliferate [13]. This genetic instability allows for the production of a variety of genetic and phenotypically different clones. Some of these clones may be resistant to radiation, hyperthermia, chemotherapy, or androgen ablation.

In the normal prostate, the androgen axis is a highly regulated pathway. Access to androgen is a paracrine interaction in which the stromal cells are the site for androgen-receptor engagement leading to the production of peptide growth factors [13,14]. These growth factors diffuse across the basement membrane into the epithelial components, affecting the growth and survival of the basal and luminal cells where the stem cells are located. Androgen receptors in the epithelium are involved in producing differentiation markers, including PSA and human glandular kallikrein-2 (HK2). Evidence suggests that the androgen receptor acts as a negative regulator of growth in the luminal cells, preventing them from continuing to grow [14]. During prostatic carcinogenesis, the androgen receptor is captured as a direct regulator for the proliferation and growth of malignant cells. This represents a major change; the paracrine two-cell interaction in the normal prostate is converted to an autocrine pathway in which the cancer cell itself engages the androgen receptor.


Because androgen ablation does not eliminate androgen-independent cancer cells in advanced prostate cancer, scientists have been studying methods to therapeutically target cells in a proliferation-independent manner. Many investigators believe that the ‘Achilles’ heel’ for prostate cancer may be the activation of apoptosis of HRPC cells without the requirement of cellular proliferation [2]. There are several therapeutic approaches that, theoretically, can activate apoptosis in cells that are not highly proliferative.


In the cancer setting, the body's immune system is unable to normally manage the cellular environment; thus, there is increased production of new cells. Suppression of the immune system results in a lack of full response to foreign antigens (i.e. cancerous cells) in part due to the way in which the antigens are presented. The idea behind vaccine therapy is to provide what are in essence growth factors that stimulate production of a systemic antitumour effect. One vaccine that has been studied is GVAX®[15]; this consists of irradiated, allogeneic, prostate carcinoma cell lines transduced ex vivo with the human granulocyte-macrophage colony-stimulating factor (GM-CSF) gene. GM-CSF stimulates dendritic cells to heighten their immune response to the cancer cells [16]. There was a decrease in the PSA level in patients with prostate cancer treated with immune therapy, but the question remains as to how much prostate cancer can actually be eliminated with vaccine therapy. A phase III registration trial to produce such data is presently being conducted, supported by Cell Genesys, within the USA.


Many strategies using gene therapy are under evaluation in oncology; this method is an exciting possibility for targeting HRPC cells in the nonproliferating state [17,18]. Briefly, gene therapy takes advantage of the fact that normal and malignant prostate cells express unique genes, including PSA, HK2, and prostate-specific membrane antigen (PSMA). These genes have been cloned for their expression; moreover, the promoter/enhancer sequences for them are now known. In one type of gene therapy, an adenovirus is prepared using prostate-specific promoters or enhancers. When taken up by non-prostatic cells the virus is not transcribed. However, when taken up by prostatic cells, the virus is transcribed and a cytolytic response is generated. The prostate cells do not have to be replicating to be infected with the adenovirus, making gene therapy a rational strategy for combating HRPC. CV706, a replication-competent, E3-deleted, cytolytic Ad5 adenovirus, has been associated with decreasing PSA levels in men with recurrent prostate cancer previously treated with definitive external beam radiation therapy [18].


The inhibition of angiogenesis offers an innovative approach to treating HRPC. The endothelial cells, which act as the ‘supply train’ for the malignancy, represent an attractive target for therapy; in theory, the angiogenic response can be blocked. A further discussion of this treatment approach, which is beyond the scope of this review, is provided in this supplement by Petrylak.


An exciting series of approaches has been undertaken to try to activate programmed cell death in HRPC. The rational development of agents with this mechanism of action is premised on the fact that proliferation and survival of cells depend on a series of signals that the cell actually receives. Some of these signals originate at the plasma membrane via receptors interacting with appropriate ligands. The sum of these signals allows for regulation of both proliferation to replace cells that have undergone apoptosis and minimisation of apoptosis. If the particular survival and proliferative pathways that are uniquely required for a cell type were identified, molecules could be made that act at different levels of inhibition. For example, a molecule could be constructed to block the ligand, the receptor signalling, or a process even further downstream. Although innovative and exciting, this approach requires identification of survival proliferation pathways specific to cancer. Targeting general survival pathways present in all epithelial cells would cause problems. Fortunately, cancers undergo molecular and genetic changes that render them dependent upon specific pathways, making targeting slightly less cumbersome. For example, gefitinib inhibits the growth of tumours expressing the epidermal growth factor receptor, a tyrosine kinase that has been implicated in initiation and progression of a variety of cancers [19,20].


As previously mentioned, both the localized and metastatic antigen-resistant growth fraction is very low in prostate cancer cells, yet extremely high in the LNCaP model. In this model, apoptosis is induced well by treatment with paclitaxel, doxorubicin or thapsigargin [21]. However, when LNCaP tumour cells are grown in media that have been conditioned by osteoblasts, these cells grow with a much lower growth fraction, while remaining viable [21]. Therefore, the use of modified LNCaP is a better model in terms of ‘real-world’ application to humans. The modified LNCaP more accurately represents the in vivo low-proliferation rate. With this model, the death response induced by doxorubicin and paclitaxel is lost, while thapsigargin maintains its ability to induce apoptosis (Fig. 1).

Figure 1.

The percentage of apoptosis in LNCaP cells after treatment with vehicle, doxorubicin, paclitaxel or thapsigargin. The conventional, unadulterated LNCaP model was the high-proliferation culture, while modified LNCaP (cells grown in media conditioned by osteoblasts) was the low-proliferation culture. Reprinted from Denmeade et al., J Natl Cancer Inst 2003; 95: 990–1000, with permission of Oxford University Press [21].


Thapsigargin is a sesquiterpene lactone isolated from the plant Thapsia garganica (Fig. 2) [21]. Thapsigargin, which historically has been termed the ‘death carrot’, is a lipophilic molecule that has been used in Arabic medicine for 2000 years. The endoplasmic reticulum is a store of calcium; thapsigargin is a potent inhibitor of the sarcoplasmic/endoplasmic reticulum calcium ATPase (SERCA) pump. The SERCA pump is found in every mammalian cell and is physiologically required for the endoplasmic reticulum in cells to function. When the SERCA pump is inhibited by thapsigargin, a rapid 3–5-fold elevation in intracellular free Ca2+ (Cai) results, and the pool of Ca2+ in the endoplasmic reticulum is depleted. Apoptotic death of these cells occurs independent of cellular proliferation [21,22].

Figure 2.

The chemical structures of (A) thapsigargin and L12ADT, chemically modified thapsigargin and (B) thapsigargin prodrug, Mu-HSSKLQ//L12ADT, which is modified thapsigargin coupled with a carrier peptide. Reprinted from Denmeade et al., J Natl Cancer Inst 2003; 95: 990–1000, with permission of Oxford University Press [21].

Because the activity of thapsigargin is not prostate- or cell type-specific, systemic administration would be associated with significant toxicity [21]. Therefore, an inactive prodrug with a carrier peptide was formulated that does not allow the drug to enter a cell until the thapsigargin analogue is liberated via proteolytic digestion (Fig. 2) [21,22]. This prodrug, named Mu-HSSKLQ//L12ADT, is specific for PSA. PSA is a serine protease with chymotrypsin-like substrate specificity that is produced in most prostate cancer sites. This useful targeting mechanism is proteolytically active in the extracellular fluid of prostate cancer, while it is inactive in the bloodstream [21].

In a LNCaP xenograft model with a PSA-producing cancer, the Mu-HSSKLQ//L12ADT prodrug was found to be associated with a reduction in tumour growth and near complete cessation of tumour growth (Fig. 3A) [21]. The PSA-targeting effect of the prodrug was shown in a group of mice with a non-PSA-producing human SN12C RCC. In this model, the prodrug had no effect on tumour growth rate (Fig. 3B). Thus, proteolytically active PSA within the tumour is a requirement for the activation of thapsigargin's prodrug and subsequent antitumoural effect. Importantly, there was no discernible toxicity in these in vivo studies.

Figure 3.

In vivo effect of Mu-HSSKLQ//L12ADT prodrug vs vehicle control (10% DMSO) on xenograft tumours. (A) Mice carrying PSA-producing LNCaP xenografts received 28 days of continuous infusion delivered by two 14-day subcutaneously implanted osmotic mini-pumps. (B) Mice carrying non-PSA-producing SN12C human renal cell xenografts were treated with one 14-day continuous infusion of prodrug or vehicle.


On an international basis, only 10% of the pharmaceutical budget is spent on research and development in oncology. However, in the biotechnology world, 90% of the finance for cancer is spent in this area. Academic and biotechnology researchers are working diligently to discover novel treatments for prostate cancer that exploit the unique biology of the disease, e.g. PSA. However, such therapies may meet resistance when presented to pharmaceutical companies due to the fiscal constraints typifying the oncology drug development process, as well as the lengthy period required to bring a product to fruition. A drug that is potentially applicable to many cancers rather than a drug that exploits a genetic pathway particular to a single malignancy is preferred. For example, anti-angiogenic agents can be tested in many cancers because they inhibit processes that are uniformly present. Companies that are willing to develop an agent specifically targeted towards prostate cancer are making a stronger commitment and taking greater risks than companies developing a general anticancer treatment such as chemotherapy.

The cost of developing a new oncology drug is staggering, ranging up to an estimated US$170 million dollars (Table 2). Moreover, an agent being evaluated could fail at any point along the continuum from drug selection to Food and Drug Administration (FDA) review, which alone costs $25–40 million [23]. Some companies are faced with budgetary constraints that allow for long-term development of only one among many promising cancer treatments. The cost of drug development may be reduced if the processes involved in early development, that is, phase 1 (safety), could be used in a manner that yielded more information, such as appropriate dosage regimen, so long as there was an acceptable level of risk. To lower drug development costs, pharmaceutical companies have also used the strategy of withholding investment until the agent is eligible for phase 2 development. Nevertheless, better and earlier cooperation between individual researchers and pharmaceutical companies best facilitates the development of appropriate treatments.

Table 2.  Oncology drug development process: time and costs. Adapted from DiMasi et al., J Health Econ 2003; 22: 151–85
Development stageTime, yearsCost (US$ millions)
Drug selection≈2.5≈4
Preclinical toxicity 1≈2
Phase I, safety 112
Phase II, efficacy 212
Phase III, efficacy, tumour progression, quality of life 3–470–100
FDA review 1–225–40


The relationship between the rates of cellular proliferation and cell death determine the growth of a cancer [4]. When the rate of proliferation exceeds the rate of death, tumour growth continues. However, when the rate of death supersedes the rate of proliferation, the cancer regresses. Prostate cancer is initially quite responsive to androgenic blockage, but relapse to a state of unresponsiveness to further antiandrogen therapy eventually results. These factors make the diagnosis of HRPC associated with a poor prognosis; efforts directed towards developing new treatment strategies are mandated. In addition, intermediate markers other than PSA that may provide an insight into the activity of prostate cancer-targeted therapies should be investigated.

Because HRPC has a lower growth fraction than other cancers, novel treatments are needed that can activate apoptosis in cells without requiring them to be highly proliferative. Such strategies under investigation and with promising results include immune therapy, gene therapy, angiogenesis inhibition, and activation of the programmed cell-death pathway. Unique properties of the prostate allow for the development of novel prostate cancer-specific treatments and may include targeting of PSA, HK2, or PSMA in the future. Such targeted therapy may be the next generation of treatment for HRPC. One novel therapy under consideration that capitalises on the biology of prostate cancer is thapsigargin, which is linked to a carrier peptide specifically targeted to PSA.

Thapsigargin is highly toxic and able to induce apoptosis in quiescent and proliferating prostate cancer cells. However, this cytotoxicity is not likely to be prostate-specific and would be associated with significant toxicity if administered systemically. Therefore, a tight system involving the coupling of a primary amine-containing thapsigargin analogue to a peptide carrier to produce an inactive prodrug has been developed. Small amounts of this prodrug can be released in noncancerous cells without resulting in death. Preclinical data suggest the PSA-targeting abilities of this novel therapy are associated with a nearly complete cessation of tumour growth with minimal toxicity. Further study of the PSA-activated thapsigargin prodrug approach for the treatment of HRPC is warranted.