- Top of page
- MATERIALS AND METHODS
- CONFLICT OF INTEREST
Prostate cancer is the most commonly diagnosed malignancy in men in the UK, with 24 710 cases being reported in 1999 (http://www.cancerresearchuk.org/statistics, accessed 2003). At the time of diagnosis about a quarter of men with prostate cancer in the UK will have distant metastases, and ≈ 90% of these men will have bone involvement . The development of bony metastases correlates well with the clinical stage and grade of the tumour . Bone metastases are frequently symptomatic and often responsible for a considerable deterioration in the quality of life in those with advanced disease, being the most frequent cause of morbidity. Potential complications are numerous, and include intermittent or constant bone pain, pathological fractures, vertebral collapse and spinal cord compression. These are grouped together as skeletal-related events.
For some time the treatment options for the patients with bony metastases have been limited. There was therefore considerable interest generated by the recent publication of results showing a reduction in the incidence of skeletal-related events in these patients when treated with zoledronic acid . Zoledronic acid is a potent member of the bisphosphonate class of drugs; these agents inhibit bone resorption and so have become established therapeutic options for treating diseases such as Paget's disease, postmenopausal osteoporosis and general tumour-associated bone disease. Bisphosphonate were shown to reduce bone resorption by strongly binding to bone, because of their high avidity for calcium, and inhibiting osteoclast function . It was later established that the more potent nitrogen-containing bisphosphonates, including zoledronic acid, achieved this by inhibiting the mevalonate pathway [5,6]. This pathway usually provides the cell with the isoprenoid lipids farnesyl pyrophosphate (FPP) or geranylgeranyl pyrophosphate (GGPP). These are required for the prenylation of certain proteins, a form of post-translational modification. G-proteins, such as Ras, Rho and Rac, need to be prenylated before they can localize to the membrane and thus function normally . By inhibiting the processing of these proteins, which are known to be essential for normal cell survival and function, nitrogen-containing bisphosphonates disrupt intracellular processes and can cause apoptosis.
Although metastases from prostate cancer are generally sclerotic it is thought that bone resorption represents a vital initial step [8,9]. The action of bisphosphonates on osteoclasts alone represented a rationale for their use in tumour-associated bone disease, as a reduction in osteoclast activity would reduce the release of growth factors in the bony environment and make the bone less fertile ‘soil’ for the seeding of tumour cells . However, an exciting development came when various bisphosphonates were shown to have direct in vitro effects on tumour cells. They can inhibit cell proliferation and viability, and induce apoptosis in several cancer cell lines [11–15]. Work in our laboratory with breast cancer cells provided early evidence that inhibiting the mevalonate pathway was again central to the apoptotic effects of zoledronic acid , with impaired membrane localization of Ras proposed as the underlying cause of the apoptotic effect. We recently published similar results for various prostate cancer cell lines .
One criticism of work showing the apoptotic effect of bisphosphonates has been that relatively high concentrations have usually been required to detect a significant effect, e.g. 100 µmol/L zoledronic acid [13,14,16]. It is difficult to accurately assess the physiological concentrations that bisphosphonates reach in the bony microenvironment, where they will be much greater than plasma concentrations because of their great avidity for bone. While one study suggested concentrations as high as 800 µmol/L could be reached , others report values nearer 1–10 µmol/L . To show whether apoptosis could be induced at lower concentrations of zoledronic acid we used a sensitive cell-death detection ELISA in two prostate cancer cell lines, and present the results herein.
Caspases are known to be important in the most common forms of apoptosis, and zoledronic acid was previously shown to induce caspase activation in prostate and breast cancer cells [13,16]. In the present study we also assessed to what degree any observed effects of zoledronic acid on apoptosis were caspase-dependent, by co-treating cells with zoledronic acid and a broad-spectrum caspase inhibitor.
Although the precise mechanisms whereby bony metastases develop from circulating tumour cells are not known, adhesion to the mineralized matrix is likely to be a crucial event. Other bisphosphonates inhibit the adhesion of prostate cancer cells to cortical bone slices [19,20]. In the present study we also report the magnitude of effect of zoledronic acid on the adhesion of prostate cancer cells to mineralized dentine slices, and data evaluating how any observed effects of zoledronic acid on adhesion or apoptosis are altered by manipulation of the mevalonate pathway.
Of the G proteins discussed earlier, RhoA is known to be important in functions involving re-organization of the cytoskeleton, including cellular adhesion [21,22]. We therefore also investigated the consequences of directly inhibiting RhoA on cell adhesion and apoptosis, and compared these with the effects of zoledronic acid.
MATERIALS AND METHODS
- Top of page
- MATERIALS AND METHODS
- CONFLICT OF INTEREST
The DU145 prostate cancer cell line was obtained from the American Type Culture Collection (Virginia, USA), and the PC-3 cell line from the European Cell Culture Collection (Wiltshire, UK). DU145 cells were maintained in RPMI medium and PC-3 cells in DMEM. Both media were supplemented with 10% fetal calf serum, 2 mmol/L glutamine, penicillin 100 µg/mL and streptomycin 100 µg/mL. Cell cultures were maintained in subconfluent monolayers, at 37 °C in 5% CO2, and passaged using 10% trypsin, as required.
Zoledronic acid ([1-hydroxy-2-(1H-imidazol-1-yl)ethylidene] bisphosphonic acid) was obtained from Novartis Pharmaceuticals Ltd (Basel, Switzerland). Stock solutions were prepared from the hydrated disodium salt in PBS. To help determine the molecular mechanisms underlying the effects of zoledronic acid, the following reagents were used to investigate the possible roles of caspase activation and altered protein prenylation of G proteins. Z-VAD-FMK (Promega, Southampton, UK) was used as a broad-spectrum caspase inhibitor, and was reconstituted in PBS, being added 3 h before any addition of zoledronic acid.
Farnesol mixed isomers (FOH, 3,7,11-trimethyl-2,6,10-dodecatrien-1-ol) and geranylgeraniol (GGOH, 3,7,11,15-tetramethyl-2,6,10,14-hexadecatetraen-1-ol) (both from Sigma, Poole, UK), as cell-permeable analogues of FPP and GGPP, were added to cells 3 h before treating with zoledronic acid, to bypass any suppression of prenylation caused by inhibition of the mevalonate pathway. Both were reconstituted in 100% ethanol.
Manumycin A (Sigma) and GGTI-298 (Calbiochem, Nottingham, UK) were used to directly inhibit protein prenylation, and compare the results with zoledronic acid. Manumycin A is a naturally occurring farnesyltransferase inhibitor (FTI); GGTI-298 potently inhibits geranylgeranyltransferase; both were reconstituted in DMSO. C3 exoenzyme (Cytoskeleton Inc., Denver, USA), reconstituted in double-distilled water, was used to directly inhibit RhoA function. All reagents were diluted in culture medium before adding to cell cultures.
DETECTION OF APOPTOSIS
The cell-death detection ELISA kit (Roche Diagnostic Ltd., Lewes, UK) was used according to the manufacturer's instructions. This assay measures cytoplasmic histone-associated DNA fragments (mono- and oligonucleosomes), produced after cell death, using mouse monoclonal antibodies against histones and DNA. Briefly, cells were lysed in situ after being treated in 96-well plates, and aliquots of lysate transferred to a streptavidin-coated multiplate. Lysates were incubated for 2 h with immunoreagent containing the antibodies against histones and DNA, which are conjugated to materials that allow for simultaneous capture of the immunocomplex to the plate and photometric quantification of the DNA. The optical density was read at 405 nm, with a reference wavelength of 490 nm, with samples measured in duplicate.
CELL ADHESION TO MINERALIZED MATRIX
Cells were seeded out in T25 or T75 flasks, using numbers such that control cells reached 80–90% confluence after 48 h. After leaving overnight the medium was replaced with that containing the test reagent(s). After a further 24 h cells were harvested nonenzymatically from T25 flasks, using 0.05% EDTA solution. Cells were seeded at a density of 2 × 104 cells/mL into a 96-well plate containing mineralized dentine slices, which were shaped to fit into individual wells (kindly supplied by Dr Tim Arnett, University College London, UK). Three discs were used per treatment. After 24 h the slices were washed in PBS, and cells fixed, stained, mounted on slides and counted. The fixing agent used was 2% sucrose/3% paraformaldehyde; 1% toluidine blue in 1% sodium borate was used as the dye.
Statistical significance of any differences was determined using Anova.
- Top of page
- MATERIALS AND METHODS
- CONFLICT OF INTEREST
Zoledronic acid caused apoptosis in both cell lines, as measured by the ELISA, related to time and concentration (Fig. 1). Results are expressed as an enrichment factor (EF), which represents a ratio of the amount of histone-associated DNA detected in treated cells to that in control cells. The manufacturer's instructions state that an EF of > 2 can be considered significant. There was significant apoptosis with the 100 µmol/L zoledronic acid after 24 h of treatment in both cell lines, and in DU145 cells after 72 h treatment with 1 µmol/L (Fig. 1).
Figure 1. Induction of apoptosis in prostate cancer cells by zoledronic acid; cells were treated with zoledronic acid (1, green; 10, red; 100 µmol/L, light red) or vehicle control (C, blank) for 1–3 days and apoptosis assessed using an ELISA. The results are expressed as the EF for (A) DU145 and (B) PC-3. For each period *P < 0.05 and ** < 0.01, vs control.
Download figure to PowerPoint
The broad-spectrum caspase inhibitor Z-VAD-FMK at 20 µmol/L reduced the apoptotic effect of 10 µmol/L zoledronic acid by 76.3% (P = 0.011) in DU145 cells and 90.8% (P = 0.025) in PC-3 cells (Table 1). When co-treating with zoledronic acid and the isoprenoid lipid analogues, GGOH had a greater effect than FOH. In DU145 cells, GGOH attenuated apoptosis by 94.4%, compared to 67.7% with FOH (P = 0.05 vs zoledronic acid + FOH). In PC-3 cells, GGOH abrogated apoptosis by 99.7%, compared to 34.3% with FOH (P = 0.02). Compared with cells treated with zoledronic acid alone the reduction in apoptosis was highly significant with GGOH (P < 0.005 in both lines). With FOH it was less significant (P < 0.05 for DU145 and 0.08 for PC-3 cells; Table 1). Both inhibitors of prenylation caused significant degrees of apoptosis. Manumycin-A and GGTO-298 treatment gave greater EFs (P < 0.01 and 0.02) than in the respective controls, but inhibiting RhoA with C3 exoenzyme did not cause significant apoptosis (Table 1).
Table 1. The effects of zoledronic acid, modulators of protein prenylation and C3 exoenzyme on the induction of apoptosis and on cellular adhesion to mineralised matrix in prostate cancer cells. Cells were exposed for 72 h (apoptosis) to various reagents, or pre-treated with reagents of interest for 24 h (adhesion) (FOH/GGOH added 3 h before zoledronic acid)
|Mean (Sd) variable||EF||Adhering cells, % of control|
|Control||1.00 (0.230)||1.00 (0.129)|| || |
|Z-VAD||1.039 (0.133)||0.913 (0.143)|| || |
|Zoledronic acid||5.255 (0.287)†||7.724 (1.281)*|| || |
|+ Z-VAD||2.007 (0.379)**||1.619 (0.519)**|| || |
|Control||1.00 (0.161)||1.00 (0.080)||100 (8.669)||100 (12.03)|
|Zoledronic acid||4.116 (0.092)†||5.075 (0.346)†|| 31.63 (8.755)‡|| 20.06 (5.144)‡|
|+ FOH||2.007 (0.266)††||3.676 (0.497)|| 29.34 (10.696)|| 37.15 (7.760)**|
|+ GGOH||1.174 (0.131)††||1.011 (0.192)††|| 89.48 (12.554)††|| 71.19 (12.08)††|
|Manumycin-A||8.484 (0.701)†||8.594 (0.840)†|| 44.41 (9.674)†|| 76.88 (10.42)|
|GGTI||6.068 (0.930)*||7.148 (0.257)†|| 40.80 (9.337)†|| 49.36 (10.17)†|
|C3-exoenzyme||1.262 (0.353)||1.932 (0.231)|| 29.04 (4.874)‡|| 50.71 (4.867)†|
ADHESION TO MINERALIZED MATRIX
Representative micrographs of DU145 cells adhering in various treatment conditions are shown in Fig. 2. There was a clear dose-response relationship for zoledronic acid in both cell lines (Fig. 3). Lower concentrations of zoledronic acid had a more significant inhibitory effect on adhesion than was shown for apoptosis, e.g. 1 µmol/L reduced the number of adhering cells by 38.3 (6.5)% (P = 0.012 vs control) and 25.3 (9.3)% (P = 0.021) in DU145 and PC-3 cells, respectively.
Figure 2. Representative micrographs of DU145 cells adhering to mineralised dentine slices. Cells were treated with agents for 24 h before being seeded onto dentine slices for 24 h, and then fixed and stained: (A) control; (B) zoledronic acid, 100 Umol/L; (C) C3 coenzyme 5 µmol/L. Original × 100.
Download figure to PowerPoint
Table 1 also shows the effects of zoledronic acid, various modulators of protein prenylation and C3 exoenzyme on cellular adhesion. GGOH was more effective in attenuating zoledronic acid-induced inhibition of adhesion than FOH. In DU145 cells, GGOH reduced the inhibitory effect of zoledronic acid by 84.6%, while FOH failed to reduce the effect (P < 0.005 vs zoledronic acid + FOH). In PC-3 cells, GGOH abrogated the effects of zoledronic acid by 64.0%, compared to 21.0% with FOH (P = 0.01). Compared with cells treated with zoledronic acid alone, co-treating with GGOH significantly abrogated the effects of zoledronic acid (P < 0.005 in both lines). FOH was much less effective (no significant effect in DU145 cells, P = 0.03 for PC-3 cells).
For direct inhibition of protein prenylation both GGTI-298 and manumycin-A inhibited adhesion in DU145 cells (P < 0.01 vs controls), while GGTI had a greater effect than manumycin-A in PC-3 cells (P = 0.008 and 0.09 vs control, respectively). Treatment with C3 exoenzyme caused significant reductions in the number of adhering cells in both lines (P = 0.002 and 0.009; Table 1).
- Top of page
- MATERIALS AND METHODS
- CONFLICT OF INTEREST
Zoledronic acid induced significant apoptosis in two prostate cancer cell lines at lower concentrations and after a shorter treatment duration than previously reported in several other studies using prostate cancer or other cancer cells [11,13,23,24]. There was significant apoptosis in DU145 cells with 1 µmol/L zoledronic acid after 72 h exposure; 10 µmol/L caused significant apoptosis in both cell lines after 48–72 h, and 100 µmol/L induced apoptosis after 24 h. We found only one study showing significant apoptosis with these concentrations, in myeloma cells .
Previous work in our laboratory showed that inhibition of cell growth in DU145 cells exposed to zoledronic acid was characterized by DNA fragmentation, an altered Bcl-2 to Bax ratio, and caspase activation, all features of apoptotic cell death . We used a sensitive apoptosis ELISA to confirm apoptotic effects at lower concentrations, by directly detecting increased levels of histone-associated DNA. Such concentrations are more likely to be physiologically relevant, as not only are they expected to be exceeded locally in bony resorption lacunae , but are also within the peak plasma concentrations of zoledronic acid that may be achieved in vivo.
As previously shown for viability in DU145 cells , we showed that the apoptotic effect of zoledronic acid depends on caspase activation, as it is considerably attenuated when a broad-spectrum caspase inhibitor is added. In previous work we reported that zoledronic acid specifically activates caspase-3 .
It is now widely accepted that NBPs principally act via inhibition of the mevalonate pathway . More specifically, NBPs specifically inhibit the FPP synthase enzyme , which is required for producing FPP and subsequently GGPP. These isoprenoid lipids usually allow prenylation of G proteins, required for their membrane localization; some are normally farnesylated, others are geranylgeranylated . Thus, NBPs cause accumulation of unprenylated proteins, disrupting various downstream signalling pathways.
Ras is known to transduce important cell survival signalling, e.g. via the Ras/MAPK/ERK pathway . It was previously postulated that inhibition of Ras function may underlie the apoptotic effects of zoledronic acid, and impaired membrane localization of Ras was reported in breast and prostate cancer cells [13,16]. The various forms of Ras are usually farnesylated to allow their normal processing .
The present study indicates that inhibiting farnesylation activity by treatment with the FTI manumycin-A is sufficient to induce apoptosis in both DU145 and PC-3 cells, suggesting a critical role for farnesylated proteins such as Ras in maintaining cell survival. However, GGTI-298 also induced apoptosis, implying that signalling via geranylgeranylated G proteins is also involved in sustaining survival.
Against the proposal that Ras is the relevant target protein for zoledronic acid-induced apoptosis in prostate cancer cells was the finding that the effects of zoledronic acid were attenuated more by co-treating with the isoprenoid GGOH than FOH. These cell-permeable analogues of GGPP and FPP allow protein prenylation, despite inhibition of the mevalonate pathway. This finding implies that inhibiting geranylgeranylation with zoledronic acid is more crucial for its apoptotic effect than inhibiting farnesylation. Furthermore, ras mutations appear to be very rare in prostate cancer, having been found in only one of 19 tumour samples, and not in DU145 or PC-3 cells . However, PC-3 cells can over-express Ha-Ras .
We also showed a concentration-dependent inhibition of adhesion of prostate cancer cells to mineralized matrix with zoledronic acid treatment. There was significant inhibition with 1 µmol/L zoledronic acid, supporting previous findings with other NBPs in prostate and breast cancer [19,20]. To identify if common pathways are involved in the actions of zoledronic acid on cell death and adhesion, the consequences of modulating protein prenylation were again assessed. GGTI-298 inhibited adhesion of both prostate cancer cell lines by 50–59% but manumycin-A induced similar inhibition of adhesion in DU145 cells and less so in PC-3 cells. Several other studies that have directly compared effects of FTIs and GGTIs on various cell functions in other cell types have found GGTIs to be more potent [15,31–33]. However, it is to be expected that effects of prenylation inhibitors on cell dynamics will be complex, and may vary among cell lines according to the inherent relative expression of various G proteins. For example, myeloma cell lines with certain activated ras mutations are more sensitive to FTIs than counterparts with wild-type ras.
As with apoptosis, the present results show that effects of zoledronic acid on prostate cancer cell adhesion were abrogated to a greater degree by co-treating cells with GGOH than with FOH. This is consistent with other studies examining the effects of NBPs on various cell functions in osteoclasts [35,36], and myeloma , breast  and prostate cancer cells [13,39]. That FOH has partial rescuing effects may be a result of farnesylated proteins also normally contributing to the observed cell functions, or it has been postulated that FOH may be converted intracellularly to GGPP via FPP .
However, the suggestion that apoptosis and adhesion in these prostate cancer cell lines are both more susceptible to inhibition of geranylgeranylation than farnesylation is at variance with the observed effects of manumycin-A on these processes. The choice of FTI in the present study may explain the unexpected magnitude of its effects. Manumycin-A was the first FTI to be reported and while it potently inhibits its target enzyme , there are suggestions that it has a broad range of other activities distinct from its farnesyltransferase activity , which may contribute to the overall observed effects in this study. Previous work with a different FTI (L-744 832) in prostate cancer cell lines showed inhibition of cell growth, but no apoptosis . We could find no other studies that have reported effects of GGTIs on prostate cancer cells.
Given that inhibition of geranylgeranylation may be central to the effects of zoledronic acid on apoptosis and adhesion, we also examined the consequences of directly inhibiting RhoA, a G protein which is known to be geranylgeranylated . This was achieved using C3 exoenzyme, a specific exotoxin derived from Clostridium botulinum. A concentration of 5 µg/mL was chosen, as this has been shown to cause significant effects in breast cancer cells, inhibiting invasion . This concentration of C3 exoenzyme caused significant inhibition of prostate cancer cell adhesion, but not apoptosis. This implies that RhoA is central in adhesion, consistent with its known function in the assembly and organization of the cytoskeleton [21,42]. RhoA may well be the crucial target G protein for the effects of zoledronic acid on cellular adhesion. While others reported a small degree of apoptosis in endothelial cells with higher concentrations of C3 exoenzyme , our results show that apoptosis in prostate cancer cells is not sensitive to the same degree of RhoA inhibition that suppresses adhesion. The apoptotic effect with zoledronic acid is likely to involve inhibition of function of other geranylgeranylated G proteins, whose normal activities maintain cell survival when RhoA is specifically inhibited with 5 µg/mL C3 exoenzyme. Inhibited processing of one such protein, Rap1A, has been shown in myeloma cells with concentrations of zoledronic acid as low as 1 µmol/L .
Zoledronic acid is being administered to an increasing number of patients with prostate cancer and hormone-refractory disease that has metastasized to bone, following the publication of data showing a reduction of skeletal-related events in this group . The present data provide further in vitro evidence to account for this effect, and have further unravelled the underlying molecular mechanisms. We show that inhibiting geranylgeranylation of G proteins appears to be pivotal to the observed effects of zoledronic acid, and present data on the effects of directly inhibiting this process with a GGTI in prostate cancer cells. This class of drugs is also being developed for use in a clinical setting, and our results suggest a potential role in prostate cancer treatment.
The findings of significant apoptosis and inhibition of cellular adhesion to mineralized matrix with low doses of zoledronic acid also contribute to the rationale for new randomized controlled trials with this drug in high-risk patients who have not yet developed metastases, and which are now underway.