The in Vitro Evaluation of 25-Hydroxyvitamin D₃ and 19-nor-1α,25-Dihydroxyvitamin D₂ as Therapeutic Agents for Prostate Cancer¹ Free

Prostate cancer is the second leading cause of cancer deaths among U.S. men (after lung cancer), with 37,000 deaths projected in 1999 (1). For tumors that are ineligible for, or that fail to respond to, surgery or radiation, the mainstay of prostate cancer therapy is androgen deprivation. About 75% of men respond to androgen deprivation, but the median duration of response is only about 2 years (2). There are no effective therapies for prostate cancers that fail to respond to androgen deprivation.

In addition to androgens, it is now clear that prostatic cells are responsive to another class of steroid hormones, namely, vitamin D (3). Most human prostate cells contain specific intracellular receptors (commonly called VDRs)³ for 1α,25(OH)₂D, the active hormonal form of vitamin D (4, 5). Numerous studies have shown that in response to 1α,25(OH)₂D₃,prostate cancer cells show an increase in differentiation and a decrease in proliferation, invasiveness, and metastasis (6, 7, 8, 9). These findings strongly support the use of vitamin D-based therapies for prostate cancer, e.g., as differentiation therapy and/or as a second-line therapy once androgen deprivation has failed. However, the use of 1α,25(OH)₂D-based therapies for prostate cancer is limited by the risk of hypercalcemia and hypercalciuria (10, 11). Thus, less calcemic or noncalcemic analogues of 1α,25(OH)₂D₃ with potent antiproliferative activity would be attractive therapeutic agents.

Recently, our group has shown that human prostate cancer cells in primary culture and several prostate cancer cell lines possess 1α-OHase, the enzyme that converts the major circulating, prohormonal form of vitamin D, 25(OH)D, to 1α,25(OH)₂D (12, 13). Because the conversion from prohormone to active hormone occurs within the cell,the problem of systemic hypercalcemia should be avoided. 25(OH)D would be an attractive candidate for human clinical trials in prostate cancer because this drug has been approved by the FDA for human use(e.g., for treating vitamin D deficiency due to liver disease; Ref. 14).

Similarly,19-nor-1α,25(OH)₂D₂, a synthetic analogue of 1α,25(OH)₂D₂, has recently been approved by the FDA for the treatment of secondary hyperparathyroidism. Several randomized controlled clinical trials have shown that 19-nor-1α,25(OH)₂D₂ is noncalcemic (15, 16). The structural similarity of 19-nor-1α,25(OH)₂D₂ to 1α,25(OH)₂D₃ suggested to us that the behavior of 19-nor-1α,25(OH)₂D₂ in prostatic cells might be similar to that of 1α,25(OH)₂D₃.

In this report, we investigated the effect of 25(OH)D₃,1α,25(OH)₂D₃, and 19-nor-1α,25(OH)₂D₂ on the proliferation of primary cultures and cell lines of human prostate cancer. In addition, we evaluated the abilities of these three vitamin D compounds to transactivate the VDR in a prostate cancer cell line,PC-3, that was stably transfected with VDR.

25(OH)D₃ and 1α,25(OH)₂D₃ were a generous gift from Dr. M. Uskokovic (Hoffman-La Roche, Nutley, NJ). 19-nor-1α,25(OH)₂D₂ was a gift from Tetrionics (Madison, WI).

Prostate cancer cell lines, LNCaP and PC-3 cells, were obtained from the American Type Culture Collection (Rockville, MD) and were grown on 24-well culture dishes with DMEM (Life Technologies, Inc.) supplemented with 5% FBS (Life Technologies, Inc.). Cells were fed three times per week. Primary cultures of human prostate epithelial cells were prepared as described previously (17). Prostate epithelial cells were cultured in a serum-free defined-growth medium (Prostate Epithelial Growth Medium BulletKit, Clonetics, San Diego, CA). Prostate cells used for this study were at their second passage.

[³H]thymidine incorporation into DNA was used as an index of cell proliferation as described previously (18). Briefly, when LNCaP cells or the second passage primary culture cells reached about 50% confluency, FBS (in the case of LNCaP) or growth factors (in the case of primary cultures) were removed from the media, and cells were grown for an additional 24 h in the absence of FBS or growth factors. Cells were then treated with and without different concentrations of 25(OH)D₃,1α,25(OH)₂D₃, or 19-nor-1α,25(OH)₂D₂ as indicated in the figure legends. Eighteen h later, the dosing medium was replaced with 0.5 ml of fresh basal medium containing[methyl-³H]thymidine (New England Nuclear,Boston, MA) and incubated for 3 h at 37°C.[³H]Thymidine incorporation into DNA was stopped by placing the 24-well plates on ice. Unincorporated[³H]thymidine was then removed, and the cells were washed three times with ice-cold PBS. DNA labeled with[³H]thymidine and other macromolecules were first precipitated with ice-cold 5% perchloric acid for 20 min and then extracted with 0.5 ml of 5% perchloric acid at 70°C for 20 min as described previously (19). The radioactivity in the extracts was determined by a liquid scintillation counter. The results were expressed as percent of control.

The second-passage primary culture cells were subcultured in the complete medium into 35-mm dishes for the morphological studies (18). Two days after the initial plating, triplicate plates of cells were incubated with complete media without insulin but containing 25(OH)D₃,1α,25(OH)₂D₃, or vehicle. Cells were dosed again with 25(OH)D or 1α,25(OH)₂D₃ 2 and 4 days later. Three days after the last dosing, the media were removed from cultures. The attached cells were then trypsinized for 30 min with 0.1% EDTA and 0.1% trypsin at 37°C and then neutralized with basal medium. The detached cells were spun down and resuspended in a known volume of basal medium. Triplicate aliquots were applied to a hemocytometer for cell counting.

The reporter plasmid MOPVDREtkCAT was constructed as described previously (19) and consisted of two copies of the VDRE of the MOP gene linked 5′ to the tk promoter and CAT gene of the vector pBLCAT2 (20). MOPVDREtkCAT was transfected into a PC-3/VDR clone (clone 3B2) using the calcium phosphate method. CMV-β-gal, which encodes theβ-gal gene driven by the CMV promoter, was included in all of the transfections to normalize for differences in transfection efficiency. Clone 3B2 (PC-3/VDR) was generated by transfecting PC-3 cells with the VDR cDNA expression vector pRc CMV-VDR followed by the selection and expansion of stable clonal isolates as described previously (21). The PC-3/VDR and LNCaP cell line expressed comparable levels of VDR (approximately 25 fmol/mg protein). The PC-3/VDR cells transfected with the reporter gene were then cultured in the presence or absence of 10⁻⁸m 25(OH)D₃,1α,25(OH)₂D₃, or 19-nor-1α,25(OH)₂D₂ in RPMI containing 10% FBS. Cells were harvested about 40 h after transfection, and cell extracts were prepared for analyzing β-gal and CAT activity. Cell extracts containing equivalent amounts of β-gal activities were used for an analysis of CAT using an adaptation of the method of Gorman et al. (22) The percentage of conversion of [¹⁴C]chloramphenicol to acetylated forms on thin-layer chromatograms was quantified using a Molecular Dynamics Phosphorimager and image Quant software (Sunnyvale,CA).

Comparisons of the antiproliferative and transactivation activities between controls and drug-treated groups, and between two different drugs, were performed using one-way ANOVA. Differences between groups were considered statistically significant when Ps were ≤ 0.05.

The effects of 1α,25(OH)₂D₃ and 19-nor-1α,25(OH)₂D₂ on the [³H]thymidine incorporation into cultured LNCaP prostate cells are shown in Fig. 1. The figure demonstrates that 1α,25(OH)₂D₃ and 19-nor-1α,25(OH)₂D₂ at 10⁻⁷ m inhibited[³H]thymidine incorporation into DNA 21 ±4% and 18 ± 5%, respectively, as compared with the controls in the absence of 1α,25(OH)₂D₃ or 19-nor-1α,25(OH)₂D₂. At 10⁻⁶ m,1α,25(OH)₂D₃ and 19-nor-1α,25(OH)₂D₂inhibited [³H]thymidine incorporation 63 ± 1% and 60 ± 1%, respectively. No significant inhibition was detected in the presence of 10⁻⁸ mof 1α,25(OH)₂D₃ or 19-nor-1α,25(OH)₂D₂. Thus, in LNCaP cells,19-nor-1α,25(OH)₂D₂ was as active as 1α,25(OH)₂D₃in inhibiting [³H]thymidine incorporation at 10⁻⁷ and 10⁻⁶m.

Fig. 2 shows a dose-dependent antiproliferative effect of 1α,25(OH)₂D₃ and 19-nor-1α,25(OH)₂D₂ in the primary cultures of human prostate cancer cells. 1α,25(OH)₂D₃ caused an 18 ± 3%, 41 ± 1%, 77 ± 1%, and 86 ± 1%inhibition of [³H]thymidine incorporation into DNA at 10⁻⁹, 10⁻⁸,10⁻⁷, and 10⁻⁶m, respectively; and 19-nor-1α,25(OH)₂D₂caused a 20 ± 5%, 39 ± 2%, 80 ± 1%, and 88 ±1% inhibition of [³H]thymidine incorporation into DNA at the same concentrations, respectively. The data show that the activities of 1α,25(OH)₂D₃ and 19-nor-1α,25(OH)₂D₂ in the primary prostate cultures are very similar. Both compounds were more effective in inhibiting [³H]thymidine incorporation in the primary cultures than in LNCaP cells.

We have previously established the presence of 1α-OHase mRNA and its enzyme activity in cultured prostate cancer cell lines and in primary cultures derived from normal, benign prostatic hyperplasia and prostate cancer cells (PC-3 and DU145; 12, 13). Unlike the primary cultures, DU145 and PC-3 cells, very little 1α-OHase activity was detected in LNCaP cells. We reasoned that 25(OH)D₃ should be converted to 1α,25(OH)₂D₃ by prostatic cancer cells that possess 1α-OHase and that the 1α,25(OH)₂D₃ would cause an inhibition in the proliferation of these cells. Fig. 3 demonstrates that when primary cultures of prostatic cells were treated with 25(OH)D₃ or 1α,25(OH)₂D₃ for 7 days,the two vitamin D₃ metabolites caused a similar inhibition of cell proliferation. At 10⁻⁶m, 25(OH)D₃ and 1α,25(OH)₂D₃ inhibited cell proliferation by decreasing cell number 60 ± 3% and 60 ± 1%, respectively, and at 10⁻⁷ m,25(OH)D₃ and 1α,25(OH)₂D₃ inhibited cell proliferation by decreasing cell number 24 ± 1% and 25 ± 5% in primary cultures of prostate cancer cells, respectively. At 10⁻⁸ m, neither compound exhibited significant activity. Using the [³H]thymidine incorporation assay for cell proliferation, we also demonstrated that 25(OH)D₃ was highly active in inhibiting primary prostate cell proliferation (Fig. 4). 25(OH)D₃ caused 33 ± 7%, 46 ± 4%,and 70 ± 3% inhibition at 10⁻⁹,10⁻⁸, and 10⁻⁷m, respectively, as compared with 55 ± 4%, 60 ± 4%, and 73 ± 4% inhibition by 1α,25(OH)₂D₃. Thus, in these assays, 25(OH)D₃ and 1α,25(OH)₂D₃ were equipotent in inhibiting the proliferation of prostate cancer cells at 10⁻⁷ m.

We next compared the capacity of 25(OH)D₃,19-nor-1α,25(OH)₂D₂, and 1α,25(OH)₂D₃ to transactivate VDR in PC-3 cells stably transfected with VDR. PC-3 cells are well-characterized models of androgen-independent prostate cancer and were previously shown to contain 1α-OHase activity (12). However, unlike many human prostate cancer cells,PC-3 cells express extremely low levels of VDR. Thus, we chose a derivative of these cells that have been stably transfected with a VDR cDNA (PC-3/VDR). To assess the ability of VDR to be activated by endogenously synthesized 1α,25(OH)₂D₃, we performed CAT reporter gene transactivation assays in PC-3/VDR cells cultured in 25(OH)D₃. DPPD was added to inhibit the auto-oxidation of 25(OH)D₃to 1α,25(OH)₂D₃. The CAT reporter gene plasmid contains two tandem copies of the VDRE found in the MOP promoter. 1α,25(OH)₂D₃ caused a 2-,16-, 37-, and 15-fold increase, and 19-nor-1α,25(OH)₂D₂caused a 3-, 21-, 40-, and 32-fold increase over the controls in CAT activity at 10⁻¹¹, 10⁻¹⁰,10⁻⁹, and 10⁻⁸m, respectively (Fig. 5). Incubation of PC-3/VDR cells with 5 × 10⁻⁸ m25(OH)D₃ induced a 29-fold increase over the controls in CAT activity, which was similar to that induced by 10⁻⁸ m1α,25(OH)₂D₃ (Fig. 6).

During the past two decades, the actions of 1α,25(OH)₂D₃ have extended far beyond its classical role on intestine, bone, kidney, and parathyroid glands to regulate serum calcium levels. 1α,25(OH)₂D₃ has been shown to have important antiproliferative and prodifferentiating activities in a variety of tissues or cells that possess VDRs,including prostatic cells (4, 5, 6, 7, 8, 9). The epidemiological similarities between prostate cancer and vitamin D deficiency (23, 24, 25) and the impressive anticancer effects of 1α,25(OH)₂D₃ on prostatic cells has led to great interest in the use of this hormone as a therapeutic agent for prostate cancer. However, a major drawback of using vitamin D-based therapies for prostate cancer is hypercalcemia (10, 11). Many different types of non- or less-calcemic vitamin D analogues have been investigated for their effects on prostate cancer cell proliferation in vitro. Potent inhibitors of prostate cancer cell proliferation include 19-nor-hexafluoride vitamin D₃ and 20-cyclopropyl-vitamin D₃ analogues (26, 27). Hisatake et al. (28) recently reported that 5,6-trans-16-ene-Vitamin D₃ analogues were more potent than 1α,25(OH)₂D₃ in inhibiting LNCaP cells in vitro and were about 40-fold less calcemic than 1α,25(OH)₂D₃ in normal mice in vivo. However, the antiproliferative effects of these analogues has yet to be demonstrated in vivo.

Recently, it was demonstrated that EB1089, an analogue of 1α,25(OH)₂D₃, was as effective as 1α,25(OH)₂D₃in inhibiting metastasis in an in vivo model of androgen-insensitive prostate cancer, the rat Dunning MAT LyLu prostate cancer model (9). Although EB1089 was significantly less calcemic than 1α,25(OH)₂D₃, it still caused an 18% increase in serum calcium level (versus a 34% increase by 1α,25(OH)₂D₃). Thus, less calcemic or noncalcemic analogues of 1α,25(OH)₂D₃ are still needed. Llach et al. (15) reported that 19-nor-1α,25(OH)₂D₂ was as effective as 1α,25(OH)₂D₃ in suppressing parathyroid hormone secretion in hemodialysis patients with secondary hyperparathyroidism without inducing hypercalcemia or hyperphosphatemia. In the current study, we examined the antiproliferative activity of 19-nor-1α,25(OH)₂D₂ in LNCaP cells and in primary cultures of prostate cancer cells. In these two different cultures,19-nor-1α,25(OH)₂D₂showed antiproliferative effects similar to those of 1α,25(OH)₂D_3, as determined by [³H]thymidine incorporation(Figs. 1 and 2). Both compounds had a greater effect in the primary cultures of prostate cancer cells than in the LNCaP prostate cancer cell line, which suggests that primary cultures may be a more sensitive system to differentiate the effectiveness of different vitamin D compounds in vitro. 1α,25(OH)₂D₃ was previously shown (21) to decrease cyclin-dependent kinase 2 activity, resulting in decreased retinoblastoma protein phosphorylation and accumulation of LNCaP cells in G₁ phase of the cell cycle. Because a functional retinoblastoma pathway seems to be required for the maximal antiproliferative effects of 1α,25(OH)₂D₃, primary prostatic cultures may exhibit increased growth inhibition by 1α,25(OH)₂D₃,25(OH)D₃, and 19-nor-1α,25(OH)₂D₂because, compared with the cell lines, these cultures are less likely to have mutations in this pathway.

Because the clinical use of 1α,25(OH)₂D₃ in cancer therapy is limited by the risk of hypercalcemia, many investigators have attempted to duplicate the antiproliferative effects of 1,25(OH)₂D₃ in vivousing analogues of 1α,25(OH)₂D₃ that are less calcemic, such as 19-nor-1α,25(OH)₂D₂,16-ene-23-yne-1α,25(OH)₂D₃(29), and EB1089 (30), or they have used a combination of 1α,25(OH)₂D₃ or 1α,25(OH)₂D₃ analogues with other drugs (31, 32). However, our discovery that prostate cells in primary culture express high 1α-OHase activity and can synthesize 1α,25(OH)₂D₃ from 25(OH)D₃ suggests that 25(OH)D₃ may offer another potential solution to the problem of hypercalcemia caused by the systemic administration of 1α,25(OH)₂D_3. This is because 1α,25(OH)₂D₃would be synthesized intracellularly, act in an autocrine fashion, and be degraded, and would not be expected to leak into the systemic circulation and cause hypercalcemia. Our results confirm that 25(OH)D₃ is highly active in inhibiting prostate cells proliferation in vitro (Figs. 3 and 4). Similar antiproliferative effects are observed when 25(OH)D₃ is administered to primary cultures of prostate cells in clonogenic assays.⁴ Because 25(OH)D₃ binds to the VDRs with only 0.001 to 0.002 of the binding affinity of 1α,25(OH)₂D₃(33), we consider it unlikely that these results are due to the direct actions of 25(OH)D₃ on the VDRs. Rather, we suggest that these results reflect the conversion of 25(OH)D₃ to 1α,25(OH)₂D₃ by 1α-OHase present in prostate cells. This conclusion is further supported by experiments in which LNCaP cells were transfected with 1α-OHase cDNA. These cells, and not untransfected cells, responded to 25(OH)D₃ by an inhibition of[³H]thymidine incorporation.⁵ Experiments to determine whether similar effects can be produced in vivousing human tumor cells xenografted into athymic mice are presently underway.

Although kidney cells are the “classic” cells that possesses 1-OHase, 1α,25(OH)₂D levels produced by the kidney are very tightly regulated by serum levels of parathyroid hormone (34). Thus, in normal individuals, even large increases in serum 25(OH)D will not result in increased systemic levels of 1α,25(OH)₂D (35, 36). However,the extrarenal synthesis of 1α,25(OH)₂D is generally unregulated (13, 37). This suggests that increases in systemic levels of 25(OH)D could result in increased local production of 1α,25(OH)₂D in some extrarenal sites (i.e., prostate) without producing hypercalcemia.

Two commonly used assay methods were used to study the antiproliferative activity of 25(OH)D₃ in the primary cultures of prostate cells:[³H]thymidine incorporation and cell count. The former involved a short-term, 18-h incubation with the hormone and studied the effect of the drugs on DNA synthesis.[³H]thymidine incorporation is an index of cell division. In certain cell types, decreases in[³H]thymidine incorporation may also reflect increases in cell differentiation (38). Conversely, cell count, which requires longer-term incubation (7 days) with the hormone,reflects cell growth only. The difference in dose-response curves that we observed in the two assays may reflect differences in the incubation time of these assays because: (a) with longer incubation times, more 25(OH)D₃ would be converted to 1α,25(OH)₂D₃intracellularly, and, thus, the effects of 25(OH)D₃ would be similar to the effects of the direct addition of 1α,25(OH)₂D₃; and(b) more exogenously added 1α,25(OH)₂D₃ would likely increase its own degradation by inducing the expression of 24-hydroxylase (34).

Most of the antiproliferative effects of 1α,25(OH)₂D₃ and its analogues are believed to be mediated through the functional expression of VDR. We, therefore, compared the transactivation activity of 1α,25(OH)₂D₃ and 19-nor-1,25(OH)₂D₂ in PC-3 cells that were stably transfected with VDR (PC-3/VDR), using the reporter plasmid MOPVDREtkCAT. This plasmid, which contains two copies of the VDRE found in the MOP gene, was chosen because previous studies in other well-characterized prostate cancer cell lines indicated that VDR transcriptional activity could be detected using this VDRE-containing reporter, even in cell lines that expressed extremely low levels of VDR (9, 21). Using this system, we demonstrated that 1α,25(OH)₂D₃ and 19-nor-1α,25(OH)₂D₂ had almost identical transactivation activity (Fig. 5) in agreement with our [³H]thymidine incorporation data (Figs. 1 and 2). 25(OH)D₃ at 5 ×10⁻⁸ m showed a comparable transactivation activity as that caused by 10⁻⁸m1α,25(OH)₂D₃, consistent with our finding that PC-3 cells have the capacity to convert 25(OH)D₃ to 1α,25(OH)₂D₃(12).

In summary, this report demonstrates that, like 1α,25(OH)₂D₃,19-nor-1α,25(OH)₂D₂, and 25(OH)D₃ possess potent antiproliferative effects on human prostate cancer cell lines and on primary cultures of human prostate cancer cells. Both 19-nor-1α,25(OH)₂D₂ and 25(OH)D₃ are equipotent to the parent hormone in their ability to transactivate the VDR. Although both of these vitamin D compounds act ultimately on the VDR, their proximal biological targets are different: 25(OH)D₃ requires the presence of 1α-OHase, whereas 19-nor-1α,25(OH)₂D₂ does not. Our studies of prostate cancer cell lines have shown a large variation in the expression of 1α-OHase. For example, although LNCaP cells showed profound growth inhibition by 1α,25(OH)₂D₃, these cells do not express measurable levels of 1α-HOase message and activity (12, 13) and, accordingly, are not growth-inhibited by 25(OH)D₃ (39). This suggests that prostatic tumors that do not express 1α-OHase should be treated with 1α,25(OH)₂D₃analogues, such as 19-nor-1α,25(OH)₂D₂ or EB1089. Because both 25(OH)D₃ and 19-nor-1α,25(OH)₂D₂ are known to be noncalcemic within a wide dosing range (14)and both are approved for human use (for other indications), these vitamin D compounds may be excellent candidates for human clinical trials in prostate cancer, especially for prostate cancers that have failed conventional therapies such as androgen deprivation.

We thank Kelly Persons and Carol Maorino for technical assistance, Drs. Jennifer M. Grad and David Jackson for the graphics,and Dr. Hector F. DeLuca (University of Wisconsin, Madison, WI) for the 19-nor-1α,25(OH)₂D₂.