Increased expression of CYP24A1 correlates with advanced stages of prostate cancer and can cause resistance to vitamin D₃-based therapies

Abstract

A major limitation of exogenous vitamin D₃ administration for the treatment of prostate cancer is the marginal, if any, clinical efficacy. We dissected the basis for the resistance to the vitamin D₃ antitumor properties and specifically examined the effect of its major catabolic enzyme, CYP24A1, in prostate cancer. Local CYP24A1 expression levels and the effect of selective modulation were analyzed using tissue microarrays from needle core biopsy specimens and xenograft-bearing mouse models. CYP24A1 mRNA was elevated in malignant human prostate tissues compared to benign lesions. High CYP24A1 protein levels were seen in poorly differentiated and highly advanced stages of prostate cancer and correlated with parallel increase in the tumor proliferation rate. The use of CYP24A1 RNAi enhanced the cytostatic effects of vitamin D₃ in human prostate cancer cells. Remarkably, subcutaneous and orthotopic xenografts of prostate cancer cells harboring CYP24A1 shRNA resulted in a drastic reduction in tumor volume when mice were subjected to vitamin D₃ supplementation. CYP24A1 may be a predictive marker of vitamin D₃ clinical efficacy in patients with advanced prostate cancer. For those with up-regulated CYP24A1, combination therapy with RNAi targeting CYP24A1 could be considered to improve clinical responsiveness to vitamin D₃.—Tannour-Louet, M., Lewis, S. K., Louet, J.-F., Stewart, J., Addai, J. B., Sahin, A., Vangapandu, H. V., Lewis, A. L., Dittmar, K., Pautler, R. G., Zhang, L., Smith, R. G., Lamb, D. J. Increased expression of CYP24A1 correlates with advanced stages of prostate cancer and can cause resistance to vitamin D₃-based therapies.

Prostate cancer is the most commonly diagnosed malignancy in men. Treatments include radical prostatectomy followed by radiation or hormone ablation therapy. Unfortunately, durable clinical responses are not always obtained, and, as prostatic carcinomas progress, systemic metastasis to distant areas may lead to patient morbidity and mortality. Therefore, development of effective therapeutic strategies for the management of castrate-resistant prostate cancer is needed. One promising alternative is the use of the hormonal form of vitamin D₃, 1,25-dihydroxyvitamin D₃ [1,25(OH)₂D₃] or calcitriol, which exerts prodifferentiative, antiproliferative, antiangiogenic, and proapoptotic effects on prostate cancer cells in vitro and in vivo (for review, see ref. 1). Phase I and II trials of calcitriol, either alone or in combination with other therapeutic agents, were conducted in patients with castrate-resistant prostate cancer (2–8). However, the clinical use of vitamin D₃-based therapies was hampered by limited antitumor benefit. Therefore, to improve the clinical efficacy of exogenous vitamin D₃ in the treatment of prostate cancer, it is useful to define predictive factors for the growth inhibitory response of vitamin D₃ in patients.

In studying vitamin D₃ action, it is critical to examine the pathways that control local tissue levels of 1,25(OH)₂D₃. The cytochrome P450 enzyme, 24-hydroxylase, encoded by CYP24A1 is critical for the catabolism of 1,25(OH)₂D₃. While its constitutive expression level is low, CYP24A1 is strongly induced by 1,25(OH)₂D₃ administration to produce less active vitamin D metabolites. Overexpression of CYP24A1 occurs in several human solid tumors, including colon (9–11), cervical (12), ovarian (9, 12), esophageal (13), lung (14), and basal cell carcinomas (15). Increased expression of this rate-limiting enzyme is likely to affect the bioavailability of 1,25(OH)₂D₃ at the tumor site and abrogate the antiproliferative action of vitamin D₃. In this present study, we specifically stressed this hypothesis in the pathological environment of human prostate cancer. The goal was to assess local CYP24A1 expression levels, to determine the association with tumor cell proliferation and the effect of their modulation on prostate tumor progression. Such predictions may ultimately allow for individualized treatments and higher clinical efficacy of adjuvant vitamin D₃ therapy in patients with advanced prostate cancer.

MATERIALS AND METHODS

Cell lines and chemicals

Human prostate cancer cell lines (LNCaP, PC3, and DU145) were obtained from the American Type Culture Collection (ATCC; Rockville, MD, USA) and grown in RPMI, DMEM, or MEM containing 10% fetal bovine serum, respectively. All cells were maintained in a humidified 37°C incubator with 5% CO₂. Calcitriol (Sigma-Aldrich, St. Louis, MO, USA) was reconstituted in 100% ethanol, stored at −20°C, and protected from light.

Transfection with small interfering RNA (siRNA)

Cells were plated in 6-well plates (1×10⁵/well) for RNA extraction or in 96-well plates (1–1.5×10³/well) for cell proliferation assays. The next day, cells were transfected for 24 h with 40 nM of nonspecific (scramble) siRNA or siRNA for human CYP24A1 (Dharmacon, Lafayette, CO, USA) using the Lipofectamine 2000 reagent (Invitrogen, Calrsbad, CA, USA) according to the manufacturer's procedure. Following transfection, the cells were treated with either vehicle (0.1% v/v ethanol) or calcitriol (10–100 nM), for 24 h (RNA extraction) or 6 d (for cell proliferation).

Cell proliferation assays

Briefly, 1–1.5 × 10³ cells/well were seeded into 96-well plates and treated with or without CYP24A1 siRNA for 24 h, then cultured in the presence or absence of 1,25(OH)₂D₃ at concentrations ranging from 1 to 100 nM. Viable cell number was determined after 3, 6, or 9 d of treatment using the WST1 Cell proliferation (Roche Applied Science, Indianapolis, IN, USA) according to the manufacturer's instructions. The plates were read using a microplate spectrophotometer at a wavelength of 450 nm and corrected to 630 nm. Each independent experiment was performed ≥3 times.

Creation of stable cell lines harboring CYP24A1 shRNA

To generate stable prostate cancer cells expressing human CYP24A1 or nontargeting shRNA, cells were incubated with lentiviral transduction particles producing a nontarget control or human CYP24A1 shRNA (Sigma-Aldrich), in presence of polybrene, according to the manufacturer's protocol. Stable clones were selected with 1 μg/ml puromycin for 3 wk.

Quantitative reverse transcription polymerase chain reaction (qRT-PCR) analysis of CYP24A1 gene expression

Total RNA was purified using the RNeasy Kit including an optional DNase I treatment according to the manufacturer's instructions (Qiagen, Valencia, CA, USA). cDNA was prepared from 500 ng of total RNA by reverse transcription. Patient cDNAs were prepared from well-documented cancer biopsy samples, normalized and assembled into ready-to-use gene expression panels (Origene, Rockville, MD, USA). The supplier provided clinicopathological information for each patient. Real-time PCR was performed with TaqMan PCR Master Mix on an ABI StepOnePlus Realtime PCR System (Applied Biosystems, Foster City, CA, USA). PCR conditions were 50°C for 2 min, 94°C for 2 min, followed by 40 cycles of 94°C for 15 s and 60°C for 30 s. For each experimental sample, the relative abundance value was normalized to the value derived from the endogenous control (β-actin or GAPDH) of the same sample. Relative mRNA levels were quantified by the comparative ΔΔC_T method.

CYP24A1 immunofluorescence

Nonconfluent prostate cells were fixed in 4% paraformaldehyde for 5 min on ice. Labeling with CYP24A1 antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) was carried out in the presence of 0.1% Triton X-100, followed by Alexa Fluor 488 goat anti-rabbit IgG secondary antibody. Cells were counterstained with 4′,6-diamidino-2-phenylindole (DAPI) for cell nuclei visualization.

Tissue microarrays and immunohistochemical detection

Tissue microarrays containing 112 prostate cancer cores, derived from 56 patients with localized and metastatic prostate cancer disease and Gleason scores ranging from 4 to 10 (Folio, Columbus, OH, USA) were used in this study. Slides of paraffin-embedded adenocarcinoma samples were deparaffinized and dehydrated. Antigen retrieval was performed by heat inactivation in 0.1 M sodium citrate for 30 min. CYP24A1 (Santa Cruz Biotechnology) or Ki67 (Dako, Carpinteria, CA, USA) primary antibodies were used for immunodetection. The staining was performed using the avidin-biotin peroxidase system (ABC-peroxidase), and positive signals were visualized as brown precipitates using 3,3′-diaminobenzidine tetrahydrochloride. Control staining was conducted by omission of the primary antibody. Hematoxylin was used for counterstaining. Two investigators, blinded to the data, performed light microscopy (Axioplan 2; Carl Zeiss, Thornwood, NY, USA) and scored semiquantitatively the quantity of protein expression in the whole section (0=none, 1= moderate, 2=strong). Scoring for semiquantitative analysis for Ki67 staining was evaluated by the percentage of cells showing positively labeled nuclei: 0 = 0–1% positivity, 1 = 1–5% positivity, 2 = 5–10% positivity, 3 = 10–100% positivity.

Animals and diets

All animal procedures were approved by the Institutional Animal Care and Use Committee of Baylor College of Medicine and conducted in compliance with the U.S. National Institutes of Health Guide for the Care and Use of Laboratory Animals in an American Association for Accreditation of Laboratory Animal Care (AAALAC)-accredited facility. Male Fox Chase SCID Beige mice (CB17.B6-Prkdc^scidLyst^bg/Crl; Charles River, Wilmington, MA, USA) were used for orthotopic (intraprostatic) xenograft experiments. Homozygous (nu/nu) nude mice (Crl:NU-Foxn1^nu; Charles River) were used for subcutaneous (s.c.) xenograft experiments. All animals were housed in sterile cages including bedding, water, and feed (irradiated). Mice were maintained on a vitamin D₃-deficient diet containing 0.5% calcium (Modified AIN-93M; TestDiet, Richmond, IN, USA) for 4 wk prior to orthotopic or s.c. injections, and until the animals were euthanized. DU145 prostate cancer cells stably transduced with shRNA control (n=20 orthotopic; n=30 s.c.) or CYP24A1 shRNA (n=20 orthotopic; n=30 s.c.) were injected. Mice were treated with vehicle (sesame oil with 2% ethanol) or calcitriol (0.5 μg/kg mouse in sesame oil with 2% ethanol) every other day by oral gavage with a 20-gauge intragastric feeding tube (Kent Scientific, Torrington, CT, USA). Treatments were administered for 90 d (orthotopic injections) or 60 d (s.c. injections). All mice were euthanized by cervical dislocation prior to tissue collection.

Orthotopic xenograft injections

SCID mice were anesthetized using Rodent III CCM combination anesthetic DEA-III (per milliliter contains 37.5 mg ketamine, 1.9 mg xylazine, and 0.37 mg acepromazine and given intramuscular at 0.02 ml/30 g body weight). One site on the dorsal prostate was inoculated with 1 × 10⁶ cells in 20 μl delivered using a Hamilton microsyringe (Hamilton, Reno, NV, USA), with a 30-gauge needle [cells suspended in Hanks balanced salt solution (HBSS)/Matrigel at a ratio of 5:1 (Matrigel:cells)]. Animals were observed daily for signs of distress or illness and weekly for weight loss.

Magnetic resonance imaging (MRI) of orthotopic xenograft mouse models

To monitor tumor growth, MRI was performed using a Bruker Biospin Pharmascan 7.0 T spectrometer (Bruker Biospin, Billerica, MA, USA) in the Mouse Phenotyping Core at Baylor College of Medicine. MRI of SCID mice was done prior to orthotopic injection of cells, and then at d 15, d 30, d 60, or d 90 after xenograft injection. T₂-weighted sequences were acquired in the axial and coronal planes using a repetition time (T_R) = 3000 ms and echo time (T_E) = 36.61 ms; slice thickness of 0.5 mm (no interslice gaps); field of view (FOV) of 40 × 40 mm; matrix = 256 × 256 mm. Each sequence acquisition total time was ∼4 min 48 s using Paravision software (Bruker BioSpin). Data were analyzed using Amira 3D image processing software (Mercury Computer Systems, Chelmsford, MA, USA) to calculate the tumor volume corresponding to the pixels in each slice. Tumor volume was defined as the product of the number of pixels in a defined tumor region relative to the FOV and matrix sizes using the equation: V_pixel = 0.5 × (40/256) × (40/256) mm.

Subcutaneous xenograft injections

Homozygous nu/nu nude recipient mice were inoculated with 1 × 10⁶ cells in 50 μl delivered to the medial dorsal side using a syringe with a 30-gauge needle [cells suspended in HBSS/Matrigel at a ratio of 5:1 (Matrigel:cells)]. Treatments with calcitriol were initiated when tumor growth reached 300 mm³. Three measurements of the tumors were taken (using digital calipers) every week and averaged to get the weekly mean tumor volume for each animal. Tumor volume was calculated using the modified ellipsoid formula (1/2 length×width²; refs. 16, 17).

Statistical analysis

The data were obtained from ≥2 different experiments and are presented as means ± sem. All statistical analyses were performed using the 1-way or 2-way analysis of variance (ANOVA). Contingency tables with χ² analysis were done using GraphPad (La Jolla, CA, USA). Values were considered statistically significant if P < 0.05.

RESULTS

CYP24A1 is overexpressed in patients with prostate cancer

To define the mRNA levels of endogenous CYP24A1 in adenocarcinomas of the human prostate, cDNAs obtained from pathologist-verified human prostate biopsies representing different pathological stages of prostate cancer, as well as benign prostatic hyperplasia, were analyzed by quantitative PCR. CYP24A1 mRNA was overexpressed in prostate cancer samples (n=29) compared with benign tissues (n=7) (4-fold; P=0.03, Wilcoxon rank-sum test; Fig. 1A). Consistently, malignant tissues displayed higher levels of CYP24A1 protein than the paired adjacent normal tissues (n=8; Fig. 1B).

Figure 1.

Overexpression of CYP24A1 in poorly differentiated and highly advanced stages of human prostate cancer. A) Determination of CYP24A1 gene expression levels in men with prostate cancer. TaqMan qPCR was performed on patient cDNAs prepared from well-documented cancer biopsy samples, normalized and assembled into ready-to-use gene expression panels. Data were analyzed using the comparative C_T method with the values normalized to β-actin levels and presented as means ± se (P=0.03, Wilcoxon rank-sum test). B) Representative immunohistochemical detection of CYP24A1, revealing higher staining in malignant prostate tissue than in paired cancer-adjacent normal tissue. View ×400. C) Distribution of the relative expression levels of CYP24A1 and Gleason score and tumor stage categories. Semiquantitative scoring of the quantity of CYP24A1 expression in 112 prostate cancer cores, derived from 56 patients with localized and metastatic prostate cancer disease, was performed (score: 0=none; 1= moderate; 2=strong). Cases are shown as percentages of total immunoreactive cancer cores on the y axis. D) Immunostaining of CYP24A1 in representative adenocarcinomas of different grades. 1–4) Low-grade lesions of prostate cancer showing moderate intensity of the staining of the catabolic enzyme (1, Gleason 2+2 and T2NxM0; 2, Gleason 2+3 and T2NxM0; 3, Gleason 2+3 and T2NxM0; 4, Gleason 1+1 and T2bNxM0). a–d) intense staining in malignant prostate tissues presenting with high Gleason score and late stage tumor progression (a, Gleason 9 and T3bN0M0; b, Gleason 7 and T3bN0M0; c, Gleason 7 and T4N0M0; d, Gleason 9 and T3bN0M0). View ×400.

To gain insight into specific changes associated with cancer progression, tissue samples from 112 prostate cancer cores, derived from 56 patients with localized and metastatic prostate cancer with Gleason scores ranging from 4 to 10 were analyzed. The percentage of CYP24A1 immunoreactive cases was high (98% of cancer cores). Nevertheless, CYP24A1 staining intensity varied between patients, ranging from intense immunoreactivity of the entire cytoplasm to more discrete granular staining. Notably, high protein levels of CYP24A1 positively correlated with high-grade cancer (P<0.0001, Pearson's χ²; Fig. 1C) as well as with late stage tumor progression (P<0.0001, Pearson's χ²; Fig. 1C). Increased CYP24A1 staining intensity was, indeed, frequently observed in cancer cores with Gleason scores > 7 and in lesions presenting tumor, nodes, and metastasis (TNM) stages 3 or 4 (Fig. 1C, D).

Next, the relationship of increased CYP24A1 expression with increased tumor proliferation (reflected by the expression of the proliferation marker Ki67) was determined. High levels of expression of CYP24A1 positively correlated with increased staining of Ki67 in the same human prostate cancer tissues (P<0.0002, Pearson's χ²; Fig. 2A, B). This observation supports the hypothesis that increased CYP24A1 expression in prostate cancer may decrease the intratumor 1,25(OH)₂D₃ levels, effectively counteracting its antiproliferative effects.

Figure 2.

Distribution of the percentage of Ki67-positive cells between cancer cores expressing high and moderate levels of CYP24A1. A) Semiquantitative scoring of the CYP24A1 staining was independently performed by 2 investigators who were blinded to the data. Scoring for semiquantitative analysis for Ki67 staining was evaluated by the percentage of cells showing positively labeled nuclei: 0 = 0–1% positivity, 1 = 1–5% positivity, 2 = 5–10% positivity, 3 = 10–100% positivity). View ×400. B) High expression of CYP24A1 positively correlated with parallel increased staining of Ki67 in the same human prostate cancer tissues. View ×400.

Specific inhibition of CYP24A1 expression by siRNA enhances the antiproliferative effect of 1,25(OH)₂D₃ in human prostate cancer cells

Consistent with a previous report on the enzymatic activity of CYP24A1 in prostate cell lines (18), an inverse correlation between the constitutive cellular levels of CYP24A1 mRNA and the growth response to vitamin D₃ was observed in human prostate cancer cell lines (Supplemental Fig. S1). Next, CYP24A1 expression was targeted with siRNA in human vitamin D₃- resistant prostate cancer cells (PC3 and DU145) to examine whether the reduction of CYP24A1 mRNA levels by this approach would enhance the growth inhibition mediated by vitamin D₃. As measured by qRT-PCR, the relative levels of CYP24A1 mRNA were efficiently knocked down by siRNA in the absence or presence of vitamin D₃ (Fig. 3A). When compared to cells transfected with nontargeting siRNA, the reduction of 24-hydroxylase expression by CYP24A1 siRNA in DU145 and PC3 cells was about 90 and 80% in the presence of vehicle (ethanol) and 60 and 74% in the presence of 10 nM vitamin D₃, respectively (Fig. 3A). The knockdown of CYP24A1 mRNA translated into reduced levels of CYP24A1 protein, as reflected by the attenuation of the punctuated staining seen in cells transfected with CYP24A1 siRNA compared to control cells transfected with the scrambled control construct (Fig. 3B). Next, cell proliferation assays were performed to define the consequences of siRNA-mediated silencing of CYP24A1 gene expression. While PC3 and DU145 cells were unresponsive to vitamin D₃-mediated growth inhibition (Supplemental Fig. S1C, D), CYP24A1 siRNA significantly enhanced the antiproliferative action of vitamin D₃ (Fig. 3C). Knockdown of CYP24A1 gene expression by siRNA renders prostate cancer cells more sensitive to the growth-suppressive effect of vitamin D₃.

Figure 3.

Effect of transient inhibition of CYP24A1 gene expression on prostate cancer cell growth response to vitamin D_3. A) qRT-PCR analysis of CYP24A1 mRNA levels in presence of 10 nM Vitamin D₃ or ethanol (EtOH) vehicle in DU145 and PC3 transiently transfected with CYP24A1 siRNA or a nontargeting siRNA (scramble). GAPDH served as the internal control. B) Immunofluorescent staining of CYP24A1 protein of DU145 and PC3 transiently transfected with CYP24A1 siRNA or a nontargeting siRNA (scramble) and incubated in presence of 10 nM vitamin D₃. DAPI was used for visualization of cell nuclei. C) Cell proliferation assays performed in presence of 10 or 100 nM calcitriol or ethanol after 6 d of treatment in DU145 and PC3 transiently transfected with CYP24A1 siRNA or nontargeting siRNA (scramble).

To assess the effect of a sustained expression of CYP24A1 siRNA on vitamin D₃-mediated growth inhibition, DU145 and PC3 cell lines stably expressing CYP24A1 shRNA or nontargeting shRNA were generated. The random integration into the genome of the vector-based shRNA led to a significant knockdown of the CYP24A1 expression in the absence and presence of calcitriol (Fig. 4A and Supplemental Fig. S2A). When compared to cells with nontargeting shRNA, relative CYP24A1 mRNA levels were reduced by 75% in PC3 and 85% in DU145 cells in the presence of the vehicle, and the knockdown was ∼90% for PC3 and 60% for DU145 cells in the presence of vitamin D₃. Under these clonal conditions, PC3 cells showed improvement in their growth inhibitory response to vitamin D₃, reaching significance at 100 nM (Supplemental Fig. S2B). In DU145 cells, a significant reduction of cell proliferation was achieved at 10 nM of vitamin D₃ (Fig. 4B). Hence, prolonged inhibition of CYP24A1 gene expression enhanced the 1,25(OH)₂D₃-mediated growth inhibition in prostate cancer cells and essentially changed vitamin D₃-resistant cells into vitamin D₃-sensitive cell lines.

Figure 4.

Tumor growth abrogation in calcitriol-supplemented mice engrafted ectopically with prostate cancer cells stably expressing CYP24A1 shRNA. A) qRT-PCR analysis of CYP24A1 mRNA levels in presence of 10 nM vitamin D₃ or ethanol (EtOH) vehicle in DU145 stably transfected with CYP24A1 shRNA or a nontargeting shRNA (scramble). GAPDH served as the internal control. B) Cell proliferation assays performed in presence of 10 or 100 nM calcitriol or ethanol for 6 d in stable DU145 expressing CYP24A1 shRNA or nontargeting shRNA (scramble). Values represent means of ≥3 separate experiments. Data are presented as means ± sem. **P < 0.01, ***P < 0.001; ANOVA. C) Tumor volumes of nude mice injected with DU145 cells stably transfected with control shRNA or CYP24A1 shRNA and treated with calcitriol (vitamin D) or sesame oil (vehicle). Measurable tumors were not observed before the fourth week of injection. Treatment was initiated when tumor volume was ∼300 mm³. Tumor volume was measured 3×/wk for 9 wk. Data are presented as means ± se. **P < 0.01, ***P < 0.001, ****P < 0.0001; ANOVA. D) Immunohistochemical analysis of the proliferation marker Ki-67 in situ using sections of subcutaneously engrafted tumors after 9 wk of treatment with or without calcitriol. Positive staining for Ki-67 (human) indicated by nuclear brown precipitate. Sections were counterstained with hematoxylin. View ×100.

Enhancement of the antitumor action of calcitriol in ectopic and orthotopic xenografts of prostate cancer cells harboring CYP24A1 shRNA in mice

To examine in vivo the efficacy of selective targeting of CYP24A1 on the anticancer action of calcitriol, tumors were produced by s.c. injection of immunodeficient nude mice with DU145 cells stably expressing CYP24A1 shRNA (n=29) or a nontargeting control shRNA (n=27). Mice were subsequently given calcitriol or sesame oil (vehicle) by oral gavage when tumor volume reached ∼300 mm³. After 7 wk of treatment, tumors derived from DU145 cells stably expressing CYP24A1 shRNA and treated with calcitriol showed a drastic reduction in tumor volume compared to all experimental groups (Fig. 4C). At wk 9 of treatment, there was a significant difference of 148% in tumor volume between tumors derived from CYP24A1 shRNA-expressing cells treated with calcitriol and the control shRNA-expressing xenografts treated with calcitriol (Fig. 4C; P<0.0001). To determine whether the reduction in tumor growth was associated with an antiproliferative effect, immunohistochemical analysis for the proliferation marker Ki-67 was performed on excised engrafted tumor tissues of each experimental group. A striking decrease in the number of nuclei positively stained for the Ki-67 protein was observed in CYP24A1 shRNA xenografts receiving calcitriol compared to all other groups (Fig. 4D). Accordingly, selective targeting of CYP24A1 may provide an efficient alternative to significantly enhance the antiproliferative action of calcitriol in ectopic xenografts of prostate cancer.

To monitor the therapeutic response of the combinatorial treatment of calcitriol and CYP24A1 shRNA in a microenvironment closely resembling that of localized growth and distal dissemination of human prostate cancer, orthotopic xenografts were established by injecting DU145 cells stably expressing CYP24A1 shRNA or control shRNA into the dorsal prostate of SCID beige mice. At 2 wk after injection, treatment with calcitriol or sesame oil was initiated. Tumor-take efficiency was lower in orthotopic xenografts (control shRNA: 65%, n=20; CYP24A1 shRNA: 74%, n=19) than in subcutaneous xenografts (control shRNA: 93%, n=27; CYP24A1 shRNA: 83%, n=29) (Supplemental Table S1). Similar to subcutaneous engraftments, tumors established in CYP24A1 shRNA xenograft-bearing mice treated with calcitriol showed greater decrease in volume and proliferation at the end of treatment (d 90) than did the other experimental groups (Fig. 5 and Supplemental Figs. S3 and S4). Of note, CYP24A1 shRNA xenograft-bearing animals treated with calcitriol displayed fewer metastatic tumors (33%, n=9) than did the other experimental groups (Supplemental Table S1). However, this difference did not reach significance (P=0.349, Fisher's exact test, Supplemental Table S1). Collectively, orthotopic tumors responded efficiently to the combinatorial treatment of calcitriol and CYP24A1 shRNA by slowing their growth. In both ectopic and orthotopic xenograft experiments, calcitriol treatments were not associated with any significant toxicity, as evidenced by the absence of calcium phosphate deposits in the kidneys of all experimental groups (Supplemental Fig. S5).

Figure 5.

Tumor growth abrogation in calcitriol-supplemented mice engrafted orthotopically with prostate cancer cells stably expressing CYP24A1 shRNA. A) Representative MRI images of coronal sections through the lower abdomen of SCID mice after 60 and 90 d of treatment with calcitriol in mice injected intraprostatically with DU145 cells stably transfected with nontargeting (control shRNA) or CYP24A1 shRNA. Top panels: one of 20 coronal MRI slices at d 60 and 90 of treatment of the same animal. Bottom panels: enlarged view with 3D surface generation of the bladder (gray), ventral prostate (red), dorsal prostate (blue), and tumor (green). Green arrows indicate tumor; red arrows indicate ventral lobe of mouse prostate. B, bladder. B) Immunohistochemical analysis of the proliferation marker Ki-67 in orthotopically engrafted tumors following 90 d of treatment with or without calcitriol. Positive staining for Ki-67 indicated by nuclear brown precipitate. Arrows indicate prostate. T, tumor. Magnification ×100.

These data demonstrate that the selective inhibition of CYP24A1 by shRNA enhances the growth-inhibiting actions of calcitriol in xenograft mouse models, implying a potential use of CYP24A1 siRNA technology or a specific CYP24A1 small inhibitor in repressing the progression or perhaps preventing the development of prostate cancer.

DISCUSSION

This study provides evidence of CYP24A1 overexpression at the mRNA and protein levels in our cohort of patients with prostate cancer, particularly those with advanced disease. Previously, changes of the expression of this enzyme were associated with poor differentiation and prognosis for patients with other types of cancer, including colon (9–11), cervical (12), ovarian (9, 12), esophageal (13), lung (14), and basal cell carcinoma (15). The chromosomal locus 20q13.2-20q13.3 that encompasses CYP24A1 was amplified in breast (19), ovarian (20), gastric (21, 22), and colorectal cancers, suggesting that this protein can serve as an oncogene (23). Despite the existence of a slight degree of heterogeneity, we, notably, highlight the positive correlation between elevated CYP24A1 protein levels, poor differentiation, advanced stages, and increased proliferation rate of human prostate tumor lesions. Elevated expression of this rate-limiting catabolic enzyme is thought to restrict in situ 1,25(OH)₂D₃ growth inhibitory activity, presumably by decreasing local levels of 1,25(OH)₂D₃. Our data provide significant arguments in favor of the in situ role of CYP24A1 in the control of vitamin D₃ responsiveness in the prostate. Indeed, cultured human prostate cells that showed the highest constitutive levels of CYP24A1 mRNA were the least growth inhibited by calcitriol. Moreover, suppression of CYP24A1 gene expression led to an enhanced 1,25(OH)₂D₃ antiproliferative action in prostate cancer cells in vitro, as well as in xenograft-bearing mouse models. Therefore, variable levels of CYP24A1 in prostatic tumors may not lead to similar clinical responses in distinct patients subjected to exogenous vitamin D₃. Assessment of CYP24A1 prostate levels could serve to monitor treatment efficacy when using exogenous vitamin D₃ in patients with advanced prostate cancer. For individuals with up-regulated CYP24A1, a combination therapy with specific inhibitors of CYP24A1, directly administered to the tumor tissues, could be considered in order to help improve clinical responsiveness to calcitriol. Additionally, CYP24A1 may represent a tissue biomarker of prostate tumorigenesis. Indeed, consistent with the notion that a selective growth advantage is provided to the prostate tumor cells expressing high levels of CYP24A1, overexpression of this catabolizing enzyme was found associated with a more aggressive tumor behavior. Hence, in addition to increasing the therapeutic efficacy of vitamin D₃-based therapies, a preliminary histological examination of CYP24A1 expression levels in the patient's biopsies might be combined with existing tests to specifically stratify prostate cancer and predict the biological behavior of tumors.

Our findings indirectly point to the potential role for vitamin D₃ as a hormone acting to inhibit and, perhaps, delay or prevent the progression to cancer. Epidemiological studies remain controversial with regard to an inverse association between sun exposure, serum levels of 25(OH)-vitamin D, and risk of developing prostate cancer. Nevertheless, none of these studies directly assessed the intraprostatic 1,25(OH)₂D₃ levels. Our examination of the tissue expression of the main enzyme responsible for the breakdown of vitamin D₃ shed light on the in situ dynamic importance of calcitriol bioavailability, thus enriching a large body of preclinical research and evidence from clinical trials supporting the biological relevance of vitamin D in cell growth restriction and prostate cancer treatment (24).

This study offers a rationale for the use of RNA interference (RNAi) to selectively inhibit CYP24A1 in the management of prostate cancer. The inverse relationship between CYP24A1 expression and differentiation status in prostate cancer calls for the development of CYP24A1-selective inhibitors that can be used either as single entities delivered specifically to the tumor sites to increase or extend endogenous calcitriol function or in combination with exogenous doses of 1,25(OH)₂D₃ to achieve localized high concentrations of vitamin D₃, resulting in optimal growth inhibition of advanced tumors with minimal systemic toxicity. Mouse xenograft models for prostate cancer proved that modulating the intratumoral expression of CYP24A1 delayed tumor growth in vivo and significantly enhanced calcitriol-mediated tumor cell kill in target tumors. Previously, nonspecific compounds, liarazole (25) or ketoconazole (26), synergistically enhanced the antiproliferative action of vitamin D₃ in prostate cancer cell lines, through the broad inhibition of P450 enzymes. However, all these compounds lack the necessary selectivity and specificity and are likely to inhibit other cellular targets including the enzymes involved in the steroidogenic pathways for testosterone, cortisol, and aldosterone biosynthesis. For some patients, administration of these drugs may lead to adrenal insufficiency that would exacerbate the hypercalcemia associated with vitamin D treatment. This present work explored, for the first time, the application of RNAi technology to selectively silence the expression of CYP24A1 in prostate cancer cells. RNAi-mediated degradation of homologous mRNA offers a high specificity that is missing from current drug treatments. Local delivery of CYP24A1 siRNA or specific small inhibitory molecules obtained from screens of chemical libraries may be rendered clinically possible at the time of prostatectomy in patients and holds the great advantage of overcoming the risk of hypercalcemia. Targeting prostate tumors could also be a tangible therapeutic option in the future with gene therapy or nanotechnology developments.

Altogether, this study highlights the endogenous anticancer action of vitamin D₃ in human prostate and underlines the importance of CYP24A1 as a predictive marker for clinical efficacy of adjuvant vitamin D₃ therapy in patients with advanced prostate cancer. This study also provides, for the first time, a proof of principle that targeting of CYP24A1 by siRNA, a more selective and specific approach and, as a result, a potentially less toxic alternative, is efficient to enhance the antiproliferative action of vitamin D₃ in human prostate cancer cells, as well as in xenograft-bearing mouse models, offering a promising combination approach for vitamin D-based therapies for advanced prostate cancer.

Supplementary Material

This article includes supplemental data. Please visit http://www.fasebj.org to obtain this information.