19-Nor-1,25(OH)₂D₂ (a Novel, Noncalcemic Vitamin D Analogue), Combined with Arsenic Trioxide, Has Potent Antitumor Activity against Myeloid Leukemia

Cancer therapies are usually toxic for normal cells. Induction of differentiation is sometimes a useful, less toxic cancer therapy that can supplement more aggressive approaches. For example, the use of all-trans retinoic acid for the treatment of acute promyelocytic leukemia (APL) often induces complete remissions as well as its use in the treatment of head and neck cancers and neuroblastomas (1–3).

1,25-Dihydroxyvitamin D₃ [1,25(OH)₂D₃] is a member of the seco-steroid hormone family, which controls calcium homeostasis and bone metabolism. It can inhibit the growth of various types of malignant cells, including breast, prostate, colon, skin, and brain cancers as well as myeloid leukemia, by inducing differentiation, apoptosis, and cell cycle arrest. In a clinical study, orally administrated 1,25(OH)₂D₃ showed some activity for preleukemic patients (4). Because of the calcemic side effect of 1,25(OH)₂D₃, the dosage that could be safely given to these individuals was less than theoretically required for an anticancer effect as noted in vitro (5, 6). Therefore, novel, potent, but less calcemic analogues of vitamin D are being synthesized and tested (7–15).

19-Nor-1,25(OH)₂D₂ (paricalcitol) is a synthetic analogue of 1,25(OH)₂D₂ that is currently approved by the Food and Drug Administration for the clinical treatment of secondary hyperparathyroidism. This compound has very little calcemic potential as shown by several randomized clinical trials (16, 17). Antiproliferative effects of paricalcitol were reported recently against human prostate, leukemia, myeloma, and colon cancer cells in vitro and in vivo (18–20). These studies suggest that clinical trails of this analogue are reasonable; to this end, we are examining its effects in individuals with myelodysplastic syndrome.

In this study, we investigated the combinations of paricalcitol and either chemotherapeutic or related agents against human myeloid leukemia, myeloma, prostate, breast, and colon cancer in vitro. We found that the combination of paricalcitol and As₂O₃ had a strong antiproliferative effect against myeloid leukemia in vitro, and we studied the mechanism of action of this combination.

Cells and Samples. Cell lines were obtained from American Type Culture Collection (Rockville, MD) and maintained according to their recommendations. Myeloid leukemia cell lines (HL-60, NB-4, and U937), myeloma cell lines (RPMI-8226, ARH-77, and NCI-H929), lung cancer cell lines (NCI-H125 and NCI-H520), and prostate cancer cell lines (LNCaP and PC-3) were grown in RPMI 1640 with 10% FCS. Breast cancer cell lines (MCF-7 and MDA-MB-231), colon cancer cell lines (HT-29 and SW837), a prostate cancer cell line (DU145), and endometrial cancer cell lines (Ishikawa, HEC59, and HEC1B) were maintained in DMEM with 10% FCS. The U937 cell line (PR9) that stably expressed PML-retinoic acid receptor α (RARα) in a Zn²⁺-inducible fashion was described previously (21) and was kindly provided by Dr. Pelicci (Perugia University, Perugia, Italy). For induction of PML-RARα in PR9 cells, 0.1 mmol/L ZnSO₄ was added to the culture medium. Viable cell numbers were counted using trypan blue solution (Sigma-Aldrich, St. Louis, MO). The paricalcitol was generously provided by Scott Toner (Abbott Laboratories, Chicago, IL). PD98059 was obtained from LC Laboratories (Woburn, MA). Other compounds were obtained from Sigma-Aldrich.

3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium Bromide Assays. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTS, 5 mg/mL, Sigma-Aldrich) was used to measure cellular proliferation. Briefly, 10³ cells were incubated in culture medium containing test compounds for 96 hours followed the MTS assay as described before (19).

Soft Agar Colony Assay. Trypsinized and washed single-cell suspensions were enumerated and plated into 24-well flat-bottomed plates using a two-layer soft agar system with a total of 1 × 10³ cells per well in a volume of 400 μL/well as described previously (11).

Cell Cycle Analysis. After treatment of 5 × 10⁴ of cells with experimental compounds, cell cycle analysis was done as described previously (19).

Western Blot Analysis. Western blot was done as noted previously (19). Briefly, cells were washed and suspended in lysis buffer. After centrifugation, the suspension was collected. Protein (40 μg) was used for SDS-PAGE, transferred to a polyvinylidene difluoride membrane, and probed sequentially with antibodies (Santa Cruz Biotechnology, Inc., Santa Cruz, CA). The blots were developed using the SuperSignal West Pico Chemiluminescent Substrate Kit (Pierce, Rockford, IL).

Measurement of Apoptosis. Terminal deoxynucleotidyl transferase–mediated dUTP nick end labeling (TUNEL) assay was done as described (19) using the In situ Cell Death Detection POD kit (Roche, Indianapolis, IN).

Measurement of Cell Surface CD14 Antigen on Acute Myeloid Leukemia Cells. Myelocytic leukemia cell lines (HL-60, NB-4, and U937) were treated with paricalcitol and/or As₂O₃ and examined for CD14 expression by flow cytometry using CD14 antibody (DAKO, Carpinteria, CA) as described previously (15). Murine IgG1 antibody (DAKO) was used as a control.

Nitroblue Tetrazolium Test. Nitroblue tetrazolium (NBT) test was done as noted previously (19). Briefly, the cell suspension was mixed with an equal volume of NBT solution and incubated at 37°C for 30 minutes. The cells were cytocentrifuged onto glass slides, washed in cold methanol, and stained with safranin. The percentage NBT-positive cells were enumerated by light microscopy.

Real-time Quantitative PCR. RNA was prepared using a RNeasy kit obtained from Qiagen (Valencia, CA). Gene expression was quantified using real-time quantitative PCR (TaqMan, iCycler, Bio-Rad, Hercules, CA). The sequences of the primer sets used for this analysis were as follows: PCR primers used to amplify 25-hydroxyvitamin D₃-24-hydroxylase (CYP-24) were 5′-AGGCCACGTTGAAGACTTGT-3′ (forward) and 5′-TTCTTCTGGAGAAGCCCAAA-3′ (reverse), CCAAT/enhancer binding protein β (C/EBPβ): 5′-GACAAGCACAGCGACGAGTA-3′ (forward) and 5′-GTGCTGCGTCTCCAGGTT-3′ (reverse), and those used to amplify 18S rRNA were 5′-AAACGGCTACCACATCCAAG-3′ (forward) and 5′-CCTCCAATGGATCCTCGTTA-3′ (reverse). Primers were synthesized by Invitrogen (Carlsbad, CA).

Probes were purchased from Applied Biosystems (Foster City, CA) and were labeled with the reporter dye FAM in the 5′ end and the quencher dye TAMRA in the 3′ end. Amplification reactions were done with the Universal TaqMan PCR Master Mix (Applied Biosystems) in triplicates in an iCycler iQ system. The thermal cycling conditions using TaqMan method were as follows: 2 minutes at 50°C and 10 minutes at 95°C followed by 45 cycles of 95°C for 15 seconds and 60°C for 1 minute. To determine the relative expression level of each sample, additional reactions with four serial 5-fold dilutions of cDNA from HL-60 cells treated with paricalcitol were done to generate a standard curve, which related the threshold cycle to the log input amount of template.

In case of using SYBR Green method, a melting curve analysis was done following PCR to identify the correct product by its specific melting temperature, because SYBR Green intercalates nonspecifically into PCR products. Melting curve analysis included 95°C for 5 seconds, 65°C for 15 seconds, and heating to 95°C at a rate of 0.1°C/2 seconds with continuous reading of fluorescence. The signal in this study was generated by the binding of the fluorophore SYBR Green (Biozym, Hess. Oldendorf, Germany) to dsDNA. Baseline and threshold calculations were done with the iCycler software. The chemical composition of the PCR assays was according to the descriptions of the Taq polymerases. The amplification followed a three-step PCR with denaturation at 94°C for 20 seconds, annealing at 60°C for 10 seconds, elongation at 65°C for 25 seconds, and 20 seconds at melting temperature. PCR products were separated on a 2% agarose gel, stained with ethidium bromide, and photographed.

Determination of Mitochondrial Membrane Potential. Mitochondrial membrane potential was analyzed with a lipophilic cationic dye 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazol carbocyanine iodide (JC-1, Sigma-Aldrich) using a method described previously (22). JC-1 selectively incorporates into mitochondria. Intact mitochondrial membranes display a negative charge, which attracts positively charged JC-1 molecules. JC-1 aggregates form as a result of accumulation in the mitochondria. The monomeric form emits green fluorescence at 527 nm, whereas aggregated JC-1 molecules emit red fluorescence at 590 nm. Therefore, red fluorescence emission, assayed by fluorescence-activated cell sorting, is interpreted as a function of mitochondrial membrane potential. Cells (HL-60 and NB-4) that had been incubated either with or without paricalcitol and/or As₂O₃ were washed with PBS and incubated with 10 μg/mL JC-1 for 10 minutes at 37°C. Cells were transferred on ice for fluorescence-activated cell sorting analysis. Forward and side light scatters were used to gate and exclude cellular debris. Cells were excited at 488 nm, and JC-1 emission was collected on FL-1 and FL-2 channels at 530 and 590 nm, respectively. Ten thousand cells were analyzed per sample.

Measurement of CYP-24 Activity. Cells treated with paricalcitol (0.1 μmol/L for NB-4 and 0.01 μmol/L for HL-60) and/or As₂O₃ (0.6 μmol/L for NB-4 and 0.8 μmol/L for HL-60) for 4 days were pelleted and washed twice in serum-free RPMI 1640. Cells (1 × 10⁶) per sample were added to 1.5 pmol [³H]25-hydroxyvitamin D₃ in a glass tube and incubated for 4 hours at 37°C. All assays were done in triplicate for each treatment condition. Following incubation, methanol 500 (μL) was added to the cell suspension, and the mixture was mixed thoroughly. Chloroform (5 mL) was added to each tube, and the mixture was vortexed for 10 seconds. Samples were centrifuged at 1,800 rpm for 10 minutes, and the chloroform layer was isolated and evaporated under N₂ at room temperature. Sterols were resuspended in dichloromethane (60 μL) by vortexing and separated on silica TLC plates in dichloromethane/isopropanol (9:1, v/v). Conversion of [³H]25-hydroxyvitamin D₃ to [³H]24,25-dihydroxyvitamin D₃ was measured on a Bioscan System 200 Imaging TLC plate scanner (Edmonds, WA) and expressed as fmol converted hour⁻¹ 1 × 10⁶ cells⁻¹.

Paricalcitol, in Combination with As₂O₃, Has a Prominent Antiproliferative Activity against Human Myeloid Leukemia Cells. The antiproliferative effects of paricalcitol in combination with other clinically useful therapeutic agents were examined against various cancer cell lines in vitro. Initial screening was done using the rapid MTS assay with a relative short exposure of 4 days to the various agents. These experiments for each agent were done to identify a concentration that produced modest growth inhibition to test subsequently its combination with paricalcitol to identify those pairs that produced enhanced activity (data not shown). Paricalcitol slightly enhanced both daunorubicin's inhibition of growth when tested against myeloid leukemia cells (HL-60 and U937) and Adriamycin's effect on proliferation of MCF-7 breast cancer cells (Fig. 1A). Paricalcitol also slightly augmented the antiproliferative effects of Adriamycin and taxol against breast (MCF-7 and MDA-MB-231) and prostate (PC-3 and DU145) cancer cells. The combination of paricalcitol and dexamethasone had prominent antiproliferative activity against multiple myeloma cells (NCI-H929 and RPMI-8226; Fig. 1A). Among the various treatment combinations, paricalcitol and As₂O₃ had the most notable antiproliferative effects against myeloid leukemia cells (HL-60 and NB-4) compared with either drug alone. Each drug suppressed the clonal growth of both cell lines in a dose-dependent manner, but the two drugs together had markedly enhanced antiproliferative effects on these cells (Fig. 1B).

We focused our attention on the paricalcitol and As₂O₃ combination and its antiproliferative activity against myeloid leukemia cells. Time course assays of viable cell numbers counted by trypan blue assays were done during which the cells were treated with paricalcitol (0.1 μmol/L for NB-4 and 0.01 μmol/L for HL-60) and As₂O₃ (0.6 μmol/L for NB-4 and 0.8 μmol/L for HL-60; Fig. 1C). Results showed that the numbers of viable cells in the population treated with both drugs was significantly less (P < 0.05) compared with diluant controls or the leukemic cells exposed to each drug alone for 5 (NB-4) or 6 (HL-60) days. We cultured prostate (LNCaP, PC-3, and DU145), breast (MCF-7), colon (HT-29), endometrial (Ishikawa, HEC59, and HEC1B), and lung (NCI-H125 and NCI-H520) cancer cell lines with paricalcitol (0.1 μmol/L) and As₂O₃ (1 μmol/L) and measured cell growth (MTS assay) on day 4 of culture. This combination showed some additive antiproliferative effects on PC-3 and slight effects on DU145 prostate cancer cells (Fig. 1D).

Paricalcitol, Combined with As₂O₃, Markedly Enhanced Monocytic Differentiation of HL-60 and NB-4 Myeloid Leukemia Cells with Subsequent Increased Apoptosis. We have shown previously that paricalcitol induces monocytic differentiation and apoptosis of human myeloid leukemia cells (19). Here, we explored the effect of paricalcitol in combination with As₂O₃ on differentiation and apoptosis of NB-4 and HL-60 cells (Fig. 2A). After treatment with paricalcitol (0.1 μmol/L for NB-4 and 0.01 μmol/L for HL-60) and As₂O₃ (0.6 μmol/L for NB-4 and 0.8 μmol/L for HL-60) for 3 days, CD14, a marker of monocytic differentiation, was measured by flow cytometry in NB-4 and HL-60 cells. Neither paricalcitol nor As₂O₃ alone induced CD14 expression in NB-4 cells (Fig. 2A). In HL-60 cells, paricalcitol alone, but not As₂O₃, increased levels of CD14. The combination of both compounds markedly induced CD14 expression on both NB-4 and HL-60 cells compared with either drug alone (Fig. 2A). Differentiation was also measured by induction of superoxide production as monitored by percentage of cells able to reduce NBT. Cells were treated with either paricalcitol (0.1 μmol/L for NB-4 and 0.01 μmol/L for HL-60) and/or As₂O₃ (0.6 μmol/L for NB-4 and 0.8 μmol/L for HL-60; Fig. 2B). For NB-4, paricalcitol combined with As₂O₃ markedly increased the NBT reduction (38% positive cells), whereas each compound alone had no effect. For HL-60 cells, NBT reduction was increased by paricalcitol (14% positive cells), and As₂O₃ enhanced the activity of paricalcitol (48% positive cells).

We examined the effect of the combination of both agents on the induction of apoptosis using two measurements: (a) percentage of cells in the sub-G₁ population of the cell cycle and (b) TUNEL assay. NB-4 and HL-60 cells were cultured with paricalcitol (0.1 μmol/L for NB-4 and 0.01 μmol/L for HL-60) and/or As₂O₃ (0.6 μmol/L for NB-4 and 0.8 μmol/L for HL-60) for 4 days. For NB-4 cells, As₂O₃ alone induced apoptosis as shown by an increased percentage of cells in the sub-G₁ population (33%), whereas 4% and 5% of control and paricalcitol-treated cells, respectively, were in the sub-G₁ population. The combination of both compounds dramatically increased apoptotic cell death, with 97% of NB-4 cells in the sub-G₁ population (Fig. 2C). We also determined the percentage of apoptotic NB-4 cells by TUNEL assay after treatment. The combination increased apoptosis (68% of the cells) compared with either agent alone (paricalcitol, 3%; As₂O₃, 14%; Fig. 2D). Apoptosis was also prominent in HL-60 cells exposed to both compounds as measured by the percentage of cells in the sub-G₁ population (46%) compared with those treated with either paricalcitol (12%) or As₂O₃ (20%) alone and by TUNEL assay (46% apoptotic cells in the presence of both agents compared with 7% and 20% in the presence of paricalcitol or As₂O₃ alone, respectively; Fig. 2C and D).

Modulation of Gene Expression by Paricalcitol and As₂O₃ in Myeloid Leukemia Cells. The enzyme CYP-24 catalyzes the first step in the catabolism of 1,25(OH)₂D₃ (23). Expression of CYP-24 is activated by the binding of its ligand 1,25(OH)₂D₃ to vitamin D receptor (VDR). The VDR-ligand complex (VDR-retinoid X receptor) binds to the vitamin D response element in the CYP-24 promoter and initiates transcription (24, 25). Because we have reported recently that the activity of paricalcitol is mediated through VDR (19), we examined the expression of CYP-24 as an early vitamin D target gene by real-time PCR. The levels of CYP-24 mRNA increased 2 × 10⁴-fold after NB-4 cells were cultured with paricalcitol (0.1 μmol/L, 24 hours) compared with control NB-4 cells (Fig. 3A). When the cells were exposed to both paricalcitol (0.1 μmol/L) and As₂O₃ (0.6 μmol/L), the transcriptional level of CYP-24 increased even more (4 × 10⁴-fold compared with control), suggesting that As₂O₃ enhanced transcriptional activation through the VDR (Fig. 3A). The level of expression of VDR and retinoid X receptor was not altered by exposure of NB-4 cells to either paricalcitol (0.1 μmol/L) and/or As₂O₃ (0.6 μmol/L; Fig. 3B; data not shown).

A recent study showed that C/EBPβ is a target gene of VDR and has an important role in CYP-24 transcription (26, 27). We examined the mRNA and protein expression of C/EBPβ by real-time reverse transcription-PCR (Fig. 3A) and Western blot analysis (Fig. 3B). In NB-4 cells, expression of C/EBPβ was low. Exposure to either paricalcitol or As₂O₃ for 24 hours increased the level of C/EBPβ mRNA by 4.3- and 2.5-fold compared with control cells, respectively. In contrast, the combination of both drugs increased expression by 90-fold compared with control (Fig. 3A). These results suggest that As₂O₃ increased the transcriptional activity of some VDR target genes. Expression of C/EBPβ protein also increased after exposure of NB-4 to both drugs (Fig. 3B). In addition, C/EBPβ protein levels increased when HL-60 cells were exposed to paricalcitol and these levels increased even more in the presence of both paricalcitol and As₂O₃ (Fig. 3C). Western blot analysis showed that As₂O₃ either alone or with paricalcitol resulted in the dramatic loss of the PML-RARα fusion protein (Fig. 3B).

As shown above, in association with enhancement of differentiation by the combination of paricalcitol and As₂O₃, HL-60 and NB-4 cells underwent apoptosis (Fig. 2). Expression of the antiapoptotic proteins Bcl-2 and Bcl-x_L markedly decreased (48% and 38%, respectively) after combined treatment of HL-60 cells with paricalcitol and As₂O₃; however, the level of expression of the proapoptotic protein Bax was unchanged by either drug alone or in combination (Fig. 3C).

Vitamin D has been suggested to act through genomic and nongenomic pathways (28). Activated extracellular signal-regulated kinase (ERK) seems to have an important role in monocytic differentiation induced by vitamin D₃ in HL-60 cells (29). We found that the levels of phosphorylated ERK increased to a similar degree after exposure of HL-60 to either paricalcitol and/or As₂O₃. Highest levels of phosphorylation expression occurred on the first day of exposure to either of the compounds (Fig. 3D). Because the expression of phosphorylated ERK might be important for differentiation, we treated HL-60 cells with the selective mitogen-activated protein kinase/ERK inhibitor, PD98059 (25 μmol/L), in combination with paricalcitol and/or As₂O₃; CD14 expression, a marker of monocytic differentiation, was measured by flow cytometry (Fig. 3E). When HL-60 cells were treated with paricalcitol and/or As₂O₃ in the presence of PD98059, the differentiation induced by paricalcitol decreased by 23% compared with paricalcitol alone, and the magnitude of inhibition was not significantly changed by the combination of both compounds. Therefore, monocytic differentiation induced by paricalcitol, and the enhanced differentiation by the addition of As₂O₃, could be blocked by inhibiting the ERK pathway.

Paricalcitol, in Combination with As₂O₃, Overcomes the Block of Differentiation Mediated by the PML-RARα Fusion Protein Present in APL Cells. NB-4 promyelocytic leukemia cells express the PML-RARα fusion gene, which has an important role in the pathogenesis of APL. As₂O₃, which is used as a therapeutic agent for this disease, is reported to enhance the degradation of this fusion protein (30). We found that As₂O₃ caused a profound decrease in the protein level of PML-RARα in NB-4 cells after 3 days such that we could not detect its expression by Western blot (Fig. 4A).

To examine the association between levels of the fusion protein and the effect of the combination treatment, we used an engineered U937 monocytic leukemia cell line (U937-PR9) that had a stably integrated PML-RARα cDNA expression vector under the control of the Zn²⁺-inducible murine metallothionein 1 promoter (21). U937 cells transfected with the empty vector (U937-PMT) were used as a control. When we cultured wild-type U937 with paricalcitol and/or As₂O₃, paricalcitol increased monocytic differentiation, and the combination of both agents further increased the percentage of CD14-expressing cells. As₂O₃ alone did not induce CD14 expression. This suggested that As₂O₃ enhanced paricalcitol-induced monocytic differentiation of U937 cells. Similarly, the combination also enhanced paricalcitol-induced monocytic differentiation of THP-1 monocytic leukemia cells (data not shown).

Next, we cultured either U937-PR9 or U937-PMT cells with paricalcitol and As₂O₃. Treatment of U937-PMT control cells with paricalcitol induced monocytic differentiation, and paricalcitol in combination with As₂O₃ enhanced their monocytic differentiation in either the presence or the absence of Zn²⁺ (Fig. 4B). Next, we cultured U937-PR9 experimental cells with Zn²⁺, resulting in these cells expressing PML-RARα (Fig. 4A), and the differentiation induced by paricalcitol in these cells was reduced by 80% compared with PR9 without Zn²⁺ (Fig. 4B). Addition of As₂O₃ to the medium decreased in a dose-dependent manner the protein levels of PML-RARα in the Zn²⁺-stimulated U937-PR9 cells. At 0.4 μmol/L As₂O₃, PML-RARα was rarely detected in U937-PR9 cells (Fig. 4A). The combination of paricalcitol and As₂O₃ enhanced the differentiation of U937-PR9 cells in parallel with the As₂O₃-induced decrease of PML-RARα levels (Fig. 4B).

As₂O₃, Induced the Damage of Mitochondria and Decreased CYP-24 Activity in Acute Myeloid Leukemia Cells. The mitochondrial enzyme CYP-24 is a transcriptional target gene of vitamin D. We showed that CYP-24 mRNA was markedly increased in the presence of paricalcitol (19) and that this was prominently enhanced by As₂O₃ in NB-4 cells (Fig. 3B). This enzyme catalyzes the initial step in the conversion of the active molecule 1,25(OH)₂D₃ into its less active metabolite 1,24,25(OH)₂D₃. Therefore, induction of this enzyme by 1,25(OH)₂D₃ and other active vitamin D compounds has a negative feedback on levels of the active vitamin D (31–34). On the other hand, As₂O₃ induces apoptosis of many cancers through mitochondrial membrane collapse (35–37). This raises the possibility that the As₂O₃ inactivates the mitochondrial enzyme CYP-24 leading to an increase in intracellular levels of the active vitamin D compound and thus explaining the enhanced activity of paricalcitol in the presence of As₂O₃.

To explore this hypothesis, we first examined if the mitochondria were damaged by either As₂O₃ and/or paricalcitol in NB-4 cells. Mitochondrial potential was measured using the JC-1 dye in conjunction with flow cytometry (Fig. 5A). The geometric mean of red fluorescence intensity was recorded as a function of mitochondrial potential. The control cells showed a geometric mean of 133.9 for red fluorescence intensity. Cells cultured with paricalcitol alone (10⁻⁸ mol/L) and As₂O₃ (0.6 μmol/L) alone displayed similar mitochondrial potential levels, with geometric means of red fluorescence of 147.4 and 140, respectively. However, when treated with both compounds together, the level of red fluorescence decreased to a geometric mean of 69.2, indicating a marked reduction in mitochondrial membrane potential of these NB-4 APL cells compared with control and either agent alone.

The CYP-24 is a mitochondrial enzyme. Because As₂O₃ either alone or in combination with paricalcitol induces mitochondria damage, we hypothesized that As₂O₃ would decrease the activity of this enzyme. Activity of CYP-24 in leukemia cell lines was monitored by measuring the conversion of 25(OH)₂D₃ to 1,24,25(OH)₂D₃ using TLC analysis. HL-60 and NB-4 cells were cultured with paricalcitol and/or As₂O₃ for 3 days. Control cells were treated with vehicle alone. As expected, paricalcitol alone increased the activity of CYP-24 in HL-60 and NB-4 cells, whereas As₂O₃ alone decreased activity compared with control cells (Fig. 5B). When cells were treated with paricalcitol and As₂O₃, CYP-24 activity was significantly less in cells cultured with paricalcitol alone (HL-60, P < 0.05; NB-4, P < 0.05; Fig. 5B). These results indicated that As₂O₃ decreased the enzyme activity of CYP-24. This would be expected to result in higher intracellular levels of biologically active paricalcitol in leukemia cells. Mitochondrial injury induced by As₂O₃ may be the cause of this CYP-24 inactivation.

In this study, we show that treatment with paricalcitol, when combined with clinically achievable concentrations of As₂O₃ (0.6-0.8 μmol/L; ref. 38), results in strong antiproliferative effects on the human myeloid leukemia cell lines HL-60 and NB-4. NB-4 cells were established from a patient with APL and HL-60 was derived from a patient with acute myeloid leukemia M2 with promyelocytic features (39). This combination also produced effects against PC-3 and DU145 prostate cells. Paricalcitol combined with dexamethasone also significantly decreased the growth of myeloma cells (NCI-H929 and RPMI-8226). This deserves further investigations.

We reported previously that the induction of macrophage differentiation by paricalcitol may be mediated through VDR (19). The CYP-24 is the enzyme responsible for the first step in the catabolism of 1,25(OH)₂D₃ (23). Transcriptional induction of expression of CYP-24 is dependent on ligand activation of VDR and binding of the complex to the vitamin D response element of the promoter of the CYP-24 gene (24, 25).

Transcription of CYP-24 was activated by paricalcitol and enhanced by As₂O₃, although As₂O₃ alone did not induce expression of the enzyme in NB-4 cells (Fig. 3A). In addition, the mRNA and protein expression of C/EBPβ was slightly increased by paricalcitol, and the combination of both drugs also enhanced the mRNA and protein expression of this transcriptional factor in NB-4 cells (Fig. 3A and B). In addition, in HL-60 cells, paricalcitol increased the protein expression of C/EBPβ, and the combination of both drugs enhanced its expression (Fig. 3C). These result suggested that As₂O₃ enhance the transcriptional activity of paricalcitol-activated VDR. In contrast, the activity of CYP-24 was decreased by As₂O₃ in both HL-60 and NB-4 cells. This mitochondrial enzyme catalyzes the initial step in the conversion of the active molecule 1,25(OH)₂D₃ into the less active metabolite 1,24,25(OH)₂D₃, resulting in the inhibition of antiproliferative effects of vitamin D (31–34). In fact, CYP-24 has been suggested to be an oncogene (34). We believe that As₂O₃ inhibits intracellular CYP-24 activity by disrupting mitochondrial integrity, resulting in higher intracellular levels of the vitamin D analogue. Therefore, although higher intracellular levels of the ligand can stimulate transcription of CYP-24, As₂O₃ blocks its activity. The combination of vitamin D and inhibitors of CYP-24, such as genistein and isoflavone, have been suggested to be a useful approach to cancer therapy (40, 41). Our studies extend these observations to two clinically relevant compounds.

Recently, studies related to the identification of coactivators of VDR-mediated transcription have been a major focus of vitamin D research (26, 27). C/EBPβ is induced by 1,25(OH)₂D₃ in renal and osteoblastic cells and can enhance the transcription of CYP-24 induced by 1,25(OH)₂D₃. Thus, C/EBPβ seems to be transcriptionally stimulated by vitamin D compounds and to have a cooperative effects with VDR in enhancing the 1,25(OH)₂D₃-induced transcription (26, 27).

During the last decade, the efficiency of As₂O₃ has been established for both newly diagnosed and relapsed patients with APL. It can be used as a single agent and induces complete remissions with only minimal myelosuppression (42). The NB-4 promyelocytic leukemia cell line is a good model for APL having the prerequisite t(15;17) chromosomal translocation specific for APL. In this study, paricalcitol when combined with As₂O₃ greatly suppressed the growth of NB-4 cells by inducing monocytic differentiation and subsequent apoptosis, whereas paricalcitol alone was unable to induce monocytic differentiation of these cells.

The PML-RARα fusion protein impairs differentiation in response to vitamin D in vitro and in vivo (43–45). This block of differentiation is thought to occur by the aberrant recruitment of corepressor proteins and histone deacetylases and perhaps by direct sequestering the vitamin D₃ receptor (45, 46).

As₂O₃ may be efficacious in APL by several mechanisms. Arsenic rapidly inhibits the interaction of the SMRT corepressor with PML-RARα leading to the reduction of the repression of retinoid target genes (47). Secondly, arsenic enhances the degradation of the PML-RARα by promoting the binding of SUMO to the fusion protein (48, 49). A recent study suggests that arsenic increases acetylation of histones H3 and H4 leading to gene activation (38).

Our study showed that the As₂O₃-enhanced differentiation mediated by paricalcitol was associated with a decreased protein expression of PML-RARα in NB-4 and U937-PR9 cells containing a Zn²⁺-inducible PML-RARα expression vector (Figs. 3A and 4), suggesting that the degradation of this fusion protein by As₂O₃ has an important role for the enhanced response to paricalcitol. Therefore, targeting of the fusion protein might be important in allowing vitamin D compounds to induce differentiation. This is further supported by the fact that arsenic does not lead to degradation of the PLZF-RARα fusion protein and is not effective in this subtype of APL (50, 51).

In addition to its VDR-mediated genomic actions, 1,25(OH)₂D₃ can activate other signal pathways, including ERKs, in a nongenomic manner, which may effect diverse processes, including cell proliferation and differentiation (52). A rapid but transient activation of ERK1/2 has been reported following induction of monocytic differentiation of HL-60 cells by high concentrations (10⁻⁶ mol/L) of 1,25(OH)₂D₃ (29, 53). On the other hand, rapid inactivation of ERK1/2 has been observed when U937 leukemia cells were treated with As₂O₃, which was associated with apoptosis (54). We treated HL-60 cells with a moderate concentration (10⁻⁷ mol/L) of paricalcitol and/or As₂O₃ (4 × 10⁻⁷ mol/L). Paricalcitol induced maximal expression of phosphorylated ERK after 1 day (Fig. 3D). As₂O₃ also activated ERK1/2, which was also maximal at 1 day of exposure (Fig. 3D). The selective mitogen-activated protein kinase/ERK kinase inhibitor PD98059 inhibited paricalcitol-induced differentiation of HL-60 cells; the combination of paricalcitol and As₂O₃ could not overcome this inhibition. These data suggest that the activation of mitogen-activated protein kinase/ERK kinase is necessary for the differentiation induced by paricalcitol as well as the combination of paricalcitol and As₂O₃.

Higher concentrations of arsenic have been shown to induce apoptosis of a variety of human hematopoietic malignancies and solid tumors (55, 56). Concerning the pathways whereby arsenic compounds induce apoptosis, attention has focused on their potential effects on the mitochondria. A leading mechanism involves alteration of transmembrane potential by modulating the permeability of the transition pore cortex leading to the release of cytochrome C, culminating in the cleaving of caspases, including caspase-3 and caspase-9 (38). In our study, As₂O₃ combined with paricalcitol decreased mitochondrial potential. This mitochondrial damage induced by the combination of both agents impaired the activity of the mitochondrial enzyme CYP-24 leading to a higher sensitivity to paricalcitol.

In this study, we have found that paricalcitol and As₂O₃ potently slowed growth and induced differentiation of myeloid leukemia cells, and this probably occurred by As₂O₃ decreasing levels of the PML-RARα fusion protein in APL and inactivating CYP-24 activity in mitochondria resulting in the increased activity of paricalcitol. The combination of both of these Food and Drug Administration–approved drugs should be considered for all-trans retinoic acid–resistant APL patients and as maintenance therapy for some types of acute myeloid leukemia. Furthermore, the combination of dexamethasone and paricalcitol deserves further exploration for the therapy of multiple myeloma.

Grant support: NIH, Parker Hughes Fund, and Abbott Laboratories.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

We thank Scott Toner (Abbott Laboratories) for helpful suggestions.