Antitumor Activity Against Leukemia

Home , Paricalcitol

Vitamin D2 Analog 19-nor-1,25-Dihydroxyvitamin D2: Antitumor Activity Against Leukemia, Myeloma, and Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 Colon Cancer Cells

Takashi Kumagai, James O’Kelly, Jonathan W. Said, H. Phillip Koeffler

A member of the secosteroid hormone family, 1,25- Background: 1,25-Dihydroxyvitamin D3 inhibits growth of dihydroxyvitamin D3 [1,25(OH)2D3], controls calcium homeo- several types of human cancer cells in vitro, but its thera- stasis and bone metabolism. 1,25(OH)2D3 can inhibit the growth peutic use is hampered because it causes hypercalcemia. 19- of various types of malignant cells, including breast, prostate, nor-1,25-Dihydroxyvitamin D2 (paricalcitol) is a noncalce- colon, skin, and brain cancer cells, as well as myeloid leukemia mic vitamin D analog that is approved by the Food and Drug cells. In a clinical study, orally administered 1,25(OH)2D3 was Administration for the treatment of secondary hyperpara- partially useful for preleukemic patients (1), but because thyroidism. We investigated the antitumor activity and 1,25(OH)2D3 also causes hypercalcemia, the dose that could be mechanism of action of paricalcitol in vitro and in vivo. Meth- given to these patients was less than that theoretically required ods: Effects of paricalcitol on proliferation, the cell cycle, for an anticancer effect (2,3). Consequently, new analogs of differentiation, and apoptosis were examined in cancer cell vitamin D that are potent but less calcemic are being synthesized lines. Effects on tumor growth were examined with colon and tested (4–12). cancer cell xenografts in nude mice (five in the experimental Paricalcitol (19-nor-1,25-dihydroxyvitamin D2) is a synthetic group and five in the control group). The interaction of pari- analog of vitamin D that is approved by the Food and Drug calcitol with the vitamin D receptor (VDR) in mononuclear Administration for the clinical treatment of secondary hyper- spleen cells and myeloid stem cells from wild-type and VDR parathyroidism. This compound has very little calcemic activity, knockout mice was examined. All statistical tests were two- as demonstrated in randomized controlled clinical trials (13,14). sided. Results: Paricalcitol inhibited the proliferation of my- Antiproliferative activity of paricalcitol has been demonstrated eloid leukemia cell lines HL-60, NB-4, and THP-1 cells at an (15) in human prostate cancer cells in vitro. In this study, we effective dose that inhibited growth 50% (ED50)of2.4–5.8 × investigate the activity of paricalcitol in leukemia and myeloma 10–9 M by inducing cell cycle arrest and differentiation. Pari- cells in vitro and colon cancer cells in vitro and in vivo and calcitol inhibited the proliferation of NCI-H929 myeloma explore its mechanism of action. –10 cells at an ED50 of 2.0 × 10 M by inducing cell cycle arrest and apoptosis. Paricalcitol also inhibited the proliferation of MATERIALS AND METHODS –8 colon cancer cell lines HT-29 (ED50 =1.7×10 M) and –8 Cells and Reagents SW837 (ED50 =3.2×10 M). HT-29 colon cancer xenografts in paricalcitol-treated nude mice were smaller (1044 mm3 Cell lines used in this study were obtained from American and 1752 mm3, difference = 708 mm3, 95% confidence in- Type Culture Collection (Manassas, VA) and were maintained terval = 311 to 1104 mm3; P = .03) and weighed less (1487 mg according to their recommendations. Myeloid leukemia cell and 4162 mg, difference = 2675 mg, 95% confidence interval = 2103 to 3248 mg; P<.001) than those in vehicle-treated mice. Paricalcitol induced committed myeloid hematopoietic Affiliation of authors: T. Kumagai, J. O’Kelly, H. P. Koeffler (Division of stem cells from wild-type but not from VDR knockout mice Hematology/Oncology, Department of Medicine, Cedars-Sinai Medical Center), to differentiate as macrophages. Conclusion: Paricalcitol has J. W. Said (Department of Pathology, Center for Health Science), University of California at Los Angeles School of Medicine, Los Angeles, CA. anticancer activity against myeloid leukemia, myeloma, and Correspondence to: Takashi Kumagai, M.D., Cedars-Sinai Medical Center/ colon cancer cells that may be mediated through the VDR. UCLA, School of Medicine, Davis Bldg. 5068, 8700 Beverly Blvd., Los Ange- Because it has been approved by the Food and Drug Admin- les, CA 90048 (e-mail: [email protected]). istration, clinical trials of this agent in certain cancers are See “Notes” following “References.” reasonable. [J Natl Cancer Inst 2003;95:896–905] Journal of the National Cancer Institute, Vol. 95, No. 12, © Oxford University Press 2003, all rights reserved.

896 ARTICLES Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 lines (HL-60, NB-4, THP-1, and U937), lymphoma cell lines (Immobilon; Amersham, Arlington Heights, IL), and probed se- (Raji, Ramos, Daudi, Jurkat, Jeko-1, and JUDHL), and myeloma quentially with antibodies against the following proteins: cell lines (RPMI-8226, ARH-77, and NCI-H929) were grown in p21WAF1, p27KIP1, PTEN, BCL2, BAX, C-MYC, cyclin D1, RPMI-1640 medium with 10% fetal calf serum (FCS). Breast cyclooxygenase (COX-1), COX-2, and glyceraldehyde-3- cancer cell lines (MCF-7 and MDA-MB-231), brain cancer cell phosphate dehydrogenase (GAPDH) (Santa Cruz Biotechnol- lines (U343, U118, U138, U373, and U87), and colon cancer cell ogy, Santa Cruz, CA). The blots were developed with the Su- lines (HT-29, SW837, SW480, SW620, and HCT116) were perSignal West Pico Chemiluminescent Substrate kit (Pierce, maintained in Dulbecco’s modified Eagle medium (DMEM) Rockford, IL). with 10% FCS. The endometrial carcinoma cell line AN-3 was Measurement of Apoptosis maintained in the ␣ modification of minimal essential medium (␣-MEM; Gibco BRL, Grand Island, NY) with 10% FCS. Pari- The cells undergoing apoptosis were detected by morpho- calcitol and 1,25(OH)2D3 were obtained from Abbott Laborato- logic changes including nuclear shrinkage and cellular fragmen- ries (Abbott Park, IL). tation by staining cells with Hoechst 33258 (Bisbenzimide H 3-(4,5-Dimethylthiazol-2-yl)-2,5-Diphenyltetrazolium 33258; Sigma). Cells in the sub-G1 population, an indication of Bromide Assays for Cell Proliferation and Viability the number of apoptotic cells present, were detected by cell Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 cycle analysis with flow cytometry. The terminal deoxynucleo- To measure cell proliferation and viability, 3-(4,5- tidyl transferase-mediated uridine 5Ј-triphosphate nick end la- dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; beling (TUNEL) assay was also used to detect and quantify Sigma, St. Louis, MO) was dissolved in phosphate-buffered sa- apoptosis (In Situ Cell Death Detection kit; Roche, Indianapolis, line (PBS) at 5 mg/mL. Approximately 104 cells per well were IN). incubated in culture medium for 96 hours in 96-well plates, and Measurement of Cell Surface CD14 Antigen on then 10 ␮L of the MTT solution was added. After a 4-hour HL-60 Cells incubation, 100 ␮L of solubilization solution (20% sodium do- decyl sulfate [SDS]) was added, and the mixture was incubated HL-60 myelocytic leukemia cells were treated with1×10–7 –7 at 37 °C for 16 hours. In this assay, MTT is cleaved to an orange M 1,25(OH)2D3, 1×10 M paricalcitol, or a vehicle control formazan dye by metabolically active cells. The absorbance of (10% FCS in RPMI-1640 medium containing 0.01% ethanol) for the formazan product is measured with an enzyme-linked im- 4 days and examined for CD14 expression with CD14 antibody munosorbent assay reader at 540 nm. (DAKO, Carpinteria, CA) by flow cytometry, as described pre- Soft Agar Colony Assay viously (12). Murine immunoglobulin G1 antibody (DAKO) was used as a control antibody. The two-layer soft agar system (10) was used. Cells were Murine Studies removed from culture plates with trypsin and washed. Cells in single-cell suspension were counted. Approximately1×103 BNX nu/nu mice were purchased from Harlan (Indianapolis, cells in 400 ␮L of medium per well were plated into 24-well IN) at 8 weeks of age and cared for in accordance with the flat-bottom plates and incubated for 14 days at 37 °C in a hu- guidelines of Cedars-Sinai Research Institute. They were main- midified atmosphere containing 5% CO2/95% air, and then the tained in pathogen-free conditions and fed irradiated chow. A colonies were counted. total of1×106 HT-29 cells in 0.1 mL of Matrigel (Collaborative Cell Cycle Analysis Biological Products, Bedford, MA) was injected subcutaneously into bilateral flanks of each mouse, resulting in the formation of –7 Cultured cells were treated with1×10 M 1,25(OH)2D3,1 two tumors per mouse. The mice were assigned blindly and ×10–7 M paricalcitol, or a vehicle control (10% FCS in RPMI- randomly to the experimental group or the control group. Treat- 1640 medium containing 0.01% ethanol) for 3 or 4 days. All ment started the day after HT-29 cells were injected and con- cells (those in suspension and those attached to the culture dish) tinued for 4 weeks. The five control mice received vehicle (100 were collected, washed, suspended in ice-cold PBS, fixed in ␮L of PBS containing 0.024% ethanol per day, intraperitoneally, 75% chilled methanol at 4 °C, and stained with propidium io- 3 days per week on Monday, Wednesday, and Friday) only, and dine. Cell cycle status was analyzed on a Becton Dickinson Flow the five experimental mice received paricalcitol (100 ng in 100 Cytometer (BD Biosciences, Franklin Lakes, NJ). ␮L of PBS containing 0.024% ethanol per day, intraperitoneally, 3 days per week on Monday, Wednesday, and Friday). Tumor Western Blot Analysis sizes were measured every week and calculated by the formula Cells were washed twice in PBS, suspended in lysis buffer A × B × C × 0.5236, where A is the length, B is the width, and (50 mM Tris [pH 8.0], 150 mM NaCl, 0.1% SDS, 0.5% sodium C is the height, all measured in millimeters. Tumor size and deoxycholate, 1% Nonidet P-40, phenylmethylsulfonyl fluoride weight from each mouse were calculated as the total of the two at 100 ␮g/mL, aprotinin at 2 ␮g/mL, pepstatin at 1 ␮g/mL, and bilateral tumors per mouse. After 4 weeks, blood was collected leupeptin at 10 ␮g/mL), and placed on ice for 30 minutes. After to measure the level of serum calcium. All mice were killed at centrifugation at 15 000g for 15 minutes at 4 °C, the suspension the end of the study, and tumors were fixed in 10% neutral was collected. Protein concentrations were quantitated by using buffered formalin and embedded in paraffin for histologic analy- the Bio-Rad Protein Assay Dye Reagent Concentrate (Bio-Rad sis. Laboratories, Hercules, CA), according to the manufacturer’s Measurement of Serum Calcium in Mice recommendation. Proteins in whole lysates (40 ␮g) were re- solved by SDS–polyacrylamide gel electrophoresis in a 4%– To measure the serum calcium levels in mice, we used Sigma 15% gel, transferred to a polyvinylidene difluoride membrane Diagnostics calcium reagent (Sigma) containing o-cresolph-

Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 ARTICLES 897 thalein, which interacts with calcium to form a purple-colored RESULTS complex. The colored complex was directly quantified with an enzyme-linked immunosorbent assay reader at 575 nm and ab- Paricalcitol and Clonal Proliferation of Leukemia and sorbance (A) of the blank, standard, and sample were measured. Colon Cancer Cell Lines In Vitro Calcium concentration of the sample (mg/dL) was calculated as We examined whether paricalcitol had antitumor activity in (Asample – A blank)/(Astandard – A blank) × 10. human cancer cell lines derived from breast, brain, colon, and Vitamin D Receptor Knockout Mice and Soft Gel uterine cancers, as well as from myeloid leukemia, lymphoma, Colony-Forming Assay and myeloma. We first used the rapid MTT assay with a rela- tively short 4-day exposure to paricalcitol (data not shown) to Five vitamin D receptor (VDR) knockout mice were gener- show that paricalcitol had antiproliferative activity in myeloid ated, and genotypes were determined by Southern blot analysis leukemia cells (HL-60, NB-4, and THP-1), myeloma cells (NCI- as described previously (16). Five wild-type littermates were H929), and colon cancer cells (HT-29 and SW837). We then used as controls. Mice were killed by cervical dislocation. Bone tested whether paricalcitol affected the ability of the same cell marrow was flushed from isolated femurs with ␣-MEM contain- lines to form colonies in soft agar (Fig. 1). The concentrations of ing 10% FCS by the use of a 26-gauge needle and syringe. Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 paricalcitol that caused 50% inhibition (ED50) of clonal growth Isolated spleens were injected with DMEM (Gibco BRL) con- in soft agar were as follows: for HL-60 cells, 2.4 × 10–9 M; for taining 10% FCS and crushed with forceps to release cells. NB-4 cells, 3.4 × 10–9 M; for THP-1 cells, 5.8 × 10–9 M; for Mononuclear cells from bone marrow or spleen were separated HT-29 cells, 1.7 × 10–8 M; for SW837 cells, 3.2 × 10–8 M; and by Ficoll–Hypaque density centrifugation (Amersham Pharma- –10 for NCI-H929 cells, 2.0 × 10 M. These ED50 values were cia, Uppsala, Sweden). comparable to those obtained with the physiologically active For the soft gel assay, resuspended mononuclear bone mar- form of vitamin D3 1,25(OH)2D3 (Fig. 1). row cells (2 × 104 cells per mL) and growth factors to final concentrations of 1% methylcellulose, 30% FCS, granulocyte– Paricalcitol and Myeloid Leukemia Cells In Vitro macrophage colony-stimulating factor (10 ng/mL), interleukin 3 –7 –4 For cell cycle analysis, HL-60 cells were treated with1×10 (10 ng/mL), 1% bovine serum albumin,1×10 M 2-mercap- M paricalcitol or vehicle for 72 hours and examined by flow toethanol, and 2 mM L-glutamine, as described previously, were cytometry. Paricalcitol-treated HL-60 cells accumulated in the added to methylcellulose medium M3234 (StemCell Technolo- G0/G1-phase populations (64% [mean], a 16% [95% confidence -to 21%] increase compared with vehicle 11% ס {gies, Vancouver, British Columbia, Canada) (16). Cells were interval {CI plated in six-well plates in 1 mL of medium and incubated at treated cells [48%]) and in the G2/M-phase populations (27%, a -to 28%] increase compared with vehicle 6% ס C in a humidified atmosphere containing 5% CO2/95% air. 17% [95% CI° 37 Colonies were counted after 2 weeks. Colony type was estab- treated cells [10%]). A concomitant decrease was observed in lished by morphology. To ensure accurate determination, repre- the proportion of cells in the S-phase population (9%, a 33% -to 44%] decrease compared with vehicle 22% ס sentative colonies were plucked from the plates, centrifuged [95% CI onto slides, stained with Wright–Giemsa stain, and examined by treated cells [42%]) (Fig. 2, A). light microscopy. The number of granulocyte colonies, macro- To explore the mechanism of action of paricalcitol, we ex- phage colonies, and mixed granulocyte/macrophage colonies amined the expression of cyclin-dependent kinase inhibitors was counted. (CDKIs) p21WAF1 and p27KIP1. Increased expression of WAF1 KIP1 Polymerase Chain Reaction Analysis p21 and p27 was observed in paricalcitol-treated cells and in 1,25(OH)2D3-treated cells but not in vehicle-treated cells. RNA extraction and reverse transcription were done with –7 Both paricalcitol and 1,25(OH)2D3 (each at1×10 M for 72 TRIzol (Invitrogen, Carlsbad, CA) and reverse transcriptase hours) increased the expression of p21WAF1 by approximately (Promega, Madison, WI); 20 ␮L of cDNA was prepared from 1 sevenfold and p27KIP1 by approximately sixfold in the HL-60 ␮g of RNA. cDNAs were amplified by polymerase chain reac- cells (Fig. 2, B). Thus, paricalcitol increased the expression of tion (PCR) with specific primers for 24-hydroxylase and 18S CDKIs, and this result may reflect the cell cycle arrest induced rRNA, with 25 PCR cycles for 18S rRNA and 32 cycles for by paricalcitol. 24-hydroxylase. The PCR primers used to amplify 24- Several analogs of vitamin D have been shown to increase the hydroxylase were 5Ј-GCTTACGCCGAGTGTACCAT-3Ј (for- expression of the potential tumor suppressor gene PTEN in a ward) and 5Ј-ATGAGCACTGTTCCTTTGGG-3Ј (reverse), and leukemic cell line, indicating that this protein may help to me- those used to amplify 18S rRNA were 5Ј-AAACGGCTACCA diate the antitumor activity of vitamin D and its analogs (12). CATCCAAG-3Ј (forward) and 5Ј-CCTCCAATGGATCCTC PTEN is a phosphatase that targets activated phosphatidylinosi- GTTA-3Ј (reverse). Annealing temperatures for the PCRs were tol 3-kinase (PI3K). We measured the levels of PTEN by west- 56 °C for 24-hydroxylase and 58 °C for 18S rRNA. PCR prod- ern blot analysis. After treatment with1×10–7 M paricalcitol for ucts were separated on a 2% agarose gel, stained with ethidium 72 hours, PTEN expression increased sevenfold in HL-60 cells bromide, and photographed. and 25-fold in NB-4 myeloid leukemia cells, compared with the increase in levels in vehicle-treated cells, essentially as observed Statistical Analysis –7 when these cells were treated with1×10 M 1,25(OH)2D3 for For in vivo studies in mice, the differences of tumor sizes and 72 hours (Fig. 2, B). weights between mice in the control group and mice in the Because vitamin D and its analogs induce monocytic differ- experimental group at the end of the study were analyzed by entiation of some myeloid leukemia cells as a result of their Student’s t test. These data met the assumptions for Student’s t antiproliferative activity (1,3,5), we examined the potency of test. All statistical tests were two-sided. paricalcitol to stimulate monocytic differentiation of HL-60

898 ARTICLES Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 Fig. 1. Paricalcitol and the clonal proliferation of human cancer cell lines in soft agar. Shown are dose–response curves from clonogenic assays of HL-60, NB-4, and THP-1 leukemia; HT-29 and SW837 colon cancer; and NCI-H929 myeloma cell lines cultured with either paricalcitol or

1,25(OH)2D3, as indicated. Colonies were counted after 14 days of culture. Results are the mean percentage of colonies relative to untreated Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 control cultures and 95% confidence intervals of three experiments, with each point representing the mean of triplicate experimental points.

cells. Paricalcitol induced monocyte/macrophage-like differen- in Fig. 3, A, these treatments also caused these cells to accumu- tiation of HL-60 cells, as measured by the increased expression late in the sub-G1 population, which is typical of cells undergo- of cell surface marker CD14. Paricalcitol (1 × 10–7 M for 96 ing apoptosis. Thus, the antiproliferative effect of paricalcitol in hours) induced 65% of HL-60 cells to express CD14, and NCI-H929 cells may be associated with induction of apoptosis. –7 1,25(OH)2D3 (1×10 M for 96 hours) induced 54% of HL-60 Apoptosis was next examined with the TUNEL assay. NCI- cells to express CD14 (Fig. 2, C). Morphologic examination H929 cells were treated with 1 × 10–7 M paricalcitol, 1 × 10–7 M clearly showed monocytic differentiation of HL-60 cells treated 1,25(OH)2D3, or vehicle for 96 hours and subjected to the 23% ס with1×10–7 M paricalcitol for 120 hours. Untreated control TUNEL assay (Fig. 3, B). Thirty-one percent (95% CI HL-60 cells are large with round or oval nuclei, prominent to 39%) of cells in paricalcitol-treated cultures and 20% (95% CI ס nucleoli, and amphophilic cytoplasm, whereas paricalcitol- 16% to 24%) of cells in 1,25(OH)2D3-treated cultures were treated cells showed monocytoid differentiation with oval, ir- apoptotic. The percentage of apoptotic cells in both cultures was regular, or indented nuclei and abundant vacuolated cytoplasm statistically significantly higher than that of vehicle-treated con- ,001. ס and P 003. ס to 9%; P 5% ס Fig. 2, D). trol cultures (7%, 95% CI) respectively). The differences between paricalcitol-treated cul- Paricalcitol and Human Myeloma Cells In Vitro tures and control cultures and between 1,25(OH)2D3-treated cul- ס Dose-dependent antiproliferative activity of paricalcitol was tures and control cultures were 24% (95% CI 16% to 30%) to 16%), respectively. Under the same 10% ס and 13% (95% CI higher in NCI-H929 myeloma cells than in RPMI-8226 and KIP1 ARH-77 myeloma cells (data not shown), and so the effect of conditions, the expression of p27 increased 3.4-fold or 2.8- fold, respectively, in paricalcitol-treated or 1,25(OH)2D3-treated paricalcitol on the cell cycle of NCI-H929 myeloma cells was WAF1 examined by flow cytometry. NCI-H929 cells were treated with cells (Fig. 3, C), but the expression of p21 was essentially –7 –7 unaffected (data not shown). When re-treated with paricalcitol or 1×10 M paricalcitol,1×10 M 1,25(OH)2D3, or vehicle for –7 72 hours, and cell cycle analysis was performed. In paricalcitol- 1,25(OH)2D3 (each at 10 M for 72 hours), NCI-H929 cells expressed approximately 40% less antiapoptotic BCL-2 protein treated cultures, 58% of the cells were in the G1/G0-phase population and 21% of the cells were in the sub-G population. In but expressed approximately the same amount of the proapop- 1 totic protein BAX (Fig. 3, C). 1,25(OH)2D3-treated cultures, 62% of the cells were in the G1/ G0-phase population and 16% of cells were in the sub-G1 popu- Paricalcitol and Colon Cancer Cells In Vitro and In Vivo lation. In vehicle-treated cultures, 52% of the cells were in the G1/G0-phase population and 4% of the cells were in the sub-G1 We next investigated the antiproliferative activity of parical- population (Fig. 3, A). citol on various colon cancer cell lines. Cells were incubated The morphologic examination showed increased changes with1×10–7 M paricalcitol or vehicle for 96 hours, and pro- typical of apoptosis, including nuclear shrinkage and cellular liferation was assessed with the MTT assay. When expressed as fragmentation, of the NCI-H929 cells treated with1×10–7 M a percentage of the vehicle control, proliferation of HT-29 cells ס 7– paricalcitol or1×10 M 1,25(OH)2D3 for 96 hours compared (50%, 95% CI 37% to 63%) and SW837 cells (70%, 95% CI to 76%) was more sensitive to paricalcitol than that of 64% ס with the morphology of control cells (data not shown). As shown

Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 ARTICLES 899 Fig. 2. Paricalcitol and HL-60 leukemia cell line. A) Cell cycle analysis of HL-60 cells. HL-60 cellswereculturedwith1×10–7 M paricalcitol or vehicle alone (control) for 72 hours, harvested, stained with propidium iodine, and examined by flow cytometry. Results are the mean ± 95% confidence intervals of three experiments, showing the percentage of total cells in the G0/G1-phase, G2/M-phase, and S-phase populations. B) West- ern blot analysis of HL-60 and NB-4 cells. Cells were treated with1×10–7 M paricalcitol,1×10–7

M 1,25(OH)2D3, or vehicle alone (control); cell lysates were harvested 72 hours later and examined by western blot analysis with antibodies against p21WAF1, p27KIP1, PTEN, and glyceraldehyde-3-phosphate dehydrogenase (GAPDH). Levels of protein, normalized to the level of Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 GAPDH, are presented in the text. C) Analysis of CD14 in HL-60 cells. Cells were cultured with 1 –7 –7 ×10 M paricalcitol,1×10 M 1,25(OH)2D3, or vehicle control for 96 hours and analyzed for expression of CD14 with flow cytometry by the use of fluorescein isothiocyanate-conjugated CD14 antibody. The x-axis shows the intensity of fluorescence and the y-axis shows the number of negative control ס cells. Dashed line/open peak CD14 antibody. CD14- ס antibody; solid peak positive cells are in the area shown by the bar M2. The area depicted by the other bar shows the CD14-negative cells. D) Morphologic analysis of HL-60 cells. Cells were cultured with1×10–7 M paricalcitol or vehicle control for 120 hours, fixed, and stained with Wright–Giemsa stain. Left) HL-60 control cells. Right) Paricalcitol- .␮m 5 ס treated HL-60 cells. Scale bar

Fig. 3. Paricalcitol and NCI-H929 myeloma cell line. A) Cell cycle analysis of NCI-H929 cells by flow cytometry. NCI-H929 cells were cultured with1×10–7 M paricalcitol,1×10–7 M

1,25(OH)2D3, or vehicle control for 72 hours, harvested, and stained with propidium iodine. Cell cycle analysis was performed by flow cytometry. The percentage of total cells is shown the G0/G1-phase, G2/M-phase, and S-phase populations. The x-axis and y-axis show DNA content and cell number, respectively. B) Quantitative analysis of apoptosis with the terminal deoxy- nucleotidyl transferase-mediated uridine 5Ј- triphosphate nick end labeling (TUNEL) assay. NCI-H929 cells were exposed to1×10–7 M pari- –7 calcitol,1×10 M 1,25(OH)2D3, or vehicle control for 96 hours and analyzed. Results are the mean ± 95% confidence intervals of four experiments. C) Western blot analysis of NCI-H929 cells. Cells were treated with1×10–7 M parical- –7 citol,1×10 M 1,25(OH)2D3, or vehicle control, and cell lysates were made 72 hours later and used for western blot analysis with antibodies against p27KIP1, BCL-2, and BAX, sequen- tially. Levels of protein, normalized to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH), are presented in the text. to 93%), and HCT116 cells further studied in HT-29 cells, the colon cancer cell line most 77% ס SW480 cells (85%, 95% CI × to 109%) sensitive to paricalcitol. HT-29 cells were incubated with 1 81% ס were resistant to paricalcitol (95%, 95% CI –7 –7 (Fig. 4, A). The mechanism of this antiproliferative effect was 10 M paricalcitol,1×10 M 1,25(OH)2D3, or the vehicle

900 ARTICLES Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 Fig. 4. Paricalcitol and colon cancer cell lines. A) HT-29, SW837, SW480, and HCT116 colon cancer cells were treated for 96 hours with1×10–7 M pari- –7 calcitol,1×10 M 1,25(OH)2D3, or vehicle control. Cell proliferation was measured by the 3-(4,5- dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay and expressed as percentage of control. Optical density was read at 540 nm. Results are the mean ± 95% confidence intervals of more than three experiments with triplicate dishes. B) HT-29 cells wereexposedto1×10–7 M paricalcitol,1×10–7 M

1,25(OH)2D3, or vehicle control, and cell lysates were made after 72 hours of culture for western blot analysis with antibodies against p27KIP1, p21WAF1, cyclin

D1, c-myc, and E-cadherin. C) Western blot analysis Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 for cyclooxygenase 1 (COX-1) and COX-2. HT-29 and SW837 cells were cultured with1×10–7 M pari- –7 calcitol, 1 × 10 M 1,25(OH)2D3, or vehicle control for 72 hours, and cell lysates were made and used for western blot analysis with antibodies against COX-1 and COX-2. In panels B and C, levels of protein presented in the text were normalized to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH).

control for 72 hours, and p21WAF1 and p27KIP1 proteins were measured. Levels of p21WAF1 and p27KIP1 protein in paricalcitol-treated cells and 1,25(OH)2D3-treated cells increased 5.2- fold and 5.8-fold, respectively, relative to control cells (Fig. 4, B). At the same time, expression of cyclin D1 and MYC protein decreased approximately 0.5-fold and 0.4-fold, respectively, after treatment with either paricalcitol or 1,25(OH)2D3. Expression of the cell adhesion protein E-cadherin, which is also associated with differentiation, increased 6.1-fold or 6.4-fold, respectively, in cells treated with paricalcitol or 1,25(OH)2D3 relative to control cells. COX-1 and COX-2 participate in the conversion of arachidonic acid to prostaglandins. Elevated expression of COX-2 has been detected in a variety of malignancies including colon cancer (17,18) and has become a target for the chemoprevention of such cancers. HT-29 and SW837 colon cancer cells were treated –7 –7 Fig. 5. Paricalcitol and HT-29 colon cancer cell xenografts in nude mice. HT-29 with1×10 M paricalcitol,1×10 M 1,25(OH)2D3, or vehicle cells were injected bilaterally and subcutaneously into nude mice to form two for 72 hours. In paricalcitol-treated cells, the expression of tumors per mouse. Tumor volume and tumor weight for each mouse was cal- COX-2 decreased 0.4-fold relative to that in vehicle-treated con- culated as a total of both tumors. Mice were divided randomly into control (five animals) and experimental (five animals) groups. Paricalcitol (100 ng per mouse) trol cells (Fig. 4, C), and in 1,25(OH)2D3-treated cells, levels of COX-2 decreased 0.5-fold. The expression of COX-1, relative to was administered intraperitoneally for 3 days per week in the experimental that in vehicle-treated control cells, did not change in cells sub- groups. A) Sequential measurements of tumor volume. Tumor volume was measured every week. The combined mean volume and 95% confidence interval of jected to either treatment. each group of tumors is shown. Tumor volumes were statistically significantly We then evaluated the effect of paricalcitol on HT-29 human different between the experimental and control groups after 4 weeks of treatment B) Tumor weights at autopsy. After 4 weeks of therapy, tumors were .(03. ס colon cancer xenografts growing in nude mice. Beginning the (P day after cells were injected into the mice, paricalcitol was in- removed from all mice and weighed. Weights of tumors (shown as mean and jected intraperitoneally 3 times per week on Monday, Wednes- 95% confidence interval) were statistically significantly different between the day, and Friday. Tumor volumes were measured weekly, and the two groups (P<.001). All statistical tests were two-sided. volumes from both bilateral tumors were added to give the total volume for the mouse. All mice were killed during the 5th week. to 3248 2103 ס mg, 95% CI 2675 ס Tumors were dissected and weighed. Those in paricalcitol- respectively; difference treated mice compared with those in vehicle-treated mice were mg; P<.001) (Fig. 5, B). Serum calcium levels were 9.5 mg/dL statistically significantly smaller (1044 mm3 and 1752 mm3, in control mice and 10.2 mg/dL in paricalcitol-treated mice (dif- ס to 1.40 mg/dL; P 0.16– ס mg/dL, 95% CI 0.62 ס to 1104 ference 311 ס mm3, 95% CI 708 ס respectively; difference .Fig. 5, A) and lighter (1487 mg and 4162 mg, .17), both in the normal range) (03. ס mm3; P

Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 ARTICLES 901 Paricalcitol and the VDR control cultures and 66 in paricalcitol-treated cultures; the total number of colonies from murine bone marrow cells isolated We examined the association between the expression of VDR from VDR knockout mice averaged 110 in untreated control and the antiproliferative effect of paricalcitol. All colon cancer cultures and 122 in paricalcitol-treated cultures. Thus, mono- cell lines examined—HT-29, SW837, SW480, and HCT116— cytic differentiation of committed myeloid stem cells appears to expressed VDR protein (Fig. 6, A). No association was observed require the VDR (Fig. 6, D). Myeloid stem cells from wild-type between the level of VDR protein in the various cell lines and mice were tested in the soft gel assay to determine the distribu- their sensitivity to paricalcitol, as determined by the MTT assay tion of differentiated colonies. Paricalcitol treatment increased ס Fig. 4, A). the percentage of macrophage colonies from 32% (95% CI) The enzyme 25-hydroxyvitamin D3-24-hydroxylase catalyzes 22% to 43%) in vehicle-treated control cultures to 69% (95% CI to 76%) in paricalcitol-treated cultures, decreased the 63% ס the first step in the catabolism of 1,25(OH)2D3 (19). Expression to 25% ס of 24-hydroxylase is transcriptionally regulated and is activated percentage of mixed colonies from 43% (95% CI to 38%) in 13% ס by the binding of its ligand, 1,25(OH)2D3, or its analog to the 62%) in control cultures to 25% (95% CI VDR. The VDR–ligand complex then binds to the vitamin D paricalcitol-treated cultures, and decreased the percentage of response element in the 24-hydroxylase promoter and activates to 33%) in Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 14% ס granulocyte colonies from 23% (95% CI ס its transcription (20,21). When HT-29 cells were incubated with –7 control cultures to 6% (95% CI –3% to 15%) in paricalcitol- 1×10 M paricalcitol, expression of 24-hydroxylase mRNA treated cultures. When myeloid stem cells from the knockout was induced within 6 hours and continued to increase for at least mice were tested in the soft gel assay, paricalcitol did not alter 24 hours (Fig. 6, B). the percentage of macrophage colonies in control cultures (30%, ,to 52%) and in paricalcitol-treated cultures (25% 8% ס We next investigated whether paricalcitol activity required an 95% CI /to 38%), the percentage of mixed granulocyte 13% ס intact VDR by using isolated mononuclear cells from the spleen 95% CI ס of a wild-type mouse or a VDR knockout mouse and culturing –8 macrophage colonies in control cultures (56%, 95% CI 29% ס these cells with1×10 M paricalcitol. Within 12 hours, 24- to 84%) and in paricalcitol-treated cultures (61%, 95% CI hydroxylase mRNA was expressed in paricalcitol-treated mono- 41% to 81%), or the percentage of granulocyte colonies in con- -to 21%) and in paricalcitol 6% ס nuclear cells from the wild-type mouse but not in paricalcitol- trol cultures (13%, 95% CI .(to 23% 4% ס treated cells from the VDR knockout mice (Fig. 6, C). treated cultures (13%, 95% CI We previously observed (16) that 1,25(OH)2D3 stimulated myeloid stem cells to differentiate via the macrophage pathway. To determine whether paricalcitol stimulated bone marrow cells DISCUSSION to differentiate via the same pathway, we added1×10–8 M paricalcitol to soft gel cultures of murine bone marrow cells We observed that the vitamin D analog paricalcitol inhibited from VDR knockout mice or wild-type mice and scored colonies proliferation of myeloid leukemia, myeloma, and colon cancer 2 weeks later as macrophage, granulocyte, or mixed granulo- cell lines in vitro by modulating cell cycle progression, differ- cyte/macrophage colonies. The total number of colonies from entiation, and apoptosis. Paricalcitol also inhibited the in vivo wild-type murine bone marrow cells averaged 87 in untreated growth of HT-29 human colon cancer xenografts in nude mice.

Fig. 6. Vitamin D receptor (VDR) and the activity of paricalcitol. A) VDR protein expression in cell lysates from HT-29, SW837, SW480, SW620, and HCT116 colon cancer cells was measured by western blot analysis. Levels of protein, normalized to that of glyceraldehyde-3-phosphate dehydrogenase (GAPDH), are presented in the text. B) 24- Hydroxylase mRNA expression in HT-29 colon cancer cells. Cells were treated with1×10–7 M paricalcitol for 0, 6, 12, or 24 hours. Expression of 24- hydroxylase mRNA was analyzed by reverse transcription–polymerase chain reaction. Levels of mRNA, normalized to that of 18S rRNA, are presented in the text. C) 24-Hydroxylase mRNA expression in mononuclear cells from spleens of wild-type and VDR knockout mice treated with1×10–8 M paricalcitol for 12 or 24 hours. Expression of 24- hydroxylase mRNA was analyzed by reverse transcription–polymerase chain reaction. Levels of mRNA, normalized to that of 18S rRNA, are presented in the text. D) Colony formation by mononuclear bone marrow cells from VDR knockout (VDR-KO) mice and wild-type mice. Mononuclear cells were obtained from femoral bone marrow plugs and grown in methylcellulose culture medium containing1×10–8 M paricalcitol or vehicle. Colonies were counted on day 10 of culture. The average numbers of total colonies were 87 (control) and 66 (1 × 10–8 M paricalcitol) in wild-type mice and 110 (control) and 122 (1 × 10–8 M paricalcitol) in VDR-KO mice. Data for each colony type are expressed as a percentage of the total number of colonies. We used triplicate wells for each mouse and studied a total of three knockout and ;mixed granulocyte/macrophage colonies ס granulocyte colonies; G/M ס three wild-type mice. Data are presented as the mean and 95% confidence interval. G .macrophage colonies ס M

902 ARTICLES Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 The antiproliferative activity of paricalcitol was accompanied by E-cadherin allows ␤-catenin to interact with cytoplasmic adeno- cell cycle arrest and changes in the expression of p21WAF1 and matous polyposis coli (APC), which helps to mediate the ubiq- KIP1 ␤ p27 . Previous studies (22–24) showed that vitamin D3 ana- uitination and degradation of -catenin. Mutation of the APC logs induce cell cycle arrest in G1/G0 phases; the arrest may also gene, which occurs frequently in the development of colon can- be mediated by p21WAF1 and p27KIP1. Paricalcitol arrested in cer, can result in ␤-catenin accumulating in the nucleus and

HL-60 leukemia cells in G1/G0 phases and G2/M phases, ar- acting as a co-stimulatory protein for the T-cell factor (TCF) rested NCI-H929 myeloma cells in G1/G0 phases, and induced family of transcription factors. Activation of these transcrip- the expression of p21WAF1 and p27KIP1 in leukemia, myeloma, tional factors increases the expression of a number of pro-growth and colon cancer cell lines. Thus, the CDKIs may play a role in genes, including cyclin D1 and c-myc (52–56). the antiproliferative effects of paricalcitol and other vitamin D A recent study (57) suggested that ligand-activated VDR analogs by reducing the ability of the tumor cells to enter S competes with TCF-4 for binding to ␤-catenin in the nucleus; ␤ phase (25,26). The block at the G2/M-phase checkpoint has been this competition releases -catenin so that it can translocate previously reported in HL-60 cells treated with 1,25(OH)2D3 from the nucleus to the E-cadherin complex at the plasma mem- (24). The vitamin D3-mediated retardation of cells at the G2/M- brane and thus inhibit the transcriptional regulatory activity of phase checkpoint that is accompanied with decreased levels of TCF. We found that paricalcitol-treated HT-29 cells had in- Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 p34(cdc) has been reported (27); we have not yet examined the creased levels of E-cadherin and decreased levels of cyclin D1 expression of p34(cdc) in paricalcitol-treated cells. and MYC, the latter two being subject to TCF/␤-catenin activa- Paricalcitol induced the expression of several tumor suppres- tion in such cells. Our results are consistent with the anticancer sor genes including PTEN and E-cadherin. The PTEN phospha- effects of paricalcitol being associated with its modulation of the tase can block PI3K/Akt signaling pathways, which contribute to E-cadherin/␤-catenin/TCF pathway. Future studies should ex- both cell death and the inhibition of cell proliferation (28,29). amine the levels and cellular location of ␤-catenin in these pari- PTEN mutations have been found in many human cancers (30– calcitol-treated cells. 36). In mice, homozygous germline deletion of PTEN resulted in Paricalcitol-induced apoptosis in NCI-H929 myeloma cells early embryonic lethality, but heterozygous germline deletion of was accompanied by the decreased expression of BCL-2 protein PTEN was associated with an increased incidence of malignant without an alteration in the level of BAX protein. In another neoplasms. Thus, PTEN appears to behave like a tumor suppres- study (58), treatment with the vitamin D3 analog EB1089 inhib- sor gene by depressing the pro-growth signals of the PI3K/Akt ited the proliferation of NCI-H929 cells, decreased the expres- pathway (37,38). We previously noted that 1,25(OH)2D3 and sion of BCL-2, increased caspase 3 activity and p38 kinase one of its analogs, 21-[3-methyl-3-hydroxy-butyl]-19-nor vita- activity, and suppressed p44 extracellular signal-related kinase min D3, increased the expression of PTEN in HL-60 cells (12). activity during apoptosis. The effect of paricalcitol on p38 and In this study, we showed that paricalcitol and 1,25(OH)2D3 in- extracellular signal-related kinase was not investigated. duced PTEN expression in HL-60 and NB-4 myeloid leukemia Epidemiologic studies suggest that the use of nonsteroidal cells. anti-inflammatory drugs decreases the risk of developing certain It remains to be determined whether paricalcitol or other vi- malignancies, including colon cancer (59,60). The major targets tamin D compounds can inhibit growth of cells with PTEN of these drugs are COX-1 and COX-2, which participate in the deletions. We suspect that these compounds may do so. Previ- conversion of arachidonic acid to prostaglandins (61). COX-1 is ously, an intragenic deletion including MMAC1/PTEN exons expressed more or less ubiquitously in the body and has many 2–5 in the myeloblastic leukemia cell line HL-60 and a four- physiologic activities, including maintenance of the gastrointes- nucleotide insertion in exon 5 in the acute monocytic leukemia tinal mucosa and various renal and platelet functions (62,63). In cell line U937 were identified (39). Vitamin D compounds can contrast, COX-2 is induced by various inflammatory stimuli, induce these cells to undergo terminal differentiation. Further- including cytokines, growth factors, and carcinogens, and may more, PTEN mutations are present in prostate cancer cell lines promote the growth of cancerous and precancerous cells (64). (30), and 1,25(OH)2D3 can inhibit the growth and induce the COX-2 expression is elevated in various malignancies (17,18) differentiation of prostate cancer cells (40–42). Low levels of and is, therefore, a reasonable target for cancer chemopreven- PTEN are detected in cells with methylated PTEN promoter tion. Selective COX-2 inhibitors suppress carcinogenesis in ro- regions. Because low levels of PTEN are expressed by some dent models, germline disruption of the COX-2 gene inhibits tumor cells, including endometrial, breast, colon, and prostate polyp formation in mice with a genetic predilection to develop cancer cell lines (30,43,44), and several vitamin D analogs that these tumors (65), and a selective COX-2 inhibitor reduces the slow the growth of these cells, it should be determined whether polyp burden in patients with familial adenomatous polyposis the antiproliferative activity of paricalcitol in these cells is as- (66). Furthermore, the number of polyps decreased after sociated with the concomitant demethylation of the PTEN pro- 1,25(OH)2D3 or its analog was given to mice with the same moter and the increased expression of PTEN. germline alternation as found in the patients with familial ade- Paricalcitol-treated colon cancer cells had increased expres- nomatous polyposis (67,68). sion of E-cadherin, a transmembrane linker protein located in These reports prompted us to investigate the effect of pari- intercellular adherent junctions, which maintain the adhesive calcitol on the expression of COX-2 in colon cancer cells. In and polarized phenotype of epithelial cells (45,46). Loss of E- HT-29 and SW837 colon cancer cells, treatment with paricalci- cadherin expression occurs as cells acquire the capacity to in- tol or 1,25(OH)2D3 decreased expression of COX-2, compared vade during the transition from adenoma to carcinoma (47,48). with vehicle-treated control cells, but not of COX-1, suggesting E-cadherin is a tumor suppressor gene, and its loss is associated that these vitamin D analogs acted as selective COX-2 inhibi- with poor prognosis (49–51). E-cadherin also regulates tors. The mechanism by which vitamin D inhibits COX-2 ex- ␤-catenin by holding it in place at the cell membrane. Loss of pression in colon cancer cells remains to be elucidated.

Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 ARTICLES 903 We also examined whether paricalcitol mediated its effects 1,25(OH)2–16-ene-vitamin D3 is a potent antileukemic agent with low through the VDR. We found no association between the level of potential to cause hypercalcemia. Leuk Res 1994;18:453–63. VDR expression in a colon cancer cell line and its sensitivity to (7) Anzano MA, Smith JM, Uskokovic MR, Peer CW, Mullen LT, Letterio JJ, paricalcitol (Fig. 6, A). This result was not surprising because et al. 1 alpha,25-Dihydroxy-16-ene-23-yne-26,27-hexafluorocholecalcif- erol (Ro24–5531), a new deltanoid (vitamin D analogue) for prevention of we (69) and others (70) have previously found little association breast cancer in the rat. Cancer Res 1994;54:1653–6. between overall cellular levels of VDR and responsiveness to (8) Pakkala S, de Vos S, Elstner E, Rude RK, Uskokovic M, Binderup L, et al. vitamin D3 analogs, and the activity of some vitamin D3 com- Vitamin D3 analogs: effect on leukemic clonal growth and differentiation, pounds may be mediated independently of the VDR through a and on serum calcium levels. Leuk Res 1995;19:65–72. nongenomic pathway. We did, however, show that the expres- (9) Koike M, Elstner E, Campbell MJ, Asou H, Uskokovic M, Tsuruoka N, et al. 19-nor-hexafluoride analogue of vitamin D3: a novel class of potent sion of 25-hydroxyvitamin D3-24-hydroxylase, a target gene of activated VDR, was induced in paricalcitol-treated cells. We inhibitors of proliferation of human breast cell lines. Cancer Res 1997;57: also showed that paricalcitol required the VDR to mediate mac- 4545–50. (10) Kubota T, Koshizuka K, Koike M, Uskokovic M, Miyoshi I, Koeffler HP. rophage differentiation of myeloid hematopoietic stem cells. 25- 19-nor-26,27-bishomo-vitamin D3 analogs: a unique class of potent inhibi- Hydroxyvitamin D3-24-hydroxylase was not induced by parical- tors of proliferation of prostate, breast, and hematopoietic cancer cells.

citol in the cells from VDR knockout mice but was induced by Cancer Res 1998;58:3370–5. Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 paricalcitol in cells from wild-type mice. Thus, the VDR appears (11) Hisatake J, Kubota T, Hisatake Y, Uskokovic M, Tomoyasu S, Koeffler to be necessary, but not sufficient, to ensure that cancer cells will HP. 5,6-trans-16-ene-vitamin D3: a new class of potent inhibitors of pro- be sensitive to paricalcitol. liferation of prostate, breast, and myeloid leukemic cells. Cancer Res 1999; 59:4023–9. In this study, we showed that paricalcitol and 1,25(OH)2D3 at similar concentrations had similar biologic activities in vitro. (12) Hisatake J, O’Kelly J, Uskokovic MR, Tomoyasu S, Koeffler HP. Novel vitamin D(3) analog, 21-(3-methyl-3-hydroxy-butyl)-19-nor D(3), that The serum half-life of both analogs is also similar (http:// modulates cell growth, differentiation, apoptosis, cell cycle, and induction www.fda.gov/cder/foi/). However, because paricalcitol is much of PTEN in leukemic cells. Blood 2001;97:2427–33. less likely to cause hypercalcemia in patients, physicians can (13) Llach F, Keshav G, Goldblat MV, Lindberg JS, Sadler R, Delmez J, et al. administer higher doses, resulting in higher serum concentra- Suppression of parathyroid hormone secretion in hemodialysis patients by ␮ tions. Thus, a typical dose of 1,25(OH)2D3 is 0.5–1.0 g every a novel vitamin D analogue: 19-nor-1,25-dihydroxyvitamin D2. Am J Kid- other day, for a peak serum level of 40–60 pg/mL, whereas a ney Dis 1998;32(2 Suppl 2):S48–54. typical dose of paricalcitol is 2.8–7.5 ␮g every other day (and up (14) Martin KJ, Gonzalez EA, Gellens M, Hamm LL, Abboud H, Lindberg J. to 16.8 ␮g has been safely given), for a peak serum level of 1850 19-Nor-1-alpha-25-dihydroxyvitamin D2 (Paricalcitol) safely and effec- tively reduces the levels of intact parathyroid hormone in patients on he- pg/mL. Therefore, paricalcitol can be given at higher doses than modialysis. J Am Soc Nephrol 1998;9:1427–32. 1,25(OH)2D3 to obtain higher serum levels without toxicity and (15) Chen TC, Schwartz GG, Burnstein KL, Lokeshwar BL, Holick MF. The in to result in greater efficiency in the treatment of vitamin D3- vitro evaluation of 25-hydroxyvitamin D3 and 19-nor-1alpha,25- sensitive cancers. dihydroxyvitamin D2 as therapeutic agents for prostate cancer. Clin Cancer In conclusion, paricalcitol has potent anticancer activity Res 2000;6:901–8. against myeloid leukemia and myeloma cells in vitro and colon (16) O’Kelly J, Hisatake J, Hisatake Y, Bishop J, Norman A, Koeffler HP. cancer cells in vitro and in vivo through modulation of the cell Normal myelopoiesis but abnormal T lymphocyte responses in vitamin D receptor knockout mice. J Clin Invest 2002;109:1091–9. cycle, differentiation, and apoptosis. Because paricalcitol has (17) van Rees BP, Ristimaki A. Cyclooxygenase-2 in carcinogenesis of the been approved by the Food and Drug Administration for the gastrointestinal tract. Scand J Gastroenterol 2001;36:897–903. treatment of secondary hyperparathyroidism, innovative clinical (18) Ristimaki A, Sivula A, Lundin J, Lundin M, Salminen T, Haglund C, et al. trials of this agent for preleukemia (myelodysplastic syndrome), Prognostic significance of elevated cyclooxygenase-2 expression in breast maintenance therapy for acute myeloid leukemia and multiple cancer. Cancer Res 2002;62:632–5. myeloma, and adjuvant therapy for colon cancer are reasonable (19) Jones G, Strugnell SA, DeLuca HF. Current understanding of the molecular to consider. actions of vitamin D. Physiol Rev 1998;78:1193–231. (20) Zierold C, Darwish HM, DeLuca HF. Identification of a vitamin D- response element in the rat calcidiol (25-hydroxyvitamin D3) 24- REFERENCES hydroxylase gene. Proc Natl Acad Sci U S A 1994;91:900–2. (21) Ohyama Y, Ozono K, Uchida M, Shinki T, Kato S, Suda T, et al. Identi- (1) Koeffler HP, Hirji K, Itri L. 1,25-Dihydroxyvitamin D3: in vivo and in fication of a vitamin D-responsive element in the 5Ј-flanking region of the vitro effects on human preleukemic and leukemic cells. Cancer Treat Rep rat 25-hydroxyvitamin D3 24-hydroxylase gene. J Biol Chem 1994;269: 1985;69:1399–407. 10545–50. (2) Staquet MJ, Byar DP, Green SB, Rozencweig M. Clinical predictivity of (22) Wang QM, Jones JB, Studzinski GP. Cyclin-dependent kinase inhibitor transplantable tumor systems in the selection of new drugs for solid tumors: p27 as a mediator of the G1-S phase block induced by 1,25- reply to a commentary. Cancer Treat Rep 1985;69:1339–40. dihydroxyvitamin D3 in HL60 cells. Cancer Res 1996;56:264–7. (3) Koeffler HP, Amatruda T, Ikekawa N, Kobayashi Y, DeLuca HF. Induc- (23) Munker R, Kobayashi T, Elstner E, Norman AW, Uskokovic M, Zhang W, tion of macrophage differentiation of human normal and leukemic myeloid et al. A new series of vitamin D analogs is highly active for clonal inhi- stem cells by 1,25-dihydroxyvitamin D3 and its fluorinated analogues. bition, differentiation, and induction of WAF1 in myeloid leukemia. Blood Cancer Res 1984;44:5624–8. 1996;88:2201–9. (4) Abe J, Nakano T, Nishii Y, Matsumoto T, Ogata E, Ikeda K. A novel (24) Jiang H, Lin J, Su ZZ, Collart FR, Huberman E, Fisher PB. Induction of vitamin D3 analog, 22-oxa-1,25-dihydroxyvitamin D3, inhibits the growth differentiation in human promyelocytic HL-60 leukemia cells activates of human breast cancer in vitro and in vivo without causing hypercalcemia. p21, WAF1/CIP1, expression in the absence of p53. Oncogene 1994;9: Endocrinology 1991;129:832–7. 3397–406. (5) Zhou JY, Norman AW, Akashi M, Chen DL, Uskokovic MR, Aurrecoe- (25) Doglioni C, Pelosio P, Laurino L, Macri E, Meggiolaro E, Favretti F, et al. chea JM, et al. Development of a novel 1,25(OH)2-vitamin D3 analog with p21/WAF1/CIP1 expression in normal mucosa and in adenomas and ad- potent ability to induce HL-60 cell differentiation without modulating cal- enocarcinomas of the colon: its relationship with differentiation. J Pathol cium metabolism. Blood 1991;78:75–82. 1996;179:248–53. (6) Jung SJ, Lee YY, Pakkala S, de Vos S, Elstner E, Norman AW, et al. (26) Yang ZY, Perkins ND, Ohno T, Nabel EG, Nabel GJ. The p21 cyclin-

904 ARTICLES Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 dependent kinase inhibitor suppresses tumorigenicity in vivo. Nat Med (51) Yoshida R, Kimura N, Harada Y, Ohuchi N. The loss of E-cadherin, alpha- 1995;1:1052–6. and beta-catenin expression is associated with metastasis and poor prog- (27) Harrison LE, Wang QM, Studzinski GP. 1,25-dihydroxyvitamin D(3)- nosis in invasive breast cancer. Int J Oncol 2001;18:513–20. induced retardation of the G(2)/M traverse is associated with decreased (52) Polakis P. Wnt signaling and cancer. Genes Dev 2000;14:1837–51. levels of p34(cdc2) in HL60 cells. J Cell Biochem 1999;75:226–34. (53) Porfiri E, Rubinfeld B, Albert I, Hovanes K, Waterman M, Polakis P. (28) Cantley LC, Neel BG. New insights into tumor suppression: PTEN sup- Induction of a beta-catenin-LEF-1 complex by wnt-1 and transforming presses tumor formation by restraining the phosphoinositide 3-kinase/AKT mutants of beta-catenin. Oncogene 1997;15:2833–9. pathway. Proc Natl Acad Sci U S A 1999;96:4240–5. (54) Hovanes K, Li TW, Munguia JE, Truong T, Milovanovic T, Lawrence (29) Di Cristofano A, Pandolfi PP. The multiple roles of PTEN in tumor sup- Marsh J, et al. Beta-catenin-sensitive isoforms of lymphoid enhancer fac- pression. Cell 2000;100:387–90. tor-1 are selectively expressed in colon cancer. Nat Genet 2001;28:53–7. (30) Li J, Yen C, Liaw D, Podsypanina K, Bose S, Wang SI, et al. PTEN, a (55) Tetsu O, McCormick F. Beta-catenin regulates expression of cyclin D1 in putative protein tyrosine phosphatase gene mutated in human brain, breast, colon carcinoma cells. Nature 1999;398:422–6. and prostate cancer. Science 1997;275:1943–7. (56) He TC, Sparks AB, Rago C, Hermeking H, Zawel L, da Costa LT, et al. (31) Teng DH, Hu R, Lin H, Davis T, Iliev D, Frye C, et al. MMAC1/PTEN Identification of c-MYC as a target of the APC pathway. Science 1998; mutations in primary tumor specimens and tumor cell lines. Cancer Res 281:1509–12. 1997;57:5221–5. (57) Palmer HG, Gonzalez-Sancho JM, Espada J, Berciano MT, Puig I, Baulida J, et al. Vitamin D(3) promotes the differentiation of colon carcinoma cells (32) Obata K, Morland SJ, Watson RH, Hitchcock A, Chenevix-Trench G, Downloaded from https://academic.oup.com/jnci/article/95/12/896/2520282 by guest on 30 September 2021 by the induction of E-cadherin and the inhibition of beta-catenin signaling. Thomas EJ, et al. Frequent PTEN/MMAC mutations in endometrioid but J Cell Biol 2001;154:369–87. not serous or mucinous epithelial ovarian tumors. Cancer Res 1998;58: (58) Park WH, Seol JG, Kim ES, Hyun JM, Jung CW, Lee CC, et al. Induction 2095–7. of apoptosis by vitamin D3 analogue EB1089 in NCI-H929 myeloma cells (33) Sakai A, Thieblemont C, Wellmann A, Jaffe ES, Raffeld M. PTEN gene via activation of caspase 3 and p38 MAP kinase. Br J Haematol 2000;109: alterations in lymphoid neoplasms. Blood 1998;92:3410–5. 576–83. (34) Vlietstra RJ, van Alewijk DC, Hermans KG, van Steenbrugge GJ, Trapman (59) Fosslien E. Biochemistry of cyclooxygenase (COX)-2 inhibitors and mo- J. Frequent inactivation of PTEN in prostate cancer cell lines and xeno- lecular pathology of COX-2 in neoplasia. Crit Rev Clin Lab Sci 2000;37: grafts. Cancer Res 1998;58:2720–3. 431–502. (35) Dahia PL, Aguiar RC, Alberta J, Kum JB, Caron S, Sill H, et al. PTEN is (60) Gupta RA, Dubois RN. Colorectal cancer prevention and treatment by inversely correlated with the cell survival factor Akt/PKB and is inactivated inhibition of cyclooxygenase-2. Nat Rev Cancer 2001;1:11–21. via multiple mechanisms in haematological malignancies. Hum Mol Genet (61) Seibert K, Zhang Y, Leahy K, Hauser S, Masferrer J, Isakson P. Distribu- 1999;8:185–93. tion of COX-1 and COX-2 in normal and inflamed tissues. Adv Exp Med (36) Liu TC, Lin PM, Chang JG, Lee JP, Chen TP, Lin SF. Mutation analysis Biol 1997;400A:167–70. of PTEN/MMAC1 in acute myeloid leukemia. Am J Hematol 2000;63:170–5. (62) Seibert K, Masferrer J, Zhang Y, Leahy K, Hauser S, Gierse J, et al. (37) Di Cristofano A, Pesce B, Cordon-Cardo C, Pandolfi PP. PTEN is essential Expression and selective inhibition of constitutive and inducible forms of for embryonic development and tumour suppression. Nat Genet 1998;19: cyclooxygenase. Adv Prostaglandin Thromboxane Leukot Res 1995;23: 348–55. 125–7. (38) Suzuki A, de la Pompa JL, Stambolic V, Elia AJ, Sasaki T, del Barco, et (63) Smith CJ, Zhang Y, Koboldt CM, Muhammad J, Zweifel BS, Shaffer A, et al. High cancer susceptibility and embryonic lethality associated with mu- al. Pharmacological analysis of cyclooxygenase-1 in inflammation. Proc tation of the PTEN tumor suppressor gene in mice. Curr Biol 1998;8:1169–78. Natl Acad Sci U S A 1998;95:13313–8. (39) Aggerholm A, Gronbaek K, Guldberg P, Hokland P. Mutational analysis of (64) Williams CS, Mann M, DuBois RN. The role of cyclooxygenases in in- the tumour suppressor gene MMAC1/PTEN in malignant myeloid disor- flammation, cancer, and development. Oncogene 1999;18:7908–16. ders. Eur J Haematol 2000;65:109–13. (65) Oshima M, Dinchuk JE, Kargman SL, Oshima H, Hancock B, Kwong E, (40) Skowronski RJ, Peehl DM, Feldman D. Actions of vitamin D3 analogs on et al. Suppression of intestinal polyposis in Apc delta716 knockout mice by human prostate cancer cell lines: comparison with 1,25-dihydroxyvitamin inhibition of cyclooxygenase 2 (COX-2). Cell 1996;87:803–9. D3. Endocrinology 1995;136:20–6. (66) Steinbach G, Lynch PM, Phillips RK, Wallace MH, Hawk E, Gordon GB, (41) Danielpour D, Kadomatsu K, Anzano MA, Smith JM, Sporn MB. Devel- et al. The effect of celecoxib, a cyclooxygenase-2 inhibitor, in familial opment and characterization of nontumorigenic and tumorigenic epithelial adenomatous polyposis. N Engl J Med 2000;342:1946–52. cell lines from rat dorsal-lateral prostate. Cancer Res 1994;54:3413–21. (67) Huerta S, Irwin RW, Heber D, Go VL, Koeffler HP, Uskokovic MR, et al. (42) Campbell MJ, Reddy GS, Koeffler HP. Vitamin D3 analogs and their 1alpha,25-(OH)(2)-D(3) and its synthetic analogue decrease tumor load in 24-oxo metabolites equally inhibit clonal proliferation of a variety of can- the Apc(min) mouse. Cancer Res 2002;62:741–6. cer cells but have differing molecular effects. J Cell Biochem 1997;66:413–25. (68) Bilger A, Shoemaker AR, Gould KA, Dove WF. Manipulation of the (43) Guanti G, Resta N, Simone C, Cariola F, Demma I, Fiorente P, et al. mouse germline in the study of Min-induced neoplasia. Semin Cancer Biol Involvement of PTEN mutations in the genetic pathways of colorectal 1996;7:249–60. cancerogenesis. Hum Mol Genet 2000;9:283–7. (69) Koeffler HP, Bishop JE, Reichel H, Singer F, Nagler A, Tobler A, et al. (44) Tashiro H, Blazes MS, Wu R, Cho KR, Bose S, Wang SI, et al. Mutations Lymphocyte cell lines from vitamin D-dependent rickets type II show in PTEN are frequent in endometrial carcinoma but rare in other common functional defects in the 1 alpha,25-dihydroxyvitamin D3 receptor. Mol gynecological malignancies. Cancer Res 1997;57:3935–40. Cell Endocrinol 1990;70:1–11. (45) Takeichi M. Morphogenetic roles of classic cadherins. Curr Opin Cell Biol (70) Norman AW, Nemere I, Zhou LX, Bishop JE, Lowe KE, Maiyar AC, et al. 1995;7:619–27. 1,25(OH)2-vitamin D3, a steroid hormone that produces biologic effects (46) Gumbiner BM. Cell adhesion: the molecular basis of tissue architecture and via both genomic and nongenomic pathways. J Steroid Biochem Mol Biol morphogenesis. Cell 1996;84:345–57. 1992;41:231–40. (47) Perl AK, Wilgenbus P, Dahl U, Semb H, Christofori G. A causal role for E-cadherin in the transition from adenoma to carcinoma. Nature 1998;392: NOTES 190–3. H. P. Koeffler was supported by Public Health Service grants 5R01 (48) Christofori G, Semb H. The role of the cell-adhesion molecule E-cadherin CA26038–25, 5P01 CA42710, and 5P01 AT00151–03 from the National Insti- as a tumour-suppressor gene. Trends Biochem Sci 1999;24:73–6. tutes of Health, Department of Health and Human Services, the Bruce Havner (49) Umbas R, Isaacs WB, Bringuier PP, Schaafsma HE, Karthaus HF, Ooster- Fund, and Abbott Laboratories. H. P. Koeffler holds the Mark Goodson En- hof GO, et al. Decreased E-cadherin expression is associated with poor dowed Chair of Oncology Research and is a member of the Jonsson Cancer prognosis in patients with prostate cancer. Cancer Res 1994;54:3929–33. Center and Molecular Biology Institute, University of California at Los Angeles. (50) Katagiri A, Watanabe R, Tomita Y. E-cadherin expression in renal cell We thank Scott Toner of Abbott Laboratories for helpful discussions. cancer and its significance in metastasis and survival. Br J Cancer 1995; Manuscript received December 16, 2002; revised March 31, 2003; accepted 71:376–9. May 2, 2003.

Journal of the National Cancer Institute, Vol. 95, No. 12, June 18, 2003 ARTICLES 905