Cooperative antitumor effects of vitamin D3 derivatives and rosemary preparations in a mouse model of myeloid leukemia



1α,25-dihydroxyvitamin D3 (1,25D3) is a powerful differentiation agent, which has potential for treatment of myeloid leukemias and other types of cancer, but the calcemia produced by pharmacologically active doses precludes the use of this agent in the clinic. We have shown that carnosic acid, the major rosemary polyphenol, enhances the differentiating and antiproliferative effects of low concentrations of 1,25D3 in human myeloid leukemia cell lines (HL60, U937). Here we translated these findings to in vivo conditions using a syngeneic mouse leukemia tumor model. To this end, we first demonstrated that as in HL60 cells, differentiation of WEHI-3B D murine myelomonocytic leukemia cells induced by 1 nM 1,25D3 or its low-calcemic analog, 1,25-dihydroxy-16-ene-5,6-trans-cholecalciferol (Ro25-4020), can be synergistically potentiated by carnosic acid (10 μM) or the carnosic acid-rich ethanolic extract of rosemary leaves. This effect was accompanied by cell cycle arrest in G0+G1 phase and a marked inhibition of cell growth. In the in vivo studies, i.p. injections of 2 μg Ro25-4020 in Balb/c mice bearing WEHI-3B D tumors produced a significant delay in tumor appearance and reduction in tumor size, without significant toxicity. Another analog, 1,25-dihydroxy-16,23Z-diene-20-epi-26,27-hexafluoro-19-nor-cholecalciferol (Ro26-3884) administered at the same dose was less effective than Ro25-4020 and profoundly toxic. Importantly, combined treatment with 1% dry rosemary extract (mixed with food) and 1 μg Ro25-4020 resulted in a strong cooperative antitumor effect, without inducing hypercalcemia. These results indicate for the first time that a plant polyphenolic preparation and a vitamin D derivative can cooperate not only in inducing leukemia cell differentiation in vitro, but also in the antileukemic activity in vivo. These data may suggest novel protocols for chemoprevention or differentiation therapy of myeloid leukemia. © 2006 Wiley-Liss, Inc.

Acute myeloid leukemia (AML) is the most common acute leukemia in adults and is characterized by a block of terminal differentiation of the hematopoietic progenitors at early stages in myelopoiesis, followed by repression of normal hematopoiesis by the expanding immature blasts.1 Despite advances in our understanding of the pathophysiology of AML, current therapeutic approaches in treatment of this disease have not yet led to major improvements in disease-free and overall survival of AML patients. Only about one-third of those aged between 18 and 60 who are diagnosed with AML can be cured; disease-free survival among patients is poor and chemotherapy devastating, particularly in older adults.2, 3 Current chemotherapy protocols for AML commonly use drugs that interfere with DNA replication and induce apoptosis, primarily in replicating cells.2, 3 However, these regimens may not effectively discriminate between normal and malignant cells, thus causing substantial damage to normal tissues. For this reason, it is important to develop treatments that can more specifically target the leukemic cell population.

Differentiation therapy is an alternative or complementary approach to standard cytotoxic drug therapy of cancer, and aims at arresting the growth of malignant cells by inducing normalization of cellular phenotype. Several compounds are known to induce differentiation of neoplastic cells. For example, all-trans retinoic acid (ATRA), alone or combined with other drugs, has proven extremely valuable in the treatment of acute promyelocytic leukemias4 and potentially other myeloid leukemias.5, 6 Similarly, vitamin D derivatives (VDDs)—1,25-dihydroxyvitamin D3 (1,25D3) and its analogs—are currently under investigation as agents that induce differentiation, growth arrest and apoptosis in a variety of tumor types7, 8, 9, 10, 11 (see also refs.12 and13 for recent reviews). Myeloid leukemia cells have been found to be remarkably sensitive to 1,25D3 and its analogs.14, 15, 16, 17 Although a full understanding of the mechanisms of terminal differentiation of these cells is still not available, we, and others, demonstrated that the cyclin-dependent kinase inhibitors p27Kip1 and p21Cip1 have an important role in the G1 cell cycle arrest of vitamin D-treated leukemia cells.18, 19, 20 There is substantial epidemiological evidence that serum levels of 1,25D3 near the top of the physiological range are associated with a lower incidence and better prognosis of the common human cancers, which affect the female breast, the prostate, the colon and other organs, and may thus represent a form of naturally occurring chemoprevention.21, 22

The well-known limitation to the therapeutic use of 1,25D3 is its hypercalcemic effect. Attempts to overcome this problem have recently focused on the synthesis of vitamin D analogs, which retain the pro-differentiation activities but have lower calcemic effects.17, 23 Various Phase I/II trials have been conducted with 1,25D3 and some of its analogs.24, 25, 26, 27, 28 However, in contrast to their efficient induction of differentiation and growth arrest in cultured human leukemia cells14, 17 and other cancer cell lines,7, 9, 11, 17 these compounds have not as yet been successfully applied to the treatment of myeloid leukemias or other cancers.12, 29, 30 Thus, the alternative strategy for developing vitamin D-based therapy of cancer by enhancing its activity using other compounds appears to be particularly promising.13

A variety of plant antioxidants (e.g., dietary and herbal polyphenols and carotenoids) with anticancer activity have been identified in both epidemiologic studies and experimental model systems.31, 32, 33 Several of these phytochemicals have been shown to synergistically enhance the differentiating and antiproliferative effects of low concentrations of 1,25D3 in human myeloid leukemia cell lines.34, 35, 36, 37, 38 (see also ref.13 for a recent review). However, despite numerous in vitro data, the ability of VDDs and plant antioxidants to cooperate in in vivo tumor models has not as yet been demonstrated.

Recently, we have shown that carnosic acid (Fig. 1), the major polyphenolic antioxidant found in the plant Rosmarinus officinalis L. (rosemary), markedly enhances 1,25D3-induced differentiation and inhibits proliferation of HL60 and U937 myeloblastic cells in liquid culture.36, 38, 40, 41 In the present study, we translated these findings to the in vivo syngeneic murine myeloid leukemia model, using WEHI-3B D murine myelomonocytic leukemia cell line known to differentiate into monocyte/macrophage lineage in response to 1,25D3.16 Both carnosic acid and carnosic acid-rich ethanolic extract of rosemary leaves significantly potentiated the in vitro differentiating and antiproliferative effects of 1,25D3 and its low-calcemic analog, 1,25-dihydroxy-16-ene-5,6-trans-cholecalciferol (Ro25-4020), (see Fig. 1 for VDD structures) in WEHI-3B D cells. Furthermore, we showed for the first time that oral treatment with rosemary extract, which alone had a significant antitumorigenic effect, in combination with a low dose of i.p. Ro25-4020 resulted in a strong cooperative inhibition of WEHI-3B D tumor growth without toxicity. The results of our study suggest that the use of polyphenol-rich rosemary preparations together with low calcemic vitamin D3 analogs may represent a novel effective and low-toxic approach to combination differentiation therapy of acute myeloid leukemias.

Figure 1.

Structures of the vitamin D3 derivatives and the rosemary polyphenol (carnosic acid) used in this study. 1,25D3 is the parent vitamin D compound. Ro25-4020 is the synthetic analog (1,25-dihydroxy-16-ene-5,6-trans-cholecalciferol) with modification of the A-ring and introduction of C-16 double bond to prevent C-3 epimerization and C-24 hydroxylation initiated modifications of side chain, respectively. Ro26-3884 is the fluorinated synthetic analog (1,25-dihydroxy-16,23Z-diene-20-epi-26,27-hexafluoro-19-nor-cholecalciferol) with introduction of C-16 and C-23 double bonds, and replacement of hydrogens on C-26 and C-27 with fluorines to prevent C-24 and C-26 hydroxylation, respectively (see ref.39 for further details). Carnosic acid is the polyphenolic diterpene with antioxidant properties derived from the plant Rosmarinus officinalis L. (rosemary).

Material and methods


The vitamin D derivatives, 1,25D3, 1,25-dihydroxy-16-ene-5,6-trans-cholecalciferol (Ro25-4020)17, 42 and 1,25-dihydroxy-16,23Z-diene-20-epi-26,27-hexafluoro-19-nor-cholecalciferol (Ro26-3884), were prepared at BioXell, Inc. (Nutley, NJ). Carnosic acid was purchased from Alexis Biochemicals (Laufenfingen, Switzerland). Dried ethanolic extract of rosemary leaves (lot # LR-04-14-02) containing 30.62% (w/w) carnosic acid and 11.86% (w/w) carnosol was provided by LycoRed Natural Products Industries (Beer Sheva, Israel). Fluorescein isothiocyanate-conjugated anti-F4/80 antibody (CI:A3-1) was purchased from Serotec (Oxford, UK). Rhodamine-conjugated anti-CD14 (MY4-RD-1) and fluorescein isothiocyanate-conjugated anti-CD11b (MO1-FITC) antibodies were obtained from Beckman Coulter (Miami, FL). Stock solutions of carnosic acid (10 mM), rosemary extract (10 mM equivalent of carnosic acid), 1,25D3 (0.1–0.3 mM) and Ro25-4020 (0.1–0.3 mM) were prepared in absolute ethanol. The precise concentration of 1,25D3, Ro25-4020 and Ro26-3884 in stock solutions was verified spectrophotometrically at 264 (ε = 19,000), 272–273 (ε = 20,600) and 260 nm (ε = 25,000), respectively.43

Cell culture and proliferation assay

WEHI-3B D murine myelomonocytic leukemia cells were obtained from Dr. A. Rice (Yale University School of Medicine, New Haven, CT). This cell line is sensitive to 1,25D3 but is practically unresponsive to ATRA.16 WEHI-3B D cells were grown at 37°C in McCoy's 5a medium supplemented with heat-inactivated 10% fetal calf serum and 1% penicillin–streptomycin (Beit Haemek, Israel) in 95% air/5% CO2. HL60-G cells, a subclone of human promyelocytic leukemia HL60 cells,44 were cultured in RPMI 1640 medium (Mediatech, Washington, DC) supplemented with 1% glutamine and 10% heat-inactivated, iron-enriched bovine calf serum (HyClone, Logan, UT). Both cell lines were passaged 2–3 times weekly to maintain log phase growth. For all experiments, WEHI-3B D cells were seeded at 5 × 103 cells/ml in 6-well plates, and HL60-G cells were seeded at 2 × 105 cells/ml in 25 cm2 tissue culture flasks in fresh respective growth medium. Test compounds were added to the cultures immediately after cell seeding, followed by incubation for 48–96 h. Cell proliferation was estimated by counting cells with a Coulter counter after dilution in Isoton-II (Coulter Electronics, Hialeah, FL). Cell viability was determined by trypan blue (0.2%) exclusion assay.

Cell cycle distribution

Aliquots of 1 × 106 cells were washed twice with ice-cold PBS and the cells were fixed in 75% ethanol at −20°C for at least 24 h, washed twice with PBS and incubated in 1 ml of PBS containing 0.1% Triton X-100 and 50 μg of RNAse (Roche Molecular Biochemicals, Mannheim, Germany) at 37°C for 40 min. Propidium iodide (10 μg/ml, Sigma) was then added for 20 min and the cells were analyzed using a FACSCalibur flow cytometer (Becton Dickinson, Franklin Lakes, NJ). Cell cycle distribution was determined by a ModFit LT computer program (Verity Software House, Topsham, ME).

Analysis of apoptosis

Cell apoptosis was detected using MEBCYTO® Apoptosis Kit (MBL, Nagoya, Japan), according to manufacturer's recommended protocol with minor modifications. Briefly, 2 × 105 cells were harvested, washed once with PBS and resuspended in 40 μl of binding buffer. Four microliters of Annexin V-FITC and 2 μl of propidium iodide were then added to the resuspended cells. After incubation at room temperature for 15 min in the dark, 200 μl of binding buffer was added and the stained cells were analyzed by flow cytometry at 488 nm.

Superoxide anion generation assay

Superoxide production in WEHI-3B D cells was determined by the nitroblue tetrazolium (NBT) reduction test, essentially as described previously.45 Briefly, approximately 1 × 106 cells were harvested and resuspended in 1.0 ml complete McCoy's 5a medium containing 0.1% NBT and 2 μM 12-O-tetradecanoylphorbol 13-acetate (TPA, Sigma). The cell suspension was incubated at 37°C for 30 min, and the percentage of cells containing blue-black formazan deposits, indicative of a TPA-stimulated respiratory burst, was determined by counting of at least 200 cells, using a hemocytometer under light microscope.

Determination of differentiation markers

The expression of F4/80 (murine monocyte/macrophage marker46) in WEHI-3B D cells and the expression of CD14 (monocytic marker) and CD11b (general myeloid marker) in HL60-G cells was assessed by flow cytometry, as described previously.36 Aliquots of 1 × 106 cells were harvested, washed twice with PBS and then incubated for 45 min at room temperature with 5 μl (0.5 μg) anti-F4/80-FITC (for WEHI-3B D cells) or 0.5 μl MY4-RD-1 and 0.5 μl MO1-FITC, each containing 0.5 μg of the antibody, to detect the expression of surface cell markers CD14 and CD11b, respectively (for HL60-G cells). The cells were then washed 3 times with ice cold PBS and resuspended in 1 ml PBS. One parameter (F4/80) or two parameters (CD14/CD11b) analysis was performed using a FACSCalibur flow cytometer, with CELLQuest software (Becton-Dickinson, San Jose, CA). Monocytic differentiation of HL60-G cells was also monitored by cytochemical determination of the cytoplasmic enzyme monocyte-specific esterase (MSE) activity, as described previously.47

Experimental animals

Experiments were carried out in the animal facility of Soroka University Medical Center (Beer-Sheva, Israel), in accordance with the Ben-Gurion University Ethical Committee for Animal Research approved protocols. Six-to eight-week-old female Balb/c mice were purchased from Harlan Laboratories (Jerusalem, Israel). Mice were housed in sterilized cages and provided with autoclaved water ad libitum. Experimental protocols were initiated following a 7-day acclimatization period.

Experimental groups and diet

During acclimatization, mice were fed ad libitum a standard powdered rodent diet. Then according to the experimental protocols, the mice receiving i.p. injections of vitamin D3 derivatives alone and the control animals injected with vehicle (ethanol) were kept on the standard diet.43, 48 Animals allotted to the groups for oral administration of rosemary extract alone received the same diet supplemented with 1% (w/w) rosemary extract and ethanol injections. For combination treatment, mice were fed diet containing rosemary extract and injected with Ro25-4020.

Administration of vitamin D derivatives

Stock solutions of 1,25D3, Ro25-4020 and Ro26-3884 in ethanol were diluted in ice-cold sterile PBS. The vehicle contained PBS and appropriate concentration of ethanol. The final ethanol concentration in the injection volume did not exceed 0.1%. Vitamin D derivatives or vehicle were injected i.p. in 100 μl PBS, 3 times a week (Sunday, Tuesday and Thursday), according to the published protocol.15, 43, 48

Cell inoculation and tumorigenicity assay

Following acclimatization, mice were inoculated i.p. with 1 × 105 WEHI-3B D cells in 100 μl sterile PBS15 and randomly assigned to control and treatment groups (10 animals per group). Twenty-four hours later, the mice were started on vehicle, Ro25-4020, rosemary extract or combination treatment. Inoculation of cells resulted in the development of a well-palpable tumor located on the anterior abdominal wall near the site of the injection (see Fig. 7). Tumor incidence was assessed by palpation every 3 days. Tumor growth was determined by measuring the smallest and the largest tumor diameters with calipers, and tumor volume was calculated using the formula: volume (mm3) = length × width2/2, according to standard procedures.49, 50 At the end of the experiment, all mice were killed by the institutionally approved method (CO2 suffocation), tumors were excised and fixed in 10% buffered formalin for later morphological analysis of tumor sections and immunohistochemical determination of cell proliferative markers and apoptosis. To assess a possible development of systemic leukemia, blood samples were collected weekly from the retro-orbital sinus of each animal under anesthesia. The complete blood count and differential leukocyte count and morphology analysis were performed using standard procedures in the Hematology Laboratory, Soroka University Medical Center (Beer-Sheva, Israel).

Figure 7.

Effects of Ro25-4020, rosemary extract and their combination on tumor incidence (a) and growth (b). Mice were inoculated intraperitonially with 1 × 105 WEHI-3B D cells and subjected to treatments with Ro25-4020 (1 μg, i.p., 3 times per week), rosemary extract (RE, 1% w/w, mixed with food) alone and in combination. Tumor incidence (a) and volume (b) were evaluated 3 times a week, as described under Materials and Methods. (b) Note that on days 25–27, the antitumor effect of the combined treatment with Ro25-4020 and RE (Group 4; triangles) is at least additive, as it is similar to or greater than the calculated sum of the effects of single treatments (Group 2 + Group 3; diamonds). The data are the means ± SE of 2 experiments performed in groups of 10 animals. * (p < 0.05) and ** (p < 0.01): statistically significant differences from untreated control; § synergistic effect of the Ro25-4020/RE combination (p < 0.05).

Evaluation of toxicity

General toxicity was assessed by clinical measures, such as weight loss, changes in appearance and behavior, lethargy and death. Animal weight was evaluated 3 times a week. For serum calcium, as well as liver and kidney tests (serum albumin, total bilirubin, aspartate aminotransferase, alanine aminotransferase, and urea), blood samples were drawn from the retro-orbital sinus of 5 animals randomly selected from each treatment group. The tests were performed by routine automatic blood chemistry analysis in the Biochemistry Laboratory, Soroka University Medical Center (Beer-Sheva, Israel).

Statistical analysis

In vitro experiments were repeated at least 3 times. Dose-response curves for 1,25D3 and Ro25-4020 in the cell proliferation assay were generated by nonlinear regression analysis. In vivo experiments were repeated twice in groups of 10 mice, if not indicated otherwise. Two compounds (A and B) were considered to show synergy in the particular experiment if the effect of their combination (AB) was larger than the sum of their individual effects (AB > A + B), the data being compared after subtraction of the respective control values from A, B and AB.36 Statistically significant differences between AB and A + B were estimated with the use of the nonparametric Wilcoxon matched pairs test. Statistically significant differences among the multiple groups were tested using nonparametric Kruskal-Wallis one-way analysis-of-variance test, followed by Dunn's adjustment for individual groups versus control, as described previously.36 All statistical analyses were performed using a GraphPad Prism 3.0 program (GraphPad Software, San Diego, CA). p < 0.05 was considered statistically significant.


Carnosic acid and rosemary extract similarly enhance the antiproliferative and differentiating effects of vitamin D3 derivatives in murine and human leukemic cells

Our previous studies have shown that carnosic acid markedly enhances the in vitro antileukemic activity of low concentrations of 1,25D3 in human HL60 and U937 myeloid leukemia cells.36, 40, 41 The objective of the present study was to translate these findings to an animal model of AML. For this purpose, we chose the 1,25D3-responsive WEHI-3B D murine myelomonocytic leukemia cell line,51 which can induce leukemic tumor deposits when injected in syngeneic Balb/c mice.52 To demonstrate the suitability of this model to study the in vivo antileukemic effects of VDD/plant antioxidant combinations, we first examined whether these compounds would cooperate in inducing the in vitro WEHI-3B D cell differentiation and growth arrest. In the proliferation and differentiation assays, we tested both 1,25D3 and its low-calcemic analog (Ro25-4020), which may be a good candidate for the non-toxic differentiation therapy of AML. To study the enhancement of the VDDs' effects, 1,25D3 or Ro25-4020 at low concentrations were combined with either carnosic acid or its “natural” preparation, the ethanolic extract of rosemary leaves, which was intended for the oral treatment in vivo. Human HL60-G cells were used for comparison in the in vitro assays as the standard system in which carnosic acid is known to enhance 1,25D3 effects.36, 38, 41

Incubation of WEHI-3B D cells for 96 h with either 1,25D3 or Ro25-4020 alone resulted in a concentration-dependent reduction in cell number, concomitant with the induction of F4/80, a mouse monocyte/macrophage differentiation marker46 (Fig. 2). Analysis of dose-response curves indicated that 1,25D3 and Ro25-4020 were approximately equipotent in both proliferation (IC50 = 3.41 ± 0.31 and 3.65 ± 0.21 nM, respectively) and differentiation (EC50 = 0.76 ± 0.08 and 0.81 ± 0.06 nM, respectively) assays.

Figure 2.

1,25D3 and Ro25-4020 similarly inhibit proliferation and induce differentiation in WEHI-3B D cells. Cells were incubated for 96 h with or without increasing concentrations of 1,25D3 (open symbols) or Ro25-4020 (solid symbols). Cell number (circles) and the expression of F4/80 differentiation marker (triangles) were measured by a Coulter counter and flow cytometry, respectively, as described under Materials and Methods. The data are the means ± SE of 3 independent experiments performed in duplicate.

To compare the enhancing action of carnosic acid in VDD-treated WEHI-3B D and HL60-G cells, 1,25D3 or Ro25-4020 at a low concentration (1 nM) were incubated with cultures of these cells in the absence or presence of 10 μM carnosic acid for 96 h, followed by cell counting and differentiation tests. As shown in Figure 3a, carnosic acid alone produced a small inhibitory effect on WEHI-3B D cell proliferation; however, it strongly cooperated in this activity, with both 1,25D3 and Ro25-4020 in a synergistic manner. A similar potentiating effect of carnosic acid was observed in parallel experiments on HL60-G cells (Fig. 3b). The marked synergistic reduction in WEHI-3B D cell number by the combination of VDDs and carnosic acid was associated with cell cycle arrest in G0 + G1 phase (Fig. 3c) but not with cell death, as no significant cytotoxic effect of these combinations was observed in trypan blue exclusion assay. Furthermore, measurement of apoptotic markers (the appearance of cell surface Annexin V binding sites and sub-G1 cell population) did not reveal significant apoptotic cell death (data not shown). In contrast, parallel treatment with 10 μM arsenic trioxide, a known antileukemic agent,53 used in this assay as the positive control, produced massive apoptosis and necrosis in WEHI-3B D cells, following a 24 h incubation (˜90% of dead cells).

Figure 3.

Carnosic acid and rosemary extract potentiate the antiproliferative effects of 1,25D3 and Ro25-4020 in murine and human myeloid leukemia cells. WEHI-3B D (a) and HL60-G (b) cells were incubated for 96 h with the indicated compounds. Cell number was determined by cell counting, as described under Materials and Methods. (c) WEHI-3B D cells were treated with the indicated compounds for 48 h, followed by flow cytometric analysis of cell cycle distribution (see Materials and Methods). The data are the means ± SE of 3 independent experiments performed in duplicate. * (p < 0.05), ** (p < 0.01) and *** (p < 0.001): statistically significant synergistic effects of the VDD/carnosic acid and VDD/rosemary extract combinations.

When added alone, carnosic acid did not induce any significant effect on differentiation of either WEHI-3B D (Figs. 4a and 4b) or HL60-G (Figs. 4c and 4d) cells. However, in both cell lines, this polyphenol synergistically augmented the effects of 1 nM 1,25D3 and Ro25-4020, as demonstrated by the measurement of surface differentiation markers (Figs. 4a, 4c and 4d) and superoxide production (NBT reduction) (Fig. 4b). The differentiation of HL60-G cells induced by 10 nM 1,25-dihydroxy-16,23Z-diene-20-epi-26,27-hexafluoro-19-nor-cholecalciferol (Ro26-3884), which became available for this study more recently, was also potentiated by carnosic acid in a similar manner. However, the overall effect was much weaker than that produced by either the 1,25D3/carnosic acid or Ro25-4020/carnosic acid combination (Figs. 4c and 4d).

Figure 4.

Carnosic acid and rosemary extract similarly potentiate differentiation of WEHI-3B D (a and b) and HL60-G (c and d) cells induced by 1,25D3 or Ro25-4020. Cells were incubated for 96 h (WEHI-3B D) or 48 h (HL60-G) with the indicated compounds. The expression of F4/80 (a), CD11b (c) and CD14 (d) was determined by flow cytometry. The percentage of NBT-reducing cells, containing black formazan deposits (b), was determined by counting 200 cells, as described under Materials and Methods. The data are the means ± SE of 3 independent experiments performed in duplicate. (a) and (b) Statistically significant synergistic effects of the VDD/carnosic acid and VDD/rosemary extract combinations (p < 0.01) were observed in both F4/80 and NBT assays. (c) and (d) * (p < 0.05), ** (p < 0.01) and *** (p < 0.001): statistically significant synergistic effects of the VDD/carnosic acid and VDD/rosemary extract combinations.

The ability of rosemary extract (added at the amount corresponding to 10 μM final concentration of carnosic acid) to synergize with either 1,25D3 or Ro25-4020 in the growth inhibitory (Fig. 3) and differentiating (Fig. 4) activity was comparable in WEHI-3B D cells and slightly greater in HL60-G cells than the activity of the purified carnosic acid. These data suggest that the enhancing activity of rosemary extract is mainly, but in HL60 cells perhaps not exclusively, due to the presence of carnosic acid.

Taken together, the in vitro data indicate that similar to our previous findings in human leukemia cells,36, 40 the enhanced antiproliferative effect of VDD/carnosic acid combinations in WEHI-3B D murine leukemia cells is due to a cytostatic rather than cytotoxic action, and is probably related to the ability of such combinations to interfere with G1-S cell cycle traverse associated with cell differentiation.54

Establishment of WEHI-3B D tumor model in syngeneic Balb/c mice

WEHI-3B cells are myeloid leukemia cells originally derived from abdominal tumor deposits developed in Balb/c mice after i.p. injection of carcinogen (mineral oil).55 In our hands, i.p. inoculation of 1.0×105 WEHI-3B D cells in syngeneic Balb/c mice resulted in the development of an easily-palpable solid tumor located on the anterior abdominal wall near the site of cell injection (Fig. 5). In untreated mice, measurable tumors appeared on days 9–12 postinoculation, and after 28–32 days, they reached 2.0–2.5 g (wet weight), which is about 10% of the Balb/c mouse body weight. Signs of illness (e.g., loss of appetite, dullness and sleepiness) were also noted clinically at this time. Therefore, approximately at the end of the fourth week experiments were terminated because of ethical considerations, and all mice were sacrificed by the institutionally approved method. Ninety percent or more mice survived during the experimental period. Tumor growth was not accompanied by significant ascites or cell deposits in the internal organs (both evident only in about 3% of tumor-bearing mice). Analysis of Wright-Giemsa-stained peripheral blood smears revealed only few, if any, large blast-like cells in untreated or treated tumor-bearing animals (0–3 blasts per 100 white blood cells), irrespective of the applied treatments, without significant changes in white blood cell differential count (data not shown), suggesting that i.p. inoculation of WEHI-3B D cells does not result in a significant systemic leukemia. Therefore, in this study, “leukemogenicity” was evaluated by the incidence and the size of the solid tumors that developed at the implantation sites.

Figure 5.

Tumorigenicity of WEHI-3B D cells in Balb/c mice. (a) A solid tumor formed on the anterior abdominal wall of the untreated control mouse 29 days after i.p. inoculation of 1 × 105 WEHI-3B D cells. (b) Tumors excised from mice treated for 29 days with either vehicle (0.1% ethanol, i.p.) or the combination of Ro25-4020 (1 μg, i.p.) and rosemary extract (RE, 1% w/w, in diet).

Effects of vitamin D3 derivatives on serum levels of calcium

Since the major toxicity of VDDs is their hypercalcemic effect, we first determined the calcemic activity of 1,25D3, Ro25-4020 and Ro26-3884 in healthy Balb/c mice. When injected intraperitoneally, 1,25D3 at ≥0.1 μg (3 times a week) induced hypercalcemia of ≥12 mg/dl (Fig. 6a), and the mice receiving 0.4 μg 1,25D3 died within 2 weeks (data not shown). In contrast, Ro25-4020 at up to 2 μg (3 times a week) did not induce elevation of serum calcium above normal values (8.5–10.5 mg/dl for Balb/c mice) and caused no general toxicity for at least 6 weeks. A higher dose (4 μg) induced a slight hypercalcemia of 10.9–11.2 mg/dl on weeks 2–4, without a significant decrease in the weight gain of the treated animals (data not shown). Thus, in agreement with previously reported data,17 the above results indicate that Ro25-4020 is at least 20 times less calcemic than 1,25D3 (e.g., 9.1 ± 0.5 mg/dl at 2 μg Ro25-4020 vs. 12.3 ± 0.8 mg/dl at 0.1 μg 1,25D3 on week 4, n = 10). On the other hand, injections of 2 μg Ro26-3884 produced a strong time-dependent hypercalcemia (Fig. 6a).

Figure 6.

Effects of VDDs on serum calcium (a), tumor growth (b) and animal weight (c). (a) Healthy mice were injected i.p. with 1,25D3, Ro25-4020 or Ro26-3884 at the indicated doses 3 times per week. Serum calcium levels were measured weekly, as described under Materials and Methods. (b) Mice were inoculated i.p. with 1 × 105 WEHI-3B D cells, followed by injections i.p. of 1,25D3 (0.0625 μg), Ro25-4020 (2 μg) or Ro26-3884 (2 μg) 3 times per week. Tumor volume was evaluated 3 times a week, as described under Materials and Methods. (c) Mice from the experiment shown in (b) were weighed 3 times per week. The data are the means ± SE of one (a) or 2 (b) experiments performed in groups of 10 animals. (a) SE less than 10%. * (p < 0.05) and ** (p < 0.01): statistically significant differences from untreated control.

Effects of vitamin D3 derivatives and rosemary extract on tumorigenicity of WEHI-3B D cells

In the in vivo tumorigenicity assays, we first determined the effects of 1,25D3 and Ro25-4020 alone at non-calcemic doses (0.065 and 2 μg, respectively). A slightly higher 1,25D3 dose of 0.0625 μg was used here instead of 0.05 μg tested in healthy mice (Fig. 6a), as 0.0625 μg has been reported by Shiohara et al.43 to be non-calcemic when injected in Balb/c mice 3 times per week for 5 weeks. In parallel, despite its high calcemic activity, Ro26-3884 was tested at the same dose as Ro25-4020 (2 μg) to compare the antitumor efficiency of these 2 analogs. To this end, 40 mice were inoculated intraperitoneally with 1.0 × 105 WEHI-3B D cells and randomly divided into 4 groups. Three groups of 10 mice were treated intraperitoneally with 1,25D3, Ro25-4020 or Ro26-3884 at the doses indicated above (3 times a week) and the other 10 mice (control group) received vehicle (0.1% ethanol). All the control mice and the animals receiving 1,25D3 or Ro26-3884 developed tumors, while 3 animals in the Ro25-4020-treated group were tumor-free throughout the experiment (about 4 weeks). As illustrated in Figure 6b, injections of 0.065 μg 1,25D3 had no significant effect on tumor growth, while treatment with 2 μg Ro25-4020 resulted in both a significant delay in tumor development and reduction in tumor size, as compared to the untreated animals or those injected with 1,25D3. Ro26-3884 also inhibited tumor growth; however, its effect became evident only after ∼20 days of treatment and was overall less pronounced than that of Ro25-4020. As expected, treatment with non-calcemic doses of 1,25D3 and Ro25-4020 was not accompanied by significant changes in the animal body weight as compared to control (Fig. 6c). Furthermore, Ro25-4020-treated animals displayed less noticeable clinical symptoms of illness than untreated or 1,25D3-treated mice. On the other hand, Ro26-3884 substantially reduced weight gain (Fig. 6c), which was accompanied by severe general toxicity. Based on its low toxicity and marked antitumor activity, Ro25-4020 was chosen for the combination experiment, using oral rosemary extract as the carnosic acid-rich “enhancer” preparation.

Several studies have reported that treatment with rosemary extracts mixed with food at 0.5–2.0% (w/w) produces anticarcinogenic effects in rats,56, 57, 58 inhibits the uterotropic effect of endogenous estrogens in mice59 and is well tolerated by animals. Thus, to determine whether Ro25-4020 and rosemary extract can cooperate in the antitumorigenic activity, a lower dose of the analog (1 μg) was combined with oral 1% (w/w) rosemary extract. In each of 2 similar experiments, 40 mice fed on a standard powdered rodent diet were inoculated with 1.0 × 105 WEHI-3B D cells and assigned to 4 treatment groups (10 mice in each). Group 1 (control) received vehicle (0.1% ethanol) i.p. while kept on the standard diet without additions. Group 2 received 1 μg Ro25-4020 i.p. and the standard diet. Group 3 received 0.1% ethanol and fed on the diet mixed with 1% (w/w) rosemary extract. Group 4 received 1 μg Ro25-4020 and oral 1% (w/w) rosemary extract. Injections of either Ro25-4020 or vehicle were then repeated 3 times a week and the mice were fed continuously with respective diets.

Overall, in the 2 experiments performed as above, 90% of both the control animals and those treated with 1 μg Ro25-4020 alone developed tumors (Fig. 7a). Tumor incidence was less pronounced in animals treated with 1% rosemary extract alone (75%) and was particularly reduced in the combination-treated groups (55%). The control mice developed relatively fast growing tumors, whereas each of the treatment regimens (Groups 2–4) slightly delayed tumor appearance (Fig. 7b). Injections of 1 μg Ro25-4020 alone (Group 2) were minimally effective during most of the treatment period, except on day 29. On the other hand, diet supplemented with 1% (w/w) rosemary extract (Group 3) resulted in a time-dependent inhibition of tumor growth. In this group, tumors became significantly smaller than those in the untreated mice when measured in the final stage of the experiment (on days 27–29). Importantly, the combined treatment (Group 4) induced a marked cooperative antitumor effect, which was found to be synergistic as determined on day 27 (Fig. 7b). On days 29–30, both experiments were terminated because tumor weight in control mice exceeded 10% of the body weight. All the treatment regimens were well tolerated, as indicated by only insignificant changes in animal weight gain in the treated groups, as compared to the untreated control (Table I). Furthermore, the absence of significant general toxicity was indicated in weekly tests of serum albumin, total bilirubin and urea in retro-orbital sinus blood samples. Importantly, while strongly cooperating with Ro25-4020 in the antitumor effect, rosemary extract did not increase the calcemic activity of this vitamin D3 analog (Table I).

Table I. Effects of RO25-4020 and Rosemary Extract Treatments on Toxicity Parameters in Balb/c Mice Inoculated with WEHI-3B D Cells1
Treatment (week 4)Animal weight (g)2Calcium (mg/dl)3Albumin (g/dl)3Total bilirubin (mg/dl)3Urea (mg/dl)3
  • RE, rosemary extract.

  • 1

    Data of 2 independent experiments are presented. The mean ± SE of the indicated number of measurements is shown.

  • 2

    n = 18–20.

  • 3

    n = 10.

  • 4

    Statistically significant differences (p < 0.05) as compared to control.

Control25.4 ± 1.58.83 ± 1.42.71 ± 0.210.23 ± 0.0137.2 ± 4.0
Ro25-4020 (1 μg)23.0 ± 2.89.23 ± 1.32.52 ± 0.160.24 ± 0.0332.3 ± 3.5
RE (1% w/w)23.8 ± 1.98.55 ± 1.82.92 ± 0.220.18 ± 0.02440.3 ± 3.9
Ro25-4020 + RE22.7 ± 1.39.67 ± 2.12.79 ± 0.310.16 ± 0.03438.6 ± 4.1


Numerous in vitro and in vivo studies have consistently documented that 1,25D3 potently induces differentiation and inhibits proliferation in different cancer cell types,12 suggesting its potential applications in cancer prevention and treatment. It has also been established that various pharmacological, herbal and dietary compounds can markedly enhance cell growth inhibition and the induction of differentiation induced by low concentrations of 1,25D3in vitro,13 thus indicating a possible way of decreasing calcemic toxicity of vitamin D3-based therapies. Obviously, the risk of vitamin D3-induced hypercalcemia can be even more reduced upon combining an enhancing agent with a synthetic vitamin D3 analog, which retains the anticancer properties of the parent compound but is less calcemic.

Plant polyphenolic antioxidants, such as curcumin derived from turmeric, silibinin from milk thistle and carnosic acid from rosemary, have been proposed as good candidates for combination with vitamin D3 derivatives in differentiation therapy of myeloid leukemias because of their ability to markedly potentiate the in vitro differentiating and antiproliferative effects of 1,25D3 in human HL60 and U937 myeloblastic cells.34, 35, 36, 37, 39 Here we translated these findings to in vivo conditions, using a syngeneic mouse leukemia tumor model. Our study shows for the first time that carnosic acid-rich rosemary extract and a low dose of a low-calcemic Ro25-4020 strongly cooperate in inhibiting tumorigenicity of WEHI-3B D cells in Balb/c mice. This is consistent with the fact that as in human leukemia cells, the in vitro antileukemic effect of VDDs in WEHI-3B D murine myelomonocytic cells can also be markedly and comparably enhanced by both rosemary extract and purified carnosic acid.

The mechanisms by which plant polyphenolic anioxidants potentiate the activity of low concentrations of 1,25D3 in leukemic cells have been extensively studied in ours and other laboratories (see ref.13 for a recent review). We have shown that carnosic acid treatment of HL60-G cells results in a decrease in the intracellular levels of reactive oxygen species and elevation of glutathione content while glutathione depletion suppresses differentiation induced by both 1,25D3 and its combination with carnosic acid.41 These findings raise the possibility that the antioxidant action of carnosic acid and other polyphenols amplifies the differentiation signal provided by 1,25D3 by generating a reducing intracellular environment, which can modulate redox-sensitive cellular systems (e.g., the transcription factors NFκB and AP-160, 61) involved in the regulation of cell differentiation and proliferation. Indeed, Sokoloski et al.34, 62, 63 have demonstrated that plant antioxidants downregulate NFκB and that inactivation of this transcription factor results in the enhancement of 1,25D3-induced differentiation. We have found that 1,25D3-induced activation of mitogen-activated protein kinase (extracellular signal-regulated kinase and c-Jun N-terminal kinase) pathways, which target AP-1, is strongly potentiated by antioxidants.38, 41 Furthermore, both AP-1 binding to its cognate DNA response element and transcriptional activity were markedly enhanced by different plant polyphenols.38, 41

Taken together, our in vivo and in vitro data suggest that at least some polyphenol-rich dietary or herbal extracts and, probably, purified plant polyphenolic compounds with differentiation enhancing activity in vitro can be employed in the VDD-based combination differentiation therapy of at least some types of myeloid leukemias. Furthermore, various plant polyphenols themselves have been shown to inhibit proliferation in leukemic and other types of cancer cells both in vitro and in vivo,50, 64, 65, 66 which would further augment the anticancer activity of VDD/polyphenol combinations. Moreover, both serum levels of vitamin D3 at the high end of the physiological range21, 22 and consumption of polyphenol-rich diets are consistent with lower risk of certain cancers.32, 33 This is supported by the findings that both VDDs67, 68, 69 and rosemary extract and its purified components56, 57, 58 suppressed experimental carcinogenesis. Thus, it seems plausible to hypothesize that these compounds may also cooperate in their chemopreventive action.

Ro25-4020 has been reported to be at least 20-fold less calcemic than 1,25D3 in Balb/c mice, as shown by Hisatake et al.17 and this was confirmed in the experiments described here. Yet, as found in the present study, these 2 VDDs displayed comparable potencies in induction of differentiation and growth inhibition of both HL60-G and WEHI-3B D cells in vitro. Similarly, Ro25-4020 has been previously found to inhibit the growth of human breast (MCF-7)17 and prostate (LNCaP)17, 42 cancer cells, as well as HL60 leukemia cells,17, 70 with a similar or higher potency than the parent compound. However, the in vivo effects of 1,25D3 and Ro25-4020 on WEHI-3B D tumor growth could not be compared at the same concentrations in this study, because the maximal tolerated (non-calcemic) dose of 1,25D3 (0.0625 μg) was ineffective. Nevertheless, even if the antitumor potencies of these 2 VDDs are theoretically similar, Ro25-4020 is undoubtedly a therapeutically more desirable compound, since it can be administered at much higher doses. Clearly, Ro26-3884 has no potential for therapeutic use, since it is hypercalcemic and highly toxic (Fig. 6), perhaps due to its hexafluorinated side chain (Fig. 1). Conversely, Ro25-4020 appears to be a credible candidate for clinical trials, as in this study its effects on murine WEHI-3B D and human HL60-G cells were similar (Figs. 3 and 4).

Importantly, despite the enhanced antitumor effect of the Ro25-4020/rosemary extract combination, food supplementation with this extract did not result in any potentiation of the calcemic effect of Ro25-4020. Furthermore, this combination did not induce a significant general toxicity in treated animals. Moreover, a significant (p < 0.05) decrease in total serum bilirubin levels and slightly higher levels of serum albumin in mice fed on rosemary-containing diet as compared to control animals suggests a possible hepatoprotective effect. This is consistent with the previously reported data showing antihepatotoxic effect of oral administration of rosemary extracts.71, 72

The study of the molecular mechanism(s) underlying the in vivo antitumorigenic effects of Ro25-4020, rosemary extract and their combination in the WEHI-3B D murine tumor model is now in progress in our laboratory, employing histochemical and immunohistochemical analysis of tumor sections obtained in the experiments described here. The results of our in vitro study show the cytostatic rather than cytotoxic nature of the VDD/carnosic acid and VDD/rosemary extract combinations. This suggests that there would not be any direct tumor cell killing by these agents in vivo but rather cessation of proliferation associated with cell maturation. Indeed, preliminary microscopic examination of the hematoxylin/eosin stained sections of tumors obtained from both control and treated animals did not reveal significant histological signs of cell death (data not shown).

In conclusion, the results of this study clearly demonstrate that a vitamin D3 derivative and a plant polyphenolic preparation can strongly cooperate not only in the induction of differentiation and growth arrest of myeloid leukemia cells in vitro, but also in the antitumor/antileukemic activity in vivo. These findings may suggest novel protocols for combination differentiation therapy of myeloid leukemia and further clarify the cancer preventive activities of vitamin D and polyphenol-rich diets.


This research was supported by the USA-Israel Binational Science Foundation grant 2001-041 (to M.D., Y.S., J.L., and G.P.S.), Israel Cancer Association grant through the Boaz Adar Memorial foundation (to M.D.), NIH grant RO1 CA 44722-15 (to G.P.S.), and the American Institute for Cancer Research grant 05A022 (to G.P.S and M.D.). We thank Dr. Zohar Nir (LycoRed Natural Products Industries) for donating rosemary extract and Dr. S. Shany (Biochemistry Laboratory, Soroka University Medical Center) for providing blood chemistry assays.