Granulocytic Differentiation of Human NB4 Promyelocytic Leukemia Cells Induced by All-trans Retinoic Acid Metabolites¹

Retinoids (vitamin A or retinol, and its derivatives) are a family of molecules involved in many physiological processes such as embryogenesis, cell growth, vertebrate development, differentiation,and homeostasis (reviewed in Ref. 1). The retinoids effects are known to be mediated by two classes of nuclear receptors,the RARs³and the RXRs, each comprising three subtypes (α, β, γ), with various isoforms of each subtype (2). These receptors belong to the superfamily of nuclear hormone receptors and act as ligand-activated transcription factors of numerous genes(reviewed in Refs. 2 and 3). Retinoid receptors interact as either homo- or heterodimers on specific hormone response elements (RA responsive elements or RX responsive elements) of target gene promoters (4, 5, 6). The transcriptional activity of these receptors is further positively (holo-receptor) or negatively(apo-receptor) modulated by ligand binding (reviewed in Refs.7 and 8). In addition, intracellular CRABPs play a regulatory function in RA signaling (9). Biological responses to ATRA may, therefore, be modulated by the type of nuclear receptors present in cells, by nuclear cofactors, by the CRABPs,and by the availability of specific retinoid ligands (reviewed in Ref.10).

The differentiating properties of ATRA have led to the development of therapies for the prevention and treatment of various human cancers(11) including APL, in which retinoids are particularly active (12). APL is a subtype of myeloid leukemia (M3),characterized by the accumulation of cells blocked at the promyelocytic stage. This leukemia exhibits a specific chromosomal translocation t(15;17) involving the PML locus on chromosome 15 and the RARαlocus on chromosome 17, thus generating a chimeric gene PML-RARα, translated into a chimeric nuclear receptor PML-RARα (13, 14, 15, 16, 17). PML-RARα functions as an aberrant receptor and is considered to be the cause of APL (18).

Although several clinical studies have established that ATRA can induce leukemia cell differentiation and remission in almost all patients(12, 19), resistance to this therapy develops rapidly(20). This acquired resistance appears to be attributable in part to the decline in ATRA plasma levels below therapeutic concentrations after repeated administration, presumably caused by the induction of hepatic cytochrome P450s that may increase its clearance (21, 22, 23). It has also been proposed that increased CRABPs contribute to resistance by decreasing nuclear availability of ATRA (24). More recently, mutations in the E-ligand-binding region of the RARα chimeric protein PML-RARα have been linked to ATRA resistance (25, 26, 27, 28, 29).

ATRA metabolism is complex and only partially understood at the present time (reviewed in Ref. 30). It can be oxidized by cytochrome P450s to various metabolites including 4-OH-RA, 4-oxo-RA,18-OH-RA, and 5,6-epoxy-RA, which may be further metabolized to water-soluble glucuronides. Although ATRA metabolites are often considered to be catabolic products possessing low biological activity,some metabolites have been reported to be active in embryogenesis and to inhibit the growth of certain cancer cell lines(31, 32, 33, 34, 35, 36). However, metabolism also seems to be involved in the response of certain cancer cells to ATRA, as shown by a study of breast cancer cell lines in which cell lines sensitive to ATRA growth inhibition (T47D and MCF7) could metabolize ATRA to polar metabolites(e.g., 4-oxo-RA and 4-OH-RA), whereas resistant cell lines(MDA-MB231 and MDA-MB418) metabolized ATRA poorly (35). Similar results involving ATRA metabolism to polar compounds by other sensitive cancer cells have also been reported (37).

As the metabolism of ATRA appears to be involved in its mechanism of action in certain cancer cell lines, the purpose of this study was to evaluate the in vitro effects of the principal oxidized ATRA metabolites using the human NB4 promyelocytic leukemia cell line that expresses the characteristic APL t(15;17) chromosomal translocation(38). Our results show that NB4 cells can metabolize ATRA to polar metabolites (4-OH-RA, 4-oxo-RA, 18-OH-RA, and 5,6-epoxy-RA),and that these metabolites are active retinoids in vitrowith regard to cell growth inhibition, cell cycle arrest in G₁ phase, cell differentiation, NB reorganization, and degradation of the chimeric protein PML-RARα. In addition, when using a specific RARα antagonist, ATRA metabolites are shown to act via the RARα nuclear receptor signaling pathway, as does the parent compound.

ATRA, its isomers 13-cis-RA and 9-cis-RA, NBT,PMA, and propidium iodide were purchased from Sigma-Aldrich(St-Quentin-Fallavier, France). 4-OH-RA, 4-oxo-RA, 18-OH-RA, and 5,6-epoxy-RA were kindly donated by Dr. Michael Klaus (Hoffmann-La Roche Ltd., Basel, Switzerland) and Dr. Louise H. Foley (Hoffmann-La Roche Inc., Nutley, NJ). The RARα antagonist BMS614 was synthesized by Bristol-Myers Squibb and kindly donated by Dr. H. Gronemeyer(Institut de Génétique et de Biologie Moléculaire et Cellulaire, Illkirch, France). Stock solutions of retinoids were prepared in ethanol at 10⁻²m and stored protected from light at −20°C.

The NB4 human APL cell line, developed in our laboratory(38), was maintained in RPMI 1640 (Life Technologies,Inc.) supplemented with 10% FCS (Bayer Diagnostics, Puteaux,France), 2 mm l-glutamine, 50 units/ml penicillin G, and 50 μg/ml streptomycin (Life Technologies, Inc.). The cells were cultured at 37°C in a humidified atmosphere containing 5% CO₂. Cell density was assessed with an electronic particle counter and size analyzer (Coulter Electronics,Hialeah, FL). Stock solutions of retinoids (at 10⁻² m in ethanol) were further diluted in culture medium to the final concentration indicated in each experiment. Final concentration of ethanol in culture medium did not exceed 0.01%. Stock solutions of retinoids were regularly checked for purity using reversed-phase HPLC analysis and did not show any contamination by ATRA. Cell cultures were diluted with fresh medium every 48 h, and drug concentrations were adjusted accordingly. All of the experimental procedures were light protected.

Cells (10⁷) were incubated at 37°C in 2 ml of the above medium with 1 μm ATRA for 24 h. After acidification (HCl, 1%), cells and medium were extracted with ethyl acetate (39), the organic phase recuperated by centrifugation, and evaporated under a gentle nitrogen stream. The dry residue was reconstituted in 100 μl of methanol, and 50 μl was injected onto a HPLC system composed of a Varian 5000 chromatograph, a C₁₈ reversed-phase column (Beckman, Ultrasphere 5μm, 4,6 mm × 250 mm), protected by a precolumn, a UV detector set at 340 nm (Waters 484), and a peak integrator (Varian 4400). Elution was accomplished using a 25-min linear gradient starting at 65% methanol/35% water with 1% ammonium acetate, and ending at 90% methanol/10% water pumped at 1 ml/min.

Cells (5 × 10⁵) were added to 450μl of 1 mg/ml NBT solution containing 3 × 10⁻⁷ m PMA in PBS buffer, and incubated at 37°C for 25 min. Cytospin preparations of 200 μl of the cell suspensions (Cytospin; Shandon) were allowed to air-dry. The NBT-positive and -negative cells were scored under phase contrast microscopy examination (Leica) at ×40. A minimum of 400 cells per slide, scored with at least two scores, were performed per experimental condition.

Cytospin preparations of 2 × 10⁵cells were allowed to air-dry, incubated in pure May-Grünwald solution for 5 min, then in 50% May-Grünwald/water for 10 min,washed in water, and incubated in a 20% Giemsa/water solution for 20 min. The slides were then washed in water, air-dried, and examined under immersion microscopy (Leica, ×63).

Cells (1 × 10⁶) were washed in PBS and resuspended in 200 μl of PBS. The cells were fixed in 2 ml of 75% ethanol/water for 2 min, centrifuged, and resuspended in 1 ml of PBS solution with 40 μg of propidium iodide and 100 μg of RNase A for 30 min at 37°C. Propidium iodide (red) fluorescence was measured using a FACScan flow cytometer (Becton Dickinson) with 10⁴ cells acquired per sample.

Cells (1 × 10⁶) were incubated with an anti-CD11c-PE, or the isotype control IgG2a-PE (Becton Dickinson), for 20 min on ice in the dark. The cells were washed once with PBS and resuspended in 500 μl of PBS. PE (red)fluorescence was measured using a FACScan flow cytometer (Becton Dickinson) with 10⁴ cells acquired per sample.

Cytospin preparations of 10⁶ cells were allowed to air-dry for 24 h, fixed with acetone for 10 min at 4°C, and air-dried for 20 min. The slides were rehydrated for 10 min in PBS and then incubated with a polyclonal rabbit antiserum directed against a PML recombinant protein (40) at a dilution of 1:500 in PBS for 1 h at room temperature. The slides were then incubated with fluorescein-coupled antirabbit secondary antibody (Sigma) at a dilution of 1:200 for 30 min at room temperature in the dark. All of the incubations were followed by three washes in PBS for 5 min. Preparations were examined under fluorescence microscopy (Leica, ×63).

Crude cytoplasmic and nuclear extracts were prepared from untreated and treated NB4 cells. The cells were collected, washed once with cold PBS,centrifuged at 1300 rpm at 4°C, and resuspended in buffer A [10 mm HEPES (pH 7.9), 1.5 mm MgCl2, 10 mm KCl, and 0.5 mm DTT] for 10 min at 4°C. After centrifugation for 1 min at 5000 rpm at 4°C, cells were lysed for 1 min on ice in buffer A containing 0.2% NP40, 1.5 μg/ml protease inhibitor cocktail, and 0.2% phenylmethylsulfonyl fluoride, and centrifuged 1 min at 5000 rpm. Supernatants corresponding to cytoplasmic proteins were collected, and the nuclear proteins contained in the pellets were extracted with buffer B [20 mm HEPES (pH 7.9), 0.2 mm EDTA, 0.5 mm DTT, 0.42 m NaCl, 25% v/v glycerol, 1.5μg/ml protease inhibition cocktail, and 0.2 mmphenylmethylsulfonyl fluoride) for 20 min on ice, and centrifuged 1 min at 5000 rpm at 4°C. Cytoplasmic and nuclear extracts were frozen at−80°C. Nuclear protein extracts (10 μg) were boiled in the presence of 5% β-mercaptoethanol and loaded onto 8%SDS-polyacrylamide gels, electrophoresed, and blotted onto polyvinylidene difluoride membranes (Immobilon P; Millipore). The membranes were blocked with 4% nonfat milk in PBS for 3 h at room temperature, then incubated with primary human antibody anti-RARα[RPα(F); 1:5000; Ref. 41] in PBS-T (PBS and 0.1%Tween 20) and 0.5% milk for 18 h at 4°C. Membranes were washed five times for 5 min each with PBS-T, incubated with horseradish peroxidase-coupled antirabbit antibody (1:10000; Jackson Laboratories) for 30 min in PBS-T at room temperature, and washed five times for 5 min with PBS-T. Detection was performed as described in the ECL kit (Amersham; Ref. 42).

NB4 cells (2 × 10⁵ cells/ml) were exposed to ATRA or its metabolites at concentrations ranging from 7 nm to 1 μm, the NBT test was performed at 72 h, and the positive cells were scored as described above. The data from three separate experiments were plotted on a semilogarithmic scale to determine the EC₅₀ that elicited half maximal NBT-positive cells. Nonlinear regression analysis was carried out using a sigmoidal dose-response curve and GraphPad software.

We first evaluated the ability of human NB4 promyelocytic leukemia cells to metabolize ATRA. Fig. 1 shows the ATRA metabolites formed after 24 h in the presence of 1μ m ATRA. The metabolites formed were 4-oxo-RA, 4-OH-RA,18-OH-RA, and 5,6-epoxy-RA, which presented concentrations of 0.06,0.03, 0.08, and 0.39 μm, respectively. The stereoisomers 13-cis-RA and 9-cis-RA presented concentrations of 0.07 and 0.16 μm, respectively, and ATRA of 0.19 μm.

NB4 cell growth in the presence of exogenously added ATRA metabolites was assessed at a concentration of 1 μm for 120 h(Fig. 2,A). Compared with untreated cells, the growth inhibition was found to be similar for ATRA and the 4 metabolites tested, yielding 50% growth inhibition between 72 and 96 h. At 120 h, cell growth was still similar for all compounds, and ranged from 22 to 31%of untreated cells. Cell growth declined to 11–13% of controls for ATRA and its metabolites at 168 h (data not shown). In parallel experiments, the influence of ATRA or its metabolites on the cell cycle distribution was assessed for 5 days (Fig. 2 B). A time-dependent cell cycle accumulation in the G₁phase was observed for all compounds tested, yielding 82–84%G₁ accumulation at day 5, at the expense of the S and G₂-M phases. These data are in agreement with the observed growth arrest.

Because the ATRA metabolites appeared to be as effective as the parent compound with regard to cell growth inhibition and cell cycle arrest,we assessed the differentiating potential of these retinoids in NB4 cells using cell morphology, NBT reduction, and CD11c expression. Morphological differentiation using May-Grünwald Giemsa staining of NB4 cells exposed to either ATRA or its metabolites was carried out every day for 5 days. The first signs of morphological differentiation were observed at 72 h for all compounds, and the differentiation was more pronounced for 18-OH-RA and 5,6-epoxy-RA than for ATRA,4-oxo-RA or 4-OH-RA (data not shown). At 96 h, the granulocytic maturation pattern was similar for all retinoids, with polylobular nuclei, decreased nuclear:cytoplasmic ratio, and decreased cytoplasm staining (Fig. 3,A). ATRA and its metabolites induced similar kinetics of differentiation with comparable efficacy at 1μ m reaching 82–88% NBT positive cells in 5 days (Fig. 3,B). Moreover, ATRA and its metabolites induced the expression of the maturation-specific cell surface marker CD11c in NB4 cells 24 h after retinoid exposure (Fig. 3,C); at 48 or 72 h, there was little change in the percentage of CD11c positive cells compared with that at 24 h, and no difference was noted between ATRA and its metabolites with regard to the level of CD11c expression (Table 1).

The intracellular distribution of PML and PML-RARα proteins in NB4 promyelocytic leukemia cells exposed to ATRA or its four metabolites at 1 μm for 5 days was analyzed by immunocytochemistry. The nuclei from control cells presented fine and numerous labeled dots containing PML (Fig. 4,A), corresponding to the micropunctated NBs typically found in APL cells. When the cells were exposed to ATRA, the morphological reorganization was characterized by an increase in NB size corresponding to macropunctated NBs, and a decrease in their number. A similar pattern of NB reorganization to that seen with ATRA was observed for all of the metabolites tested (Fig. 4,A). As early as 24 h, NB reorganization could be seen,which reached a maximum at 72 h. No major difference in NB size was observed between 72 h and 120 h for ATRA or its metabolites (data not shown). Concomitant to NB reorganization,degradation of the PML-RARα chimeric protein by ATRA metabolites was also assessed. ATRA metabolites treatment induced PML-RARαdegradation as did ATRA (Fig. 4,B). The protein became barely detectable at 48 h, with the emergence of a lower molecular weight band corresponding to degradation of the chimeric protein (Fig. 4 B).

NB4 cells were treated by ATRA or its metabolites at concentrations ranging from 7 nm to 1 μm for 72 h (Fig. 5). ATRA was more efficient than its four metabolites in the induction of differentiation, with effective concentrations to elicit 50%(EC₅₀) of NBT positive cells of 15.8 ± 1.7 nm. The metabolites yielded EC₅₀ values of 38.3 ± 1.3 nm for 4-oxo-RA, 55.5 ± 1.8 nmfor 18-OH-RA, 79.8 ± 1.8 nm for 4-OH-RA,and 99.5 ± 1.5 nm for 5,6-epoxy-RA. However, at higher concentrations (from 0.25 to 1 μm),ATRA and its metabolites showed similar differentiation efficiency of about 80% NBT positive cells.

Because the ATRA metabolites could induce NB4 promyelocytic leukemia cell differentiation, it was of interest to investigate the nuclear receptor pathways involved in this response. Because differentiation of NB4 cells by ATRA is mediated through the RARα receptor, we therefore blocked this pathway using the RARα-specific antagonist BMS614. NB4 cells were simultaneously treated with ATRA or its metabolites at 0.1μ m, in the absence or presence of the RARα-specific antagonist BMS614 (1 μm) for 48 h (Fig. 6). As expected, BMS614 abolished ATRA-induced CD11c expression as shown by the decrease in fluorescence intensity to near control values. Likewise, induction of CD11c expression by the ATRA metabolites was blocked by BMS614, which indicated that the metabolites induced maturation through the RARα signaling pathway, as does the parent compound.

Although ATRA is considered to be the best single agent available for APL therapy, it is limited in that clinical resistance can develop rapidly (20). One cause of ATRA resistance is attributable to its increased metabolism with rapid clearance often observed in relapsing patients (43). ATRA has, therefore, been considered to be the unique effective compound, and its metabolites have been considered as catabolic products. However, several authors have proposed that ATRA metabolism may be required for cell growth inhibition of various cancer cell lines in vitro, although no causal relationship has been shown (35, 37, 44). Because very little information is presently available concerning the biological activity of the principal Phase I ATRA metabolites, the purpose of this work was to evaluate the in vitro effects of the main oxidized metabolites of ATRA using the NB4 human promyelocytic leukemia cell line (38). The NB4 cell line is a relevant model for this study in that it exhibits the chromosomal translocation t(15;17) specific for APL, with the resultant production of the chimeric protein PML-RARα. It can be differentiated by ATRA treatment and may be used to elucidate mechanisms of differentiation in APL, a paradigm of a disease sensitive to differentiation therapy.

Our results show that NB4 cells metabolize ATRA to the more polar metabolites 4-oxo-RA, 4-OH-RA, 18-OH-RA, and 5,6-epoxy-RA. ATRA metabolism is cytochrome P450-dependent and appears to involve several isoforms including CYP2C8 (45) and CYP26 (46, 47; reviewed in Ref. 48). More recently, Marill et al. (49) have identified several other human cytochrome P450s significantly involved in the metabolism of ATRA.

ATRA and its metabolites were found to be equally active in terms of growth inhibition of NB4 leukemia cells at 1-μmconcentration. These data are in agreement with those of van der Leede et al. (35) and van Heusden et al. (33), who also observed growth inhibition of breast cancer cell lines with ATRA metabolites (4-oxo-RA,4-OH-RA, and 5,6-epoxy-RA). Rat rhabdomyosarcoma cells have been reported to be sensitive to growth inhibition by 4-oxo-RA, 4-OH-RA, and 5,6-epoxy-RA (34). However, the NB4 leukemia cell line used in this report appears to be more sensitive than solid tumor cell lines, because ∼90% growth inhibition was obtained in these cells compared with 40–50% reported in MCF-7 cells or rhabdomyosarcoma cells (33, 34, 35). To our knowledge, this is the first report on the growth inhibitory activity of the principal Phase 1 metabolites of ATRA in a promyelocytic leukemia cell line and, also, the first account of the activity of the 18-OH-RA.

The ATRA metabolites were also shown to be active in regard to differentiation induction in the NB4 cells as evaluated by several criteria including cell morphology, NBT reduction test, and CD11c integrin expression. These metabolites were as active as the parent compound at concentrations between 0.2 and 1 μm. At lower concentrations, differences were seen between ATRA and its metabolites in their relative potency (EC₅₀) to induce maturation of NB4 cells in the NBT reduction test. Although ATRA was found to be the most potent retinoid tested(EC₅₀, 15.8 nm), 4-oxo-RA(EC₅₀, 38.3 nm) is also very active,inasmuch as only a 2.4-fold difference was observed between this metabolite and the parent compound. It is of interest to note that the EC₅₀ found for 4-oxo-RA is achievable in vivo in humans after ATRA administration, because concentrations in the range of 19–30 ng/ml (61–96 nm) can be reached (50). In some studies, 4-oxo-RA has also been reported to be more active than ATRA in the modulation of positional specification in Xenopus laevis embryos (31). With regard to the EC₅₀ found for 4-oxo-RA,4-OH-RA, and 5,6-epoxy-RA, similar relative potency data have been reported in the differentiation of rat rhabdomyosarcoma cells(34). The relative potency of these natural retinoids may,therefore, be dependent on the cell type and on the physiological phenomenon considered. Comparison of the relative potency of ATRA metabolites may also be biased in the sense that the metabolites were applied exogenously to the cells, as opposed to the physiological situation in which the metabolites are formed intracellularly, via the cytochrome P450s.

Concomitant to the induction of differentiation in the NB4 cells by the ATRA metabolites, we observed a reorganization of the nuclear PML-containing bodies, and a degradation of the PML-RARα chimeric protein, as has been previously observed with ATRA(51, 52, 53). These data suggest that the metabolites share the same signaling pathway as ATRA. To verify this, we used a RARα-specific antagonist (BMS614) in combination with the metabolites. CD11c expression analysis showed a complete inhibition of CD11c expression induction when NB4 cells were cotreated with BMS614,which indicated that ATRA metabolites act through the RARα pathway for differentiation induction.

In summary, these data demonstrate new biological properties of the ATRA metabolites and indicate that they are not ineffective catabolic products. These metabolites may participate in the ATRA mechanism of action on cell growth inhibition and differentiation via the RARαreceptor signaling pathway. In conclusion, the principal ATRA Phase I metabolites are active retinoids which, provided that manageable toxicity and in vivo pharmacological concentrations were achievable, could be of interest for differentiation therapy of sensitive human cancers.

We thank the following persons who made these studies possible: Drs. Michael Klaus (Hoffmann-La Roche Ltd., Basel,Switzerland) and Louise H. Foley (Hoffmann-La Roche Inc., Nutley, NJ),who kindly donated the ATRA metabolites used in this study; Prof. Pierre Chambon [Institut de Génétique et de Biologie Moleculaire et Cellulaire (IGBMC, Illkirch, France)] for providing the anti-RARα antibody; Dr. Marcel Koken (Unité Propre de Recherche 9051 CNRS, Paris) for the PML antibody; Dr. H. Gronemeyer (IGBMC, Illkirch, France) for kindly providing the RARα antagonist BMS614; and Christelle Doliger and Michel Schmid of the Cellular and Molecular Imaging Service of our Institute for expert assistance in image analysis.