Transcriptional activation of the human p21(WAF1/CIP1) gene by retinoic acid receptor. Correlation with retinoid induction of U937 cell differentiation.

We reported previously that the induced differentiation of the myelomonocytic cell line U937 by vitamin D3 is facilitated by the transcriptional induction of the p21(WAF1/CIP1) gene by the vitamin D3 receptor (Liu, M., Lee, M.-H., Cohen, M., and Freedman, L. P. (1996) Genes Dev. 10, 143-153). Retinoic acid (RA), a physiological metabolite of vitamin A, is also a potent inducer of differentiation of several cell types, including myeloid leukemic cells. Like vitamin D3, RA acts through a subfamily of nuclear hormone receptors, RARs and RXRs (retinoid X receptors), which regulate the expression of target genes by binding to specific DNA elements and modulating transcription initiation. In this report we demonstrate that the gene encoding p21 is also a RA-responsive target gene, and we describe a functional RA response element in this gene's promoter which is required to confer RA induction through RAR.RXR heterodimers. These results correlate the RA induction of monocytic differentiation of U937 cells with the transcriptional activation of the p21 gene and suggest a role for this cyclin/cyclin-dependent kinase complex inhibitor in facilitating this differentiation pathway.


of vitamin A, is also a potent inducer of differentiation of several cell types, including myeloid leukemic cells. Like vitamin D 3 , RA acts through a subfamily of nuclear hormone receptors, RARs and RXRs
(retinoid X receptors), which regulate the expression of target genes by binding to specific DNA elements and modulating transcription initiation. In this report we demonstrate that the gene encoding p21 is also a RAresponsive target gene, and we describe a functional RA response element in this gene's promoter which is required to confer RA induction through RAR⅐RXR heterodimers. These results correlate the RA induction of monocytic differentiation of U937 cells with the transcriptional activation of the p21 gene and suggest a role for this cyclin/cyclin-dependent kinase complex inhibitor in facilitating this differentiation pathway.
Retinoic acid (RA) 1 is a physiological metabolite of vitamin A which has potent effects on cell differentiation, proliferation, and development. This ligand functions as a putative morphogen in regulating fetal development during vertebrate embryogenesis and influences adult epithelial cell growth and differentiation (for review, see Refs. 1 and 2). It also suppresses malignant transformation both in vivo and in vitro. Its most widely studied effects have been on its ability to induce the terminal differentiation of human promyelocytic leukemia cells and mouse embryonal carcinoma cells (3). Significantly, patients suffering from acute promyelocytic leukemia carry a 15:17 chromosomal translocation that fuses the RA receptor ␣ (RAR-␣) gene on chromosome 17 to a putative transcription factor, PML, on 15 (4,5). The resulting fusion protein, called PML-RAR, has as yet unknown functions, but nearly all acute promyelocytic leukemia patients who are treated with alltrans-retinoic acid (ATRA) achieve complete remission for a given time period, presumably through an induction of differentiation of the leukemic cells (6,7).
In normal as well as transformed cells, the biologic effects of RA are mediated by two related members of the nuclear receptor superfamily: RA receptors (RAR-␣, -␤, and -␥) and retinoid X receptors (RXR-␣, -␤, and -␥) (for review, see Ref. 8). These proteins act as ligand-inducible transcription factors by interacting with specific DNA target sequences as well as other regulatory factors (i.e. corepressors, coactivators) in promoter/ enhancer regions to regulate the expression of linked genes. RARs bind ATRA and a RA isomer, 9-cis-RA, with high affinity, whereas RXRs only bind 9-cis-RA. The two ligands often elicit similar effects on differentiation, but it has been difficult to ascertain the exact identities of the active metabolites in responsive cells because of the high propensity of RA to isomerize and because RARs and RXRs typically are coexpressed in most cells. On the other hand, RAR-and RXR-specific analogs have been developed and used to demonstrate, for example, that RAR-specific ligands suffice to induce the myeloid leukemic cell line HL-60 to differentiate to granulocytes, but RXR-specific ligands are required to elicit the apoptosis of these cells, suggesting that each receptor mediates distinct processes in this cell line (9).
Although the effects of RA on cell differentiation have been studied extensively, the mechanism by which the ligand induces this effect is poorly understood. Scores of genes and gene products are induced or repressed by RA, but only a handful have been identified as target genes for RAR and/or RXR. Fewer still have been associated with the induction of myeloid differentiation by the ligand. A candidate gene whose protein's function suggests an important role in this process is the cell cycle inhibitor p21 WAF1/CIP1 . p21 belongs to a group of recently characterized small proteins that interact with and inhibit the activity of cyclin-cyclin-dependent kinase (CDK) complexes; as such, they have been termed CDK inhibitors. Typically, CDK inhibitors cause cells to arrest in G 1 and withdraw from the cell cycle (10). Evidence is accumulating that CDK inhibitors are targets of extra-and intracellular signals that regulate cell growth, differentiation, and apoptosis. For example, radiationinduced DNA damage elevates p53 levels, and this protein in turn transcriptionally induces p21 expression, resulting in cell cycle arrest (11,12). Transforming growth factor-␤ acts as an anti-mitogenic factor by inducing p15 and/or p27 expression (13)(14)(15). MyoD up-regulates p21 expression, which is correlated with muscle cell differentiation (16). Finally, several agents * This work was supported in part by National Institutes of Health Grants DK45460 (to L. P. F.) and CA08748 (to Sloan-Kettering). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We described previously (20) a differential screening approach to discover vitamin D-responsive target genes in the myelomonoblastic cell line U937 which are directly responsive to 1,25(OH) 2 D 3 and the vitamin D 3 receptor (VDR), another member of the nuclear receptor superfamily. We reported that the p21 WAF1/CIP1 gene is in fact such a target: it is transcriptionally induced by 1,25(OH) 2 D 3 , and this induction facilitates the differentiation of U937 cells to monocytes. Since U937 cells are also induced to differentiate to monocyte/macrophages by RA, we speculated that p21 might also be a transcriptional target of RARs. In this report we demonstrate that p21 is indeed a RA-responsive target gene, and we describe a functional RA response element (RARE) in this gene's promoter which is required to confer RA induction through RAR⅐RXR heterodimers. These results correlate the RA induction of monocytic differentiation of U937 cells with the transcriptional activation of the p21 gene and suggest a role for this CDK inhibitor in facilitating this differentiation pathway.

MATERIALS AND METHODS
Cell Lines and Growth Media-U937 cells (clone 4) were a gift from K. Nilsson (Uppsala University, Sweden). Cells were routinely maintained in RPMI 1640 medium supplemented with L-glutamine and 10% heat-inactivated, charcoal-stripped, fetal bovine serum (lot 1468, Bio Gemini, CA). HaCaT cells, a human immortalized keratinocyte line (kindly provided by Dr. Norbert Fusenig, EMBL, Heidelberg, Germany), were grown in minimum essential medium supplemented with nonessential amino acids and 10% fetal bovine serum.
FACS and Cell Cycle Analysis-For FACS analysis of CD14 and CD11b cell surface antigen markers, early log-phase growing U937 cells were treated with 1 ϫ 10 Ϫ7 M ATRA or ethanol for 72 h. Approximately 1.0 ϫ 10 6 cells were harvested and washed twice with ice-cold PBS (magnesium-and calcium-free), resuspended in 50 l of PBS containing 0.5% bovine serum albumin and 0.5% sodium azide. 5.0 l of fluorescein isothiocyanate/phosphatidylethanolamine-conjugated CD11b or CD14 antibodies (Caltag) was added to the cell suspension; an isotypematched antibody control was included for each sample as a control for nonspecific binding. Cells were incubated in the dark on ice for 30 -45 min and then washed twice with ice-cold PBS, 1% bovine serum albumin and resuspended in 0.5 ml of this same buffer in a 5-ml polystyrene tube. Stained cells were collected by a Becton Dickinson FACScan, and data were analyzed by the LYSYS II program.
For cell cycle analysis, early log-phase growing U937 cells were treated with 1.0 ϫ 10 Ϫ7 M ATRA or ethanol for the indicated times. 2 ϫ 10 6 U937 cells were washed in ice-cold 1 ϫ PBS (magnesium-and calcium-free, 1% bovine serum albumin) and permeabilized with 70% ethanol at 4°C for at least 1 h. After centrifugation, cell pellets were resuspended in 1 ϫ 10 6 /ml of 1 ϫ PBS, treated with RNase A (1.0 g/ml), and stained with propidium iodide (50 g/ml) at 37°C for 30 min. DNA fluorescence was collected and analyzed using the Multicycle program from Phoenix Flow System. Cell viability remained greater than 95% in all cultures.
RNA and Protein Analysis-Total RNA was extracted from exponentially growing cells treated with ATRA for the indicated times, and Northern blot analysis for p21 was done as described (20). For immunoblot analysis, cells were lysed in 50 mM Tris-Cl, pH 7.4, 200 mM NaCl, 2 mM EDTA, 0.5% Nonidet P-40, 0.3 mM sodium orthovanadate, 50 mM NaF, 0.5 mM dithiothreitol, and protease inhibitors. Equal amounts of lysates (50 g) were boiled in SDS sample buffer, resolved by SDSpolyacrylamide gel electrophoresis, and immunoblotted with a polyclonal p21 antibody (Pharmingen). To measure Cdk2-associated retinoblastoma protein (RB) kinase activity, cell extracts from each time point were immunoprecipitated with a Cdk2 antibody (Pharmingen), and the immunoprecipitates were assayed for RB kinase activity as described previously (21).
Reporter Constructs and Transient Transfections-p21-pGL2 was generated by inserting a 2.4-kb HindIII fragment corresponding to the human p21 promoter derived from WWP-Luc (12) into the luciferase reporter plasmid pGL2-Basic (Promega). A series of p21 promoter de-letions (Ϫ2280, Ϫ1435, Ϫ837, Ϫ539, and Ϫ45; see Fig. 4A) was very generously provided as chloramphenicol acetyltransferase reporters by B. Vogelstein (Johns Hopkins Medical School) and has been described previously (34). This deletion series was transferred to pGL2-Basic by generating XhoI-HindIII fragments, blunt ending, and inserting them into a blunt-ended HindIII site of pGL2-Basic. 1.3p21 (also called ⌬1.1) was a gift of X.-F. Wang (Duke University Medical Center).
For transient transfections, HaCaT cells were used; DNA was introduced using the DEAE/dextran method. Cells seeded in 100-mm dishes were rinsed with serum-free minimal essential medium; 2.5 ml of this medium containing 25 l of 10 mM chloroquine was then added and allowed to sit for 2 min at room temperature. 500 l of a mix containing DEAE/dextran (12 mg/ml) and luciferase reporter DNA (10 -15 g/ transfection) was added to the cells and allowed to incubate for 3 h at 37°C. The medium was then removed, and 10% dimethyl sulfoxide in PBS was added (3 ml/plate) and incubated for 2 min. Cells were then rinsed with minimal essential medium ϩ 10% fetal bovine serum. 2 h later, cells were fed with fresh medium containing 10% charcoalstripped fetal bovine serum and incubated with 1 ϫ 10 Ϫ6 M ATRA or its solvent (ethanol) for 48 h. Cells were harvested and lysed. Total cellular extracts were then split and assayed for luciferase and ␤-galactosidase activities, as described previously (20).
Deletion Plasmids and Oligonucleotide-directed Mutagenesis-For oligo-directed mutagenesis, a 39-base oligonucleotide was synthesized as the reverse complement of the p21 sense strand excluding 47 bp corresponding to the putative p21 RARE region (Ϫ1212 to Ϫ1166) (5Ј-CAGAAATGAGTGATGTGTCTATTGCCTCCTTTCTGTGCC-3Ј) and used to anneal and extend the wild type single-stranded template generated from p21-pGL2. A second oligonucleotide that destroys a unique MluI site in pGL2 was also simultaneously annealed and extended. The extended products were used to transform the excision repair-defective Escherichia coli strain BMH 71-18; plasmid DNA was then prepared and subjected to restriction digestion by MluI. Mutagenized plasmids carrying the abolished MluI site remain supercoiled and therefore transform at much greater efficiencies than wild type plasmids with intact MluI sites which are linearized upon digestion. Digested DNA was used to transform JM109, and individual colonies were picked and analyzed by polymerase chain reaction sequencing using a p21 promoter-specific internal primer. The selection of mutant candidates as described results in a mutagenesis efficiency typically greater than 50%. In addition to this internal deletion, a smaller deletion mutant was generated which removed sequences corresponding to Ϫ1212 to Ϫ1194 using an oligonucleotide containing the sequence 5Ј-CTACCTCACACCCCTGACCTTGCCTCCTTTCTGTGCC-3. p21-pGL2 and the deletion derivatives, called p21⌬(Ϫ1212 to Ϫ1166)-Luc and p21⌬(Ϫ1212 to Ϫ1194)-Luc, respectively, were then used to transiently transfect HaCaT cells to test for ATRA responsiveness.
DNA Mobility Shift Assays--Gel shift assays with HaCaT nuclear extracts and purified RAR and RXR were carried out essentially as described (22,23). RAR-␣ was overexpressed and partially purified from baculovirus-infected Sf9 cells and was a gift from K. Ozato and J. Blanco (NIH); RARdbd was generously provided by D. Gerwirth and P. B. Sigler (Yale University). Flag-RXR was overexpressed in insect cells and affinity purified as described. 2 HaCaT nuclear extracts were a gift from A. Hata (Memorial Sloan-Kettering Cancer Center). The p21 RARE probe is a double-stranded oligonucleotide containing the sequence (top strand) 5Ј-AGCTTGGAGGCAAAGGTGAAGTCCAGGG-GAGGTCAGGGGTGA-3Ј. A positive control probe for RAR⅐RXR binding, called DR5, was also included (top stand): 5Ј-AGCTT-CAGGTCAACAGTAGGTCAGAGTGA-3Ј. This oligonucleotide was also used as a specific competitor in the HaCaT nuclear extract shifts; a glucocorticoid response element-containing oligonucleotide (23) was included as a control for nonspecific binding in these experiments.

RESULTS
Effect of all-trans-RA on U937 Cell Growth and Differentiation-U937 cells are induced to differentiate to monocytes by several agents, including 12-O-tetradecanoylphorbol-13-acetate, dimethyl sulfoxide, interferon-␥, RA, and 1,25(OH) 2 D 3 . To investigate the nature of the retinoid effect on U937 cells, we set out initially to examine the differentiation response of these cells to one retinoid, ATRA, which is specific for a subgroup of nuclear hormone receptors, RAR-␣, -␤, and -␥. To do so, we first determined the expression of monocyte/macrophage-specific markers expressed by differentiating U937 cells following addition of 10 Ϫ7 M ATRA. As indicated in Table I, continuous exposure to the ligand for 72 h elicited a very strong response by cells in the expression the cell surface antigen markers CD14 and CD11b, which are specific for mature, but not immature, monocyte/macrophages (24,25). This induction of differentiation markers was preceded by the growth arrest of these cells, which was observed within 30 h after the addition of ATRA as a block in G 1 (Table II). 9-cis-RA, a pan-agonist for both RARs and RXR, also resulted in a cell cycle arrest at G 1 , as did the RAR-specific analog LG272 and RXR-specific analog LG153 (data not shown). These synthetic compounds, however, appeared to induce a slightly weaker cell cycle block than ATRA or 9-cis-RA (data not shown).
Induction of p21 WAF1/CIP1 by ATRA-We demonstrated previously that 1,25(OH) 2 D 3 -induced U937 differentiation is mediated in part by the transcriptional induction of the CDK inhibitor p21 (20) by VDR. We therefore chose to examine p21 expression in U937 cells in response to ATRA. As shown in Fig.  1A, p21 mRNA levels are strongly induced by ATRA after 8 h of exposure to the ligand; this induction is sustained throughout a 72-h time course. Likewise, p21 protein levels are elevated within 30 h of the addition of ATRA (Fig. 1B). The effect of p21 induction by the ligand was also examined functionally by examining cyclin-CDK activity using a COOH-terminal construct of RB as a substrate for Cdk2. Cdk2-associated kinase activity appeared very high from 0 to 30 h following the addition of ATRA to U937 cells, after which it declined rapidly to essentially undetectable levels (Fig. 1C). This inhibition of Cdk2 activity correlated with the cell cycle data demonstrating an accumulation of cells in G 1 within 30 h of the ATRA addition (Table II).
Transcriptional Activation of p21 by ATRA--The relatively rapid kinetics of p21 mRNA induction in response to ATRA suggested that this up-regulation might be a transcriptional effect of RAR on the p21 promoter. It appears that p21 is transcriptionally regulated by a variety of factors, including p53, MyoD, Sp1, VDR, and STAT1 (12,16,19,20,26,27). To determine if ATRA stimulated transcription from the p21 promoter, we analyzed the responsiveness of a series of p21 promoter/reporter fusions following the transient transfection of a human keratinocyte cell line called HaCaT (28). The efficacy of using this cell line was 3-fold. First, its origins as a keratinocyte line made it a relevant cellular background with which to study the actions of ATRA, since it contains relatively high levels of retinoid receptors, and skin is a major target tissue of retinoids. Second, it is p53Ϫ/Ϫ, thereby eliminating any contribution of p53 effects on p21 transcription. Finally, it is more readily transfectable than U937 cells. As shown in Fig. 2, a 2.4-kb genomic fragment that includes the p21 promoter resulted in a 6-fold ATRA induction of luciferase expression following transfection of HaCaT cells. This induction was elevated further by cotransfecting a RAR-␣ expression plasmid; overexpression of RAR in HaCaT cells conferred a 13-fold induction in response to ATRA (Fig. 2). These results suggest that the p21 promoter is a transcriptional target of ligand-induced RARs.
Delineation of a p21 RARE-ATRA signals through RARs; these receptors act as direct transducers of the ligand by binding specifically to characteristic DNA sequences within promoters of regulated genes. RARs typically bind to direct repeat hexameric elements spaced by five nucleotides (DR5), called RA response elements (for review, see Ref. 29). To localize a putative RARE within the p21 promoter, a series of deletion constructs of the promoter was analyzed for luciferase inducibility in response to ATRA in HaCaT cells. The deletion series spanned the 2.4-kb p21 promoter region (Fig. 3A). ATRA induction was reasonably strong up to Ϫ1300 (defining ϩ1 as the start site of p21 transcription) but was lost in constructs from Ϫ837 to the TATA box, indicating that an element between Ϫ837 and Ϫ1300 was conferring the RA up-regulation (Fig. 3,  B and C).
A visual sequence scan of the promoter region corresponding to Ϫ1.3 to Ϫ0.8 kb revealed a putative RARE centered at Ϫ1.2 kb (Fig. 3A). Two internal deletions were generated in the context of the 2.4-kb promoter fragment: one 46 bp in length,   1. ATRA induces the expression of p21 mRNA and protein and inhibits Cdk2-associated RB kinase activity of U937 cells. Total RNA and whole cell extracts were prepared from U937 cells treated for the indicated times (in h) with 10 Ϫ7 M ATRA. Panel A, Northern blot analysis for p21 and actin (used as a control for loading). Panel B, immunoblot analysis with a specific antibody against p21. Panel C, Cdk2 immunoprecipitates were subjected to a RB kinase assay, loaded on SDS-polyacrylamide gel, and autoradiographed. which removed the DR5 and additional surrounding sequence (⌬1212-1166), and a second 17-bp deletion that removed only the suspected RARE (⌬1212-1194). In both cases, ATRA-mediated induction of the p21 promoter was abolished completely (Fig. 3C). These results suggest that the p21 sequence from Ϫ1212 to Ϫ1194 defines a functional RARE that is required to mediate RA induction of p21 transcription.
To determine if RAR can bind to this sequence element, in vitro gel mobility shift assays were carried out with both nuclear extracts, as well as purified RAR and RXR overexpressed in insect cells. To demonstrate that HaCaT cells indeed contain a binding activity that specifically recognizes the sequence centered at Ϫ1200 of the p21 promoter, gel shifts were carried out with nuclear extracts prepared from HaCaT cells and incubated with an oligonucleotide probe containing p21 promoter sequences from Ϫ1212 to Ϫ1194 (Fig. 4A). Shifted complexes were competed by an excess of a consensus RARE oligonucleotide, comprised of a AGGTCA repeat element spaced by five nucleotides (DR5; lane 4), but not a nonspecific competitor (glucocorticoid response element; lane 3), indicating that the complex contained RARs that are able to recognize and bind the p21 RARE selectively. Consistent with this, purified RAR and RXR bound to the p21 RARE (Fig. 4B). RAR alone bound quite weakly (Fig. 4B, lane 2); this pattern is typical of the fact that nuclear receptors such as RAR bind as high affinity species to properly spaced direct repeat elements primarily as heterodimeric complexes with RXR. A similar binding pattern was seen with RAR and RXR binding to a consensus DR5 element (lanes 5-8). That RAR dimerizes to the p21 RARE was apparent when gel shifts were carried out with the RARdbd (Fig. 4B). In contrast to the full-length receptor, dbd derivatives exist exclusively as monomers in solution and dimerize only on DNA (30). Thus, monomeric and dimeric species of the RARdbd were resolved upon binding to the p21 RARE and consensus DR5 positive control (lanes 1-3 and 4 -6, respectively), but not to a vitamin D response element (VDRE) we localized previously at Ϫ770 in the p21 promoter (20) (lanes 7-9). Although also organized as an imperfect direct repeat, the VDRE is spaced by three nucleotides and confers DNA binding specificity to a VDR⅐RXR heterodimer (31,32). Thus, although the RARdbd can recognize and bind to two properly spaced half-sites comprising the p21 RARE at Ϫ1200, it cannot dimerize to the VDRE because a 3-bp spacing confers only VDR dimer specificity (31). DISCUSSION Several inducers of myeloid cell differentiation appear to up-regulate p21 expression (10). These agents include phorbal esters, okadaic acid, sodium butyrate, dimethyl sulfoxide, vita- min D 3 , and retinoids. Therefore, the p21 promoter may contain multiple response elements for several differentiation signals, and RA-mediated growth arrest and the subsequent differentiation that follows may be due in part to the G 1 block of the cell cycle mediated by p21. By Northern analysis we detected strong induction of p21 mRNA 8 h after the addition of ATRA to U937 cells. This induction is most likely mediated at the level of transcription since the fusion of the p21 promoter to a luciferase reporter gene conferred RA responsiveness to the reporter gene in HaCaT cells. Importantly, this effect is p53independent since HaCaT are a p53Ϫ/Ϫ cell line. Utilizing a nested set of p21 promoter deletions beginning at Ϫ2.4 kb and extending to the transcriptional start site, we delineated a RARE organized as an imperfect DR5 centered at Ϫ1200. Consistent with its identity as a RARE, oligonucleotide probes containing this site bound purified, recombinant RAR and RXR in vitro specifically.
We reported previously finding a VDRE necessary to confer the up-regulation of p21 transcription following the addition of 1,25(OH) 2 D 3 to cells (20). Using the p21-Luc reporter we have observed inductions in response to both ATRA and 9-cis-RA added in combination with 1,25(OH) 2 D 3 (data not shown), but the combined treatments appeared additive rather than cooperative in their inductions. This suggested that RAREs and/or RXREs also exist as distinct elements within the p21 promoter or that VDR binds to this VDRE in a complex with RAR or RXR and that RA effects occurred through heterodimerization with VDR. However, RAR (as opposed to RXR) is not known to form heterodimers with VDR; we in fact found that the 1,25(OH) 2 D 3 effect is distinct from the RA induction, since a reporter carrying a 15-bp deletion that removes the VDRE still conferred strong RA induction (data not shown). This is consistent with our findings reported here that separate elements conferring responsiveness to RA and 1,25(OH) 2 D 3 exist in the p21 promoter, mediated by RAR⅐RXR and VDR⅐RXR heterodimers, respectively. It is not completely clear, however, whether 9-cis-RA⅐RXR homodimers might also regulate p21 expression. As mentioned, both ATRA and 9-cis-RA induce p21 transcription and U937 differentiation. Although induction by ATRA confirms the role of RARs in mediating this effect, 9-cis-RA is a pan-agonist that can signal through both RARs and RXRs. Thus, distinct response elements for RXR homodimers (i.e. DR1) may also be present in the p21 promoter. The use of RAR-specific and RXR-specific analogs should clarify this question.
Although it is well established that RA is a potent inducer of myeloid cell differentiation, its effect on U937 cells may be complex. For example, Oberg et al. (33) reported that whereas RA blocked U937 cell proliferation and induced monocyte-specific markers such as the expression of naphthol AS-D acetate esterase, these authors did not observe the induction of CD14 expression. They did, however, observe a strong increase in CD23 expression in response to RA. Interestingly, the inverse was observed for 1,25(OH) 2 D 3 . In this report we found that CD14, which is used as a distinguishing marker for monocytic differentiation, was in fact strongly induced by ATRA; levels of this cell surface antigen in U937 cells in response to ATRA were comparable to what we have observed previously for 1,25(OH) 2 D 3 (data not shown). The inhibition of cell growth by ATRA is at the G 1 /S phase of the cell cycle (Table II) and correlates with a loss of RB phosphorylation by Cdk2 (Fig. 1C).
In contrast to U937 cells, which are committed to differentiate along a monocyte/macrophage pathway in response to both 1,25(OH) 2 D 3 and RA, promyelocytic leukemia cells, such as HL-60 cells, will differentiate to granulocytes upon exposure to retinoids and to monocytes by 1,25(OH) 2 D 3 (3). An intriguing question is whether distinct subsets of genes, in addition to common targets such as p21, are induced by 1,25(OH) 2 D 3 versus retinoids, resulting in two different differentiation pathways. An examination of the regulation of emerging target genes in these cell lines in response to these two ligands should reveal a great deal about the mechanisms mediating the initiation of this complex differentiation process.  1-4) or the consensus DR5 RARE (lanes [5][6][7][8]. Both RXR and RAR␤ were baculovirus-expressed and partially purified from insect cells. 50 ng of each receptor alone, or in combination, was used. Panel C, RARdbd dimerizes specifically on the p21 RARE but not on a VDRE. RARdbd was expressed and purified from E. coli and incubated with probes corresponding to the p21 RARE (lanes 1-3), the DR5 element (lanes 4 -6), or a previously identified VDRE localized to a distinct subregion of the p21 promoter (Ϫ770; lanes 7-9). 1 denotes monomer, and 2 denotes bound dimer. In each series 0, 25, and 50 ng of purified RARdbd were used.