Retinoic acid causes MEK-dependent RAF phosphorylation through RARα plus RXR activation in HL-60 cells

C Blackwell Wissenschafts-Verlag 2001

Differentiation (2001) 68:55–66

ORIGINAL ARTICLE

Hee-Youn Hong ¡ Susi Varvayanis ¡ Andrew Yen

Retinoic acid causes MEK-dependent RAF phosphorylation through RARa plus RXR activation in HL-60 cells

Accepted in revised form: 29 April 2000

Abstract Retinoic acid (RA) is known to cause the myeloid differentiation of HL-60 human myeloblastic leukemia cells in a process requiring MEK-dependent ERK2 activation. This RA-induced ERK2 activation appears after approximately 4 h and persists until the cells are differentiated and G0 arrested (Yen et al, 1998). This motivates the question of whether RA also activated RAF as part of a typical RAF/MEK/MAPK cascade. Retinoic acid is shown here to also increase the phosphorylation of RAF, but in an unusual way. Surprisingly, increased RAF phosphorylation is first detectable after 12 to 24 hours by phosphorylation-induced retardation of polyacrylamide gel electrophoretic mobility. The RA-induced increased RAF phosphorylation is still apparent after 72 hours of treatment when most cells are differentiated and G0 arrested. There is a progressive dose-response relationship with 10ª 8, 10ª 7, and 10ª 6 M RA. The RA-induced RAF phosphorylation corresponds to increased in vitro kinase activity. Inhibition of MEK with a PD98059 dose which inhibits ERK2 phosphorylation and subsequent cell differentiation also inhibits RAF phosphorylation. RA-induced MEK-dependent RAF phosphorylation is not due to changes in the amount of cellular MEK. The induced RAF phosphorylation, as well as anteceding ERK2 activation, depends on ligand-induced activation of both an RARa receptor and an RXR receptor. This and the slow kinetics of activation suggest a need for prior RA-induced gene expression. In summary, RA induces a MEK-de-

✉

H.-Y. Hong1 ¡ S. Varvayanis ¡ A. Yen ( ) Department of Biomedical Science, College of Veterinary Medicine, Cornell University, Ithaca, NY 14853, USA e-mail: AY13/cornell.edu Tel: π 1 607 253 3354 Fax: π 1 607 253 3317 Present address: Korea Food & Drug Administration, 1582-4 Woosan-dong, Kwangsan-ku, Kwangju 506-050, South Korea

1

U. S. Copyright Clearance Center Code Statement:

pendent prolonged RAF activation, whose slow onset occurs after ERK2 activation but still well before cell cycle arrest and cell differentiation. The RA-induced increased RAF phosphorylation thus differs from typical mitogenic growth factor signaling, features that may contribute to cell cycle arrest and differentiation instead of division as the cellular outcome. Key words retinoic acid ¡ RAF ¡ MAPK ¡ HL-60 ¡ differentiation

Introduction Retinoic acid can regulate cell division, differentiation, and other physiological processes, making its mechanism of action of significant interest (reviews: Sklan, 1987; Blomhoff et al, 1990; Friedman and Sklan, 1993; Gudas, 1994). Its role as a morphogen regulating development during embryogenesis is well known. It is also a chemotherapeutic agent used in the differentiation induction therapy of acute promyelocytic leukemia (Flynn et al, 1983). Retinoic acid is rapidly metabolized intracellularly to a variety of retinoid metabolites (Gudas, 1994) which serve as ligands for the RAR and RXR subclasses of receptors (Mangelsdorf et al, 1990, 1992). In particular, retinoic acid is a ligand for RARs and can be isomerized to 9-cis retinoic acid which is a ligand for RARs and RXRs. RAR and RXR receptors form hetero- and homo-dimers and are ligand activated transcription factors in the steroid-thyroid hormone superfamily of nuclear receptors (Mangelsdorf et al, 1994; review). Recently retinoic acid has also been shown to activate MAPK, implicating MAPK signaling as another signal transduction pathway for retinoic acid-induced regulation of gene expression (Yen et al, 1998, 1999a, 1999b; Battle et al, 2000). Activation of the ERK2 MAPK is MEK dependent, suggesting involvement of the

0301–4681/2001/6801–55$ 15.00/0

56

RAF/MEK/MAPK kinase cascade in retinoic acid signaling. However, retinoic acid has hitherto not been shown to regulate RAF activation. Retinoic acid induces HL-60 human myeloblastic leukemia cells to undergo G0 cell cycle arrest and myeloid differentiation (Collins et al, 1977; Breitman et al, 1980; Yen, 1990: review). HL-60 cells are uncommitted precursor cells that will undergo G0 cell cycle arrest and either myeloid or monocytic differentiation. While retinoic acid induces myeloid differentiation, 1,25-dihydroxy vitamin D3 induces monocytic differentiation. For either of these agents, the induced metabolic cascade leading to onset of G0 arrest and differentiation occurs over a period corresponding to two division cycles, approximately 48 hours in the common subline studied. This period segregates into two segments which are each approximately the duration of one cell cycle: ‘‘early’’ events, leading to a precommitment state where cells are primed to differentiate without lineage specificity, and ‘‘late’’ events which specify whether the lineage elicited is myeloid or monocytic (Yen et al, 1984, 1987a, 1987b, 1990). A variety of molecular markers indicative of pending cell cycle changes occur during the early events, including down regulation of c-myc (Yen and Forbes, 1990) and the RB (retinoblastoma tumor suppressor) protein, which subsequently shifts from the hyper- to the hypo-phosphorylated state (Mihara et al, 1989; Yen et al, 1992a, 1993, 1994a, 1994b). In contrast, other cell cycle regulatory transcription factors, c-fos and c-myb, are down regulated during the late events prior to onset of G0 arrest and differentiation (Yen et al, 1992b). In the case of retinoic acid-induced myeloid differentiation, activation of both an RARa and an RXR is needed to elicit these cellular responses culminating in G0 arrest and myeloid differentiation (Brooks III et al, 1996). Retinoic acid thus elicits features of a cellular response characteristic of metabolic slow-down anteceding overt changes in cell cycle or differentiation. In contrast to the molecular effects characteristic of impending cell cycle arrest and differentiation, retinoic acid also causes MAPK activation in HL-60 cells (Yen et al, 1998). This activation is necessary for retinoic acid to elicit G0 arrest and differentiation. It is also selective for the ERK2 MAPK (Yen et al, 1999b). Retinoic acidinduced ERK2 activation leading to arrest and differentiation is striking in that MAPK signaling is the prototypical mitogenic signaling sequence utilizing the RAF/MEK/MAPK kinase cascade in response to peptide growth factors. The retinoic acid-induced ERK2 activation, however, differs in several respects from the anticipations of peptide growth factor induced MAPK signaling. It is not apparent until approximately 4 h after exposure to retinoic acid, which is relatively slow compared to MAPK activation that can be vigorous in 30 minutes or less due to growth factors. It also has a long duration and is still evident after 72 h when most cells are in G0 and differentiated. This is similar to the effects

of NGF on PC12 pheochromocarcinoma cells, which at the time of discovery were also considered anomalous and have since led to the hypothesis that prolonged MAPK signaling promotes differentiation in contrast to the brief signal which promotes cell division (Traverse et al, 1992). A similar paradigm is now motivated for retinoic acid-induced HL-60 cell differentiation, extending this conceptual model to steroid hormones and hematopoietic cells. In HL-60 cells, retinoic acid caused the MEK-dependent activation of ERK2, which is the predominant ERK in these cells (Yen et al, 1998), suggesting that retinoic acid activated the RAF/MEK/ MAPK signaling complex. However, it is not explicitly known that retinoic acid caused RAF activation or if the prototypical RAF/MEK/MAPK causal sequence is invoked. The potential involvement of RAF in retinoic acid-induced arrest and differentiation has already been indicated by the finding that ectopic expression of the RAF CR3 domain promotes retinoic acid-induced arrest and differentiation (Yen et al, 1994b). But these studies did not establish if RAF was downstream of retinoic acid. There is evidence that MAPK signaling does not always utilize the prototypical RAF/MEK/MAPK sequence. For example, MAPK can phosphorylate RAF in vitro (Anderson et al, 1991; Lee et al, 1992). MEK can also activate RAF in transient transfection assays (Zimmerman et al, 1997). However, it is not clear if RAF can be downstream of MEK or MAPK in a functional in vivo cellular context. The present communication pursues these questions in the case of retinoic acid-treated HL-60 myeloblastic leukemia cells induced to undergo G0 arrest and myeloid differentiation. In this context, retinoic acid causes RAF activation. The activation is MEK dependent, but does not apparently reflect changes in the amount of available MEK. Furthermore RAF activation occurs temporally after ERK2 activation. RAF activation is maximal when both an RARa and RXR ligand are presented. Thus like retinoic acid-induced G0 arrest and myeloid differentiation, RAF activation also depends on MEK, RARa, and RXR activation. Retinoic acid signaling thus causes sequential MEK, ERK2 and then RAF activation.

Methods Cells and culture conditions Human myeloblastic leukemia HL-60 cells were grown in RPMI 1640 (GIBCO, Grand Island, NY) supplemented with 5 % fetal bovine serum (Intergen, Purchase, NY). The cultures were passaged three times weekly to maintain logarithmic growth. For experiments, cells at an initial concentration of 2 ¿ 105 cells/ml were seeded in the medium and treated with specific compounds for the indicated times. Cell viability was assessed by exclusion of 0.2 % trypan blue and was routinely more than 95 % in all cultures. In RA-treated cultures, RA (.R2625, Sigma Chemical Co., St. Louis, MO) was added from a 1 mM stock in ethanol to make a final concentration of 1 mM at 0 h. Control cells not treated with

57 RA were treated with a carrier blank instead. The carrier blank had no detectable effect on cell proliferation, differentiation or RAF phosphorylation. In experiments where cells were treated with PD98059 (.9900, New England Biolabs, Beverly, MA), cell culture and PD98059 treatment were done as previously described (Yen et al, 1998). Briefly, PD98059 was added from a 2 mM stock in DMSO to make a final concentration of 2 mM, with three successive additions at – 16, 0 and 16 h. Cells were harvested at the indicated times for Western analysis of RAF or MEK, as well as assaying RAF in vitro kinase activity. For in vitro RAF kinase activity assays and also MEK or RAF co-immunoprecipitation assays, identical cultures, manipulated with or without PD98059 or RA as described above, were initiated at staggered times to allow their simultaneous harvest. The kinase assays or immunoprecipitations were thus performed simultaneously for all time points being compared. In experiments to determine the effects of the RAR and RXR selective retinoid ligands, cells were initiated in culture with 1 mM RAR or RXR selective retinoid ligands, as well as a combination thereof. The ligands were AM580, an RARa selective retinoid; CD437, an RARg selective retinoid; Ro25 – 6603, a pan-RXR selective retinoid; and Ro41 – 5253, an RARa antagonist. In addition to being used singly, the combination of an RARa selective plus an RXR ligand (AM580 plus Ro25 – 6603) was used. Cultures were initiated at a cell density of 2 ¿ 105 cells/ml with no addition (control) or 1 mM RA or 1 mM ligands. The ligands were added to cultures from stock solutions of 5 mM in ethanol, stored at ª 70 æC, protected from light. The cells were harvested for Western analysis of RAF after 72 h. In experiments to examine the effect of PKC inhibition on the RAinduced RAF phosphorylation, three PKC inhibitors, staurosporine (.S4400, Sigma Chemical Co., St. Louis, MO), calphostin C (.EI198, Biomol Research Laboratories, Inc., Plymouth Meeting, PA) and GF109203X (.EI246, Biomol Research Laboratories, Inc., Plymouth Meeting, PA), were used. HL-60 cells were initiated in culture with or without RA or a PKC inhibitor for 48 h. The concentrations of inhibitors used were 2.5 nM staurosporine, 20 nM GF109203X, and 10 nM calphostin C. The IC50 of staurosporine (Tamaoki et al, 1986; Matsumoto and Sasaki, 1989), GF109203X (Toullec et al, 1991), and calphostin C (Kobayashi et al, 1989a, 1989b) are 0.7, 20, and 50 nM, respectively. Staurosporine was added to cultures from a stock solution of 0.1 mM in ethanol. Stocks of 1 mM calphostin C and 1 mM GF109203X were prepared in DMSO. Stock solutions were stored at – 20 æC. Before use, each stock solution was diluted with culture medium to the required concentration. The total amount of solvent in the cultures was no more than 0.15 % upon drug treatment. At the doses used, the drugs did not cause cellular toxicity during the period of observation. Cells were harvested at the indicated times for Western analysis of RAF. Western Blotting of RAF or MEK Western blotting of RAF was done using whole cell lysates from cells cultured as indicated. At indicated times, 1 ¿ 106 cells were harvested and fixed in 1 ml of 90 % methanol at – 70 æC as previously described (Yen et al, 1992a, 1993). The cells were stored at – 20 æC until analysis by SDS-PAGE. Cells were solubilized in 50 ml of loading buffer (6 % SDS, 4 M urea, 4 mM EDTA, 125 mM Tris (pH 6.9), 0.25 % bromphenol blue, 35 ml/ml b-mercaptoethanol) by immersion in a boiling water bath for 5 minutes. SDSPAGE was done using a 4 % stacking gel and a 10 % resolving gel with 37.5:1 acrylamide:bis. Samples were electrophoresed at 75 volts for 16 h. 1 ¿ 106 cells were loaded per lane. Proteins were electrotransferred (Trans Blot Cell, Bio-Rad Inc., Hercules, CA) from the gel to a nitrocellulose membrane (0.45 mm, NitroBind, Micron Separation Inc., Westborough, MA). Transfer was done at 0.4 amps for 2 h. To confirm that lanes were loaded uniformly, the membrane was stained with Ponceau S (0.1 % Ponceau S, 5 % acetic acid solution). The Ponceau S was then washed out with distilled

water. The resulting membrane was blocked by immersion overnight at 4 æC in 5 % nonfat dried milk and 0.05 % Tween 20 in PBS. The membranes were probed with an antibody detecting the phosphorylated and unphosphorylated forms of RAF. The antibody to detect RAF (C-12, .SC133 rabbit polyclonal antibody, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) was used at 0.5 mg/ml in 5 % BSA and 0.05 % Tween 20 in PBS with a 1 h incubation at room temperature. Detection was performed using a horseradish peroxidase conjugated secondary anti-rabbit antibody and enhanced chemiluminescence (ECL kit, Amersham Ltd., Piscataway, NJ) following the manufacturer’s instructions. Densitometric analyis of phosphorylated and unphosphorylated RAF bands was done using a CAMAG TLC scanner 3 interfaced to CAMAG ‘‘CATS’’ data acquisition and analysis software (CAMAG, Muttenz, Switzerland). The densitometer settings used were as follows. Slit dimension: 2 ¿ 0.1 mm. Wavelength: 595 nm. Monochronometer bandwidth: 20 nm. Measurement mode: absorption/ reflection. Integration parameter, peak threshold, slope: 5. Data was analyzed using an EXCEL spreadsheet to derive means and standard deviations. The procedure for Western blotting of total cellular MEK using whole cell lysates was the same as described for Western blotting of RAF, except for using an antibody to detect MEK (12-B, .SC436 rabbit polyclonal antibody, Santa Cruz Biotechnology, Inc., Santa Cruz. CA) which was used at 0.4 mg/ml in 5 % BSA and 0.05 % Tween 20 in PBS with a 1.5 h incubation. RAF Kinase Assay The in vitro kinase assay for RAF was performed essentially as previously described (Davis et al, 1993; Kharbanda et al, 1994). Briefly, following the indicated treatments, 4 ¿ 106 cells were washed twice in 1 ml ice-cold PBS and resuspended in 300 ml lysis buffer containing 25 mM Tris (pH 7.6), 25 mM b-glycerophosphate, 3 mM EDTA, 3 mM EGTA, 250 mM NaCl, 1 % Triton X100, 2 mM dithiothreitol, 2 mM sodium orthovanadate, 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml aprotinin, and 10 mg/ml leupeptin (added freshly before use) for 15 minutes on ice. The whole cell lysate was centrifuged at 15,000 rpm for 15 minutes at 4 æC to remove insoluble cell materials. RAF in the supernatant was immunoprecipitated (IP) for 4 hours at 4 æC by adding 10 ml of a RAF specific antibody (C-12, .SC133 rabbit polyclonal antibody, Santa Cruz Biotechnology, Inc., Santa Cruz, CA) and 10 ml of Protein A (insoluble, .P7155, Sigma Chemical Co., St. Louis, MO), with gentle mixing. The IPs were washed three times in 0.5 ml lysis buffer and once in 0.5 ml kinase buffer containing 20 mM Tris (pH 7.6), 20 mM magnesium chloride, 25 mM b-glycerophosphate, 1 mM sodium orthovanadate, and 2 mM dithiothreitol. The IP complex was then resuspended in 30 ml of kinase buffer containing 50 mM of ATP with 10 mCi of g-32P ATP (.NEG035C, 150 mCi/ml, NEN Life Science Products, Inc., Boston, MA) and 25 mg of a histone (Type III-S, .H5505, Sigma Chemical Co., St. Louis, MO). The kinase reaction proceeded for 30 minutes at room temperature with frequent mixing. The reaction was terminated by adding 3x SDS-gel sample buffer and boiling for 5 min. The reaction mixture was centrifuged for 3 min, and the supernatant was then subjected to SDS-PAGE, using a 4 % stacking gel and a 12.5 % resolving gel. The reaction products were visualized by autoradiography.

Results Retinoic acid causes an increase in the amount of phosphorylated RAF. HL-60 cells were initiated in culture with or without 10ª 6 M retinoic acid. The control cells without retinoic acid were cultured with a carrier blank.

58

Fig. 1 (A) Western blot of RAF expression in parallel retinoic acidtreated (RA) and untreated carrier blank control (C) HL-60 cells at indicated times (h) after initiation of culture. Phosphorylation of RAF retards its gel mobility. Enrichment of the slower migrating species is first discernable at ca. 12 h and persists through 72 h when the cell population is largely differentiated and G0 arrested. Lysate from equal numbers of cells were loaded in each PAGE lane and blots were stained with Ponceau S to confirm uniform lane loading. The untreated control cultures (C) contained a carrier blank instead of retinoic acid. (B) The average of the ratio of phosphorylated to unphosphorylated RAF for control (horizontal stripes) and RA treated (diagonal stripes) cells at indicated times after initiation of culture. The values are derived as the average of phosphorylated to unphosphorylated RAF from the replicate cases reported in Table 1. RA induces an increase from approximately 1 to approximately 3

At the indicated times cells were harvested and the lysates were collected for Western analysis of RAF. Figure 1A shows a Western blot of RAF in retinoic acid-treated and untreated HL-60 cells at sequential times. For each treatment and time, Table 1 summarizes the data from replicate blots and reports the mean percentage of the total RAF in either the phosphorylated or the unphosphorylated state measured with a scanning densitometer; the standard deviation; and the number of replicates. Figure 1B shows the mean of the ratio of phosphorylated to unphosphorylated RAF for untreated and treated cells at sequential times in culture. Phosphorylation of RAF retards its relative gel mobility on PAGE (Cai et al, 1993;

Morrison et al, 1993), making the phosphorylated RAF apparent as a slower migrating species. The lanes are loaded with equal numbers of cells, and the blot was stained with Ponceau S to confirm uniform loading. Enhanced RAF phosphorylation due to retinoic acid is first apparent when comparing retinoic acid-treated to untreated control cells at approximately 12 h, but not at the earlier times measured. By 24 h of RA treatment there is a clear phosphorylation-induced gel retardation of RAF showing a relative loss of unphosphorylated RAF and an increase in the amount of phosphorylated RAF, compared to control cells without retinoic acid. The enhanced amount of phosphorylated RAF in retinoic acid-treated

Table 1 Percentage of RAF in Phosphorylated and Unphosphorylated States

% Phos. RAF % Unphos. RAF S.D. n

C0

C2

R2

C6

R6

C12

R12

C24

R24

C48

R48

C72

R72

52 48 7 5

55 45 2 5

56 44 3 5

54 46 2 3

54 46 1 3

55 45 2 3

59 41 1 3

54 46 2 7

62 38 4 7

44 56 4 7

65 35 3 7

41 59 5 7

73 27 6 7

C designates carrier control and R designates RA treated at 0, 2, 6, 12, 24, 48, and 72 hours. S.D. is the standard deviation. n is the number of replicates measured. Percentages rounded to nearest integer

59 Table 2 Percentage of RAF in Phosphorylated and Unphosphorylated States

% Phos. RAF % Unphos. RAF S.D. n

Control

10ª 8 M RA

10ª 7 M RA

10ª 6 M RA

43 57 12 3

56 44 2 3

62 38 6 3

68 32 12 3

Control designates carrier control and RA designates RA treated with the indicated concentration. S.D. is the standard deviation. n is the number of replicates measured. Percentages rounded to nearest integer

Fig. 2 (A) Western blot of RAF expression in cells treated with 10ª 8, 10ª 7, 10ª 6 M retinoic acid. Cells were cultured for 72 h with varying doses of retinoic acid. There is a dose response relationship between retinoic acid and induced RAF phosphorylation revealed by retardation of gel mobility. The untreated control (C) was cultured with a carrier blank without retinoic acid. (B) The average of the ratio of phosphorylated to unphosphorylated RAF for control (Con) and 10ª 8, 10ª 7, 10ª 6 M retinoic acid treated cells. The values are derived as the average of phosphorylated to unphosphorylated RAF from the replicate cases reported in Table 2

compared to RA-untreated control cells persists through 72 h. In control cells, the amount of activated RAF changes over time as cells are placed in new medium, increasing slightly when cells were placed in new medium and then decreasing after approximately 48 h of culture as the medium become nutritionally depleted. (Unlike retinoic acid treated cells which arrest in G0, controls cells attain higher cell densities as they continue to proliferate.) The retinoic acid-induced changes are thus evidenced as deviations from the background changes due to culture. The onset of G0 cell cycle arrest and differentiation in reti-

noic acid-treated cells is not apparent until approximately 48 h (Brooks III et al, 1996; Yen et al, 1998). The retinoic acid-induced increase in phosphorylated RAF thus begins well before any apparent changes in cell cycle or differentiation. The enhanced amount of phosphorylated RAF also persists until most cells are G0 arrested and differentiated at 72 h (Brooks III et al, 1996; Yen et al, 1998). The induced enhancement in phosphorylated RAF signaling thus has a long duration and continues in cells as they G0 arrest. Compared to retinoic acid-induced ERK2 activation, the enhancement in RAF phosphorylation is not apparent until after ERK2 is activated. ERK2 activation by retinoic acid is first apparent by approximately 4 h (Yen et al, 1998), but retinoic acid-induced RAF phosphorylation occurs later. The retinoic acid-induced increase in phosphorylated RAF is dose dependent and is detectable at an in vitro biologically low dose of 10ª 8 M. HL-60 cells were initiated in culture with 10ª 8, 10ª 7, and 10ª 6 M retinoic acid and harvested after 72 h for Western analysis of RAF. Figure 2A shows a RAF Western blot of the harvested lysates. For each treatment, Table 2 summarizes the data from replicate blots and reports the mean percentage of the total RAF in either the phosphorylated or the unphosphorylated state measured with a scanning densitometer; the standard deviation; and the number of replicates. Figure 2B shows the mean of the ratio of phosphorylated to unphosphorylated RAF for each dosage. There is a progressively increasing phosphorylation-induced retardation in gel mobility as the retinoic acid concentration increases by logs from 10ª 8 M to 10ª 6 M, which is the standard dosage used to induce G0 arrest and differentiation of these cells. 10ª 8 M retinoic acid is a very weak inducer of HL-60 cell differentiation and causes almost no G0 arrest (Yen et al, 1997), while 10ª 6 M causes G0 arrest and differentiation of essentially the entire population. There is thus a rough correlation between induced RAF phosphorylation enhancement and ultimate downstream cellular effects. The retinoic acid-induced increase in RAF phosphorylation also reflects increased in vitro kinase activity. HL-60 cells were initiated in culture with or without

60

Fig. 3 RAF in vitro kinase activity for parallel retinoic acid-treated (RA) and untreated (C) cells at indicated times since initiation of culture. Kinase activity of immunoprecipitated RAF was measured using a histone substrate. Retinoic acid induces increased RAF kinase activity corroborating the induced RAF phosphorylation

10ª 6 M retinoic acid for the indicated times and harvested. RAF was immunoprecipitated and the kinase activity measured using a histone substrate. Figure 3 shows the autoradiograph of the RAF in vitro phosphorylated substrate. As with RAF phosphorylation, the RAF kinase activity increases over the 72 h treatment period in retinoic acid-treated cells compared to untreated controls. As part of a different series of experiments to be discussed below, Fig. 7 also shows that retinoic acid increases RAF kinase activity compared to controls. The retinoic acid-induced increase in RAF in vitro kinase activity thus generally corroborates the increased in vivo phosphorylated RAF observed above. RAF phosphorylation is MEK dependent in control and retinoic acid-treated cells. HL-60 cells were treated with PD98059, a MEK-specific inhibitor, using a proto-

Fig. 4 RAF Western blot for parallel retinoic acid-treated (RA) or untreated (C) cells with (π) or without the MEK inhibitor, PD98059. The PD98059 treatment (2 mM added at ª 16, 0 and 16 h) was previously shown to inhibit ERK2 phosphorylation (Yen et al, 1998). PD98059 inhibits RAF phosphorylation and blocks the retinoic acid (added at 0 h)-induced increase in phosphorylated RAF. The retinoic acid-induced increase in phosphorylated RAF is thus MEK dependent

col previously shown to inhibit MEK with minimal cellular toxicity (Yen et al, 1998). This treatment blocked retinoic acid-induced differentiation and arrest. Cells were either pretreated or not for 16 h with 2 mM PD98059 and initiated in culture with or without 10ª 6 M retinoic acid. In PD98059 treated cells, an additional 2 mM PD98059 was added to the cultures at this time and 16 h thereafter, resulting in four cases: untreated, retinoic acid-treated, PD98059-treated, and PD98059 plus retinoic acid-treated. While the PD98059 dosage was low compared to the typically used 50 mM used to suppress growth factor signaling, its continued exposure at the low dosage suppressed ERK2 activation without significantly impairing growth over the period of observation (Yen et al, 1998). Cells were harvested at the indicated times for Western analysis of RAF. Figure 4 shows the Western blot of RAF at the indicated times until 96 h. At this time the retinoic acid-treated cell population is almost completely in G0 and differentiated (Brooks III et al, 1996; Yen et al, 1998). At 0 h when cells were initiated in culture with or without retinoic acid, the amount of phosphorylated RAF is already reduced by PD98059. Inhibiting MEK thus shifts RAF from the phosphorylated toward the unphosphorylated state. In retinoic acid-treated cells, PD98059 inhibits the usually induced increase in phosphorylated RAF. This effect persists over the 96 h course of the experiment. During this time cells treated with retinoic acid without PD98059 showed an increased amount of phosphorylated RAF compared to untreated cells. Thus retinoic acid-induced increases in RAF phosphorylation are

61

MEK-dependent. Likewise RAF phosphorylation in control cells is also MEK-dependent. The simplest rationalization of the results is that retinoic acid stimulates a normally present MEK-dependent RAF phosphorylation in these cells. The inhibition of RAF phosphorylation by PD98059 correlates with the subsequent failure of these PD98059 treated cells to differentiate and arrest as usual after retinoic acid treatment (Yen et al, 1998). Retinoic acid-induced changes in RAF phosphorylation are not attributable to changes in the amount of cellular MEK. Nor does PD98059 alter the amount of cellular MEK. PD98059-treated or untreated HL-60 cells were cultured with or without retinoic acid as above. At the indicated times cells were harvested for Western analysis of MEK. Figure 5 shows the Western blot of MEK at the indicated times for the four cases: untreated, PD98059-treated, retinoic acid-treated, and PD98059 plus retinoic acid-treated. There is no apparent change in the amount of MEK comparing these cases at either 48 or 72 h when the differences in RAF phosphorylation are clearly evident. The antibody used detects both MEK1 and MEK2; however, the predominant species expressed in HL-60 cells is MEK1. The induced changes in RAF phosphorylation, although MEK-dependent, are not attributable to changes in the total amount of cellular MEK. The induced increase in RAF phosphorylation depends on presentation of both RARa and RXR receptor activating ligands. HL-60 cells were initiated in culture with 10ª 6 M RAR or RXR selective retinoid ligands, as well as a combination thereof. The ligands were AM580, an RARa selective retinoid; CD437, an RARg selective retinoid; Ro25 – 6603, a pan-RXR selective retinoid; and Ro41 – 5253, an RARa antagonist. In addition to being used singly, the combination of an RARa selective plus an RXR ligand (AM580 plus Ro25 – 6603) was used. AM580 and Ro25 – 6603 are highly selective as indicated

Fig. 5 Western blot of total cellular MEK in retinoic acid-treated (RA) and untreated (C) cells with (π) or without PD98059. The amount of MEK per cell does not change in response to either retinoic acid or PD98059, added singly or together. Increased phosphorylated RAF due to retinoic acid thus depends on MEK but does not reflect changes in total available MEK

by their IC50 values (Brooks III et al, 1996, reports the IC50 values and gives the chemical structures). The cells were harvested for Western analysis of RAF after 72 h when retinoic acid-induced increased RAF phosphorylation has long been evident. Figure 6A shows the resulting Western blot. As above, retinoic acid shifts RAF toward the phosphorylated state. The combination of both the RARa selective ligand plus the pan-RXR ligand also causes an increase in RAF phosphorylation, which is comparable to that caused by retinoic acid. In contrast the RARa selective ligand, RARg selective ligand, the pan-RXR selective ligand, and the RARa antagonist used singly are not as effective. The RARg ligand was used because HL-60 cells do not have RARg receptors, and an effect with this ligand would implicate a non-nuclear RAF activating function of retinoids. Likewise, an effect with the RARa antagonist would implicate RAF activation without transcriptional activation. Activation of both an RARa and an RXR thus leads to the RAF activation seen with retinoic acid. Consistent with the involvement of both RARa and RXR in activating this signaling pathway, presentation

Fig. 6 (A) Western blot of RAF expression in cells treated with RAR or RXR receptor selective ligands. The selective ligands were RARa (AM580), RARg (CD437), or RXR (Ro25 – 6603) selective, as well as an RARa (Ro41 – 5253) antagonist. The lanes are designated C, untreated control; RA, retinoic acid; A, RARa selective; X, RXR selective; ant., RARa antagonist; G, RARg selective. All selective ligands were used at 10ª 6 M as was retinoic acid. Treatment was for 72 h. The combination of RARa and RXR selective ligands results in increased phosphorylated RAF, as does all-trans retinoic acid; but all single ligands are much less effective. The RARg selective ligand is a control since there is no RARg receptor in HL-60 cells. (B) Western blot of activated ERK2 for similarly RA- or ligand- treated cells. The results parallel previously published data (Brooks III et al, 1996) showing that activation of both RARa and RXR is needed to elicit cell differentiation and G0 arrest. Taken together, the data suggest that the induced increase in RAF phosphorylation involves transcription since single RAR or RXR ligands are not very effective, but addition of both RAR and RXR activating ligands does work like retinoic acid to increase phosphorylated RAF

62

Fig. 7 RAF in vitro kinase activity for parallel retinoic acid-treated (RA) or untreated (C) cells with (π) or without the MEK inhibitor, PD98059. The PD98059 treatment was previously shown to inhibit ERK2 phosphorylation (Yen et al, 1998). Retinoic acid causes increased RAF kinase activity, but PD98059 does not inhibit the retinoic acid-enhanced RAF kinase activity in vitro using a histone substrate. Interestingly, the in vitro kinase activity does not correlate with the in vivo RAF phosphorylation in the case of PD98059treated cells

of both RARa and RXR activating ligands is also needed to induce MAPK activation. Retinoic acid is known to induce MEK-dependent MAPK activation (Yen et al, 1998), which antecedes RAF phosphorylation. Figure 6B shows a Western blot of activated ERK2 for cells treated with RA or the receptor selective ligands as in Fig. 6A. Activated ERK2 was detected with an antibody specific for the phosphorylated T(183) and Y(185) within the TEY motif of activated ERK2 and parallels the in vitro kinase activity as well as the phosphorylation-induced mobility shift of ERK2 (Yen et al, 1998). Since neither the RARa nor RXR ligand by itself is as effective at causing RAF or ERK2 activation, but the combination is as effective as retinoic acid, the data argue against a non-nuclear effect of the ligands in causing RAF activation or ERK2 activation. This is consistent with the ineffective RAF activation by the RARg ligand and the RARa antagonist. It was not possible to directly test the need for transcription or translation for retinoic acid to induce ERK2 or RAF activation using actinomycin D or cycloheximide because, as reported previously (Yen et al, 1998), the drugs cause loss of MAPK signaling in control cells and have toxicities, precluding their use to this end in these cells. Taken together, the data are consistent with a need for retinoidregulated transcription dependent on both RARa and RXR transcription factors, which are known to act as both hetero- and homo-dimers, in order to cause RAF activation. An ancillary observation made in the course of the experiments using retinoic acid plus PD98059 is that RAF in vitro kinase activity using a histone substrate does not correlate with the in vivo RAF phosphorylation in PD98059-treated cells. HL-60 cells were cultured with or without PD98059 or retinoic acid, generating the same four cases as for Fig. 4. The cells were harvested

and the kinase activity of immunoprecipitated RAF was assayed. Figure 7 shows the autoradiograph of the phosphorylated histone substrate. As before, retinoic acidtreated cells show greater RAF kinase activity. But the PD98059-treated cells yield immunoprecipitated RAF with no consistent reduction of in vitro kinase activity compared to their corresponding case without PD98059. In contrast to the case of retinoic acid-induced RAF phosphorylation in vivo, PD98059 causes no inhibition of the retinoic acid-induced increase in RAF in vitro kinase activity. One of the repeats of this experiment extended the measurements to 72 h, and the pattern of histone phosphorylation for C, RA, C π, and RA π was indistinguishable from that shown for 48 h. The results are enigmatic, and we can only speculate that they reflect the differences between the in vivo and in vitro assays. In particular, it appears that in vitro assays of RAF activation may not necessarily bear fidelity to in vivo assays. It is possible that the PD98059 dissociated from the putative RAF/MEK/MAPK complex during the process of cell harvesting, lysing, and immunoprecipitation, allowing the MEK to sufficiently activate RAF during the time required for these experimental manipulations. However, there is an alternative interpretation, if one accepts the fidelity of the assay. In this case, if in fact there is no correlation between either the amount of MEK and RAF phosphorylation, as shown above, or the inhibition of MEK activity and RAF in vitro kinase activity; then the data suggest that while RAF phosphorylation may depend on MEK, MEK is not the kinase activating RAF; and another kinase may be involved. In the experiments below, the data suggest that PKC, although it is well known to participate in MAPK signaling in many other contexts, is unlikely to be this other kinase. Retinoic acid-induced RAF phosphorylation is relatively insensitive to three PKC inhibitors, staurosporine, GF109203X, and calphostin C. HL-60 cells were initiated in culture with or without retinoic acid or a PKC inhibitor for 48 h. The concentrations of inhibitors used were 2.5 nM staurosporine, 20 nM GF109203X, and 10 nM calphostin C. The IC50 of staurosporine, GF109203X and calphostin C are 0.7, 20, and 50 nM respectively (Tamaoki et al, 1986; Kobayashi et al, 1989a, 1989b; Matsumoto et al, 1989; Toullec et al, 1991). The drug concentrations used were chosen to avoid overt toxicity over 48 h and also, in the case of staurosporine, to minimize induced cell adherence that would impede their harvest. At the doses used, the drugs did not significantly inhibit cell growth during the period of observation, thus avoiding any overt toxicity. Cells were harvested at the indicated times for Western analysis of RAF. Figure 8 shows the resulting Western blot of RAF. None of the three PKC inhibitors used prevent the retinoic acid-induced increase in phosphorylated RAF. Interestingly, the inhibitors by themselves had varying effects on RAF expression and phosphorylation, comparing inhibitor-treated to untreated cells, al-

63

Fig. 8 Western blot of RAF in HL-60 cells that were untreated (C) or retinoic acid-treated (RA) in the presence of no inhibitor or of the PKC inhibitors, 10 nM calphostin C (CPC), 20 nM GF109203X (GFX), or 2.5 nM staurosporine (STSP), for 48 h when retinoic acid-induced RAF phosphorylation is typically evident. None of the inhibitors block the retinoic acid-induced RAF phosphorylation

though they did not block the retinoic acid-induced phosphorylation shift. In contrast to what is often the case for growth factor-induced activation of the RAF/ MEK/MAPK complex, the retinoic acid-induced RAF activation is relatively insensitive to these PKC inhibitors.

Discussion The presented results show that retinoic acid causes a prolonged MEK-dependent enhancement in RAF activation. The retinoic acid-induced enhanced RAF activation surprisingly occurs hours after retinoic acidinduced ERK2 activation is known to occur. While retinoic acid-induced RAF phosphorylation is MEK dependent, the amount of MEK does not appear to be significantly affected by retinoic acid, and the identity of the kinase that directly activates RAF remains to be defined with certainty. This retinoic acid-induced RAF activation, as well as anteceding ERK2 activation, depend on both RARa and RXR activation, suggesting a dependence on transcriptional regulation. The process does not appear to be strongly dependent on PKC. The data are consistent with a model in which retinoic acid causes activation of the RARa and RXR ligand activated transcription factors, resulting in MEK-dependent activation of ERK2 and finally RAF. Katagiri et al (1994) reported that retinoic acid does not cause RAS activation in HL-60 cells, consistent with atypical RAF activation by another means in this context. The retinoic acid-induced activation of MAPK signaling provides another signaling route to regulate transcription in addition to the RAR and RXR retinoid-activated receptors which are usually used to explain transcriptional changes induced by retinoic acid. In addition to transcription factors, such as of the Ets family, activated by

MAPKs, there may also be other relevant targets downstream of ERK2. For example, one potentially relevant target of MAPKs is the RSK kinase family, members of which (RSK-2) can cause histone (histone H3) phosphorylation and potentially influence transcription through chromatin remodeling (Sassone-Corsi et al, 1999). Previous studies showed that MAPK signaling is necessary for retinoic acid to induce HL-60 differentiation and cell cycle arrest (Yen et al, 1998), and the present results show its unusual character. The involvement of MAPK signaling to propel retinoic acid-induced growth arrest and differentiation is contrary to expectations based on growth factor signaling by peptide growth factors, where such signaling is typically mitogenic. However, it is not without historical basis. We have previously shown that retinoic acid causes ERK2 activation. This ERK2 activation is MEK dependent and necessary for induced arrest and differentiation (Yen et al, 1998, 1999b). There have been other indications that enhanced MAPK signaling can positively regulate retinoic acid-induced arrest and differentiation. Ectopic expression of the RAF CR3 domain, the catalytic domain without the amino terminus negative regulatory regions, accelerates retinoic acid-induced arrest and differentiation (Yen et al, 1994b). Ectopic expression of a c-FMS transgene also promotes retinoic acid-induced arrest and differentiation (Yen et al, 1996, 1997). c-FMS is a transmembrane tyrosine kinase receptor for CSF-1 which initiates a signaling cascade that involves MAPK. Most of the signaling molecules activated by c-FMS can also be activated by the polyoma middle T antigen. Ectopic expression of polyoma middle T antigens also accelerates retinoic acid-induced arrest and differentiation (Platko et al, 1998; Yen et al, 1999a). While previous data indicated that MEK and ERK2 were downstream components of retinoic acid signaling, they did not explicitly demonstrate the involvement of RAF. The present data indicate that RAF activation is also downstream. Retinoic acid thus activates RAF/ MEK/MAPK signaling, but not in that prototypical sequence. These data suggest that RAF/MEK/MAPK signaling can drive either cell division or differentiation and arrest. The outcome of division or differentiation appears to depend on cellular context. In the case of retinoic acid-induced HL-60 cell differentiation, it may be the other retinoid-activated RARa and RXR signals that contribute to the cellular context, making MAPK drive cell differentiation instead of division. For example, the RARa and RXR signals may regulate the relative abundance of different MAPK substrates, thereby shifting the cellular outcome of MAPK signaling by shifting substrate availability. The retinoic acid-induced activation of the putative RAF/MEK/MAPK signaling complex has several unanticipated features that distinguish its involvement in retinoic acid signaling compared to peptide growth factor signaling. One is that the sequence is not the prototypi-

64

cal causal RAF/MEK/MAPK sequence. The temporal sequence instead appears to be MEK, ERK2 and then RAF. Our studies do not demonstrate which kinase phosphorylated RAF. However, there have been reports that at least in vitro RAF can be phosphorylated by MAPK (Anderson et al, 1991; Lee et al, 1992; Zimmerman et al, 1997). The present data demonstrate that there is an in vivo context where this may be functionally employed in retinoic acid signaling. Another unanticipated feature is the long 4 h lag from presentation of retinoic acid to activation of ERK2, and the still further (ca. 12 h) lag before activation of RAF. This may reflect the need to transcriptionally activate adaptor molecules that need to be recruited to the complex. Consistent with this possibility, retinoic acid causes the up-regulation of paxillin expression approximately within this time frame (Platko and Yen, 1997). Another possibility is that another kinase capable of phosphorylating RAF is involved. Akt, for example, can bind and phosphorylate RAF, although Akt per se is unlikely to be involved here since it appears to inhibit RAF/MEK/MAPK signaling (Rommel et al, 1999; Zimmerman and Moelling, 1999). Finally, the long duration of both ERK2 and RAF activation is unlike the signaling of peptide growth factors. In the case of retinoic acid, the signaling remains active until the cells are G0 arrested and differentiated. It has been suggested that this long duration distinguishes differentiation signaling from mitogenic signaling which typically might only be an hour long (Traverse et al, 1992). How this signaling is sustained is still enigmatic. Retinoic acid induces expression of BLR1 (Burkitt’s Lymphoma Receptor 1, a putative 7-pass heterotrimeric G protein coupled receptor also called CXCR5) mRNA; and BLR1 stimulates MAPK activation (Battle et al, 2000). Inducing BLR1 may thus contribute to the sustained MAPK activation observed. As a result, substrates not normally phosphorylated by MAPK during a brief activation may now become phosphorylated during sustained activation, thus directing cellular outcome to differentiation instead of division. Retinoic acid-induced myeloid differentiation of HL60 cells is not the only inducer or differentiation pathway utilizing MAPK activation. 1,25-dihydroxy vitamin D3, which induces monocytic differentiation, also causes MAPK activation in HL-60 cells (Marcinkowska et al, 1997), as well as the closely related NB4 cell (Song et al, 1998). Interestingly, 6-cis analogues of 1,25-dihydroxy vitamin D3 which bind poorly to the vitamin D receptor also activate p42 MAPK, indicating that MAPK activation may be a non-genomic effect of vitamin D3 (Song et al, 1998). In other cultured cells, vitamin D3 has been found to cause RAF activation as well (Lissoos et al, 1993). In contrast to retinoic acid, vitamin D3-induced RAF activation (Lissoos et al, 1993), MAPK signaling (Pan et al, 1999) and monocytic differentiation (Pan et al, 1999) appear to be PKC dependent. It is an interesting coincidence that retinoic acid and vitamin D3 are

known to induce equivalent early events leading to the ‘‘precommitment’’ state (Yen et al, 1984,1987a, 1987b, 1990), whereby cells become primed to differentiate prior to selecting a myeloid or monocytic differentiation lineage, and also that both inducers activate MAPK signaling during this time. Activation of MAPK signaling may thus be a component of the ‘‘early’’ events that prime a cell to differentiate. In summary, the present results show that in HL-60 cells retinoic acid causes MEK-dependent RAF activation, and indicate that activation of RAF/MEK/ MAPK signaling by retinoic acid differs from the conventional anticipations of growth factor-mediated signaling. Such differences may contribute to the differing outcomes of retinoic acid-induced G0 arrest with differentiation and growth factor-induced cell proliferation. Acknowledgements We are grateful to Dr. J. K. Choi, College of Pharmacy, Chonnam National University, Kwangju 500 – 757, South Korea, for his support of H-Y. H. while in his laboratory and for his helpful discussions. This work was supported in part by grants from the NIH (USPHS) and USDA to A.Y.

References Anderson, N.G., Li, P., Marsden, L.A., Williams, N., Roberts, T.M. and Sturgill, T.W. (1991) Raf-1 is a potential substrate for mitogen-activated protein kinase in vivo. Biochem J 227:573–576. Battle, T., Levine, R.A. and Yen, A. (2000) Retinoic acid-induced blr1 expression promotes ERK2 activation and cell differentiation in HL-60 cells. Exp Cell Res 254:287–298. Blomhoff, R., Green, M.H., Berg, T. and Norum, K. (1990) Transport and storage of vitamin A. Science 250:399–403. Breitman, T.R., Selonick, S.E. and Collins, S.J. (1980) Induction of differentiation of the human promyelocytic leukemia cell line (HL-60) by retinoic acid. Proc Natl Acad Sci USA 77:2936–2940. Brooks, III S.C., Kazmer, S., Levin, A.A. and Yen, A. (1996) Myeloid differentiation and RB phosphorylation changes in HL-60 cells induced by RAR- and RXR-selective retinoic acid analogs. Blood 87:227–237. Cai, H., Erhardt, P., Troppmair, J., Diaz-Meco, M.T., Sithanandam, G., Rapp, U.R., Moscat, J. and Cooper, G.M. (1993) Hydrolysis of phosphatidylcholine couples Ras to activation of Raf protein kinase during mitogenic signal transduction. Mol Cell Biol 13:7645–7651. Collins, S.J., Gallo, R.C. and Gallagher, R.E. (1977) Continuous growth and differentiation of human myeloid leukaemic cells in suspension culture. Nature 270:347–349. Davis, B.H., Coll, D. and Beno, D.W. (1993) Retinoic acid suppresses the response of platelet-derived growth factor in human hepatic Ito-cell-like myofibroblasts: A post-receptor mechanism independent of raf/fos/jun/egr activation. Biochem J 294:785– 791. Flynn, P.J., Miller, W.J., Weisdorf, J.D., Arthur, D.C., Brunning, R. and Branda, R.F. (1983) Retinoic acid treatment of acute promyelocytic leukemia: In Vitro and In Vivo observations. Blood 62:1211. Friedman, A. and Sklan, D. (1993) Vitamin A and immunity. In: Klurfeld, D.M. (ed.) Human Nutrition – A Comprehensive Treatise. Plenum Press, pp 197–216. Gudas, L.J. (1994) Retinoids and vertebrate development. J Biol Chem 22:15399–15402. Katagari, K., Hattori, S., Nakamura, S., Yamato, T., Yoshida, T.

65 and Katagiri, T. (1994) Activation of Ras and formation of GAP complex during TPA-induced Monocytic differentiation of HL60 cells. Blood 84:1780–1789. Kharbanda, S., Saleem, A., Emoto, Y., Stone, R., Rapp, U. and Kufe, D. (1994) Activation of Raf-1 and mitogen-activated protein kinases during monocytic differentiation of human myeloid leukemia cells. J Biol Chem 269:872–878. Kobayashi, E., Ando, K., Nakano, H. and Tamaoki, T. (1989a) UCN-1028A, a novel and specific inhibitor of protein kinase C, from Cladosporium. J Antibiot (Tokyo) 42:153–155. Iida, T., Kobayashi, E., Yoshida, M., Sano, M. (1989) Calphostins, novel and specific inhibitors of protein kinase C. II. Chemical structures. J Antibiot (Tokyo) 42:1475–1481. Kobayashi, E., Nakano, H., Morimoto, M. and Tamaoki, T. (1989b) Calphostin C (UCN-1028C), a novel microbial compound is a highly potent and specific inhibitor of protein kinase C. Biochem Biophys Res Com 159:548–553. Lee, R., Cobb, M.H. and Blackshear, P.J. (1992) Evidence that extracellular signal-regulated kinases are the insulin-activated Raf-1 Kinase Kinases. J Biol Chem 267:1088–1092. Lissoos, T.W., Beno, D.W.A. and Davis, B.H. (1993) 1,25-dihydroxyvitamin D3 activates raf kinase and raf perinuclear translocation via a protein kinase C-dependent pathway. J Biol Chem 268:25132–25138. Mangelsdorf, D.J., Borgmeyer, U., Heyman, R.A., Zhou, J.Y., Ong, E.S., Oro, A.E., Kakizuka, A. and Evans, R.M. (1992) Characterization of three RXR genes that mediate the action of 9-cis retinoic acid. Genes and Dev 6:329–344. Mangelsdorf, D.J., Ong, E.S., Dyck, J.A., Evans, R.M. (1990) Nuclear receptor that identifies a novel retinoic acid response pathway. Nature 345:224–229. Mangelsdorf D.J., Umesono, K.U. and Evans, R.M. (1994) The retinoid receptors. In: Sporn, M.B., Roberts, A., Goodman, D.S. (eds.) The Retinoids. 2nd edition, Chapter 7, Raven Press, NY, pp 319–349. Marcinkowska, E., Wiedlocha, A. and Radzikowski, C. (1997) 1,25-dihydroxy vitamin D3 induced activation and subsequent nuclear translocation of MAPK is upstream regulated by PKC. Bioch Biophys Res Comm 241:419–426. Matsumoto, H. and Sasaki, Y. (1989) Staurosporine, a protein kinase C inhibitor interferes with proliferation of arterial smooth muscle cells. Biochem Biophys Res Comm 158:105–109. Mihara, K., Cao, X-R., Yen, A., Chandler, S., Driscoll, B., Murphree, A.L., T’Ang, A. and Fung, Y.K.T. (1989) Cell-cycle dependent regulation of phosphorylation of the human retinoblastoma gene product. Science 246:1300–1303. Morrison, D.K., Heidecker, G., Rapp, U.R. and Copeland, R.R. (1993) Identification of the major phosphorylation sites of the RAF-1 kinase. J Biol Chem 268:17309–17316. Pan, Q., Granger, J., O’Connell, T.D., Somerman, M.J. and Simpson, R.U. (1999) Promotion of HL-60 cell differentiation by 1,25-dihydroxy-vitamin D3 regulation of protein kinase C levels and activity. Bioch Pharm 54:909–915. Platko, J.D. and Yen, A. (1997) Paxillin increases as retinoic acid or vitamin D3 induce HL-60 cell differentiation. IN VITRO Cell. Dev Biol 33:84–87. Platko, J.D., Forbes, M.E., Varvayanis, S., Williams, M.N., Brooks III, S.C., Cherington, V. and Yen, A. (1998) Polyoma middle T antigen in HL-60 cells accelerates hematopoietic myeloid and monocytic cell differentiation. Exp Cell Res 238:42–50. Rommel, C., Clarke, B.A., Zimmermann, S., Nunez, L., Rossman, R., Reid, K., Moelling, K., Yanacopoulos, G.D. and Glass, D.J. (1999) Differentiation state-specific inhibition of the Raf-MekErk pathway by Akt. Science 286:1738–1741. Sassone-Corsi, P., Mizzen, C.A., Cheung, P. and Crosio, C. (1999) Requirement of Rsk-2 for epidermal growth factor-activated phosphorylation of histone H3. Science 285:886–891. Sklan, D. (1987) Vitamin A in human nutrition. Prog Food Nut Sci 11:39–55.

Song, X., Bishop, J.E., Okamura, W.H. and Norman, A.W. (1998) Stimulation of phosphorylation of mitogen-activated protein kinase by 1a,25-dihydroxyvitamin D3 in promyelocytic NB4 leukemia cells: A structure-function study. Endocrinology 139:457– 465. Tamaoki, T., Nomoto, H., Takahashi, I., Kato, Y., Morimoto, M. and Tomita, F. (1986) Staurosporine, a potent inhibitor of phospholipid/Caπ π dependent protein kinase. Biochem Biophys Res Comm 135:397–402. Toullec, D., Pianetti, P., Coste, H., Bellevergue, P., Grand-Perret, T., Ajakane, M., Baudet, V., Boissin, P., Boursier, E., Loriolle, F., et al (1991) The bisindolyllmalleimide GF 109203X is a potent and selective inhibitor of protein kinase C. J Biol Chem 266:15771– 15781. Traverse, S., Gomez, N., Paterson, H., Marshall, C. and Cohen, P. (1992) Sustained activation of the mitogen-activated protein (MAP) kinase cascade may be required for differentiation of PC12 cells: Comparison of the effects of nerve growth factor and epidermal growth factor. Biochem J 288:351–355. Yen, A. (1990) HL-60 cells as a model of growth control and differentiation: The significance of variant cells. Hemat. Rev 4:5–46. Yen, A. and Forbes, M.E. (1990) C-myc down regulation and precommitment in HL-60 cells due to bromodeoxyuridine. Cancer Res 50:1411–1420. Yen, A., Brown, D. and Fishbaugh, J. (1987a) Precommitment states induced during HL-60 myeloid differentiation: Possible similarities of retinoic acid and DMSO induced early events. Exp Cell Res 173:80–84. Yen, A., Chandler, S., Forbes, M.E., Fung, Y-K., T’Ang, A. and Pearson, R. (1992a) Coupled down-regulation of the RB retinoblastoma and c-myc genes antecedes cell differentiation: possible role of RB as a ‘‘status quo’’ gene. Eur J Cell Biol 57:210–221. Yen, A., Cherington, V., Schaffhausen, B., Marks, K. and Varvayanis, S. (1999a) Transformation defective polyoma middle T antigen mutants defective in PLCg, PI-3 or SRC kinase activation enhance ERK2 activation and promote retinoic acid-induced cell differentiation like wild type middle T. Exp Cell Res 248:538–551. Yen, A., Forbes, M., deGala, G. and Fishbaugh, J. (1987b) Control of HL-60 cell differentiation lineage specificity: A late event occurring after precommitment. Cancer Res 47:129–134. Yen, A., Forbes, M.E., Varvayanis, S., Tykocinski, M.L., Groger, R.K. and Platko, J.D. (1993) C-fms dependent HL-60 cell differentiation and regulation of RB gene expression. J. Cell Physiol 157:379–391. Yen, A., Reece, S.L. and Albright, K.L. (1984) Dependence of HL60 myeloid cell differentiation on continuous and split retinoic acid exposures: Pre-commitment memory associated with altered nuclear structure. J Cell Physiol 118:227–286. Yen, A., Roberson, M.S. and Varvayanis, S. (1999b) Retinoic acid selectively activates the ERK2 but not JNK/SAPK or p38 MAP kinases when inducing myeloid differentiation. IN VITRO Cell Dev Biol 35:527–532. Yen, A., Roberson, M.S., Varvayanis, S. and Lee, A.T. (1998) Retinoic acid induces mitogen-activated protein (MAP)/extracellular signal-regulated kinase (ERK) kinase-dependent MAP kinase activation needed to elicit HL-60 cell differentiation and growth arrest. Cancer Res 58:3163–3172. Yen, A., Samuel, V. and Forbes, M. (1992b) Regulation of cell proliferation: late down-regulation of c-Myb preceding myelo-monocytic cell differentiation. J Cell Physiol 153:147–156. Yen, A., Soong, S., Dwon, H.J., Yoshida, M., Beppu, T. and Varvayanis, S. (1994a) Enhanced cell differentiation when RB is hypo-phosphorylated and down regulated by radicicol, a SRCkinase inhibitor. Exp Cell Res 214:163–171. Yen, A., Sturgill, R. and Varvayanis, S. (1997) Increasing C-FMS expression decreases retinoic acid concentration needed to cause cell differentiation and RB hypophosphorylation. Cancer Res 57:2020–2028.

66 Yen, A., Sturgill, R., Varvayanis, S. and Chern, R. (1996) FMS (CSF-1 receptor) prolongs cell cycle, promotes retinoic acid-induced hypophosphorylation of RB protein, G1 arrest and cell differentiation. Exp Cell Res 229:111–125. Yen, A., Williams, M., Platko, J.D., Der, C. and Hisaka, M. (1994b) Expression of activated RAF accelerates cell differentiation and RB protein down regulation but not hypophosphorylation. Eur J Cell Biol 65:103–113.

Zimmerman, S. and Moelling, K. (1999) Phosphorylation and regulation of Raf by Akt (Protein Kinase B). Science 286:1741– 1744. Zimmerman, S., Rommel, C., Ziogas, A., Lovric, J., Moelling, K. and Radziwill, G. (1997) MEK1 mediates a positive feedback on Raf-1 activity independently of Ras and Src. Oncogene 15:1503– 1511.