INTRODUCTION

The modulation of the expression of c-myb and c-myc protooncogenes has been implicated in the physiological signal pathways of growth and differentiation of erythroid precursor cells (1). Erythropoietin (Epo),¹ the principal regulator of erythropoiesis, induces a rapid up-regulation of c-myc and a simultaneous down-regulation of c-myb mRNA levels (2, 3, 4, 5). These protooncogene responses occur via two discrete signaling pathways. The effect of Epo on c-myc expression is prevented by inhibitors of protein kinase C, while the signal to c-myb is prevented by okadaic acid, an inhibitor of the serine/threonine-specific protein phosphatases 1 and 2A (5).

The precise role of the c-myb and c-myc responses in the effect of Epo on cell growth and differentiation has not yet been clarified. Treatment of Rauscher erythroleukemia cells with an antisense oligodeoxynucleotide to c-myb results in induction of hemoglobin synthesis without any significant effect on cell proliferation (6). These results suggest that the suppression of c-myb expression is related to the differentiation-inducing activity of Epo. In fact, a common feature of the action of Epo and dimethyl sulfoxide (Me₂SO)-like chemical inducers of the differentiation in erythroleukemia cells is the early down-regulation of c-myb mRNA levels (2, 3, 4, 5). Me₂SO, in contrast to Epo, induces a decrease also in c-myc expression.

In previous studies we reported that [Ca²⁺]_i-increasing agents, such as Ca²⁺-ionophores and inhibitors of the endoplasmic reticulum Ca²⁺-pump, induce in Friend erythroleukemia cells an early and transient decrease in c-myb mRNA levels without similar changes in the c-myc expression (7, 8). The early decrease in c-myb mRNA was followed by induction of -globin mRNA and hemoglobin synthesis. These results support the view that the early down-regulation of c-myb expression is involved in the induction of the erythroid differentiation pathway and provide further insight into biochemical mechanisms regulating c-myb expression in erythroid precursor cells. Since Epo has been shown in many studies to induce an increase in [Ca²⁺]_i (9, 10, 11), the question arises whether the Epo-induced down-regulation of the c-myb expression is mediated by a Ca²⁺ signal.

The aim of the present investigations was to further analyze the down-regulation of c-myb mRNA levels and induction of hemoglobin synthesis by [Ca²⁺]_i-increasing agents and to compare the Ca²⁺-induced changes with the action of Epo in the Epo-responsive murine erythroleukemia cell line ELM-I-1 (12).

MATERIALS AND METHODS

A23187, cyclopiazonic acid, and thapsigargin were purchased from Sigma, trifluoperazine and calmidazolium from Biomol (Plymouth Meeting, PA), okadaic acid from Life Technologies, Inc. and from Sigma, KN-62 and fura-2/AM from Calbiochem; cyclosporin A was obtained from Sandoz (Basel, Switzerland), FK506 from Fujisawa (Tokyo, Japan), and mouse and human recombinant erythropoietin from Boehringer (Mannheim, Federal Republic of Germany).

[-³²P]dCTP (3000 Ci/mmol) and [-³²P]ATP (4500 Ci/mmol) were obtained from Amersham (Buckinghamshire, United Kingdom) and ICN Pharmaceuticals (Irvine, CA);  minimal essential medium without nucleosides, Ham's F-12 medium, horse serum, fetal calf serum, and trypan blue solution were purchased from Life Technologies, Inc., benzidine and bovine hemoglobin from Sigma, analytical grade chemicals from Sigma, Fluka (Buchs, Switzerland), and Merck (Darmstadt, FRG).

Erythropoietin-responsive murine erythroleukemia cells, line ELM-I-1 (12, 13) and Friend erythroleukemia cells, line F4-6 (14) were kindly provided by Prof. W. Ostertag, Heinrich Pette Institute for Experimental Virology and Immunology (Hamburg, FRG). Cells were grown in  medium without nucleosides, supplemented with 2 mM glutamine, 50 units/ml penicillin, and 50 µg/ml streptomycin and with 10% horse serum (ELM-I-1 cells) or with 10% fetal calf serum (F4-6 cells) at 37 °C in a humidified atmosphere containing 5% CO₂.

For the experiments, exponentially growing cells were plated at 6-8 × 10⁴ cells/ml. Approximately 16 h later, the cells were treated with the test substances and at defined time points thereafter, cells were harvested and stored at 80 °C before RNA isolation. For protein isolation, cells were washed in ice-cold phosphate-buffered saline and suspended in phosphate-buffered saline with 6% trichloroacetic acid and stored at 80 °C. To measure hemoglobin production, ELM-I-1 cells were incubated for 3 days, F4-6 cells for 4 days with the test substances, and hemoglobin content of the cells was determined by the benzidine technique of Luftig et al. (15) using bovine hemoglobin as standard. Cell viability was examined by trypan blue exclusion. Serum-free incubations of ELM-I-1 cells were carried out in Ham's F-12 medium (13).

Total cellular RNA was isolated by the acid guanidium thiocyanate-phenol-chloroform method of Chomczinski and Sacchi (16). 30-µg samples of RNA were denaturated by glyoxylation, size-separated by electrophoresis through 1% agarose gel, and transferred to a Biodine A membrane (Pall, Portsmouth, UK) by the capillary blotting technique (17) using 20 × standard saline citrate solution (SSC, 1 × SSC contains 0.15 M sodium chloride and 0.015 M sodium citrate). RNA was immobilized by baking the membrane at 80 °C for 1.5 h. The DNA probes for hybridization were labeled with [-³²P]dCTP by the random primer method (18) using the multiprime DNA labeling system from Amersham. Prehybridization (6-8 h) and hybridization (18-20 h) were carried out at 42 °C in 50% formamide, 50 mM sodium phosphate buffer, pH 6.5, 5 × SSC, 5 × Denhardt's solution (0.1% bovine serum albumin, 0.1% Ficoll, 0.1% polyvinylpyrrolidone), 0.2% SDS, and 250 µg/ml denaturated herring sperm DNA (Boehringer). Hybridized blots were washed in 2 × SSC and 0.1% SDS at room temperature for 20 min and at 50 °C for 60 min, in 0.5 × SSC and 0.1% SDS at 50 °C for 30 min, and finally in 0.2 × SSC and 0.1% SDS at 55 °C for 30 min. The membranes were exposed at 80 °C to an x-ray film using intensifying screens.

The following hybridization probes were used: human -actin cDNA, a 1.2-kb PstI fragment was kindly provided by Dr. T. Braun (Institute of Biochemistry and Biotechnology, University of Braunschweig, Braunschweig, FRG); human c-myb cDNA (19), a 1.2-kb BamHI fragment isolated from the pHM1 construct (20) was kindly donated by Dr. F. Kalkbrenner (Institute of Pharmacology, Free University of Berlin, Berlin, FRG); mouse ^minor-globin DNA (21) was generously donated by Dr. J. Nowock (Heinrich Pette Institute of Experimental Virology and Immunology, Hamburg, FRG). A 1.0-kb BamHI fragment with coding sequences from the structural gene was used. Human c-myc exon 3, a 1.4-kb ClaI-EcoRI fragment, was purchased from Oncor (Gaithersburg, MD).

The quantitative evaluation of the blots was performed using a Hewlett-Packard ScanJet II CX/T flatbed scanner and the software NIH image 1.56 as described previously (8).

Trichloroacetic acid-precipitated cell material was collected by centrifugation, solubilized in 8 M urea, and diluted with the same volume of denaturation buffer containing 50 mM Tris phosphate, pH 6.8, 2% SDS, 2% mercaptoethanol, 20 mM EDTA, and 20% glycerin and incubated at 85 °C for 1 min. Protein content of the samples was determined according to Lowry et al. (22). Portions of 60 µg of total cell protein were separated by electrophoresis on a 10% SDS-polyacrylamide gel (23) and transferred to Biodin A membrane (Pall) using the Mini-Protean II system from Bio-Rad. c-Myb protein was detected with a rabbit anti-v/c-Myb polyclonal antibody (Medac, Hamburg, FRG) and an anti-rabbit IgG peroxidase conjugate (Sigma) as secondary antibody using the ECL Western blotting system of Amersham.

Fluorescence [Ca²⁺]_i measurements and calculations were carried out as described previously (8). Briefly, cells (2.5-5 × 10⁵/ml) were loaded in culture medium with 1.5 µM of fura-2/AM fluorescent calcium indicator at 37 °C for 30 min. Before each measurement, an aliquot (1 ml) of the loaded cell suspension was rapidly centrifuged (10 s at 12,000 × g) in an Eppendorf microcentrifuge, and the pellet was rinsed three times with the standard measuring medium, containing 120 mM NaCl, 5 mM KCl, 0.5 mM MgCl₂, 0.04 mM CaCl₂, 10 mM Hepes-Na, pH 7.4, 10 mM NaHCO₃, 5 mM NaHPO₄, and 10 mM glucose and resuspended in 2 ml of measuring medium (2-2.5 × 10⁵ cells/ml). Fluorescence was measured in a Hitachi F-4000 fluorescence spectrophotometer at 37 °C (excitation wavelength, 340 nm; emission wavelength, 500 nm; bandwidth, 5 nm). Cytoplasmic free calcium concentration was calculated by using the method of Tsien et al. (24).

Cells were treated with test substances and calcineurin phosphatase activity in hypotonic lysates was measured by the methods of Fruman et al. (25) using the RII peptide substrate (Alexis Corp., Läufelfingen, Switzerland) labeled with [-³²P]ATP and the catalytic subunit of bovine cardiac cAMP-dependent protein kinase (Sigma) (26). Assays were performed in duplicate, and calcineurin activity was determined as Ca²⁺-dependent phosphatase activity in the presence of 500 nM okadaic acid.

RESULTS

In accordance with previous results in Friend erythroleukemia cells, line F4-6 (7, 8), [Ca²⁺]_i-increasing agents (A23187, thapsigargin, cyclopiazonic acid) induced a rapid decrease in c-myb mRNA levels in ELM-I-1 cells, too. In order to study whether calmodulin (CaM) is involved in suppression of c-myb expression by [Ca²⁺]_i-increasing agents, experiments were performed with the known chemically unrelated CaM antagonists, trifluoperazine and calmidazolium (27). As shown in Fig. 1, 50 µM trifluoperazine and 5 µM calmidazolium antagonized the effect of thapsigargin (2 nM) and A23187 (1.5 µM) on the c-myb mRNA levels during a 3-h incubation.

The results with the CaM antagonists suggest that the down-regulation of c-myb mRNA levels by [Ca²⁺]_i-increasing agents is mediated by Ca²⁺/CaM-dependent enzymes. In the experiment shown in Fig. 2 we investigated the effect of KN-62, an inhibitor of the Ca²⁺/CaM-dependent protein kinase II and other members of the Ca²⁺/CaM-dependent kinase family (28, 29, 30). KN-62, at concentrations active in cell cultures, 3 and 6 µM, did not inhibit the suppression of c-myb mRNA levels by A23187, 1.5 µM. In further experiments, cyclosporin A and FK506, specific inhibitors of the Ca²⁺/CaM-dependent protein phosphatase, PP2B, or calcineurin in complex with immunophylins (31, 32) were investigated. Cyclosporin A (CsA) proved to be a potent antagonist of the Ca²⁺-induced down-regulation of c-myb expression. Fig. 3 shows the effect of CsA on the c-myb mRNA levels in ELM-I-1 cells at concentrations of 6, 15, 40, 100, and 250 nM after 3-h incubation without or with A23187, 1.5 µM. In the absence of A23187, c-myb mRNA levels were not significantly influenced by CsA (Fig. 3A). In the presence of A23187, CsA inhibited the decrease in c-myb mRNA already at the lowest concentration tested (Fig. 3B). The quantitative evaluation of the blots showed that the Ca²⁺-induced decrease in c-myb mRNA was inhibited to 80% at 40 nM CsA and completely abolished at 100 and 250 nM CsA. On the other hand, FK506 failed to antagonize the Ca²⁺-induced down-regulation of c-myb mRNA levels, even at the high concentrations of 200 and 1000 nM (Fig. 3C).

and -actin mRNA in ELM-I-1 cells after a 3-h exposure to A23187 (1.5 µ

and -actin mRNA in ELM-I-1 cells after a 3-h incubation with CsA (6, 15, 40, 100, and 250 n

ELM-I-1 cells were exposed to CsA and FK506 for 1 h, and hypotonic cell lysates were then made and assessed for PP2B activity (25). The results are summarized in Fig. 4. CsA inhibited the calcineurin phosphatase activity in a concentration range of 3-200 nM with approximately the same effectivity as the Ca²⁺-induced down-regulation of c-myb mRNA levels (Fig. 3B). FK506 at concentrations up to 1000 nM gave no appreciable inhibition correlating with its lack of ability to prevent the Ca²⁺ effect on the c-myb expression (Fig. 3C). Parallel, we measured calcineurin phosphatase activity in Jurkat T cells under the same experimental conditions and ensured that the FK506 used was active in inhibiting phosphatase activity in this cell line (data not shown). Similarly to the results of previous studies (25, 33, 34), 15-20% of the net phosphatase activity measured was resistant to 200 nM CsA in both cell lines and to 200 nM FK506 in Jurkat cells.

In order to compare the Ca²⁺-induced down-regulation of c-myb mRNA with the action of Epo in the Epo-sensitive ELM-I-1 cell line (12, 13), we studied the effects of Epo on the c-myb and c-myc expression. Fig. 5A shows the time course of the effect of Epo, 1 unit/ml, on c-myb and c-myc mRNA levels during an incubation period of 24 h. In accordance with studies in other murine erythroleukemia cell lines (2, 3, 4, 5), Epo induced a rapid decline in c-myb mRNA levels. A maximal drop (to approximately 20% of control) was reached within 2 h. c-myb mRNA levels remained decreased during the 24-h incubation period at 20-50% of control. The c-myc mRNA showed a slight transient increase at 1 h (135% of control). The Epo-induced suppression of c-myb expression was confirmed in the present study by Western blot analysis of the c-Myb protein during the first 12 h of incubation (Fig. 5B).

The effect of Epo, 1 unit/ml, on c-myb expression was also studied in serum-free medium and in the presence and absence of EGTA, 3 mM. The results shown in Fig. 6 indicate that the Epo effect on c-myb expression occurred independently of other serum factors or of extracellular Ca²⁺. In accordance with previous results in F4-6 cells (8), EGTA caused an increase in c-myb mRNA at 3 h. In the same experiment, a strong increase in -globin mRNA was demonstrated in cultures treated with Epo in serum-free medium for 36 h.

In further experiments, the CaM antagonists, trifluoperazine and calmidazolium, as well as the calcineurin inhibitor CsA were investigated for their ability to prevent the effect of Epo on c-myb mRNA levels in ELM-I-1 cells. The inhibitors were tested at concentrations known to be active in antagonizing the suppression of c-myb mRNA by [Ca²⁺]_i-increasing agents without significant influence on c-myb mRNA levels alone (Figs. 1 and 3). The results are summarized in Fig. 7. Trifluoperazine, 50 µM, calmidazolium, 5 µM, and CsA, 200 nM, failed to block the Epo effect on the c-myb mRNA levels.

In order to study the effects of the inhibitors tested on the differentiation status of ELM-I-1 cells, hemoglobin production was determined at 72 h after drug treatments. Longer incubations with calmodulin antagonists proved to be highly toxic for the cells. On the other hand, CsA was tolerated by the cells without decrease in cell viability. We tested the effect of CsA, 200 nM on the hemoglobin synthesis in ELM-I-1 cells induced by Epo, 0.5 unit/ml, or by the [Ca²⁺]_i-increasing agent cyclopiazonic acid, 2.5 and 5 µM. Cyclopiazonic acid was shown to increase hemoglobin production in Friend erythroleukemia cells with lower toxicity (8). Epo induced a 5-6-fold increase in hemoglobin production, while cyclopiazonic acid proved to be less effective with a 2-3-fold increase in hemoglobin synthesis. Similarly to the effects on c-myb mRNA levels (Figs. 3 and 7), CsA antagonized the Ca²⁺-mediated increase in hemoglobin production, while the Epo-induced hemoglobin synthesis was not affected (Table I). CsA also inhibited the decrease in cell viability in the cyclopiazonic acid-treated cultures.

Table I. Studies on the effect of CsA on the Epo or cyclopiazonic acid (CPA)-induced hemoglobin production in ELM-I-1 cells Cells were incubated with the test compounds for 3 days and analyzed as described under ``Materials and Methods.'' The CsA-treated cultures were preincubated with CsA for 30 min. Results represent the mean ± S.E. of three experiments, each with duplicate cultures. Substance Hemoglobin Cell viabilitya µg/108 cells % of control Control 29.5  ± 8.2 100.0 Epo (0.5 unit/ml) 166.9  ± 41.7 115.0  ± 7.2 CPA (2.5 µM) 59.2  ± 15.4 54.8  ± 4.2 CPA (5.0 µM) 67.5  ± 10.0 39.5  ± 8.5 CsA (0.2 µM) 33.3  ± 7.1 93.5  ± 12.1 CSA + Epo 180.9  ± 30.1 99.0  ± 1.4 CsA + CPA (2.5 µM) 31.4  ± 6.8 92.6  ± 10.5 CsA + CPA (5.0 µM) 33.0  ± 8.0 81.6  ± 5.3 a Trypan blue negative cells relative to the untreated controls.

Since cyclopiazonic acid induced only a moderate hemoglobin production in ELM-I-1 cells, experiments were performed also with the Friend erythroleukemia cell line F4-6 (14) in which [Ca²⁺]_i-increasing agents induced hemoglobin production more effectively (8). A maximum in hemoglobin production in this cell line was observed after 96-h incubation. The effect of cyclopiazonic acid, 2.5 and 5 µM, was compared in the Epo-insensitive F4-6 cells with the strong chemical inducer of differentiation, Me₂SO, 1.5% (Table II). The results show that the cyclopiazonic acid-induced hemoglobin production and decrease in cell viability was antagonized by CsA, 200 nM. On the other hand, CsA had no effect on the hemoglobin production and decrease in cell viability induced by Me₂SO. The effects of the same drug treatments on the c-myb mRNA levels was comparatively studied in F4-6 cells at 3 h. CsA antagonized the decrease in c-myb mRNA by the [Ca²⁺]_i-increasing agent cyclopiazonic acid, while the strong suppression of c-myb mRNA levels by Me₂SO treatment was not inhibited (data not shown).

Table II. Studies on the effect of CsA on the Me2SO or cyclopiazonic acid (CPA)-induced hemoglobin production in F4-6 cells Cells were incubated with the test compounds for 4 days and analyzed as described under ``Materials and Methods.'' The CsA-treated cultures were preincubated with CsA for 30 min. Results represent the mean ± S.E. of three experiments, each with duplicate cultures. Substance Hemoglobin Cell viabilitya µg/108 cells % of control Control 19.0  ± 5.3 100.0 Me2SO (1.5%) 203.7  ± 38.2 43.0  ± 8.2 CPA (2.5 µM) 131.7  ± 7.2 83.3  ± 3.7 CPA (5.0 µM) 133.6  ± 16.5 32.4  ± 2.7 CsA (0.2 µM) 15.0  ± 4.2 109.3  ± 4.1 CSA + Me2SO 215.7  ± 17.5 41.6  ± 8.1 CsA + CPA (2.5 µM) 13.5  ± 3.9 115.7  ± 8.1 CsA + CPA (5.0 µM) 5.3  ± 1.9 89.5  ± 1.5 a Trypan blue negative cells relative to the untreated controls.

Previously we demonstrated the development of Ca²⁺ signals in F4-6 cells with cyclopiazonic acid and thapsigargin (8). In the present experiments the effect of Epo on [Ca²⁺]_i was tested under similar experimental conditions in ELM-I-1 cells loaded with fura-2 fluorescent calcium indicator. As documented in Fig. 8, in unstimulated ELM-I-1 cells an approximately 110 nM resting cytoplasmic Ca²⁺ level could be measured in the presence of 0.5 mM free external Ca²⁺. Treatment of the cells with repeated addition of Epo (2 units/ml, 10 units/ml, 50 units/ml) did not induce an increase in [Ca²⁺]_i during the 20-min treatment period. In spite of the lack of an effect by Epo, 2 nM thapsigargin induced a rapid enhancement in [Ca²⁺]_i (to approximately 450 nM) in the same ELM-I-1 cells. Similar results were obtained when the external Ca²⁺ concentration was increased up to 2.5 mM, and no differences were seen when human or mouse recombinant Epo preparations were used (data not shown).

DISCUSSION

In accordance with previous studies in Friend erythroleukemia cells (7, 8), the present results demonstrate a rapid suppression of c-myb mRNA levels by [Ca²⁺]_i-increasing agents in the Epo-sensitive murine erythroleukemia cell line ELM-I-1 (12). The Ca²⁺-induced down-regulation of c-myb expression could be inhibited with the CaM antagonists, trifluoperazine and calmidazolium (27), as well as with CsA, an inhibitor of the Ca²⁺/CaM-dependent serine/threonine-specific protein phosphatase, PP2B, or calcineurin (31, 32). CsA inhibited the Ca²⁺ effect on c-myb mRNA with the same concentration dependence as the calcineurin phosphatase activity (Figs. 3B and 4). On the other hand, FK506, another immunosuppressant and well characterized inhibitor of calcineurin (31, 32), was unable to inhibit the Ca²⁺-induced c-myb mRNA decrease and PP2B activity in ELM-I-1 cells. Similar differences in the effects of CsA and FK506 were previously observed in other cell lines. In mouse bone marrow-derived progenitor mast cells, CsA inhibited calcineurin activity effectively (IC₅₀ = 8 nM), while FK506 was without effect at concentrations up to 1000 nM (34). Resistance to FK506 was associated with a deficiency in FK506-binding immunophylin, FKBP12. A decreased sensitivity of PP2B to FK506 relative to CsA has been also observed in rat pancreatic acinar cells (35). Therefore, ELM-I-1 cells used in the present study represent an additional cell line with a lack of sensitivity of calcineurin to the inhibitory action of FK506. It would be of interest to determine whether this is also a characteristic of human erythroid cells.

The effect of CaM antagonists and the close correlation of the effect of immunosuppressants on the c-myb mRNA changes and calcineurin phosphatase activity strongly suggest that the Ca²⁺-induced down-regulation of c-myb expression is mediated by calcineurin. Calcineurin has been implicated in the regulation of gene expression in different cell systems. In T lymphocytes, calcineurin was identified as a key enzyme in the T cell receptor-mediated signal transduction pathways playing a positive regulatory role in the transcription of interleukin-2 or interleukin-4 genes (36, 37, 38). In a pancreatic islet cell line, calcineurin is involved in the cAMP response element (CRE)-mediated induction of glucagon gene transcription after membrane depolarization (39). Both positive and negative signaling functions of calcineurin have been demonstrated in the transcriptional regulation of immediate early genes in PC12 cells (40). In this cell line, calcineurin mediates the activation of NGFI-B and inhibition of NGFI-A transcription in response to Ca²⁺ signals in synergism or antagonism with Ca²⁺/CaM-dependent protein kinases. The present results with ELM-I-1 cells suggest a negative regulatory role of calcineurin in the c-myb expression, similarly to the NGFI-A expression in PC12 cells (40). However, experiments in ELM-I-1 cells with KN-62, an inhibitor of Ca²⁺/CaM-dependent protein kinases (28, 29, 30), provided no evidence that a Ca²⁺/CaM-dependent kinase is involved in early Ca²⁺-induced changes in c-myb mRNA levels.

The target of action of calcineurin in regulation of c-myb expression remains to be elucidated. In T lymphocytes, NF-AT-like transcription factors operate as downstream components of the calcineurin pathway (36, 38). In ELM-I-1 cells, dephosphorylation by calcineurin may result in activation of negative regulatory proteins or inactivation of protein factors necessary for the maintenance of a high level of c-myb mRNA in these cells. Since calcineurin is known to activate PP1 in certain tissues via a protein phosphatase cascade (41, 42, 43), PP1 could also mediate these effects of calcineurin. CREB, a well characterized transcription factor, which binds to CRE in the promoter region of cAMP- or Ca²⁺-responsive genes (44, 45), can be efficiently dephosphorylated by PP2B in vitro (30) and by PP1 in vivo (46, 47). However, there is no indication that c-myb expression is regulated by a CREB-dependent mechanism. The murine c-myb gene has a weak promoter (48) and transcriptional elongation appears to be the main target of regulation (49, 50). A transcriptional arrest mechanism operates in the first intron of the mouse c-myb gene, and sequence-specific protein binding in this region could play an important role in the regulation of c-myb expression (51).

The physiological regulator of erythropoiesis, Epo, has been shown to induce a rapid down-regulation of c-myb mRNA levels in various murine erythroleukemia cell lines (2, 3, 4, 5). We now demonstrate this effect of Epo also in ELM-I-1 cells and confirmed this observation by Western blot analysis of the c-Myb protein (Fig. 5). On the other hand, Epo caused only a moderate increase in c-myc mRNA levels at an early time in ELM-I-1 cells as compared with other cell lines (2, 3, 4, 5). Experiments were undertaken to investigate whether the Epo-induced down-regulation of c-myb expression is mediated also by a Ca²⁺/CaM-dependent mechanism. The Epo-induced suppression of c-myb mRNA levels occurred also in the presence of EGTA and was not prevented by CaM antagonists or CsA, which antagonized the Ca²⁺ effect on the c-myb expression (Figs. 6 and 7). In addition, we could not observe significant changes in [Ca²⁺]_i in ELM-I-1 cells in response to treatment with Epo (Fig. 8). These results clearly show that the Epo-induced decrease in c-myb mRNA occurred independently of Ca²⁺/CaM and calcineurin. However, a common characteristic of the action of Epo and [Ca²⁺]_i increasing agents is the rapid down-regulation of c-myb mRNA without similar effect on the c-myc expression (2, 3, 4, 5, 7, 8). Patel et al. (5) have demonstrated that okadaic acid, 100-400 nM, inhibits the Epo effect on c-myb mRNA levels in Rauscher erythroleukemia cells, suggesting a role of PP2A and/or PP1. Dephosphorylation of the same regulator protein(s) by Ca²⁺/CaM-dependent or -independent serine/threonine-specific protein phosphatases may represent a common mechanism of the action of Epo and [Ca²⁺]_i-increasing agents in the down-regulation of c-myb expression. Alternatively, an activation of PP1 by Epo or by [Ca²⁺]_i-increasing agents via a PP2B-PP1 cascade (41, 42, 43) could commonly mediate Epo or Ca²⁺ signals to c-myb. In fact, our recent experiments indicate that okadaic acid above 100 nM inhibits also the Ca²⁺ effect on c-myb mRNA levels in ELM-I-1 cells (data not shown). The relative high inhibitory concentrations needed may point to a role of PP1, which is less effectively inhibited by this compound than PP2A (42). Studies are in progress in our laboratory to test this possibility.

The early down-regulation of c-myb expression has been suggested to be a prerequisite for either Epo-initiated or chemically induced erythroid differentiation in murine erythroleukemia cells (2, 4). Previously we reported that the Ca²⁺-mediated transient suppression of c-myb mRNA levels is followed by induction of globin mRNA and hemoglobin synthesis in Friend erythroleukemia cells, line F4-6 (8). The present experiments show that prevention of the early decrease in c-myb mRNA by CsA is correlated with an inhibition of hemoglobin production in ELM-I-1 or F4-6 cells incubated in the presence of the [Ca²⁺]_i-increasing agent, cyclopiazonic acid (Tables I and II). These results provide further evidence that down-regulation of c-myb expression is associated with the initiation of hemoglobin synthesis in erythroid precursor cells (6). However, cyclopiazonic acid proved to be less active than Epo in increasing hemoglobin synthesis in ELM-I-1 cells. Toxic effects may attenuate the differentiation-inducing potency of the [Ca²⁺]_i-increasing agent during longer incubation periods. Me₂SO also proved to be a rather weak inducer of hemoglobin synthesis in this cell line (12).

On the other hand, the experiments also demonstrate that the Epo-induced down-regulation of c-myb expression and stimulation of hemoglobin synthesis in ELM-I-1 cells occurred independently of Ca²⁺/CaM and calcineurin. Therefore, the relevance of the Ca²⁺ effects on c-myb expression and hemoglobin synthesis for the physiological signal pathway of erythroid differentiation remains undefined. Under the present experimental conditions, an increase in [Ca²⁺]_i by Epo could not be demonstrated. This observation is consistent with similar negative results from other laboratories (reviewed in Ref. 9). However, several other studies, which report on the [Ca²⁺]_i-increasing activity of Epo (9, 10, 11), suggest that under certain conditions, Epo receptors may generate Ca²⁺ signals. Miller et al. (52) have demonstrated that Epo stimulates a rise in [Ca²⁺]_i in single BFU-E-derived erythroblasts at specific stages of differentiation. Accordingly, the Epo effect on [Ca²⁺]_i may depend on the cell line investigated and on its developmental stage. These results leave the possibility open that Epo can influence c-myb expression and hemoglobin synthesis also via a Ca²⁺/CaM-dependent pathway. This could act synergistically to the Ca²⁺/CaM-independent Epo signaling. In addition, hormones, growth factors, or mediator substances in bone marrow, which induce Ca²⁺ signals in erythroid progenitor cells, might positively modulate the differentiation-inducing activity of Epo. Finally, the Ca²⁺ effects on c-myb expression and hemoglobin synthesis suggest the possibility of a pharmacological influence on erythroid differentiation. Further studies, especially with human erythroid precursor cells, are necessary to explore these possibilities.