cells. 15 The GATA-1 transcription factor has been shown to including -globin, PBGD, erythropoietin receptor (EPOR) Reagents

Transcription

1 Leukemia (1997) 11, Stockton Press All rights reserved /97 $12.00 BRIEF COMMUNICATION Time-course of butyric acid-induced differentiation in human K562 leukemic cell line: rapid increase in -globin, porphobilinogen deaminase and NF-E2 mrna levels B Chénais, I Molle, C Trentesaux and P Jeannesson Laboratoire de Biochimie et Biologie Moléculaire, EA2063, IFR53-Biomolécules, Faculté de Pharmacie, Université de Reims Champagne Ardenne, Reims, France Butyric acid (BA) was shown to induce hemoglobinization of dence of globin gene expression on the NF-E2 transcription K562 cells in a dose- and time-dependent manner. The maximal factor has been demonstrated in murine erythroleukemia differentiation (54% of hemoglobinized cells) was obtained with cells. 15 The GATA-1 transcription factor has been shown to the 0.5 mm concentration, which induced a 60% inhibition of cell growth at day 3 without cytotoxicity. Parallel to the kinetics play a central role in the control of several erythroid genes of hemoglobinization, a rapid increase in -globin and porphobilinogen deaminase (PBGD) mrnas was observed in BA- and GATA-1 itself. 16 including -globin, PBGD, erythropoietin receptor (EPOR) treated cells. This increase was time-dependent and higher for In view of the above, we propose in this report to study the -globin than for PBGD (six- and two-fold at day 3, time-course of BA-induced hemoglobinization and erythroid respectively). In contrast, erythropoietin receptor mrnas were mrna overexpression in human K562 cells. By so doing, we not affected by BA treatment. Analysis of erythroid transcripintend to bring new insights into the involvement of GATA-1 tion factor mrna levels during the time course of BA treatment showed, for the first time, an early and marked (up to three-fold) and, especially, NF-E2 transcription factors in the BA-triggered increase in p45 NF-E2 mrna, contrasting with that of GATA-1 differentiation process. mrna ( 1.5-fold). Taken together, these results showed the rapid differentiating effect of BA and suggest the involvement of the NF-E2 transcription factor. Materials and methods Keywords: butyric acid; NF-E2; GATA-1; erythroid genes; leukemia; differentiation Reagents All chemical reagents were purchased from Sigma (St Quentin Introduction Fallavier, France) unless otherwise stated and were of reagent grade or molecular biology grade when necessary. n-butyric Butyric acid (BA) is a widely used short-chain fatty acid which acid solution was diluted to 100 mm in water and sterilized demonstrated a variety of effects both in vivo and in tumoral through 0.22 m filters (Millipore, St Quentin en Yvelines, cell lines. BA induces cell growth arrest, changes in cell morphology, France). induction of erythroid differentiation of erythro- leukemic cell lines and especially induction of fetal hemoglobin synthesis. 1 Cell culture In the human K562 leukemic cell line, 2 BA has been reported to induce an increase, either of the activity of heme K562 cells were cultured in RPMI-1640 medium (Life Techno- biosynthesis enzymes such as -aminolevulinate synthase logies, St Quentin en Yvelines, France) supplemented with (ALAS 3 ), -aminolevulinate dehydratase (ALAD 4,5 ) and 10% heat-inactivated fetal bovine serum and 2 mm L-glutaporphobilinogen deaminase (PBGD 5 ), or of glycophorin-a mine, in a 5% CO 2 humidified atmosphere. expression, a membrane-specific protein of red blood cells. 3 Nevertheless, neither these recent results nor former report 7 mentioned short-term kinetics analysis of BA-differentiating Assays for cell growth and erythroid differentiation effect. Furthermore, although -globin mrna level has been found to increase in sodium butyrate-treated K562 cells, Cells were treated with BA, or other inducers, at the beginning owing to a transcriptional activation 6 and an enhancement of of this exponential growth phase. Cell growth and viability promoter activity, 8,9 no information is yet available concerncells. were assessed by direct counting of trypan blue dye-excluding ing BA action on erythroid transcription factors NF-E2 and Growth inhibition was calculated from GATA-1. {[(C n C 0 ) (T n T 0 )]/(C n C 0 ) 100} The heterodimeric NF-E2 transcription factor 10,11 recogwhere C n, C 0, T n, T 0 represent the numbers of cells/ml at days nizes an extended AP-1 site initially characterized in the PBGD gene promoter 12 and in the locus control region of the 0 and n, respectively. The percentage of hemoglobinhuman -globin gene. 13 Despite mild effects on erythropoiesis producing cells was determined by a benzidine staining of p45 NF-E2 suppression in knockout mice, 14 the depen- method as previously described. 17 Northern blot analysis Correspondence: B Chénais, Laboratoire de Biochimie et Biologie Moléculaire, Faculté de Pharmacie, Université de Reims Champagne Ardenne, 51 Rue Cognacq-Jay, F Reims Cedex, France Total RNAs were extracted using Trizol reagent (Life Received 30 December 1996; accepted 6 May 1997 Technologies), 10 g were submitted to formaldehyde-

2 1576 containing 1% agarose gel electrophoresis and capillary blot- with the optimal differentiation-inducing concentration, BA ted onto a nylon membrane (Hybond N; Amersham, Les Ulis, 0.5 mm, which exhibited a cytostatic effect, and BA 1 mm, France). Random priming labeling of cdna probes was per- which combined a maximal growth inhibitory effect with a formed using the Ready-To-Go DNA labeling kit (Pharmacia low cytotoxicity (90% of viable cells). Biotech, St Quentin en Yvelines, France) and - 32 P-dCTP Hemoglobinization of K562 cells was studied as a function (3000 Ci/mmol; DuPont NEN, Les Ulis, France). Membranes of time by scoring benzidine-positive cells after 2 to 72 h of were then hybridized with labeled cdna according to the treatment with 0.5 and 1.0 mm BA. The percentage of differentiated cells increased in a time-dependent manner for both supplier s instructions. The A -globin, EPOR and GATA-1 human cdnas were a kind gift from Dr S Ottolenghi concentrations, and was significantly increased as soon as (University of Milan, Italy), human PBGD and murine p45 NF- 10 h after treatment with 0.5 mm BA (Figure 2a). Cell prolifer- E2 cdnas were from Dr P-H Roméo (Hôpital Henri Mondor, ation was not affected before 12 h of treatment (Figure 2b), Créteil, France) and Dr SH Orkin (Harvard Medical School, then, as soon as 24 h, both BA concentrations induced a cell Boston, MA, USA), respectively. Finally membranes were growth inhibitory effect of 20% to reach 60% and 71%, dehybridized in boiling 0.1% SDS and reprobed with human respectively, at 72 h. glyceraldehyde-3-phosphate dehydrogenase (GAPDH) cdna In order to investigate the reversibility of the BA-differentiating effect, K562 cells were treated by 0.5 and 1 mm BA for from Clontech (Montigny le Bretonneux, France). Northern blots were quantified using a GS363 molecular imager 24 h and then incubated in complete drug-free medium for (BioRad, Ivry sur Seine, France) and submitted to autoradiography. cells. Hemin (30 M), which did not affect cell proliferation, an additional 24 or 48 h before scoring the benzidine-positive was used as a control of reversible induction as previously described. 18 Anthracyclines (aclarubicin and doxorubicin) Results and discussion have been demonstrated to be potent differentiating agents in K562 cells (reviewed in Ref. 19) and were used at optimal Effect of BA on cell growth and differentiation differentiating concentrations as irreversible inducers. The results presented in Table 1 showed that BA-induced differentiation and cell growth inhibition were irreversible, at least for Exponentially growing K562 cells were treated with mMBA for 3 days and cell hemoglobinization was studied by scoring benzidine-positive cells. As shown in Figure 1, the increase of benzidine-positive cells was gradual and reached a maximum of 54% of hemoglobinized cells with 0.5 mm BA vs 3% in control cells. In contrast, higher concentrations (1 to 2mM) resulted in a decline in the percentage of hemoglobinized cells followed by an increase of cell mortality which went up to 39% in 2 mm BA-treated cells vs 5% in both 0.5 mm BA-treated and untreated cells (Figure 1). In addition, cell proliferation was strongly affected by BA treatment, and a 60% growth inhibition was observed with 0.25 mm BA, increasing up to a plateau of 87% with 1 mm BA. Consequently, further differentiation analyses were performed both Figure 1 Dose-response of BA-induced hemoglobinization and cell growth inhibition. The percentage of benzidine-positive cells (- -) was scored at day 3 of culture in the presence of 0.1 to 2 mm of BA. For each culture the percentage of mortality (- -) was estimated Figure 2 Time-course of BA-induced hemoglobinization (a) and by using the trypan blue dye exclusion method, and the percentage cell growth inhibition (b). Hemoglobinization and cell number/ml of growth inhibition (- -) was calculated as described in Materials and were determined after 2 to 72 h incubation in the presence of 0.5 and methods. Data represent the mean ± s.d. from three independent 1mM of BA, and in untreated cells (control). Data represent the mean experiments. ± s.d. from five independent experiments.

3 Table 1 Treatment Brief communication Irreversibility of BA-induced hemoglobinization and cell growth inhibition % of benzidine-positive cells (% of growth inhibition) a Day 2 Day Continuous Pulse Continuous Pulse Control 3 ± 1 3 ± ± 1 2 ± 1 Butyric acid 0.5 mm 35 ± 3 31 ± 2 b 47 ± 9 39 ± 7 b (57 ± 7) (59 ± 6) b (60 ± 6) (51 ± 7) b Butyric acid 1 mm 23 ± 5 24 ±3 b 38 ± 4 33 ± 11.7 b (64 ± 4) (54 ± 9) b (76 ± 11) (61 ± 6) b Hemin 30 M c 75 ± 4 9 ± 7 80 ± 1 13 ± 3 Aclarubicin 20 nm 24 ± ± 3 b 49 ± ± 5 b (57 ± 11) (58 ± 7) b (52 ± 8) (35 ± 7) b Doxorubicin 40 nm 42 ± 3 41 ± 13 b 47 ± 1 50 ± 9 b (88 ± 1) (89 ± 5) b (93 ± 1) (90 ± 4) b a Cell induction was performed with the indicated drugs as continuous treatment (48 and 72 h), and pulse treatment (24 h in the presence of the drug followed by incubation for 24 h or 48 h in complete drug-free medium). Hemoglobinization of K562 cells was determined as the percentage of benzidine-positive cells at day 2 and day 3, and the percentage of growth inhibition was calculated as described in Materials and methods. Data represent the mean ± s.d. from at least three independent experiments. b Values from pulse treatment were not significantly different from those obtained from continuous treatment with P 0.05 according to the Student s t-test for paired data. c Cell growth was not affected by hemin (30 M) in all cases. 72 h, in contrast to hemin-induced hemoglobinization which on the surface of aclarubicin-treated cells was augmented. 22 requires the continuous presence of the drug. In contrast, we did not observe any increase of EPOR mrna Consequently, although more rapid, BA-inducing effects in K562 cells treated for 6 to 72 h with 0.5 and 1 mm BA share some similarities with those of anthracyclines, particularly (Figure 4), thus ruling out a putative role of EPOR in BA- aclarubicin. 19 Indeed, BA as well as anthracyclines induced differentiation of K562 cells. induce irreversible, time- and dose-dependent hemoglobinization and cell growth inhibition. Moreover, like aclarubicin, but in contrast with doxorubicin, 17,19 BA did not require the BA-induced increase of erythroid transcription factor mrna level total arrest of cell growth to induce optimal hemoglobinization of K562 cells. Effect of BA on erythroid gene expression -Globin mrna expression was studied in 0.5 mm and 1 mm BA-treated cells for 6 to 72 h. Hybridization of Northern blots with 32 P-labeled cdna showed that -globin expression increased as soon as 6 h (1.5-fold) after BA treatment (Figure 3). This effect was time-dependent with a maximum of sixfold at 72 h, in accordance with previous results. 6 This marked increase of -globin mrna correlated with cell hemoglobinization, since benzidine-positive cells appeared 4 h after the increase of -globin mrna. An upregulation of -globin mrna was also observed in cells treated with 1 mm BA with approximately the same efficiency (data not shown). Concerning the heme synthesis pathway, PBGD gene expression was increased by 1.5-fold in 0.5 mm BA-treated cells as soon as 12 h, with a maximum of two-fold at 72 h (Figure 3). According to Fowler et al, 5 this indicates that, in BA-treated cells, the PBGD mrna increase is lower than that of -globin. In addition, the time-course of PBGD mrna increase paralleled that of hemoglobinized cells. This observation is in agreement with that of Kawasaki et al 3 who reported that ALAS activity and hemoglobin levels were simultaneously increased to a similar extent and with a similar timecourse in sodium butyrate-treated K562 cells. The EPOR gene, like PBGD and -globin genes, has been previously found to be transcriptionally activated in aclarubicin-differentiated K ,21 Moreover, the EPOR expression The erythroid transcription factors NF-E2 and GATA-1 have been previously shown to be increased during aclarubicin-, but not doxorubicin-induced differentiation of K562 cells. 20,23,24 In an attempt to better understand the involvement of these transcription factors, we used p45 NF-E2 and GATA-1 cdnas to investigate their mrna expression timecourses during BA-induced differentiation of K562 cells. As shown in Figure 3, p45 NF-E2 mrna level was increased as early as 6 h (1.2-fold) to reach a maximum of three-fold at 72 h after 0.5 mm BA treatment (Figure 3). With 1mM BA the maximal increase in p45 NF-E2 mrna was obtained much earlier (as soon as 12 h after treatment) and was enhanced up to three-fold (data not shown). It should be noted that the time-course of p45 NF-E2 mrna increase paralleled that of PBGD and that of cell hemoglobinization (Figures 3b and 2a). In contrast, GATA-1 mrna slightly increased during the time-course of BA treatment, with a maximum of 1.3-fold at 12 h (Figure 3) and then decreased. The use of 1 mm BA instead of 0.5 mm did not amplify the GATA-1 mrna increase which remained below 1.5-fold (data not shown). In conclusion, these results showed the rapid effect of BA on erythroid gene expression, with the major exception of the EPOR gene, and the correlation between -globin, PBGD and p45 NF-E2 mrna expression and cell hemoglobinization time-courses. Contrasting with the mildness of the GATA-1 mrna increase, the early and marked increase of p45 NF- E2 mrna level is consistent with a transcriptional activation process and suggests an important role of this transcription

4 1578 factor. Nevertheless, the involvement in BA-triggered differentiation of NF-E2, GATA-1, as well as other transcription factors such as EKLF, remains to be directly demonstrated. Acknowledgements This work was supported by grants from Association pour la Recherche sur le Cancer and from the Ligue Nationale Contre le Cancer, comité de l Aisne. We are grateful to Hélène Bobichon for the photographs and Denis Bosson for his grammatical assistance. References 1 Kruh J. Effects of sodium butyrate, a new pharmacological agent, on cells in culture. Mol Cell Biol 1982; 42: Lozzio CB, Lozzio BB. Human chronic myelogenous leukemia cell line with positive Philadelphia chromosome. Blood 1975; 45: Kawasaki N, Morimoto K, Tanimoto T, Hayakawa T. Control of hemoglobin synthesis in erythroid differentiating K562 cells. Arch Biochem Biophys 1996; 328: Chang CS, Sassa S. -Aminolevulinate dehydratase in human erythroleukemia cells: an immunologically distinct enzyme. Blood 1985; 65: Fowler DA, Weidner DA, Sommadossi J-P. Effects of 3 -azido-3 - deoxythymidine on erythroid inducible gene expression in human K562 leukemia cells. Toxicol Lett 1995; 80: Weidner DA, Sommadossi J-P, 3 -Azido-3 -deoxythymidine inhibits globin gene transcription in butyric acid-induced K562 human leukemia cells. Mol Pharmacol 1990; 38: Anderson LC, Jokinen M, Gahmberg CG. Induction of erythroid differentiation in the human leukemia cell line K562. Nature 1979; 278: Safaya S, Ibrahim A, Rieder RF. Augmentation of -globin gene promoter activity by carboxylic acids and components of the Figure 3 Time-course of mrna expression during BA-induced human -globin locus control regions. Blood 1994; 84: 3929 differentiation. (a) Total RNA from 0.5 mm BA-treated cells and untreated cells (control) were transferred onto nylon membranes and 9 Pace B, Li Q, Peterson K, Stamatoyannopoulos G. -Amino butyric hybridized sequentially with NF-E2 and -globin or GATA-1 and acid cannot reactivate the silenced gene of the locus YAC PBGD cdna probes. Finally blots were hybridized with GAPDH transgenic mouse. Blood 1994; 84: cdna as control. Before autoradiography, each blot was submitted 10 Andrews NC, Erdjument-Bromage H, Davidson MB, Tempst P, to BioRad molecular imager scanning to allow quantification of RNA Orkin SH. Erythroid transcription factor NF-E2 is a haematopoiexpression. (b) Results were plotted as the relative mrna level etic-specific basic-leucine zipper protein. Nature 1993; 36: increases vs control for each incubation time. Data from a typical experiment representative of three. 11 Andrews NC, Kotkow KJ, Ney PA, Erdjument-Bromage H, Tempst P, Orkin SH. The ubiquitous subunit of erythroid transcription factor NF-E2 is a small basic-leucine zipper protein related to the v- maf oncogene. Proc Natl Acad Sci USA 1993; 90: Mignotte V, Eleouet J-F, Raich N, Romeo P-H. Cis- and trans-acting elements involved in the regulation of the erythroid promoter of the human porphobilinogen deaminase gene. Proc Natl Acad Sci USA 1989; 86: Ney PA, Sorrentino BP, Lowrey CH, Nienhuis AW. Inducibility of the HSII enhancer depends on binding of an erythroid specific nuclear protein. Nucleic Acids Res 1990; 18: Shivdasani RA, Orkin SH. Erythropoiesis and globin gene expression in mice lacking the transcription factor NF-E2. Proc Natl Acad Sci USA 1995; 92: Kotkow KJ, Orkin SH. Dependence of globin gene expression in mouse erythroleukemia cells on the NF-E2 heterodimer. Mol Cell Biol 1995; 15: Weiss JW, Orkin SH. GATA transcription factors: key regulators of hematopoiesis. Exp Hematol 1995; 23: Ngo Nyoung M, Trentesaux C, Aries A, Carpentier Y, Jardillier J-C, Gorisse M-C, Jeannesson P. Effect of aclacinomycin doxorubicin association on differentiation and growth of human erythroleu- Figure 4 Time-course of EPOR mrna expression in K562 cells treated with BA. Northern blot of total RNA from untreated and kemic K562 cells. Anticancer Res 1994; 14: mm and 1 mm BA-treated cells were hybridized simultaneously 18 Dean A, Erard F, Schneider AB, Schechter AN. Induction of hemowith EPOR and GAPDH radiolabeled cdnas, and submitted to auto- globin accumulation in human K562 cells by hemin is reversible. radiography. Data from a typical experiment. Science 1981; 212:

5 19 Jeannesson P, Lahlil R, Chénais B, Devy L, Gillet R, Aries A, Mor- R, Jardillier J-C. Induction of erythropoietin receptors during aclaceau F, Trentesaux C. Anthracyclines as tumor cell differentiating cinomycin-mediated erythroid differentiation of K562 leukemia agents: effects on the regulation of erythroid gene expression. Leuk cells. Leukemia 1991; 5: Lymphoma (in press). 23 Trentesaux C, Ngo Nyoung M, Aries A, Morceau F, Ronchi A, 20 Morceau F, Chénais B, Gillet R, Jardillier J-C, Jeannesson P, Tren- Ottolenghi S, Jardillier J-C, Jeannesson P. Increased expression of tesaux C. Transcriptional and post-transcriptional regulation of GATA-1 and NF-E2 erythroid transcription factors during aclacinoerythroid gene expression in anthracycline-induced differentiation mycin-mediated differentiation of human erythroleukemic cells. of human erythroleukemic cells. Cell Growth Differ 1996; 7: Leukemia 1993; 7: Morceau F, Aries A, Lahlil R, Devy L, Jardillier J-C, Jeannesson 21 Aries A, Trentesaux C, Ottolenghi S, Jardillier J-C, Jeannesson P, P, Trentesaux C. Evidence for distinct regulation processes in the Doubeikovski A. Activation of erythroid specific promoters during aclacinomycin and doxorubicin-mediated differentiation of anthracycline-induced differentiation of K562 cells. Blood 1996; human erythroleukemic cells. Biochem Pharmacol 1996; 51: 87: Jeannesson P, Trentesaux C, Ngo Nyoung M, Mayeux P, Jacquot 1579