Nuclear factor-erythroid 2 (NF-E2) expression in normal and ... - Nature

Leukemia (2002) 16, 1773–1781  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu

Nuclear factor-erythroid 2 (NF-E2) expression in normal and malignant megakaryocytopoiesis L Catani1, N Vianelli1, M Amabile1, L Pattacini1, L Valdre`1, ME Fagioli1, M Poli1, L Gugliotta3, P Moi2, MG Marini2, G Martinelli1, S Tura1 and M Baccarani1 1 Istituto di Ematologia e Oncologia Medica ‘L. e A. Sera`gnoli’, University of Bologna, Bologna, Italy; 2Istituto di Biologia e Clinica dell’Eta` Evolutiva, University of Cagliari, Italy; and 3Unita` Operativa Ematologia, Reggio Emilia, Italy

Although the transcription factor nuclear factor-erythroid 2 (NFE2) is known to be functionally linked to the megakaryocytic lineage, little is known about its role in malignant megakaryocytes. We used real-time RT-PCR and Western blotting to investigate expression of NF-E2 and its partner, MafG, in CD34derived normal (five cases) and malignant megakaryocytes from essential thrombocythemia (ET) patients (eight cases) and in megakaryoblastic cell lines. We also quantitated the mRNA of the thromboxane synthase (TXS) gene, which is directly regulated by NF-E2. Although real-time RT-PCR showed that both a and f NF-E2 isoforms were significantly reduced with respect to the normal counterpart both in ET megakaryocytes and in cell lines (P ⭐ 0.01), western blotting revealed decreased NF-E2 protein expression only in the latter. However, both the NF-E2a/MafG mRNA ratio (P ⭐ 0.01) and TXS (P ⭐ 0.01) mRNA expression were significantly reduced in megakaryocytes from ET patients and cell lines with respect to healthy subjects. These two findings provide strong indirect evidence of altered activity of the a isoform of NF-E2 in malignant megakaryocytes, raising the possibility that NF-E2 could play a role in megakaryocyte transformation. Leukemia (2002) 16, 1773–1781. doi:10.1038/sj.leu.2402597 Keywords: NF-E2; MafG; thromboxane synthase; malignant megakaryocytic cells

Introduction The transcription factor nuclear factor-erythroid 2 (NF-E2) has recently attracted much interest in the field of megakaryocytopoiesis.1–4 It is composed of two subunits (p45 and p18), each of which contains a basic region-leucine zipper DNA binding domain.3,5 The p18 subunit is a small Maf protein,6–8 namely MafG, which partners the p45 subunit (ie p45 NF-E2) in megakaryocyte (MK) development.8 The heterodimers of p45 and small Mafs can activate transcription via Mares (ie Maf recognition elements: TGCTGAC(T/GT)CAGCA), whereas the homodimers of small Mafs form repressors that act on the same sites.6,9 Acetilation by cAMP-response element-binding protein (CREB)-binding protein (CBP) augments DNA binding of p45 NF-E2.10 The p45 NF-E2 subunit is almost exclusively expressed in hematopoietic progenitors and differentiated cells of the erythroid, MK, granulocyte and mast cell lineages.5,11–14 Human p45 NF-E2 has two isoforms (a and f), which differ in the 5⬘ untranslated region.13 The gene has three exons: whereas exons 2 and 3 are common to both isoforms, exon 1 differs between the two species (exon 1a and exon 1f). The two isoforms produce the same protein but are regulated by two different promoters and are expressed in different ratios during development. Although both isoforms are always coexpressed in every human cell type, the f mRNA is more abundant in fetal liver and less abundant in adult bone Correspondence: L Catani, Istituto di Ematologia e Oncologia Medica ‘L e A. Sera`gnoli’, Ospedale S Orsola, Via Massarenti 9, 40138 Bologna, Italy; Fax: 051/6364037 Received 17 October 2001; accepted 11 April 2002

marrow (BM) compared to the a isoform.13 Mice lacking p45 NF-E2 show megakaryocytosis and severe thrombocytopenia leading to fatal hemorrhage, and also have mild to moderate anemia.2 Loss of p45 NF-E2 interferes with MK maturation at an advanced stage: ultrastructural studies reveal a markedly reduced number of granules and failure to form platelet fields.2 In contrast to control megakaryocytes, p45 NF-E2−/− megakaryocytes are larger, exhibit a paucity of granules, and show hyperlobulated nuclei.2 Therefore, genes that are regulated by NF-E2 are critical to terminal MK maturation and platelet formation. MafG-deficient mice also have impaired platelet production, but they do not show a propensity to hemorrhage; therefore loss of MafG appears less severe than that of p45 NF-E2.8 It has recently been reported that NF-E2 (or genes regulated by NF-E2) seems to play a major role in integrin ␣IIb␤3 signaling.15 Experimental evidence also suggests that the TXS (thromboxane-synthase) and ␤1-tubulin genes, which are mainly expressed in MK, are directly regulated by p45 NFE2.16,17 TXS transcripts are present at all stages of MK differentiation and increase in quantity during maturation.18 Although TXS is not MK-specific, it is totally absent from p45NF-E2 MKs and normally expressed in other hematopoietic cells.16 Essential thrombocythemia (ET) is a chronic myeloproliferative disorder whose behavior is governed principally by the abnormal proliferation of a malignant MK clone.19,20 ET is usually associated with elevated numbers of bone marrow megakaryocytes and of morphologically and functionally abnormal circulating platelets.19,21,22 Enlarged and mature megakaryocytes with hyperploid nuclei are usually observed in ET bone marrow.23 Even though ET is usually considered to be a clonal disease, it has recently been suggested that some patients do not have a clonal disorder.24 The issue is complicated by the recent demonstration of monoclonal hematopoiesis in normal elderly women.25 The cause of ET is unknown. Since few consistent cytogenetic abnormalities have been reported,26–28 little is known about the existence of any underlying molecular lesions. Due to the importance of NF-E2 in MK development, we investigated whether this gene plays a prominent role in malignant megakaryocytopoiesis. To do this, we studied in vitro the expression of NF-E2 and its partner, MafG, in MKs from ET patients. In particular, primary CD34+ hematopoietic progenitor cells taken from the BM of ET patients and healthy donors were induced to differentiate along the MK lineage in liquid suspension cultures by continuous addition of 100 ng/ml thrombopoietin. Parallel experiments were performed on megakaryoblastic cell lines. NF-E2 and MafG expression pattern were monitored at both the mRNA and protein levels by quantitative reverse transcription-polymerase chain reaction (RT-PCR) and Western blot analysis, respectively. We also quantitated TXS mRNA in normal and malignant MK cells. Our results suggest that the activity (and in cell lines also the expression) of NF-E2 is altered in malignant MK cells.

NF-E2 and malignant megakaryocytopoiesis L Catani et al

1774

Materials and methods

Patients BM samples were collected from eight patients (three males and five females) with ET. The diagnosis of ET followed the PVSG criteria.29 Median age of the patients at the time of testing was 45 years (range 21–56). Median platelet count at test was 1183 × 109/l (range 890–1681). At the time of the study no patient showed hemorrhagic or thrombotic complications. The patients were at diagnosis (five cases) or at least 3 months from any cytotoxic treatment (three cases). Informed consent to the study was obtained according to the Helsinki declaration of 1975 from ET patients and from five healthy BM donors.

Cell lines One erythroid (K-562) and five MK (MK-1, MK-2, HEL, B1647, MO7e) cell lines were also analyzed for the expression of NFE2, MafG and TXS. MK-1 and MK-2 cells30 were maintained in Dulbecco’s modified Eagle’s medium (DMEM) with 4.5 g/l dextrose supplemented with 20% fetal bovine serum (FBS Gibco-Life Technologies, Paisley, UK), 2 mM L-Glutamine (GIBCO, Grand Island, NY, USA), 10−5 M 2-mercaptoethanol (Sigma, St Louis, MO, USA) and antibiotics. HEL cells31 were cultured in RPMI 1640 (Gibco-Life Technologies) plus 10% fetal bovine serum (Gibco-Life Technologies) at an optimal cell density of 0.5–1 × 106/ml. B1647 cells32 were grown in Iscove’s modified Dulbecco’s medium (IMDM) supplemented with 10% fetal calf serum (FCS; Gibco). M07e cells33 were cultured in IMDM plus 10% FBS plus IL-3 (10 U/ml; Genzyme, Boston, MA, USA). K-562 cells were maintained in RPMI 1640 plus FBS 10%.34

Purification of CD34+ cells For CD34+ cell selection, mononuclear cells were isolated from BM aspirates (20–30 ml) by Ficoll–Hypaque (d = 1.077 g/ml; Pharmacia, Uppsala, Sweden), rinsed and adherencedepleted overnight. After removal of adherent cells, CD34+ cells were isolated using a magnetic cell sorting program Mini-MACS (Miltenyi Biotec, Auburn, CA, USA) and the CD34 isolation kit in accordance with the manufacturer’s recommendations. The purity of CD34-selected cells was determined for each isolation by using a monoclonal antibody that recognizes a separate epitope of the CD34 molecule (HPCA2); Becton Dickinson, San Jose`, CA, USA) directly conjugated to fluorescein. The percentage of CD34+ cells ranged from 89% to 97%.

Liquid suspension cultures and purification of CD34derived MKs To avoid the interference of serum, purified CD34+ cells (80 000/ml) were resuspended in a serum-free medium: Iscove’s modified Dulbecco’s medium (GIBCO) containing 10−4 mol/l bovine serum albumin (BSA)-adsorbed cholesterol, nucleosides (uridine, adenosine, deoxyadenosine, guanosine, deoxyguanosine, cytidine, deoxycytidine, and deoxythymidine at 10 ␮g/ml each), 0.5% BSA (Fraction V Chon), 10 ␮g/ml insulin, 200 ␮g/ml iron-saturated transferrin, and 5 × 10−5 Leukemia

mol/l 2-␤-mercaptoethanol (all purchased from Sigma) in the presence of 100 ng/mL of human recombinant thrombopoietin (Genzyme). Every 3 days, viable cells were scored by Trypan blue dye exclusion and cultures were amplified with fresh serum-free medium, readjusting the cell density to 80 000 cells/ml. Each well was then supplemented with 100 ng/ml thrombopoietin. After 14 to 16 days of liquid culture, CD34derived MKs were purified by means of an anti-CD41a monoclonal antibody (MoAb) (Dako, Milan, Italy) directed against the glycoproteic ␣IIb–␤3 complex and immunobeads (MPC 450 Dynabeads; Dynal, Oslo, Norway), as previously described.35 The purity of MKs was determined for each isolation by indirect immunofluorescence, using an anti-CD41b MoAb which reacts with a different epitope of ␣IIb–␤3 subunit (Immunotech, Westbrook, ME, USA), followed by a goat antimouse IgG, covalently linked to fluorescein (Becton Dickinson). Cell lines were also selectively purified for ␣IIb– ␤3+ cells by means of positive immunomagnetic selection.

Quantitative RT-PCR After purification, CD34-derived MKs of the ET patients or healthy subjects and cell lines were processed as follows. Total cellular RNA was extracted using the commercially available RNAzol TM Kit (Cinna/Biotecx, Houston, TX, USA). cDNAs were prepared from 1 ␮g RNA template in 50 ␮l reaction mixtures containing 10 mM Tris HCl pH 8.3, 50 mM KCl, 4 mM MgCl2, 2.5 ␮M random examers (Perkin Elmer, Milan, Italy), 200 ␮M each of dATP, dCTP, dGTP, dTTP (Perkin Elmer), 10 U placental ribonuclease inhibitor (Boehringer Mannheim, Milan, Italy), and 200 U Molony murine leukemia virus (M-MLV RT) (Promega, Milan, Italy). The RT reaction was carried out by incubation at 37°C for 1 h.

Real-time RT-PCR The PCR primers and TaqMan probes to amplify and detect NF-E2 (a and f isoforms), MafG or TXS were designed using the Primer Express software version 1.0 (Perkin Elmer/Applied Biosystems, Foster City, CA, USA) as follows: NF-E2a sense, CCTGCTGTGACTCCACCACA; NF-E2f sense, GGGCATTTTGCCTGGAAA; NF-E2 antisense (common to both isoforms), GCCAGAGTCTGGTCCAGGTTC. The resulting NF-E2a fragment was 68 bp, and spanned from exon 1a to exon 2. The NF-E2f transcript was 69 bp, and spanned from exon 1f to exon 2. The MafG sense and antisense sequences were CCCAATAAAGGAAACAAGGCC and GCACCGACATGGTCACCA, respectively. TXS sense and antisense sequences were ATCCCAAGCCTTCTCCTTTCA and AAACATCCGACGACCAAGATAGTAC, respectively. The TXS and MafG transcripts were 125 or 96 bp, respectively; both spanned from exon 2 to exon 3. The PCR primers were purchased from BT Biotecnica, (Saronno, Italy). The TaqMan probe for NF-E2 (AGAGCCATCTGGGCTTTCCGGG) was placed on exon 2 in order to detect both isoforms. The probe sequence for TXS (spanning exons 2 and 3) was CATTTTTCCGCCAGGGTTTTTGGGA. That of MafG (positioned on exon 3) was AAGCGGGAGCCGGGTGAGAA. The probes were purchased labeled with 6-carboxy-fluorescein (FAM), as the reporter dye, and with 6-carboxy-tetramethyl-rhodamin (TAMRE), as the quencher fluorescent (Perkin Elmer/Applied Biosystems). In order to minimize variability in the results due to differences in the RT efficiency


and/or RNA integrity among the unknown samples, the GAPDH house-keeping gene was also tested. The TaqManGAPDH control reagents (Perkin Elmer/Applied Biosystems) were used to amplify and detect GAPDH as recommended by the manufacturer. GAPDH expression was quantified using a 5⬘-JOE (2,7 dimethoxy-4,5-dichloro-6-carboxyfluorescein)labeled probe. Quantitation of each of the five genes was done using standard curves established from plasmids containing their specific sequences. NF-E2a, NF-E2f, MafG, TXS and GAPDH transcripts were amplified by single PCR from cDNA of a pure population of normal CD34-derived MK cells, using the same primers employed in the real time PCR. The 68, 69, 96, 125 and 241 bp fragments, respectively, were cloned into the pCRII-Topo vector using the TOPO TA cloning kit (Invitrogen, San Diego, CA, USA). The five plasmids of 4 kb were purified using a Qiagen Mini Prep Kit (Qiagen, Chatsworth, CA, USA), according to the manufacturer’s directions. The plasmids were then quantitated and serial dilutions were prepared starting at 107 copies (4 to 6-fold dilutions) and conserved at −20°C. The real-time PCR was performed on an ABI PRISM 7700 Sequence detector, using the ABI PRISM 7700 Sequence detector Software 1.6 (Perkin Elmer/Applied Biosystems36–38). Optimal reaction conditions for amplification of the genes were as follows: 50 cycles of a two step PCR (95°C × 15 s; 60°C × 60 s) after initial denaturation (95°C × 10 min). All the reactions for samples or standard were run in triplicate. The Raji cell line was used as the negative control.14 A standard curve was then constructed plotting the cycle threshold (Ct) against the known copy number of each standard sample. There was a linear decrease of the Ct in proportion to the log of the starting copy number with a correlation coefficient ranging from 0.97 to 1.00 in the five genes. For quantitation, expression of each genes was normalized by comparison with GAPDH expression.36–38 Normalized levels of unknown samples were calculated as the ratios between the investigated gene (NF-E2a, NF-E2f , MafG or TXS) and GAPDH. The following ratios were also calculated: NF-E2a/NF-E2f, NFE2a/MafG and NF-E2f/MafG.

Western blot analysis Total cell lysates were prepared by adding 100 ␮l lysis buffer to PBS-washed cell pellets of MK cells of normal controls, ET patients or megakaryoblastic cell lines (2 × 106 cells). The lysis buffer consisted of 50 mM NaF, 1 mM EDTA, 250 mM NaCl, 50 mM Tris, pH 7.5, 0.1% Triton X-100 in the presence of protease inhibitors (Sigma). The cell suspension was incubated at 4°C for 30 min and then centrifuged for 20 min at 13 000 g at 4°C. The supernatant was recovered and frozen at −80°C. The protein concentrations were determined by the DC Protein assay (Bio-Rad, Munchen, Germany) according to the manufacturer’s instructions. Twenty micrograms of protein from each sample were diluted to 12 ␮l with distilled water and then added to 12 ␮l 2 × sample buffer (0.75 mol/l of TrisHCl pH 8.8, 4% SDS, 20% glycerol, 1.25% 2-mercaptoethanol and 0.5% bromophenol blue). Cruz marker molecular weight standards (Santa Cruz Biotechnology, Santa Cruz, CA, USA) and samples were loaded on to 10% SDSacrylamide/bis-acrylamide (29:1, Sigma) gel, electrophoresed at 100 V for 90 min and blotted on to nitrocellulose membranes (Hybond C; Amersham, Arlington, IL, USA). The transfer buffer consisted of 0.05 mol/l of Tris-HCl, 0.038 mol/l of glycine and 20% methanol. After blocking overnight at 4°C in 5% non-fat milk and washing with TBS-0.1% Tween 20,

the membranes were incubated with a polyclonal rabbit antiNF-E2 p45 or anti-NF-E2 p18 (1:2000; Santa Cruz) for 2 h at 4°C. The membranes were then washed and incubated with horseradish peroxidase-conjugated anti-rabbit IgG (1:4000; Santa Cruz) for 2 h at RT. After washing, the membranes were incubated with the detection reagents (ECL Western blotting detection system; Amersham) and then developed by autoradiography. Membranes were re-probed for actin (as a loading control), using monoclonal goat anti-actin antibody (Santa Cruz) for detection.

1775

Statistical analysis Comparisons among groups were performed by KruskalWallis (KW) test or Wilcoxon test, as appropriate. All tests were two sided. P values ⬍ 0.05 were considered significant. Results

Characterization of the cultured CD34-derived MKs from normal donors and ET patients or cell lines In order to characterize MK maturation, CD34+ BM cells from normal donors or ET patients were cultured in serum-free liquid cultures in the presence of 100 ng/ml thrombopoietin, which was added in culture at day 0 and then every 3 days during the period of incubation. After the full maturation time (14–16 days of culture), the total number of viable cells was counted by Trypan blue dye exclusion. MK differentiation was assessed by light microscopy evaluation after May–Grunwald Giemsa staining, and by analysis of ␣IIb␤3 expression with a CD41a MoAb and flow cytometry. After 14–16 days of culture, the number of viable cells expanded 8- to 16-fold with respect to day 0, and polyploid megakaryocytes were detectable at light microscopy examination (Figure 1a). Morphological examination showed the presence of cells belonging to various maturational stages of megakaryocytopoiesis with small mononucleated megakaryoblasts together with fully developed MKs. The MK maturation was confirmed by the high-level expression of ␣IIb␤3+ cells in CD34-derived MKs. ET and normal cells showed a percentage of ␣IIb␤3+ cells ranging from 90 to 94% (Figure 1b). ␣IIb␤3 expression of cell lines was also evaluated (Figure 1c).

Expression of NF-E2a and NF-E2f isoforms in malignant MK cells We initially investigated whether NF-E2 expression was deregulated in the CD34-derived MK cells of ET. For this purpose, CD34+ cells, either from normal subjects (five cases) or from ET patients (eight cases) were cultured in a serum-free medium plus TPO (100 ng/ml) for 14–16 days. We then analyzed the NF-E2 mRNA expression in a pure population of CD34-derived MKs. In parallel experiments, we also studied NF-E2 mRNA expression in the (MK1, Mk2, B1647, HEL, M07e, K562) cell lines. Using real time RT-PCR, we quantitated the a and f isoforms of NF-E2 mRNA in each sample, expressed as normalized NF-E2a/GAPDH and NF-E2f/GAPDH ratios (Figure 2). The NF-E2a/GAPDH ratio was significantly reduced in ET patients with respect to healthy subjects (median 0.085, range 0.026–0.16 vs median 0.32, range 0.24–0.4; P = 0.007). In the cell lines, the NF-E2a/GAPDH Leukemia


1776

Figure 1 Analysis of morphology (a) and ␣IIb␤3 expression (b) of CD34+ cells from normal subjects after 16 days of liquid culture with 100 ng/ml thrombopoietin. (c) ␣IIb␤3 expression of cell lines. (a) Cytospins were observed at light microscopy after May-Grunwald staining (× 40 magnification). Note the presence of cells belonging to various maturational stages of megakaryocytopoiesis. ␣IIb␤3 expression (white area) of normal CD34-derived MK cells (b) and cell lines (c) was examined with anti-CD41a MoAb directly conjugated to phycoerytrin and fluorescence was analyzed by flow cytometry. Negative control (black area) is represented by cells treated with an isotype-matched (anti-CD8) MoAb. xaxis: fluorescence intensity; y-axis: relative number of cells.

Leukemia


Expression of MafG and TXS in malignant MK cells

1777

To further characterize the NF-E2 expression in the malignant MK cells, we investigated the expression of the MafG gene, the partner of NF-E2 in MK differentiation, and of the TXS gene, which is directly regulated by NF-E2. As shown in Figure 3, the MafG/GAPDH ratio was similar in healthy subjects (median 0.006, range 0.003–0.01) and ET patients (median 0.007, range 0.004–0.01), but was significantly lower in the cell lines (median 0.002, range 0.001–0.005) (P = 0.001 and P = 0.04 vs ET patients and normal subjects, respectively). Since NF-E2/MafG heterodimers and Maf/Maf homodimers, respectively, seem to act as positive and negative regulators of transcription,6,9,39 we analyzed the NF-E2/MafG ratio for both the a and f NF-E2 isoforms. No difference in the NFE2f/MafG ratio was observed among healthy subjects (median 1.4, range 1.0–3.1), ET patients (median 1.2, range 0.9–2.3) and cell lines (median 1.4, range 1.2–2.4) (Figure 4). However, the NF-E2a/MafG ratio was significantly lower in ET patients than in healthy subjects (median 11.6, range 7.2–14.4 vs median 22.1, range 16.5–31.7; P = 0.01); in cell lines, the ratio (median 6.2, range 1.9–14.6) was significantly lower with respect only to healthy subjects (P = 0.002) (Figure 4). Interestingly, as can be seen from Figure 5, the TXS/GAPDH ratio was significantly lower in cell lines (median 0.0004, range 0.0002–0.01) with respect to both ET patients (median 0.02, range 0.004–0.03) and normal subjects (median 0.04, range 0.03–0.06) (P = 0.001 and P = 0.0007, respectively). The reduction in the ratio between ET patients and healthy subjects was also significant (P = 0.01).

NF-E2 and Maf protein expression In parallel experiments, the MKs from healthy subjects and ET patients or cell lines were also tested for NF-E2 and MafG proteins expression by Western blotting (Figure 6). In the cell lines, NF-E2 protein expression was clearly reduced with respect to the MKs of either the healthy subjects or ET patients. No difference in NF-E2 protein expression was observable between ET patients and healthy subjects. Maf protein expression was comparable in healthy subjects and ET Figure 2 NF-E2a and f isoforms mRNA quantitation by real time RT-PCR in CD34-derived MK cells of normal subjects and ET patients, and in cell lines. (a) NF-E2a/GAPDH ratio; (b) NF-E2f/GAPDH ratio; (c) NF-E2a/NF-E2f ratio. Each point represents the mean of two experiments performed in triplicate. The black line represents the median value. Apart from the comparison between ET patients and controls of NF-E2a/NF-E2f ratio, all the other pairwise comparisons among the three groups were significant (P = 0.001 overall, KW test).

ratio was even lower (median 0.022, range 0.008–0.03; P = 0.01 vs healthy subjects; P = 0.003 vs ET patients). A similar pattern was apparent for the NF-E2f/GAPDH ratio (healthy subjects: median 0.03, range 0.025–0.042; ET patient: median 0.01, range 0.006–0.027; cell lines: median 0.0045, range 0.0025–0.01; healthy subjects vs ET patients and cell lines P = 0.007 and P = 0.01, respectively; ET patients vs cell lines, P = 0.02). Analysis of the NF-E2a/NF-E2f ratio revealed a nonsignificant decrease between healthy subjects and ET patients (median 10.1, range 8.5–12.9 vs median 6.7, range 5.4–10.6). In the cell lines, the NF-E2a/NF-E2f ratio (median 4.065, range 2.5–6.7) was significantly lower than in either the ET patients (P = 0.01) or the healthy subjects (P = 0.01).

Figure 3 MafG mRNA quantitation (MafG/GAPDH ratio) by real time RT-PCR in CD34-derived MK cells of normal subjects and ET patients, and in cell lines. Each point represents the mean of two experiments performed in triplicate. The black line represents the median value. Apart from the comparison between ET patients and controls, all the other pairwise comparisons among the three groups were significant (P = 0.01 overall, KW test). Leukemia


1778

Figure 6 Expression of NF-E2 and Maf proteins in normal and malignant MK cells by Western blot analysis. (a) Lanes 1–8: ET CD34derived MK cells; (b) lanes 9–11: normal CD34-derived MK cells; lanes 12–16: megakaryocytic cell lines (lane 12: B1647; lane 13: HEL; lane 14: M07e; lane 15: MK1; lane 16: MK2). ␤-actin signal was used to verify equivalent sample loading.

Figure 4 NF-E2f/MafG ratio (a) and NF-E2a/MafG ratio (b) in CD34-derived MK cells of normal subjects and ET patients, and in cell lines. Each point represents the mean of two experiments performed in triplicate. The black line represents the median value. Whereas no difference in NF-E2f/MafG ratio was observed among the three groups, for the NF-E2a/MafG ratio all the pairwise comparisons were significant (P = 0.002 overall, KW test) except for that between ET patients and cell lines.

Figure 5 TXS/GAPDH ratio of normal and malignant MK cells. Each point represents the mean of two experiments performed in triplicate. The black line represents the median value. All the pairwise comparisons among the three groups were significant (P = 0.001 overall, KW test).

patients (it is noteworthy that, with respect to the other four cell lines, B1647 showed lower Maf protein expression). Discussion ET is thought to be caused by molecular lesions in critical genes responsible for regulating the proliferation and differenLeukemia

tiation phases of megakaryocytopoiesis.19 The initial hypothesis that the c-Mpl oncogene receptor and thrombopoietin play fundamental roles in the genesis of all malignant MK disorders including ET19,20 has lost ground. Although thrombopoietin gene mutations are present in familial thrombocytosis, causing increased plasma levels of the ligand,40 in ET there is no evidence of structural mutations in either gene,41,42 suggesting the involvement of other molecules. Among the various transcription factors implicated in megakaryocytopoiesis, FOG, GATA-1, and NF-E2 play essential roles.43–46 In mice, the NF-E2 deficient phenotype differs markedly from other transcription knockouts affecting the MK lineage.2 For example, the thrombocytopenia encountered in GATA-1deficient mice derives not only from incomplete cytoplasmic MK maturation but also from nuclear immaturity (the anemia found in these mice is particularly severe).45 Furthermore, mice lacking FOG show failure of megakaryocytopoiesis at a much earlier stage and also suffer from arrested erythropoiesis.46 Thus, in a possible hierarchy of transcription factors involved in megakaryocytopoiesis, FOG and GATA-1 seem to affect only the early phase of MK development, while terminal MK maturation and platelet production appear to be regulated by NF-E2.46 However, NF-E2 may have a broad spectrum of action, since it also appears to be required to regulate replication of the progeny of committed MK precursors.47 In the present study, we investigated whether deregulated expression and/or activity of the NF-E2 transcription factor could play a prominent role in the malignant megakaryocytopoiesis that characterizes ET and megakaryoblastic cell lines (MK-1, MK-2, HEL, B1647, M07e). Among the various transcription factors that affect megakaryocytopoiesis we chose to study NF-E2 because: (1) the syndrome developed by NF-E2 knockout mice specially affects megakaryocytopoiesis (and only to a lesser extent erythropoiesis or granulocytopoiesis); (2) although ET megakaryocytopoiesis is abnormal at all stages of development, the increased ploidy and the prevalence of large, mature cells suggest a link with NF-E2 deregulation; (3) retroviral integration within the Fli-2 locus results in inacti-


vation of NF-E2 in Friend erythroleukemias, suggesting that loss of NF-E2 function may contribute to leukemogenesis.48 Our results indicate that the activity, but not the expression, of NF-E2 is altered in malignant MK cells from ET patients. With respect to normal cells, expression of the two isoforms (a and f) of the NF-E2 gene was reduced both in ET MKs and in malignant cell lines bearing MK features. However, decreased NF-E2 protein expression was observable only in the cell lines and not in the patients’ MKs. On the other hand, analysis of the NF-E2a/MafG mRNA ratio and TXS mRNA expression did provide strong indirect evidence that the activity of the a isoform of NF-E2 is altered in malignant MKs. Indeed, with respect to healthy subjects, both these parameters were significantly reduced in MK cells from ET patients (as well as in the cell lines). Regulation of MK transcription by NF-E2 seems to be dependent on the abundance of small Maf proteins.4,6,9,39 It is thought that when there is a lower amount of small Mafs than NF-E2, Maf-NFE2 heterodimers become the dominant species in the cells, activating transcription, whereas when the small Mafs exceed the NF-E2, they form Maf homodimers or heterodimers and suppress transcription by occupying the MARE sites.6 In our study, NF-E2 mRNA expression always exceeded that of MafG. However, the reduced NF-E2a/MafG ratio in the malignant MK cells from ET patients suggests an alteration in the balance between the a (but not the f) isoform of NF-E2 and MafG. This should have a direct effect on TXS gene expression, which is directly regulated by NF-E2.16 In fact, we recorded significantly downregulated TXS mRNA expression among the malignant MK cells from ET patients. The downregulation was even more pronounced in the cell lines. While the expression of the TXS gene in ET MKs could be referred to altered Nf-E2a activity only, the immaturity of the cell lines could further limit the production of transcripts. The reduced amount of NF-E2 mRNA and protein in the malignant MK cell lines studied by us does not seem to be related to the immaturity of the cells. NF-E2 is present early in MK ontogeny4 and Toki et al49 demonstrated that NF-E2 transcripts in both a and f isoforms were detectable in maturating MKs at day 7 of culture. In order to better characterize NF-E2 protein expression, we also checked whether the protein was differently localized in the malignant cells (from ET patients or cell lines) in comparison with the normal MKs. Fluorescence microscopy demonstrated that the protein was localized in the nucleus both in the normal and malignant MKs (data not shown). Thus, altered distribution of the protein would not appear to be responsible for NF-E2 deregulation in malignant MKs. Studying NF-E2 expression during normal human hematopoiesis in the erythroid and granulocytic compartments, Labbaye et al50 found that while the mRNA and protein were only barely detectable in quiescent hematopoietic progenitor cells, they were gradually induced in both erythroid and granulocytic cultures. During maturation, NF-E2 further accumulated in the erythroid pathway, whereas it was suppressed in the granulopoietic series. Toki et al49 recently characterized the two NF-E2 isoforms in CD34-derived MKs from human umbilical cord blood, and found a prevalence of the a isoform. To our knowledge, the present study is the first to investigate NF-E2 expression in human adult BM MKs. Our mRNA analysis indicated that in normal human MKs the a isoform is much more highly expressed than the f isoform. This finding is also consistent with the preponderance of the a isoform found by Pischedda et al13 and Moroni et al51 in adult human BM and the myeloid component of mouse BM, respectively.

In addition, the a isoform is known to predominate in human peripheral granulocytes, which abundantly express the NFE2 transcript.52 As regards the f isoform, Moroni et al51 found that it predominated (comprising about 75% of the total transcript) in a highly-enriched erythroid cell population derived from adult mouse BM. It was also recently reported that the f isoform is more expressed in the cells of the erythroid pathway than in MK cells, and is downregulated in HEL cells after exposure to maturating agents like phorbol myristate acetate.49 Therefore, the f isoform seems to play a major role in the erythroid compartment. Our finding that the NF-E2a/MafG ratio but not the NF-E2f/MafG ratio was reduced in MKs of ET patients and in the malignant cell lines is in keeping with the concept that only the a isoform plays a major role in MK development.13,49 Furthermore, while both NF-E2 isoforms (and especially NFE2a) were significantly decreased in the mRNA of malignant MKs (derived from both ET patients and cell lines) with respect to their normal counterpart, the overall prevalence of the a isoform was retained. These findings can all probably be explained by the more-erythroid specific characteristics of the f isoform. In conclusion, we provide the first example of detailed characterization of NF-E2 in human malignant MK cells. Taken together, our results clearly demonstrate that the a NFE2 isoform prevails over the f in normal and malignant human BM MK cells. They also suggest that the NF-E2 transcription factor could play a prominent role in the malignant megakaryocytopoiesis of ET and megakaryoblastic cell lines. It remains to be seen whether altered NF-E2a activity in some way influences MK malignant transformation or whether it is a consequence of the malignancy. Since retroviral integration within the Fli-2 locus results in inactivation of NF-E2 in Friend erythroleukemias, it is thought that the gene may act as a tumor suppressor whose disruption or dysfunction confers a growth advantage to the leukemic phenotype.48 Therefore, it would be interesting to evaluate NF-E2 expression and activity in cells from acute megakaryoblastic leukemia patients.

1779

Acknowledgements This work was supported by ‘Giovani Ricercatori’ Project 1999–2000, University of Bologna. We are grateful to Robin MT Cooke for helping work up the manuscript.

References 1 Andrews NC, Erdjument-bromage H, Davidson MB, Tempst P, Orkin SH. Erythroid transcription factor NF-E2 is a haematopoietic-specific basic-leucine zipper protein. Nature 1993; 362: 722–728. 2 Shivdasani RA, Rosenblatt MF, Zucker-Franklin D, Jackson CW, Hunt P, Saris C, Orkin SH. Transcription factor NF-E2 is required for platelet formation independent of the actions of thrombopoietin/MGDF in megakaryocyte development. Cell 1995; 92: 8690–8694. 3 Andrews NC. The NF-E2 transcription factor. Int J Biochem Cell Biol 1998; 30: 429–432. 4 Shivdasani RA. Molecular and transcriptional regulation of megakaryocyte differentiation. Stem Cells 2001; 19: 397–407. 5 Chan JY, Han X, Kan YW. Isolation of cDNA encoding the human NF-E2 protein. Proc Natl Acad Sci USA 1993; 90: 11366–11370. 6 Motohashi H, Shavit JA, Igarashi K, Yamamoto M, Engel JD. The world according to MAF. Nucleic Acids Res 1997; 25: 2953– 2959. Leukemia


1780

Leukemia

7 Blank V, Andrews NC. The Maf transcription factors: regulators of differentiation. Trends Biochem Sci 1997; 22: 437–441. 8 Shavit JA, Motohashi H, Onodera Ko, Akasaka J, Yamamoto M, Engel JD. Impaired megakaryopoiesis and behavioral defects in mafG-null mutant mice. Genes & Dev 1998; 12: 2164–2174. 9 Igarashi K, Kataoka K, Itoh K, Hayashi N, Nishizawa M, Yamamoto M. Regulation of transcription by dimerization of erythroid factor NF-E2 p45 with small Maf proteins. Nature 1994; 367: 568–572. 10 Hung HS, Kim AY, Hong W, Rakowski C, Blobel GA. Stimulation of NF-E2 DNA binding by CREB-binding protein (CBP)-mediated acetilation. J Biol Chem 2001; 276: 10715–10721. 11 Peters LL, Andrews NC, Eicher EM, Davidson MB, Orkin SH, Lux SE. Mouse microcytic anaemia caused by a defect in the gene encoding the globin enhancer-binding protein NF-E2. Nature 1993; 362: 768–770. 12 Bean TL, Ney PA. Multiple regions of p45 NF-E2 are required for beta-globin gene expression in erythroid cells. Nucleic Acids Res 1997; 25: 2509–2515. 13 Pischedda C, Cocco S, Melis A, Marini MG, Kan YW, Cao A, Moi P. Isolation of a differentially regulated splicing isoform of human NF-E2. Proc Natl Acad Sci USA 1995; 92: 3511–3515. 14 Lecine P, Villeval J, Vyas P, Swencki B, Xu Y, Shivdasani RA. Mice lacking transcription factor NF-E2 provide in vivo validation of the proplatelet model of thrombocytopoiesis and show a platelet production defect that is intrinsic to megakaryocytes. Blood 1998; 92: 1608–1616. 15 Shiraga M, Ritchie A, Aidoudi S, Baron V, Wilcox D, White G, Ybarrondo B, Murphy G, Leavitt A, Shattil S. Primary megakaryocytes reveal a role for transcription factor NF-E2 in Integrin ␣IIb␤3 signaling. J Cell Biol 1999; 147: 1419–1429. 16 Deveaux S, Cohen-Kaminsky S, Shivdasani RA, Andrews NC, Filipe A, Kuzniak I, Orkin SH, Romeo PH, Mignotte V. p45 NF-E2 regulates expression of thromboxane synthase in megakaryocytes. EMBO J 1997; 16: 5654–5661. 17 Lecine P, Italiano Jr JE, Kim S-E, Villeval J-L, Shivdasani RA. Hematopoietic-specific ␤1 tubulin partecipates in a pathway of platelet biogenesis dependent on the transcription factor NF-E2. Blood 2000; 96: 1366–1373. 18 Vitrat N, Letestu R, Masse` A, Lazar V, Vainchenker W, Debili N. Thromboxane synthase has the same pattern of expression as platelet specific glycoproteins during human megakaryocyte differentiation. Thromb Haemost 2000; 83: 759–768. 19 Cripe LD, Hromas R. Malignant disorders of megakaryocytes. Semin Hematol 1998; 35: 200–209. 20 Tefferi A. Pathogenetic mechanisms in chronic myeloproliferative disorders: polycythemia vera, essential thrombocythemia, agnogenic myeloid metaplasia and chronic myelogenous leukemia. Semin Hematol 1999; 36: 3–8. 21 Catani L, Gugliotta L, Cascione ML, Mattioli Belmonte M, Vianelli N, Belardinelli AR, Tura S. Platelet function and interferon alpha2a treatment in essential thrombocythaemia. Eur J Haematol 1991; 46: 158–162. 22 Green AR. The pathogenesis and management of essential thrombocythaemia. Haematologica 1999; 84: 36–39. 23 Thiele J, Kvasnicka HM, Zankovich R, Diehl V. Relevance of bone marrow features in the differential diagnosis between essential thrombocythemia and early stage idiopathic myelofibrosis. Haematologica 2000; 85: 1126–1134. 24 Harrison CN, Gale RE, Machin SJ, Lynch DC. A large proportion of patients with a diagnosis of essential thrombocythemia do not have a clonal disorder and may be at lower risk of thrombotic complications. Blood 1999; 93: 417–424. 25 Champion KM, Gilbert JG, Asimakopoulos FA, Hinshelwood S, Green AR. Clonal haematopoiesis in normal elderly women: implications for the myeloproliferative disorders and myelodysplastic syndromes. Br J Haematol 1997; 97: 920–926. 26 Sessarego M, Defferrari R, Dejana AM, Rebuttato AM, Fugazza G, Salvidio E, Ajmar F. Cytogenetic analysis in essential thrombocythemia at diagnosis and at transformation. A 12-year study. Cancer Genet Cytogenet 1989; 43: 57–65. 27 Reis MD, Sher GD, Lakhani A, Dube ID, Senn JS, Pinkerton PH. Deletion of the long arm of chromosome 5 in essential thrombocythemia. Cancer Genet Cytogenet 1992; 61: 93–95. 28 Munro LR, Stevenson DA, Culligan DJ. Translocation (X;5)

29

30

31 32

33

34 35

36

37 38

39 40 41 42

43 44 45 46 47

(q13;q33) in essential thrombocythemia. Cancer Genet Cytogenet 1999; 114: 78–79. Michiels JJ, Juvonen E. Proposal for revised diagnostic criteria of essential thrombocythemia and polycythemia vera by the Polycythemia Vera Study group. Semin Thromb Haemost 1997; 23: 339–347. Di Noto R, Luciano L, Lo Pardo C, Ferrara F, Frigeri F, Mercuro O, Lombardi ML, Pane F, Vacca C, Manzo C, Salvatore F, Rotoli B, Del Vecchio L. JURL-MK1 (c-kit(high)/CD30-/CD40-) and JURLMK-2 (c-kit(low)/CD30+/CD40+) cell lines: ‘two sided’ model for investigating leukemic megakaryocytopoiesis. Leukemia 1997; 11: 1554–1564. Martin P, Papayannopoulou T. HEL cells: a new human erythroleukemia cell line with soontaneous and induced globin expression. Science 1982; 216: 1233–1235. Bonsi L, Grossi A, Strippoli P, Tumietto F, Tonelli R, Vannucchi MA, Ronchi A, Ottolenghi S, Visconti G, Avanzi GC, Pegoraro L, Bagnara GP. An erythroid and megakaryocytic common precursor cell line (B1647) expressing both c-mpl and erythropoietin receptor (Epo-R) proliferates and modifies globin chain synthesis in response to megakaryocyte growth and development factor (MGDF) but not to erythropoietin (Epo). Br J Haematol 1997; 98: 549–559. Avanzi GC, Lista P, Giovinazzo B, Miniero R, Saglio G, Benetton G, Coda R, Cattoretti G, Pegoraro L. Selective growth response to IL-3 of a human leukaemic cell line with megakaryoblastic features. Br J Haematol 1988; 69: 359–366. Lozzio CB, Lozzio BB. Human chronic myelogenous leukemia cell line with positive Philadelphia chromosome. Blood 1975; 45: 321–334. Zauli G, Catani L, Gibellini D, Re MC, Milani D, Borgatti P, Bassini A, La Placa M. The CD4 receptor plays essential but distinct roles in HIV-1 infection and induction of apoptosis in primary bone marrow GPIIb/IIIa+ megakaryocytes and the HEL cell line. Br J Haematol 1995; 91: 290–298. Holland PM, Abramson RD, Watson R, Gelfand DH. Detection of specific polymerase chain reaction product by utilizing the 5⬘ 3⬘ exonuclease activity of Thermus aquaticus DNA polymerase. Proc Natl Acad Sci USA 1991; 88: 7276–7280. Heid CA, Stevens J, Livak KJ, Williams PM. Real time quantitative PCR. Genome Res 1996; 6: 986–994. Krauter J, Watties MP, Nagel S, Heidenech O, Krug U, Kafert S, Bunjes D, Bergmann L, Ganser A, Gerhard HL. Real-time RT-PCR for the detection and quantification of AML-1/MTG8 fusion transcripts in t(8–21)-positive AML patients. Br J Haematol 1999; 107: 80–85. Nagai T, Igarashi K, Akasaka J, Furuyama K, Fujita H, Hayashi N, Yamamoto M, Sassa S. Regulation of erythroleukemia cell differentiation. J Biol Chem 1998; 273: 5358–5365. Wiestner A, Schlemper RJ, van der Maas AP, Skoda RC. An activating splice donor mutation in the thrombopoietin gene causes hereditary thrombocythemia. Nat Genet 1998; 18: 49–53. Kiladjan JJ, El Kassar N, Hetet G, Briere J, Grandchamp B, Gardin C. Study of the thrombopoietin receptor in essential thrombocythemia. Leukemia 1997; 11: 1821–1826. Harrison CN, Gale RE, Wiestner AC, Skoda RC, Linch DC. The activating splice mutation in intron 3 of the thrombopoietin gene is not found in patients with non-familial essential thrombocythemia. Br J Haematol 1998; 102: 1341–1343. Mouthon MA, Bernard O, Mitjavilla MT, Romeo PH, Vainchenker W, Mathieu-Mahul D. Expression of tal-1 and GATA-binding proteins during human hematopoiesis. Blood 1993; 81: 647–652. Lemarchandel V, Ghysdael J, Mignotte V, Rahuel C, Romeo PH. GATA and Ets cis-acting sequences mediate megakaryocyte-specific expression. Mol Cell Biol 1993; 13: 668–704. Vyas P, Ault K, Jackson CW, Orkin S, Shivdasani R. Consequences of GATA-1 deficiency in megakaryocytes and platelets. Blood 1999; 93: 2867–2875. Tsang AP, Fujiwara Y, Hom DB, Orkin SH. Failure of megakaryocytopoiesis and arrested erythropoiesis in mice lacking the GATA 1 transcriptional cofactor FOG. Genes Dev 1998; 12: 1176–1188. Levin J, Peng J-P, Baker JR, Villeval J-L, Lecine P, Burstein SA, Shivdasani RA. Pathophisiology of thrombocytopenia and anemia in mice lacking transcription factor NF-E2. Blood 1999; 94: 3037–3047.


48 Lu SJ, Rowas S, Bani MR, Ben-David Y. Retroviral integration within the Fli-2 locus results in inactivation of the erythroid transcription factor NF-E2 in Friend erythroleukemias: evidence that NF-E2 is essential for globin expression. Proc Natl Acad Sci USA 1994; 91: 8398–8402. 49 Toki T, Arai K, Terui K, Komatsu N, Yokoyama M, Katsuoka F, Yamamoto M, Ito E. Functional characterization of the two alternative promoters of human p45 NF-E2 gene. Exp Hematol 2000; 28: 1113–1119. 50 Labbaye C, Valtieri M, Barberi T, Meccia E, Masella B, Pelosi E, Condorelli GL, Testa U, Peschle C. Differential expression and

functional role of GATA-2, NF-E2, and GATA-1 in normal adult hematopoiesis. J Clin Invest 1995; 95: 2346–2358. 51 Moroni E, Mastrangelo T, Razzini R, Cairns L, Moi P, Ottolenghi S, Giglioni B. Regulation of mouse p45 NF-E2 transcription by an erythroid-specific GATA-dependent intronic alternative promoter. J Biol Chem 2000; 275: 10567–10576. 52 Toki T, Itoh J, Arai K, Kitazawa J, Yokoyama M, Igarashi K, Yamamoto M, Ito E. Abundant expression of erythroid transcription factor p45 NF-E2 mRNA in human peripheral granulocytes. Biochem Biophys Res Commun 1996; 27: 760–765.

1781

Leukemia