The release of transforming growth factor-β1 (TGF-β1) in the bone marrow microenvironment is one of the main mechanisms leading to myelofibrosis in murine models and probably in the human idiopathic myelofibrosis (IMF). The regulation of TGF-β1 synthesis is poorly known but seems regulated by nuclear factor κB (NF-κB). We previously described the overexpression of an immunophilin, FK506 binding protein 51 (FKBP51), in IMF megakaryocytes. Gel shift and gene assays show that FKBP51's overexpression in a factor-dependent hematopoietic cell line, induces a sustained NF-κB activation after cytokine deprivation. This activation correlates with a low level of IκBα. A spontaneous activation of NF-κB was also detected in proliferating megakaryocytes and in circulating CD34+ patient cells. In normal cells, NF-κB activation was only detected after cytokine treatment. The expression of an NF-κB superrepressor in FKBP51 overexpressing cells and in derived megakaryocytes from CD34+ of IMF patients revealed that NF-κB activation was not involved in the resistance to apoptosis after cytokine deprivation of these cells but in TGF-β1 secretion. These results highlight the importance of NF-κB's activation in the fibrosis development of this disease. They also suggest that FKBP51's overexpression in IMF cells could play an important role in the pathogenesis of this myeloproliferative disorder.
Idiopathic myelofibrosis (IMF) is a very rare hematologic disorder with an annual incidence in European countries of 1 to 5 per 106 inhabitants. Although IMF is a clonal stem cell disorder, there are several lines of evidences that show that myelofibrosis is secondary to the proliferation of the malignant clone and to the marrow release of fibrotic cytokines. A marked proliferation of morphologically abnormal megakaryocytes is consistently observed in IMF, and at the initial phase of the disease, the reticulin deposition surrounds the clusters of marrow megakaryocytes (1–3). Furthermore, megakaryocytes likely play key roles in triggering fibrosis through their production of numerous cytokines, particularly in transforming growth factor-β1 (TGF-β1). Others have emphasized the role of monocytes/macrophages in the pathogenesis of myelofibrosis (4, 5).
Until quite recently, the molecular underprinnings of myelofibrosis have been unclear, but new insights have come from mouse models. Two models mimicking human primitive myelofibrosis have been recently described: one is due to a thrombopoietin overexpression (6, 7) and the other is due to low expression of GATA-1 (GATA-1low mice; ref. 8). These two models highlight the role of the megakaryocytes in the development of the fibrosis. Furthermore, it could be shown that induction of the fibrosis is related to TGF-β1 synthesis in the murine model obtained by thrombopoietin overexpression (9). These two models have allowed us to move forward towards the understanding of the myelofibrosis mechanisms but not enough to characterize the genes involved in the development of the malignant clone. Indeed, in IMF, megakaryocyte hyperplasia occurs in the absence of a recognizable cytokine stimulus, and no alteration in the level of GATA-1 has been observed (10). Recently, we have shown that an immunophilin called FK506 binding protein 51 (FKBP51) was overexpressed in spontaneously grown megakaryocytes of IMF patients (11). Overexpression of FKBP51 leads to a resistance to apoptosis induced by cytokine deprivation through sustained signal transducers and activators of transcription 5 (STAT5) activation. A similar role of STAT5 was found in IMF CD34+ cells and megakaryocytes suggesting that FKBP51 may be implicated in the pathogenesis of this disease (12).
Although it seems clear that the STAT5 pathway is involved in the proliferation of the malignant clone, other signaling pathways may be implicated in the regulation of TGF-β1 synthesis and secretion, which is directly responsible for this disorder. Rameshwar et al. have previously reported that monocytes from IMF patients presented a spontaneous activation of the nuclear factor κB (NF-κB) transduction pathway, and that NF-κB inhibition by antisense oligonucleotides decreased TGF-β1 secretion (5). Therefore, the activation of the NF-κB may be important to the disorder of the secretion of cytokines such as TGF-β1 in IMF. Because IMF is a clonal stem cell disorder and involves all hematopoietic cell lineages, we hypothesized that this pathway could also be spontaneously activated in patients with megakaryocytes and CD34+ cells. Moreover, FKBP51 overexpression recapitulates the same phenotype as in the one observed in IMF cells, that is to say the activation of STAT5 and its antiapoptotic phenotype. We examined whether FKBP51 could also be involved in NF-κB activation and TGF-β1 secretion, in light of the potential relationship between FKBP51 and the NF-κB transduction pathway, established recently (13).
Our data provides a direct demonstration that the NF-κB pathway is spontaneously activated in IMF CD34+ cells and megakaryocytes deriving from CD34+ cells. Furthermore, FKBP51 overexpression also leads to NF-κB's activation as well as the release of TGF-β1 from hematopoietic cells. This study highlights the possible role of the NF-κB pathway in this disorder and on the different possibilities to target the pathology to reduce myelofibrosis development.
Materials and Methods
Cell lines. Factor-dependent cell lines, UT7/Mpl and UT7/Mpl overexpressing FKBP51 (UT7 FKBP51) were cultured in MEM α medium (Life Technologies, Paisley, United Kingdom) supplemented with 10% heat-inactivated FCS, 10 units/mL of penicillin G, 10 μg/mL of streptomycin (all from Life Technologies), and 5 ng/mL of granulocyte macrophage colony–stimulating factor (GM-CSF, a gift from Novartis, Basel, Switzerland). The original UT7 cell line was obtained from N. Komatsu (Jichi Medical School, Tochigi-ken, Japan) and was modified by retroviral transduction of c-mpl.
Patients. Blood samples (20-50 mL) from nine IMF patients (P1-P9) were collected after informed consent. IMF diagnosis was done on bone marrow biopsy for all patients. Moreover, all nine IMF patients showed teardrop poikilocytosis, leukoerythroblastosis (Table 1) and the lack of Bcr-Abl transcript. All patients have been included in previousc publications, P2 in ref. (11) and P1 to P9 in ref. (12). Normal CD34+ cells were obtained after informed consent from either bone marrow aspiration of normal donors or peripheral blood of donors who had been treated with granulocyte colony–stimulating factor (G-CSF) before sampling.
|Patients .||Lille score at diagnosis (38) .||Transfusion requirements .||Therapy before blood sampling .||WBC (×103/mL) .||Immature cells (%) .||Hemoglobin (g/d) .||Platelets (×106/mL) .||Gel shift analysis of NF-κB|
|.||Infection with NF-κB superrepressor .|
|.||.||.||.||.||.||.||.||CD34+ .||MKs .||.|
|Patients .||Lille score at diagnosis (38) .||Transfusion requirements .||Therapy before blood sampling .||WBC (×103/mL) .||Immature cells (%) .||Hemoglobin (g/d) .||Platelets (×106/mL) .||Gel shift analysis of NF-κB|
|.||Infection with NF-κB superrepressor .|
|.||.||.||.||.||.||.||.||CD34+ .||MKs .||.|
Abbreviations: MKs, spontaneously grown megakaryocytes; HU, hydroxyurea; PU, 6-mercaptopurine.
Human CD34+ cell cultures. Mononuclear cells were separated on a Ficoll gradient (Lymphoprep; Nycomed Pharma, Oslo, Norway). Low-density cells (<1.077 g/cm3) were recovered and washed, and CD34+ cells were separated using a magnetic cell sorting system (miniMACS, Miltenyi Biotec GmbH, Bergisch Gladbach, Germany). The purity of recovered cells was determined by flow cytometry using phycoerythrin/anti-CD34 monoclonal antibody (phycoerythrin-HPCA2; Becton Dickinson, Le Pont du Claix, France) and was over 90%.
Cells (1 × 105 cells/mL) were cultured in serum-free medium as previously reported (13) with or without Pegylated recombinant human megakaryocyte growth and development factor (thrombopoietin; a generous gift from Kirin Brewery, Tokyo, Japan).
Real-time PCR. To quantify the FKBP51 and the TGF-β1 expressions in the UT7 cell lines, interexonic quantitative reverse transcription-PCR (RT-PCR) were designed. Briefly, primers and probe were designed using Perkin-Elmer software for quantitative PCR using a forward primer in exon 5 5′-AGGTTCAATGCCAAATTTAGGCT-3′ and a reverse primer in exon 6 5′-AGATATAAAATACATTGTTCTTCCCGCT-3′ for FKBP51 and in exon 6 5′-CCCTACATTTGGAGCCTGGA-3′ and a reverse primer in exon 7 5′-ACAACCAGCATAACCCGGG-3′ for TGF-β1. The probes were designed to hybridize in these two exons (5′-CCTGCCTCTCCAAAACCATATCTTGGTCC-3′ and 5′-ACGCAGTACAGCAAGGTCCTGGCC-3′ for FKBP51 and TGF-β1, respectively). Total RNA was prepared from cell lines using the RNA+ Reagent (Quantum, Appligene, Illkirch, France). RT-PCR was done with random hexamers and real-time PCR were done using the ABI Prism GeneAmp 5700 Sequence Detection System (Perkin-Elmer Applied Biosystems, Foster City, CA). FKBP51 and TGF-β1's expression levels were presented relative to a housekeeping gene, β2-microglobulin (β2m) using GeneAmp 5700 SDS software algorithm.
Calcineurin activity measurement. To determine the calcineurin (PP2B) activity in cells, we used a nonradioactive enzyme assay based on a peptide dephosphorylation kinetic. The activity is defined as the initial velocity determined on the first 6 minutes of its Hill kinetic (36). Cells were suspended in 20 μL of lysis buffer [1 mol/L KCl and 1 mol/L Tris-HCl (pH 8.3)], NP40, and Tween 20. Lysates were mixed with 140 μL of assay buffer [50 mmol/L Tris-HCl (pH 7.0), 0.1 mmol/L EGTA, 0.5 mmol/L DTT, 0.3 mg/mL bovine serum albumin, 0.1 μmol/L calmodulin, 1 mmol/L MnCl2, 1 mmol/L CaCl2, and 500 nmol/L okadaic acid]. A synthetic 19-amino acid peptide corresponding to a portion of the regulatory subunit of type II (RII) cyclic AMP–dependent protein kinase (Bachem, Bubendorf, Switzerland), was used as a substrate. The spectrophotometric detection of peptide dephosphorylation by calcineurin was done using high-performance liquid chromatography on a RP 18 column (Lichrocart 125-4 Lichrospher 100 RP-18, 5 μmol/L; Merck, Darmstadt, Germany). The mobile phase composed by a KH2PO4 buffer (10 mmol/L, pH 5.9) and acetonitrile was run under gradient conditions (initial, 87:13 KH2PO4/acetonitrile and 83:17 KH2PO4/CAN for 6-11 minutes and returned at initial conditions for 11-14 minutes).
Preparation of protein extracts. For the preparation of total cell extracts, cells were washed with PBS and incubated on ice for 30 minutes in lysis buffer [50 mmol/L Tris-HCl, 1% NP40, 0.5% SDS, 150 mmol/L NaCl, and 2 mmol/L EDTA (pH 8.0)] with protease inhibitor cocktail (Boehringer Mannheim, Mannheim, Germany) and phosphatase inhibitor (1 mmol/L sodium orthovanadate). After centrifugation, the supernatant was used for total cell extracts. Cytosol and nuclear extracts were prepared as follows: after washing with PBS, cells were incubated on ice for 10 minutes in buffer A [10 mmol/L HEPES (pH 7.8), 10 mmol/L KCl, 2 mmol/L MgCl2, 0.1 mmol/L EDTA] with protease inhibitor cocktail and NP40 10% was added. Supernatants were used as cytosol extracts. Pellets were resuspended in buffer B [50 mmol/L HEPES (pH 7.8), 50 mmol/L KCl, 300 mmol/L NaCl, 0.1 mmol/L EDTA, and protease inhibitor cocktail] and incubated on ice for 30 minutes. Nuclear debris were pelleted and the supernatants were used as nuclear extracts.
Immunoblot analysis. Immunoprecipitated proteins, total cell extracts, nuclear or cytosolic extracts were resolved on SDS-PAGE, transferred to the nitrocellulose membrane (Bio-Rad, Hercules, CA), and probed with appropriate antibodies such as IκBα (Santa Cruz Biotechnology, Santa Cruz, CA), p65 (Santa Cruz Biotechnology), p50 (Santa Cruz Biotechnology), topoisomerase IIα (Roche, Meylan, France), or β-actin (Sigma-Aldrich, St. Quentin-Fallavier, France). Bands were visualized with either peroxidase-conjugated goat, antimouse or antirabbit IgG (Amersham Biosciences, Orsay, France) and an enhanced chemiluminescence system (ECL kit; Amersham) following to the manufacturer's instructions.
Gel shift assay. Oligonucleotides were 32P-labeled with [γ-32P]ATP using T4 polynucleotide kinase (New England BioLabs, Beverly, MA). Binding reaction mixtures (20 μL) were incubated on ice for 30 minutes containing 0.5 ng DNA probe [NF-κB consensus oligonucleotide; 5′-AGTTGAGGGGACTTTCCCAGG-3′, 3′-TCAACTCCCCTGAAAGGGTCCG-5′ (Santa Cruz Biotechnology), NF-κB p65 oligonucleotide; 5′-AGCTTGGGGTATTTCCAGCCG-3′, 3′-TCGAACCCCATAAAGGTCGGC-5′] and 5 μg nuclear extract in 10 mmol/L Tris-HCl (pH 7.5) buffer with 50 μmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L DTT, 0.1% NP40, and 5% glycerol. For competition experiments, 1-, 10-, or 100-fold excess of unlabeled double-stranded oligonucleotides, were added. DNA-protein complexes were resolved by electrophoresis through 5% native polyacrylamide gels in 0.25× Tris-borate EDTA (1× TBE: 89 mmol/L Tris-HCl, 89 mmol/L boric acid, and 2 mmol/L EDTA). In a super shift assay, 2 μg of antibody against NF-κB p50 (sc-109X, Santa Cruz Biotechnology) were added to the nuclear extracts.
Retrovirus constructions and infections. The following three retroviral vector constructs were used: Migr (long terminal repeat/internal ribosome entry site/enhanced green fluorescence protein, LTR/IRES/EGFP; a generous gift from J. Miller, Philadelphia, PA), Migr-FKBP51 (the retrovirus encoding the entire cDNA from human FKBP51; LTR/FKBP51/IRES/EGFP), and Migr-NF-κB superrepressor. The construction of FKBP51 has been previously reported (11). A retroviral vector encoding an NF-κB superrepressor (NF-κBsr; refs. 15, 16), in which serine residue 32 and 36 of IκBα have been mutated to alanines, was constructed. 293T-EBNA cells were transfected with one of these retrovirus construct together with GAG-POL plasmid and VSV-G plasmid (kindly provided by J. Morgenstein, Millenium Pharmaceuticals, Boston, MA) by using the ExGen kit according to the manufacturer's protocol (Euromedex, Mundolsheim, France). Supernatants were collected and stored frozen until infection.
Infection procedures were as follows: UT7 eGFP and UT7 FKBP51 cells were cultured in MEM α medium supplemented with 50% NF-κBsr retroviral supernatants and selected by a FACS Vantage cytometer (Becton Dickinson) using the eGFP relative enhancement intensity. For human CD34+ cells, infection was done in serum-free medium with (normal control cells) or without (pathologic samples) cytokines such as thrombopoietins (10 ng/mL), stem cell factor (25 ng/mL; gift from Amgen, Thousand Oaks, CA), interleukin 3 (IL-3, 100 IU/mL; a gift from Novartis), and IL-6 (10 ng/mL, a generous gift from S. Burstein, Oklahoma City, OK). EGFP-positive cells were selected using a FACS Vantage cytometer.
Spontaneous growth tests on infected UT7 cells. Cell survival/proliferation after cytokine deprivation were tested in a standard 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay. Cells were washed with PBS thrice and dispensed in a 96-well flat-bottomed plate (Corning, Inc., Corning, NY) at a concentration of 1 × 104 cells in 100 μL per well with or without cytokine. After incubation periods of 24, 48, and 72 hours, a MTT test was done.
Flow cytometry for apoptosis analysis. UT7eGFP cells, UT7 FKBP51 cells, and those infected with Migr-NF-κBsr were deprived of cytokines for 48 hours, labeled by phycoerythrin/Annexin V (Becton Dickinson) for 15 minutes at room temperature in the dark, counterstained with 7-amino-actinomycin D, and analyzed by flow cytometry.
Reporter gene assay. Transcriptional activity of NF-κB was measured by a luciferase-based reporter gene test; 2 × 105 cells/mL of the different cell lines were transfected transiently with pNF-κB-luciferase plasmid (Strategene, La Jolla, CA) by LipofectAMINE (Invitrogen SARL, Cergy Pontoise, France). After transfection, cells were cultured with or without cytokine during 20 hours and cell lysates were subjected to the Luciferase test according to the manufacture's instruction. (Promega, Madison, WI). Relative luciferase units were normalized by protein concentration.
Statistical analysis. In all experiments, statistical analyses were done using the Student's t test. Statistical differences (P < 0.05) are indicated.
Transforming growth factor-β1 secretion. TGF-1 levels in cell culture supernatants from 105 plated cells were determined with an ELISA (human TGF-1 Quantikine Kit; R&D Systems, Oxon, United Kingdom) according to the manufacturer's instructions.
Spontaneous activation of nuclear factor κB in CD34+ cells from idiopathic myelofibrosis and in a cell line overexpressing FKBP51. This study had two main goals: (a) to determine if NF-κB was activated within CD34+ and in megakaryocytes spontaneously grew from IMF patients and (b) to investigate if the imunophilin FKBP51, a gene that we have previously reported to be overexpressed in IMF cells, could be involved in this activation. In this study, it was very important to obtain cell lines that were used as a model because IMF is a rare disease and it is difficult to perform biochemical studies only on patient cells. Thus, we have overexpressed FKBP51 into the factor-dependent cell lines UT7/Mpl by means of a bicistronic retrovirus vector containing FKBP51 and eGFP expression (11). We derived two cell lines, one with a high eGFP expression the other with a lower eGFP expression (UT7 FKBP51; Fig. 1A). Using real-time RT-PCR, these two cell lines express 6- to 12-fold higher level of FKBP51 than the wild-type UT7/Mpl (Fig. 1B). This level of overexpression was in the same order of magnitude than presented in IMF megakaryocytes compared with normal megakaryocytes (11). Inhibition of calcineurin activity was also correlated with the level of FKBP51's expression (ref. 11; Fig. 1C). Therefore, in all this study, we compared the properties of the primary CD34+ cells from IMF or of IMF megakaryocytes with the cell line of the highest expression of FKBP51 (UT7 FKBP51), which has been previously described (11). We did this dual comparison because CD34+ cells include various subsets of cells including lymphoid progenitors that may not belong to the malignant clone. In contrast, spontaneously grown megakaryocytes are derived from hematopoietic progenitors that are clonal.
We chose to study NF-κB's activation by a gel shift assay (Fig. 2). Nuclear extracts of peripheral blood CD34+ and spontaneously grown megakaryocytes from IMF patients were used. The activation of NF-κB's pathway was seen in CD34+ cells as well as in spontaneously grown megakaryocytes from patients that were studied (patients 1 and 8, respectively). In contrast, no NF-κB activation could be detected in normal marrow CD34+ cells. However, NF-κB activation was detected in mobilized CD34+ cells after G-CSF treatment, suggesting that NF-κB activation is regulated by cytokine treatment. A similar result was obtained in normal megakaryocytes derived from CD34+ cells in the presence of thrombopoietin.
To investigate the disease specificity of this constitutive NF-κB activation, we analyzed CD34+ cells from other myeloproliferative disorders. Two samples obtained from patients with essential thrombocythemia presented no activation of NF-κB, whereas two samples from patients with chronic myeloid leukemia presented a spontaneous activation of NF-κB [Fig. 2 mentioned ET and chronic myelogenous leukemia (CML)].
Then, we tried to determine the nature of these activated forms of NF-κB. Two forms were usually present in protein extracts from mobilized CD34+ cells (Fig. 2 mentioned PB), patient samples, and the UT-7 cell line stimulated by GM-CSF (Fig. 3A). These two forms were displaced by a cold competitor demonstrating their specificity (Fig. 2, PB). The nature of these two bands was precised by using samples from the UT7 cell line stimulated by GM-CSF (Fig. 3A). Both forms were supershifted by an anti-p50 antibody indicating that they both contained the p50 member. Only the upper form was displaced by a cold probe specific for p65. These results suggest that the NF-κB activated form of lower molecular weight corresponds to p50 homodimer and to the upper molecular form to a p65/p50 heterodimer.
Activated forms were different among patient samples; that is, in CD34+ cells, the two forms were present in two patients, activation of only the p50/p50 homodimer or the p65/p50 heterodimer were detected, respectively, in four and one patient. No correlation between the activated NF-κB forms and clinical data could be established. NF-κB activation was also studied in spontaneously grown megakaryocytes from two patients and in both cases the two forms of activated NF-κB were present including one patient, for which only the p50 homodimer was detected in CD34+ cells.
To investigate if a FKBP51 overexpression could lead to a spontaneous NF-κB activation, we investigated the two UT7 cell lines. Gel shift assays were done on UT7 eGFP and UT7 overexpressing FKBP51 (UT7 FKBP51). Cells were washed thrice with PBS and cultured with or without GM-CSF during 24 hours. Ten micrograms of nuclear extracts were loaded. Constitutive activation of NF-κB, consisting of the p65/p50 heterodimer and p50 homodimer was detected in UT7 FKBP51 cells, even after deprivation of cytokine (Fig. 3A), whereas as in normal cells, NF-κB activation was essentially observed in the presence of cytokines in control UT7 eGFP cells and at a lower extent than in UT7 FKBP51 cells.
To show that activation of NF-κB by FKBP51 led to a transcriptionally active complex, we used a reporter gene assay. A NF-κB–inducible reporter plasmid NF-κB/Luc was transiently transfected into each cell lines and after 20 hours of incubation with or without GM-CSF. Cellular lysates were assayed for luciferase activity. As shown in Fig. 3B, induction of NF-κB activation in UT7 FKBP51 cells was nearly 4-fold higher than in control UT7 eGFP cells, in agreement with the gel shift assays. Moreover, the ratio of NF-κB activity between UT7 FKBP51 cells and UT7 eGFP cells did not change in the presence or absence of growth factor (Fig. 3B , right). These results suggest that the significant activation of NF-κB induced by FKBP51's overexpression might be due to an over sensitivity of this transduction pathway but not to a constitutive activation.
Thus, as reported for STAT5's activation, FKBP51 lets a persistent NF-κB activation after cytokine deprivation.
FKBP51 activates nuclear factor κB by inducing IκBα degradation. The transcriptional activity of NF-κB is primarily controlled by its intracellular localization (16). In most unstimulated cells, NF-κB remains in a cytoplasmic, under the form of an inactive complex associated with the inhibitor proteins, IκBs (17). In most mammalian cells, IκBα and IκBβ represent the two major forms of the IκB family (17). Upon cell stimulation with various stimuli, IκBα is phosphorylated and degradated by proteasome, allowing active NF-κB to translocate to the nucleus (16). Therefore, we evaluated the level of cytoplasmic IκBα in UT7 eGFP and UT7 FKBP51. Overexpression of FKBP51 in UT7 induced a strong reduction of cytoplasmic IκBα in comparison with the wild-type cell line. In addition, this low level of IκBα was not increased after cytokine deprivation (Fig. 3C). To confirm that the low level of IκB in UT7 FKBP51 cells was a consequence of its degradation by proteasome, the proteasome inhibitor (lactacystin) was added before cytokine deprivation. Treatment of UT7 FKBP51 cells with this proteasome inhibitor induced a marked IκB increase following the inhibition of NF-κB nuclear translocation (data not shown), indicating that FKBP51 induced NF-κB activation was proteasome dependent.
Nuclear factor κB inhibition using a nuclear factor κB superrepressor. To examine the consequences of this NF-κB transduction activation, we used the NF-κBsr. This mutant form of IκBα contains serine-to-alanine mutations at residues 32 and 36, which inhibit signal-induced phosphorylation and subsequent proteasome-mediated degradation of IκB (15, 16). A bicistronic retrovirus containing the entire coding region of the NF-κBsr form and eGFP was constructed as described in Materials and Methods. UT7 FKBP51 and UT7eGFP cells were infected at the same time. The NF-κBsr retrovirus encodes the same reporter gene (eGFP) as the previous retroviral vectors, which were used to obtain the UT7 FKBP51 and UT7 eGFP cell lines. Therefore, to select the cells, which were infected by the NF-κBsr retrovirus, we selected the ones that had the highest level of eGFP (Fig. 4A). The expression of the NF-κBsr and the decrease of the translocation to the nucleus of NF-κB were confirmed by Western blot using anti-IκB, anti–NF-κB p65 and p50 antibodies (Fig. 4B). In agreement with this result, the transcriptional activity of NF-κB, when analyzed by reporter gene assay, was inhibited dramatically by the NF-κBsr (Fig. 4C).
The same approach was done on CD34+ cells from patients and on normal controls which were infected with the NF-κBsr retrovirus after a 5-day culture. Infected cells were sorted by flow cytometry on the dual expression of eGFP and CD34 (Fig. 5A). These infected cells were tested for viability and TGF-1 secretion capacity.
Nuclear factor κB inhibition does not affect cell survival/proliferation in UT7 M51 cells and idiopathic myelofibrosis CD34 cells. The NF-κB pathway plays a major role in gene regulating apoptosis. We have previously reported that activation of the JAK2/STAT5 pathway is responsible for the apoptotic resistance phenotype of both IMF patient cells and UT7 FKBP51 cells (12). To examine whether the NF-κB transduction pathway is also involved in this phenotype, we tested the effects of NF-κBsr on apoptosis. After culture with or without GM-CSF for 48 hours, cell survival and cell proliferation rates were analyzed by Annexin V labeling and by MTT assay. As shown in Fig. 4D and E, NF-κBsr did not significantly change either Annexin V or MTT staining in UT7 FKBP51 cells after cytokine deprivation. These results suggest that NF-κB's activation is not essential for the cell survival/proliferation phenotype induced by FKBP51's overexpression in this cell line, like in the survival effect of GM-CSF.
The same results were obtained with human CD34+ cells and IMF patients infected with the NF-κBsr; that is, a number of viable cells were counted 5 days after cell sorting, and no differences were noticed between cells with or without NF-κBsr (Fig. 5B).
Transforming growth factor-β secretion in UT7 FKBP51 cells and idiopathic myelofibrosis patient cells. NF-κB activation has also been implicated in the secretion and synthesis of numerous cytokines. Rameshwar et al. previously reported that monocytes from IMF patients showed spontaneous NF-κB activation, which was responsible for the TGF-β1 secretion (5). Therefore, we hypothesized that NF-κB activation observed in IMF CD34+ cells and spontaneously grown megakaryocytes could also be responsible for TGF-β1 oversecretion. After NF-κBsr overexpression, cells were counted and plated at the same concentrations in serum-free medium with and without growth factor (GM-CSF for the UT7 cells and thrombopoietin for human normal megakaryocytes in the absence of cytokine for pathologic cells). Twenty-four (UT7 cells) or 48 hours later (human cells), culture supernatants were collected and human TGF-β1 concentration was determined. UT7 cells were also collected for TGF-β1 transcript quantifications.
Using an ELISA assay, it is possible to distinguish active and latent forms of the TGF-β1. Indeed, the antibody only recognizes the active form. Thus, dosage of TGF-β1 before acidification permits the detection of the constitutively active form and the total TGF-β1 present in the samples after acidification that activated latent TGF-β1. The concentration in TGF-β1 protein (active and nonactive forms) was higher in the supernatants of UT7 FKBP51 cells than they were in control UT7 eGFP, whether the cells were cultured with or without GM-CSF and markedly decreased after NF-κBsr's overexpression (Fig. 5D).
The expression of NF-κBsr in the culture supernatants of both normal and IMF megakaryocytes also induced a decrease in the TGF-β1 protein (Fig. 5C). Surprisingly, a lower amount of TGF-1 was detected in the supernatant of megakaryocytes from IMF than there was from normal controls. However, this may be related to the culture conditions because normal cells were grown in the presence of thrombopoietin when the pathologic cells were grown in its absence.
In contrast, we did not find any significant differences in the expression of TGF-β1 mRNA between UT7 eGFP and UT7 FKBP51 cells with or without the NF-κBsr (Fig. 5E). These results suggest that the NF-κB pathway could regulate production of TGF-1 in these pathologic cells at a nongenomic level.
Little is known about the signal transduction pathways involved in the development of IMF. However, an activation of the Janus-activated kinase (JAK)/STAT pathway is involved in all myeloproliferative disorders for which the molecular mechanism has been characterized [i.e., CML (18) and hypereosinophilic syndrome (37)]. Recently, using a transcriptional approach, we have shown that the immunophillin FKBP51 was overexpressed in IMF megakaryocytes (11). Remarkably, FKBP51 overexpression induced a resistance to apoptosis through STAT5 pathway activation (12). In this report, we are showing that another signal transduction pathway, NF-κB, was also activated both in IMF patients cells and in FKBP51 overexpressing cell line.
The transcription factor NF-κB is established as a regulator of immune and inflammatory responses. Presently, there are five known members of the NF-κB family: p50, p52, c-Rel, RelB, and p65 (RelA). Each of these different members are distinguished by its Rel homology domain, the portion of the protein that controls DNA binding, the dimerization, but also the interactions with inhibitory factors known as the IκB proteins (17, 19, 20). They form various homodimers and heterodimers bound to DNA. In response to appropriate stimuli, the IκB proteins become phosphorylated on their serine residues, resulting in rapid ubiquitination and subsequent proteolysis by the 26S proteasome which results in the liberation of NF-κB, allowing it to accumulate in the nucleus. Recently, NF-κB activation has been connected with multiple aspects of oncogenesis, including the control of apoptosis. For example, several reports have indicated that the ectopic expression of Bcr-Abl in IL-3–dependent murine myeloid cells activates p65 (21), which blocks the apoptosis induced by cytokine deprivation. In our study, we have compared NF-κB activation in IMF cells and UT7 cells overexpressing FKBP51. There is no direct demonstration of the role of this overexpression in the IMF physiopathology. In both cases, we detected a constitutive NF-κB activation which involves the p65/p50 heterodimer and the p50 homodimers. However, the inhibition of NF-κB activity by the NF-κB superrepressor did not alter the apoptotic resistance of both cell types but inhibited secretion of TGF-β1.
It has been proved that an abnormal release of growth factors such as TGF-β1 can be responsible for the development of marrow fibrosis (22). TGF-β1 has also been reported an important cytokine involved in the development of fibrosis in several organs such as lung, liver, and kidney (23). Plasma level of TGF-β1 is increased in IMF patients (24). Moreover, we have previously shown that the bone marrow fibrosis development was also TGF-β1 dependent in a model of mice overexpressing thrombopoietin (9). Two main hematopoietic cells are responsible for TGF-β1 synthesis in human: monocytes and megakaryocytes/platelets. In addition, it has been shown that normal progenitor cells could release TGF-β1 (25). Rameshwar et al. have shown that monocytes from IMF patients present spontaneous activation of the NF-κB pathway, which leads to a monocyte adherence to plastic and an increased of TGF-β1 expression (5). In this study, we could show that spontaneous NF-κB activation could also regulate TGF-β1 secretion in megakaryocytes from IMF. On the contrary, we did not find any spontaneous activation of NF-κB in CD34+ cells from essential thrombocythemia, a megakaryocytic disease which is not associated with myelofibrosis. However, in other myeloproliferative diseases such as CML, NF-κB can be activated. This would also be relevant to the transition of bone marrow fibrosis secondary to myeloproliferative disorders because in the majority of cases, fibrosis is secondary to CML. Therefore, this study suggests that either there is a requirement for a NF-κB activation in a specific cell lineage such as the megakaryocytic cell line to develop a myelofibrosis, or that NF-κB activation is not able to induce by itself a myelofibrosis but is able to participate with other unknown genetic events in its development. The mechanism by which NF-κB is activated in IMF remains also undetermined. However, the fact that FKBP51 overexpressing cell lines have also a NF-κB activation and increased TGF-β1 secretion suggests that this immunophilin may play an important role in this disease. FKBP51 could induce NF-κB activation by at least two ways: the activation of the JAK2/STAT5 pathway and the inhibition of calcineurin. Indeed, the overexpression of this immunophilin induced a sustained activation of the JAK2/STAT5 pathway. Recently, Digicaylioglu and Lipton have reported the importance of the erythropoietin-activated JAK2, to NF-κB signaling in neuroprotection (26). Moreover, the TEL-JAK2 fusion protein is also able to activate the NF-κB pathway (27). These findings suggest that there is a cross-talk between the JAK2/STAT5 and the NF-κB pathway, JAK2 being able to directly or indirectly activate NF-κB. FKBP51 overexpression in UT7 cells also leads to an inhibition of the calcineurin phosphatase activity. Calcineurin is essentially a serine/threonine phosphatase, and phosphorylated IκBs can be its substrate (28, 29). Therefore, FKBP51 by inhibiting calcineurin may induce the degradation of IκB and NF-κB activation in UT7 FKBP51 cells. Indeed, a constitutive NF-κB activation is very common in hematologic malignancies. Bouwmeester et al. have recently shown that FKBP51 can be associated with IKKα. In addition, the knock down of FKBP51, induced by RNA interference, inhibits the activation of NF-κB by tumor necrosis factor-α, demonstrating that FKBP51 is a key regulator of the NF-κB pathway. However, this relationship between FKBP51 overexpression and the presence of a constitutive NF-κB activation by IMF CD34+ cells remain speculative.
Surprisingly, despite the increase of TGF-β1 protein secretion in cells overexpressing FKBP51 and its inhibition by NF-κBsr, we failed to show its regulation at the mRNA level. In fact NF-κB cannot directly induce an increased transcription of TGF-β1 because its promoter does not have NF-κB binding sites. Thus, the effect of NF-κB on TGF-β1 must be indirect, involving another mediator which can be cell specific. Several studies suggest that TGF-β1 production may be controlled both at the transcriptional and post-transcriptional level by different independent factors (30, 31). Therefore, in many cell types, there is a discrepancy between the mRNA and protein levels. Generally, it is known that elements within both the 5′ untranslated region (UTR) and the 3′UTR of mRNA are involved in the control of translation (32). For TGF-β1, three different transcripts have been reported, which differ by their translational efficiency (33). This may explain the discrepancy between mRNA and protein levels in some cell types, which may depend on the type of transcripts. However, TGF-β1 production may also be regulated at the level of secretion (34), but the precise mechanism of this regulation is unknown. Therefore, it can be hypothesized that NF-κB regulates TGF-β1 production, directly or indirectly via other mediators at the post-transcriptional level, by interfering with either translation or secretion mechanisms.
We also found in this study that the inhibition of NF-κB pathway was not toxic in myeloid cells from IMF patients. Rameshwar et al. found similar results in IMF monocytes (5). This suggests that NF-κB or TGF-β1 inhibitors may be useful in IMF therapy to inhibit the development of fibrosis. PS341 (Velcade; Millenium Pharmaceuticals) has been tested in myeloma patients with some success (35). PS341 is a proteasome inhibitor that can inhibit NF-κB's pathway activation through the inhibition of IκB degradation. Such a therapy could then be tested in the murine model of myelofibrosis, and in case of a success, in IMF patients.
In conclusion, we have shown that the activation of NF-κB is the main mechanism involved in TGF-β1 secretion of IMF cells in vitro. A limitation of our study could be related to the culture itself. Thus, these results have to be directly confirmed on in vivo cells especially on megakaryocytes. A parallel study done on a cell line overexpressing FKBP51 emphasizes the possible role of this immunophilin in the pathogenesis of IMF. Therefore, further studies are required to precisely understand the role of FKBP51's overexpression in the pathogeny of the disease using murine models and IMF CD34+ cells knocked down by FKBP51 small interfering RNA.
Grant support: Institut National de la Sante et de la Recherche Medicale and la Ligue Nationale contre le Cancer “équipe labellisée 2004” and fellowships from Association de Recherche contre le Cancer (E. Komura) and the Research Ministry (H. Chagraoui).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
We thank C. Lefebvre for helpful discussions and improving the English article.