| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
J. Biol. Chem., Vol. 279, Issue 49, 50874-50885, December 3, 2004
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
From the Feinberg School of Medicine and the Robert H. Lurie Comprehensive Cancer Center, Northwestern University and Chicago Lakeside Veterans Affairs Hospital, Chicago, Illinois 60611
Received for publication, May 24, 2004 , and in revised form, September 2, 2004.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In addition to transcriptional activation, ICSBP also has transcriptional repression activity. Non-tyrosine-phosphorylated ICSBP impacts transcription via direct, IRF1-independent interaction with PRDI consensus sequences (7). Only non-tyrosine-phosphorylated ICSBP binds directly to DNA (7), suggesting that post-translational modification also regulates ICSBP repression activity. The only known genuine ICSBP repression target is the gene encoding the antiapoptotic protein, BclXL (8). BclXL is expressed in mature myeloid cells in response to inflammatory mediators, such as lipopolysaccharide and interferon-
(9). In this context, BclXL expression increases survival of activated phagocytes (9).
Based upon these identified target genes, one would anticipate ICSBP deficiency to be characterized by terminal differentiation block and immune dysfunction. Indeed, this is the phenotype of the IRF1-deficient mouse (10). Instead, 100% of mice with the ICSBP gene disruption develop a myeloproliferative disorder, resembling human chronic myeloid leukemia (CML) (11). Similar to CML, the majority of ICSBP–/– mice develop blast crisis, fatal in 35% of the mice at 4 months (11). Consistent with these observations, ICSBP expression is decreased in human CML blast crisis (12), and forced overexpression of ICSBP delays blast crisis in a Bcr/Abl-expressing murine model (13). Such findings suggest that ICSBP is a leukemia anti-oncogene but do not indicate the mechanisms for this function.
Bone marrow myeloid progenitor cells from ICSBP–/– mice exhibit greater than normal colony formation, in response to low doses of GM-CSF or granulocyte colony-stimulating factor (referred to as "hypersensitivity") (14). Similar hypersensitivity to cytokines (including GM-CSF, IL-3, and stem cell factor) is also seen in bone marrow myeloid progenitor cells from mice with NF1 gene disruption (15). NF1 encodes the Ras-GAP neurofibromin 1 (Nf1), which inactivates Ras in hematopoietic cells (16). Mutations that inactivate Nf1 or activate Ras have been described in the myeloproliferative disorder, juvenile chronic myelomonocytic leukemia (17). This disorder is also characterized by GM-CSF and stem cell factor hypersensitivity (18). In Nf1-deficient hematopoietic cells, increased Ras activity increases ERK and Akt activation, increasing proliferation (19) and decreasing apoptosis (20).
Therefore, ICSBP and Nf1 deficiency states are both characterized by a myeloproliferative disorder, which includes an increased response to hematopoietic cytokines. These results suggest the possibility of a functional association between ICSBP and Nf1. Consistent with this, our current investigations identify the NF1 gene as an ICSBP activation target. Therefore, our results suggest that one mechanism of myeloproliferation in ICSBP-deficiency is impaired regulation of cytokine-induced Ras activity.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Oligonucleotides—Oligonucleotides were custom synthesized by MWG Biotech (Piedmont, NC). Oligonucleotides that were used for analysis of co-immunoprecipitated chromatin were used to amplify the human promoter regions of the following genes: CYBB (F, 5'-tcagttgaccaatgattattagccatt-3'; R, 5'-ctatgcttcttcttccaatgaccaaat-3') (2); NCF2 (F, 5'-gcagaagcattttggggaactgatcct-3'; R, 5'-aaaatccacaggaaatgtcccaccttt-3') (3); TOLL-like receptor 4 (F, 5'-gatgactaattgggataaaagccaact-3'; R, 5'-tggagaggaagtgaaagcggcaacctta-3') (4); and NF1 (F, 5'-cagaggaaaagctgggcttaaat-3'; R, 5'-agatctgtcctccccgcggccgggg-3') (21). Oligonucleotide primers were also designed to amplify a conserved region of the proximal murine NF1 promoter, including the putative ICSBP-binding site, homologous to the PRDI-IRF-binding consensus (mNF1; F, 5'-ggatcccacttccggtggggtgtcatggcggcc-3'; R, 5'-gtcctcccggcgacccgggg-3') (22). All oligonucleotides were designed with BamHI half-sites on the ends to facilitate subcloning into plasmid vectors for sequence verification of the PCR products.
Oligonucleotides used to generate truncation mutants of the human NF1 promoter were designed as follows: NF1-337 F (5'-ggatcccacttccggtggggtgtcatggc-3'); NF1-315 F (5'-catggcggcgtctcggactgtgatggctgt-3'); NF1-3' R(5'-gtcctccccgcggccgggg-3') (22). These primers were designed with BglII half-sites to facilitate subcloning into plasmid vectors.
Oligonucleotides used in quantitative real time PCR were designed to amplify the following messages: gp91PHOX (F, 5'-gagagccagatgcaggaaag-3'; R, 5'-ggtgcacagcaaagtgattg-3'); actin (F, 5'-gatgagattggcatggcttt-3'; R, 5'-caccttcaccgttccagttt-3'); Nf1 (F, 5'-agccaccacctagaatcgaa-3'; R, 5'-ggccgcatatgttcttcttt-3'); and 18 S rRNA (F, 5'-accgcggttctattttgttg-3'; R, 5'-cggtccaagaatttcacctc-3').
Double-stranded synthetic oligonucleotides were used in EMSA. Oligonucleotides were designed to the NF1 promoter sequence with homology to the IRF-binding, PRDI consensus (NF1-337 5'-ggatcccacttccggtggggtgtcatggc-3') for use as a DNA-labeled probe in EMSA. Double-stranded oligonucleotide competitors were the PRDI consensus (5'-tcactttcactttcactt-3') (7) and the CCAAT box from the 
-globin gene as an irrelevant competitor (5'-ggcggcgcttcattggctggcgcggagcccg-3') (6). The PRDI-like sequence and the PRDI-consensus in these oligonucleotides are underlined. These oligonucleotides were all designed with BamHI half-sites on the ends to facilitate subcloning into plasmid vectors.
Cell Lines and Culture Conditions—The human myelomonocytic cell line U937 (23) was obtained from Andrew Kraft (University of Colorado, Denver). Cells were maintained and differentiated as described (2). For differentiation experiments, U937 cells were treated for 48 h with 400 units/ml human recombinant IFN (Roche Applied Science) (2, 3).
Chromatin Immunoprecipitation and Cloning—U937 cells were cultured with or without IFN for 48 h, as described (2, 3). Cells for chromatin immunoprecipitation were incubated with formaldehyde prior to lysis, and lysates were sonicated to generate chromatin fragments with an average size of 2.0 kb, as described (24). Lysates underwent two rounds of immunoprecipitation with either antiserum to ICSBP (a kind gift of Dr. Stephanie Vogel) or preimmune serum, as described (24). Some immunoprecipitated chromatin was treated with Klenow fragment to create blunt ends and ligated into a plasmid vector. Plasmids were transformed into Escherichia coli, and transformants with inserts were identified and sequenced according to standard techniques. The sequences of chromatin inserts were analyzed by searching GenBankTM and the National Center for Biological Information (NCBI) Human Genome Data base.
The remainder of the U937 immunoprecipitated chromatin was analyzed by PCR for ICSBP antibody-specific co-precipitation of the CYBB, NCF2, TLR4, and NF1 genes. For these experiments, input chromatin was a positive control, and chromatin precipitated by preimmune serum was a negative control. PCR products were analyzed by acrylamide gel electrophoresis. The identity of the PCR product was verified by subcloning into a plasmid vector, followed by dideoxysequencing. In similar experiments, chromatin was co-immunoprecipitated with ICSBP antibody or irrelevant (glutathione S-transferase (GST)) control antibody from murine bone marrow mononuclear cells, cultured for 24 h in GM-CSF, followed by 72-h differentiation in M-CSF. For these experiments, wild type murine bone marrow was compared with ICSBP–/– bone marrow, which was a negative control in these experiments. Input (nonprecipitated chromatin) was a positive control, as above. PCRs were performed with primers to amplify the proximal NF1 promoter (see above).
ICSBP Knock-out Mice—Mice with homologous disruption of the ICSBP gene were a generous gift of Dr. Keiko Ozato (National Institutes of Health, Bethesda, MD). Generation of these mice has been previously described (11). Homozygous ICSBP–/– mice and wild type litter mates were sacrificed at 8 weeks of age, when ICSBP–/– mice are in proliferative phase (11), and bone marrow was harvested from the femurs according to standard methods.
Culture of ICSBP–/– Myeloid Cells—Bone marrow mononuclear cells were obtained from the femurs of ICSBP–/– and wild type litter mates. ScaI+ cells were separated using the Miltenyi magnetic bead system, according to the manufacturer's instructions (Miltenyi Biotechnology, Auburn, CA). These cells were cultured (at a concentration of 2 x 105 cells/ml) for 48 h in DMEM supplemented with 10% fetal calf serum, 1% penicillin/streptomycin, 10 ng/ml murine GM-CSF (R&D Systems Inc., Minneapolis, MN), and 5 ng/ml murine recombinant IL-3 (R&D Systems). Then cells were either maintained in GM-CSF + IL-3 (myeloid progenitor cells) or switched to DMEM supplemented with 10% fetal calf serum, 1% penicillin/streptomycin, 10 ng/ml murine recombinant M-CSF (R&D Systems) for 72 h (monocyte differentiation), with or without 200 units/ml recombinant murine IFN (Roche Applied Science) for the last 24 h. Cells were harvested, and cell lysates were used in Western blot experiments, as described below.
Some ICSBP–/– bone marrow mononuclear cells were cultured long term in DMEM supplemented with 10% fetal calf serum, 1% penicillin/streptomycin, 10 ng/ml murine GM-CSF (R&D Systems), and 5 ng/ml murine recombinant IL-3 (R&D Systems), as described (25). After 3 months of passage, these cells were evaluated by flow cytometry and found to be negative for the B-cell marker CD19 and the T-cell marker CD3, dimly positive for myeloid marker CD14, and dimly positive to negative for Mac1 (CD11b) and Gr1.
Retroviral Transduction of ICSBP–/– Murine Bone Marrow Myeloid Cells—High titer murine stem cell retroviral supernatants were produced using the pMSCV vector and PT67 cell line, as per the manufacturer's instructions (Stratagene). Filtered retroviral supernatants were used immediately or stored at –80 °C. Transductions of long term cultured murine bone marrow myeloid cells from ICSBP–/– mice were performed as previously described (26). Briefly, cells were harvested, and 4.0 x 106 cells were plated in 3 ml of DMEM, supplemented with 10% fetal calf serum, 10 ng/ml GM-CSF, and 5 ng/ml IL-3. An equal volume of retroviral supernatant was added to each dish, and Polybrene was added to a final concentration of 6 µg/ml. Cells were incubated for 8 h at 37 °C, 5% CO2, and then diluted 3-fold with media, supplemented as above. Cells were incubated overnight, and the procedure was repeated the next day. The day after transduction, puromycin was added to 1.2 ng/ml. Cells were selected in antibiotics for 96 h and then treated with cytokines as indicated. Each experiment was repeated at least three times. Expression of transduced proteins was independently verified by Western blots for each experiment.
Western Blots, Immunoprecipitation, and Ras Activity Assays—For Western blots of wild type and ICSBP–/– murine bone marrow myeloid cells, cell pellets were lysed by boiling in 2x SDS sample buffer (without Coomassie Blue). Protein assays were performed by standard methods, and equivalent amounts of protein (50 µg) were separated by SDS-PAGE. Similar experiments were performed with long term cultured ICSBP–/– cells, transduced with murine stem cell retroviral vectors for protein expression, as indicated. In these experiments, the total amount of protein was either 30 µg (Fig. 2C) or 50 µg (Fig. 2D). Proteins were transferred to nitrocellulose, and Western blots were probed with antibodies to neurofibromin 1, phospho-ERK, and ERK1/2 (obtained from Santa Cruz Biotechnology, Inc., Santa Cruz, CA). Proteins were detected by chemiluminescence using the ECL reagents from Amersham Biosciences according to the manufacturer's instructions.
|
For Ras activity assays, murine bone marrow myeloid cells or long term cultured ICSBP–/– myeloid cells were lysed in MLB buffer (25 mM Hepes, pH 7.5, 150 mM NaCl, 1% Igepal, 10% glycerol, 10 mM MgCl2, 1 mM EDTA, 10 µg/ml aprotinin, 10 µg/ml leupeptin). Lysates (150 µg) were precleared with recombinant GST conjugated to glutathione-agarose. Lysates were next incubated with glutathione-agarose conjugated to a GST fusion protein of the Ras binding domain (RBD) from Raf-1 (Upstate), or control GST-glutathione-agarose, according to the manufacturer's instructions. The beads were extensively washed in MLB buffer, and precipitated proteins were separated by SDS-PAGE and detected by probing Western blots with an anti-Ras antibody (pan-Ras; Santa Cruz Biotechnology). Total cell lysates (20 µg) were also separated by SDS-PAGE, and Western blots were probed with pan-Ras antibody to determine the abundance of total Ras protein.
Cell Proliferation Assays—Long term cultured, transduced ICSBP–/– cells were selected for 96 h in puromycin prior to proliferation assays. Cells were harvested and plated at 105 cells/200 µl in a 96-well dish in DMEM supplemented with antibiotics and serial dilutions of GM-CSF or M-CSF (10-fold from 10 to 0.01 ng/ml or no cytokine). Cells were incubated for either 24 h (GM-CSF or M-CSF) or 72 h (M-CSF) at 37 °C, 5% CO2, with [3H]thymidine for the last 8 h. Cells were harvested, and TCA-precipitated DNA was counted by scintillation counting, according to standard techniques.
RNA Isolation and Quantitative Real Time PCR—Total cellular RNA was isolated, as previously described (27). Reverse transcription reactions were performed using the ImPromII reverse transcriptase kit, according to manufacturer's instructions (Promega). Quantitative real time PCR was performed using the Platinum SYBR Green qPCR SuperMix UDG kit, according to the manufacturer's instructions (Invitrogen). Real time PCR was performed using the ABI 7900 Thermocycler (Applied Biosystems, Foster City, CA), and results were analyzed using SDS version 2.1 software (Austin Biodiversity Web site gallery).
Transfection and Reporter Gene Assays—U937 cells were cultured and transfected as previously described (2). Cells (32 x 106 cells/sample) were transfected with a vector to express ICSBP (ICSBP/pcDNAamp) or empty vector control; the pCATE reporter vector with the proximal 973 bp of the human NF1 promoter (973NF1), 337 bp (337NF1), 315 bp (315NF1), 230 bp (230NF1), or empty pCATE vector control and p-CMV
gal (as a control for transfection efficiency). Transfectants were harvested 48 h after transfection, with and without incubation with recombinant human IFN (400 units/ml). Lysates were analyzed for CAT and -galactosidase activity, as described (2).
Nuclear Protein Isolation and EMSA—Nuclear extract proteins were prepared, from U937 cells, by the method of Dignam, with protease inhibitors (as described) (6). In some experiments, U937 cells were differentiated with 400 units/ml of IFN before nuclear protein isolation. Oligonucleotides probes were prepared, and EMSA and antibody supershift assays were performed, as described (6). Antibody to ICSBP and control GST antibody, used in these experiments, are described above.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
differentiation. IFN treatment decreases U937 proliferation and results in acquisition of characteristics of mature phagocytes (23). This is associated with transcription of genes encoding the oxidase proteins gp91PHOX (the CYBB gene) (2) and p67PHOX (the NCF2 gene) (3) and the TOLL-like receptor 4 (the TLR4 gene) (4).
In initial experiments, we investigated in vivo ICSBP interaction with these previously identified target genes. Chromatin that co-precipitated with ICSBP antibody was amplified by PCR, using primers flanking ICSBP-binding sites in the CYBB, NCF2, and TLR4 genes. In these experiments, chromatin coprecipitating with preimmune serum was a negative control, and input (nonprecipitated) chromatin was a positive control. We found ICSBP-specific co-precipitation of the CYBB, NCF2, and TLR4 cis elements from IFN-differentiated U937 lysates (Fig. 1A) but not from undifferentiated U937 lysates. This is consistent with expression patterns of these genes during myeloid differentiation. Based on the success of these control experiments, we pursued identification of novel ICSBP target genes. We approached this by subcloning anti-ICSBP co-precipitated chromatin into a plasmid vector. Chromatin insert sequences were analyzed by a GenBankTM data base search (NCBI Web site, National Institutes of Health). Clones with GenBankTM sequence matches were analyzed using the Human Genome Resources Data base (NCBI Web site), to localize the sequences within known genes.
|
50 clones analyzed, one-third represent centromeric repeated sequences. These clones are less than 100 bp in length, consistent with the results of other investigators.2 An additional one-third fall outside identified genes, and one-fifth have no GenBankTM match. Of the six remaining clones, three co-precipitate from undifferentiated U937 cell lysates. These clones will be the subject of future investigations. The remaining three chromatin fragments co-immunoprecipitate from IFN-differentiated cell lysates. One of these clones includes the 5'-flank of the gene encoding neurofibromin 1 (the NF1 gene). Based on similarities in hematopoiesis in ICSBP-deficient and Nf1-deficient cells, the NF1 gene was pursued as a potentially mechanistically significant ICSBP target gene. The specificity of NF1 promoter coprecipitation with ICSBP was verified by PCR, as described above (Fig. 1A). Using this technique, we also determined that the ICSBP-NF1 gene interaction is specific to differentiated U937 cells. Since proliferation decreases during IFN-differentiation of U937 cells, this finding is consistent with a previously described function of Nf1 in hematopoietic cells, down-regulation of Ras-induced proliferation.
We also tested interaction of ICSBP with the NF1 promoter, in ex vivo primary murine bone marrow myeloid cells, a non-transformed model of myelopoiesis. Since the NF1 promoter co-precipitates with ICSBP from IFN-treated U937 lysates, murine bone marrow cells were also monocyte-differentiated for these experiments. Bone marrow mononuclear cells were isolated from the femurs of wild type mice and cultured for 24 h in GM-CSF and IL-3, followed by 72 h of culture in M-CSF, as described (29–31). M-CSF-differentiated bone marrow cells from ICSBP–/– mice were also studied, as a negative control. These mice have disruption of the ICSBP gene by homologous recombination and do not express ICSBP protein, as has been previously described (10).
Chromatin was co-precipitated from murine monocyte lysates with an anti-ICSBP or irrelevant control antibody, as described above. Co-precipitated chromatin was analyzed by PCR with primers designed to amplify a highly conserved region of the murine proximal promoter (discussed further below). For these experiments, input (nonprecipitated) chromatin was a positive control, and irrelevant antibody co-precipitation was a negative control. Consistent with U937 results, we found specific co-precipitation of the NF1 promoter from wild type bone marrow-derived murine monocytes (Fig. 1B). The NF1 promoter did not co-precipitate with ICSBP antibody from bone marrow-derived ICSBP–/– monocytes, consistent with the absence of ICSBP protein in these cells.
ICSBP Deficiency Decreases Nf1 Expression in Myeloid Cells—We next investigated the impact of ICSBP deficiency on Nf1 expression in differentiating myeloid cells. For these studies, we also used bone marrow from ICSBP-deficient and wild type mice, as above. ScaI+ cells were isolated from the bone marrow of ICSBP–/– mice or wild type litter mates and cultured for 48 h in GM-CSF + IL-3. These cells are considered "myeloid progenitors," as previously defined (29). After 48 h, cells were maintained in GM-CSF + IL-3 or differentiated into monocytes with M-CSF for 72 h (as described (30, 31)). Since IFN induces ICSBP transcription in mature myeloid cells (32), some M-CSF-differentiated cells were treated with IFN for the last 24 h. These cytokines were chosen because monocyte differentiation is blocked in ICSBP–/– cells, and many known ICSBP target genes are transcribed in mature monocytes. Additionally, we identified NF1 as a potential ICSBP target gene in two monocyte differentiation models.
Cells were harvested and lysate proteins analyzed for Nf1 expression by Western blot. We found that, in wild type bone marrow, Nf1 protein is expressed in myeloid progenitor cells (in GM-CSF and IL-3); expression significantly increases during M-CSF differentiation and is further increased by IFN (Fig. 2A). In contrast, we detected minimal Nf1 protein in lysates from ICSBP–/– myeloid progenitor cells, and Nf1 protein expression does not increase in response to M-CSF or IFN (Fig. 2A). For these Western blots, we used an Nf1 antibody that recognizes both dominant isoforms of the protein present in mammalian cells. Therefore, Nf1 protein appearing as a doublet band in most experiments. We confirmed monocyte differentiation of wild type cells by flow cytometry (myeloid progenitors, CD14dim/–Mac1–; monocytes, CD14+Mac1+; data not shown). Consistent with previous reports, we found that these cytokines do not completely differentiate ICSBP–/– cells into monocytes (myeloid progenitors, CD14dim/–Mac1–; monocytes, CD14dimMac1dim/–; data not shown) (14).
To identify a potential mechanism for ICSBP-dependant, differentiation-induced Nf1 expression, we investigated changes in ICSBP protein abundance in differentiating wild type bone marrow cells. ICSBP was immunoprecipitated (under denaturing conditions) from lysates of murine bone marrow cells in GM-CSF + IL-3, M-CSF, or M-CSF + IFN, as above. Immunoprecipitates were separated by SDS-PAGE, and Western blots were serially probed with anti-phosphotyrosine and anti-ICSBP antibodies (Fig. 2B). We found that M-CSF differentiation increases total ICSBP protein, which is further increased by IFN. Tyrosine-phosphorylated ICSBP also significantly increases during M-CSF differentiation but is not further increased by IFN.
We next determined whether ICSBP reconstitution restores Nf1 expression in ICSBP–/– myeloid cells. Previously, other investigators described immortalization of bone marrow mono-nuclear cells from ICSBP–/– mice by long term culture in GM-CSF and IL-3 (25). In those reported studies, the long term cultured cells were myeloid by immunophenotyping (25) and underwent some morphological differentiation upon transduction with an ICSBP expression vector (25). We derived a similar line from ICSBP–/– bone marrow mononuclear cells, cultured in IL-3 and GM-CSF. Consistent with previous results, these cells are an immature myeloid population by immunophenotyping (CD3–, CD19–, CD14dim, Mac1–, Gr-1–; data not shown).
We transduced long term cultured ICSBP–/– myeloid cells with a murine stem cell retroviral vector to express ICSBP or empty vector control. Transduced cells were cultured for 96 h in media supplemented with IL-3, GM-CSF, and puromycin, and lysates were analyzed by Western blot for ICSBP and Nf1 protein (Fig. 2C). Each experiment was repeated three times. Consistent with expectations, ICSBP–/– cultured myeloid cells do not express ICSBP in the absence of transduction with an ICSBP expression vector. In these cells, the abundance of overexpressed ICSBP is approximately equivalent to the abundance of endogenous ICSBP in wild type cells, treated with M-CSF + IFN (Fig. 2, B and C; note differences in the quantity of input protein). We also found that ICSBP reconstitution induces Nf1 expression in long term cultured ICSBP–/– cells, consistent with our data above.
We next investigated whether M-CSF differentiation further increases Nf1 protein expression, in ICSBP-transduced ICSBP–/– myeloid cells. ICSBP–/– cells were transduced with either ICSBP vector or empty vector, selected in puromycin, and cultured for 72 h in GM-CSF + IL-3, M-CSF, or M-CSF + IFN (for the last 24 h). Cell lysates were analyzed by Western blot for Nf1 protein. We found that ICSBP–/– cells, transduced with ICSBP vector, express significantly more Nf1 protein than vector control-transduced cells (Fig. 2D). Additionally, M-CSF increases Nf1 protein expression in ICSBP-transduced ICSBP–/– cells but not vector control-transduced cells (Fig. 2D). Nf1 protein is further increased by IFN in ICSBP-transduced, but not control vector-transduced, ICSBP–/– cells. In these experiments, neither M-CSF nor M-CSF + IFN influenced ICSBP expression by the murine stem cell retroviral vector (not shown).
ICSBP Deficiency Increases Ras Activity in Myeloid Cells— Nf1 deficiency is associated with increased Ras activity in human and murine myeloid progenitor cells (15, 17). To investigate the physiological relevance of Nf1 deficiency in ICSBP-deficient cells, we investigated Ras activity in response to the same cytokines as used in the experiments, above. First, we compared the effect of GM-CSF on freshly isolated bone marrow myeloid progenitor cells from wild type versus ICSBP–/– mice, as above. Because M-CSF (+IFN) increases Nf1 protein in wild type but not ICSBP–/– cells, the effect of these cytokines was also determined. Activated Ras was identified by affinity to a recombinant fusion protein, containing the Ras binding domain from Raf-1 (Raf-RBD/GST). This domain interacts with GTP-bound (activated) Ras but not inactive Ras. GST alone was the control for nonspecific precipitation. Active and total Ras abundance was compared by Western blot of lysate proteins.
We found slightly less activated Ras in wild type myeloid progenitor cells, treated with GM-CSF, than in ICSBP–/– cells under the same conditions. More strikingly, M-CSF and M-CSF + IFN decrease activated Ras in wild type myeloid progenitor cells but increase activated Ras in ICSBP–/– cells (Fig. 3A). These results are consistent with Nf1 protein induction by cytokine-induced differentiation in cells from wild type but not ICSBP-deficient mice. In these experiments, Ras does not interact with GST control (not shown). Total Ras abundance is not altered by cytokine treatment in either wild type or ICSBP–/– cells (Fig. 3A).
|
We found abundant activated Ras in long term cultured ICSBP–/– cells, transduced with empty vector control, in GM-CSF + IL-3 (Fig. 3B). Additionally, the abundance of activated Ras is increased by M-CSF and further increased by IFN, similar to our results with freshly isolated ICSBP–/– cells. In contrast, long term cultured ICSBP–/– myeloid cells, transduced with either ICSBP or Nf1-GRD expression vectors, exhibit less activated (but not total) Ras than control vector-transduced cells, in GM-CSF + IL-3 (Fig. 3B). Additionally, M-CSF + IFN decreases the abundance of activated (but not total) Ras in both ICSBP-transduced and Nf1-GRD-transduced ICSBP–/– cells (Fig. 3B). As above, there was no interaction between Ras and GST control (not shown). In all of these experiments, equal protein loading was verified by reprobing the blot with tubulin (not shown).
ICSBP Deficiency Increases Activation of Downstream Ras Targets in Myeloid Cells—To determine the impact of abnormal Ras activation in ICSBP-deficient cells, we investigated activation of the ERK-mitogen-activated protein kinase. ERK is a downstream Ras target, involved in mediating the proliferative response to both GM-CSF and M-CSF in myeloid cells (32, 33). To investigate whether abnormal Ras activation increases ERK phosphorylation in ICSBP-deficient cells, wild type and ICSBP–/– bone marrow cells were isolated and cultured in GM-CSF + IL-3, M-CSF, or M-CSF + IFN, as described above. Cell lysates were analyzed by Western blots, serially probed for phosphorylated (activated) ERK (p-ERK) and total ERK (ERK1/2).
We found no reproducible difference in p-ERK in wild type versus ICSBP–/– progenitor cells (in 10 ng/ml GM-CSF). However, M-CSF + IFN profoundly decreases p-ERK abundance at 72 h in wild type bone marrow progenitor cells (Fig. 2A). In contrast, 72 h of M-CSF treatment significantly increases p-ERK abundance in ICSBP–/– myeloid progenitor cells with IFN (Fig. 2A). Abundance of total ERK protein is not altered by differentiation in either ICSBP-deficient or wild type murine bone marrow cells (Fig. 2A).
To confirm these results, we determined the effect of ICSBP reconstitution or Nf1-GRD expression on ERK phosphorylation in long term cultured ICSBP–/– myeloid cells. These cells were transduced with empty vector or vectors to express ICSBP or Nf1-GRD, as in the experiments above. Transduced cells were selected in puromycin and cultured for 72 h in GM-CSF + IL-3, M-CSF, or M-CSF with IFN (for the last 24 h), as above. Cell lysates were analyzed by Western blot for total and p-ERK. We found that expression of either ICSBP or the Nf1-GRD decreases phosphorylated (but not total) ERK in GM-CSF + IL-3 in comparison with vector control-transduced ICSBP–/– cells. M-CSF, with and without IFN, further decreases p-ERK in ICSBP- or Nf1-GRD-expressing ICSBP–/– cells (Fig. 2D). In contrast, M-CSF, with or without IFN, increases the abundance of p-ERK, in vector control-transduced ICSBP–/– cells. Total ERK abundance is not significantly different in any transduced cells under these conditions. In all of these experiments, equalization of protein loading was verified by probing the blots for tubulin (not shown).
ICSBP Deficiency Results in "Hypersensitivity" to GM-CSF-induced Proliferation—Based on these results, we investigated whether ICSBP-dependant alterations in Nf1 protein, Ras activation, and ERK phosphorylation correlate with cytokine-induced proliferation. Previous investigations found bone marrow mononuclear cells from ICSBP–/– mice "hyperresponsive" to GM-CSF (14). We reconfirmed these results in long term cultured ICSBP–/– myeloid cells, transduced with empty vector control or vectors to express ICSBP, the Nf1-GRD, or a mutant form of Nf1-GRD without GAP activity (19). Cells were selected in puromycin and deprived of cytokines, and a dose-response curve for GM-CSF-induced proliferation was determined.
We found that GM-CSF (0.01–10.0 ng/ml) induces significantly more proliferation in control vector-transduced ICSBP–/– cells than in cells transduced with a vector to express ICSBP or the Nf1-GRD (p < 0.05, n = 4) (Fig. 4A). Consistent with expectations, the proliferative response to GM-CSF is not significantly different in ICSBP–/– cells transduced with empty vector or the GAP mutant Nf1-GRD (Fig. 4A).
|
At 24 h of stimulation, M-CSF induces significantly more proliferation in ICSBP–/– cells transduced with the empty control vector or GAP mutant Nf1-GRD vector than in cells transduced with ICSBP or Nf1-GRD expression vectors (p < 0.05, n = 3) (Fig. 4B). Differences in M-CSF-induced proliferation, between control vector and ICSBP expression vector-transduced ICSBP–/– cells, are statistically significant (p < 0.05, n = 3) at the lower M-CSF doses (0.01–1.0 ng/ml) and approach significance (p = 0.06, n = 3) at the highest dose (10 ng/ml). Differences in M-CSF-induced proliferation, between control vector and Nf1-GRD expression vector-transduced cells, are statistically significant at all doses tested (p < 0.05, n = 3). There was no significant difference in M-CSF-induced proliferation between ICSBP–/– cells transduced with control vector or GAP mutant Nf1-GRD expression vector.
We found sustained M-CSF-induced proliferation at 72 h in empty vector or GAP mutant Nf1-GRD-transduced ICSBP–/– cells, although proliferation is less than at 24 h (Fig. 4C). In contrast, there is insignificant M-CSF-induced proliferation at 72hin ICSBP–/– myeloid cells transduced with either ICSBP or wild type Nf1-GRD expression vectors. Differences in proliferation between control vector-transduced ICSBP–/– cells and cells transduced with ICSBP or Nf1-GRD expression vectors are statistically significant at all M-CSF doses (0.01–10 ng/ml or M-CSF; p < 0.05, n = 3). These results are consistent with Ras activity and ERK phosphorylation studies above.
ICSBP Activates the NF1 Promoter in Differentiating U937 Cells—Our studies suggest ICSBP is involved in Nf1 expression and interacts with the NF1 promoter, in vivo. To determine whether ICSBP activates the NF1 promoter, we investigated the effect of ICSBP overexpression on NF1 promoter activity in U937 transfection experiments. In initial experiments, we determined the effect of IFN on endogenous Nf1 mRNA abundance in U937 cells. These studies were essential, because regulation of Nf1 expression during myeloid differentiation has not been characterized. Using quantitative real time PCR, we found IFN differentiation of U937 cells significantly increases Nf1 mRNA at 48 h (Fig. 5A). This correlates with IFN-induced proliferation arrest in these cells (23). In control experiments, IFN also increases gp91PHOX mRNA, verifying differentiation (Fig. 5A). We found relatively insignificant change in ICSBP mRNA by this technique, consistent with our previous analysis of ICSBP mRNA by Western blot (6).
|
, and reporter gene assays were performed. We found that overexpression of ICSBP significantly increases NF1 promoter activity in undifferentiated U937 cells (p = 0.05, n = 8) (Fig. 5B). Additionally, IFN differentiation increases reporter expression from the NF1 promoter construct (Fig. 5C). Interestingly, overexpression of ICSBP significantly augments the effect of IFN on NF1 promoter activity in U937 transfectants (p = 0.02, n = 8). However, IFN did not increase ICSBP expression, consistent with our previous results (6). In these experiments, neither IFN treatment nor ICSBP overexpression alter reporter expression from pCATE control vector, which was subtracted as background.
Based on these results, we further analyzed the NF1 promoter to identify an ICSBP-binding site. Analysis of the promoter sequence did not reveal any perfect matches to previously described consensus sequences for ICSBP binding, such as composite Ets/IRF or PRDI consensus sequences. However, we found a sequence in the proximal promoter, similar to the PRDI consensus (–326 to –331 bp from the ATG: ccactt versus tcactt for the PRDI consensus) (22). Therefore, we generated a truncation mutant of the NF1 promoter, including 337 bp of 5'-flank, up to and including this sequence (referred to as 337NF1). This vector was co-transfected into U937 cells with a vector to overexpress ICSBP (or empty vector control), and transfectants were analyzed with and without IFN treatment. Consistent with our results above, ICSBP significantly increased expression of the 337NF1-pCATE construct in U937 transfectants (p < 0.05, n = 9) (Fig. 5B). As in transfection experiments with the 973-bp promoter fragment, this effect was increased by IFN treatment of the transfectants (p < 0.02, n = 9) (Fig. 5C).
Therefore, we generated a 315-bp truncation mutant, which excludes the PRDI-like NF1 promoter sequence (315NF1). U937 cells were co-transfected with the 315NF1-pCATE vector and a plasmid to overexpress ICSBP or vector control. Reporter expression was determined with and without IFN differentiation of the transfectants. In contrast to results with 973 and 337 bp of the NF1 promoter, ICSBP overexpression did not increase reporter activity from 315NF1-pCATE, with or without IFN (Fig. 5, B and C). To verify these results, an additional construct was generated with 230 bp of the proximal NF1 promoter (230NF1) and used in transfections as above. Similar to the 315-bp construct, neither ICSBP overexpression nor IFN treatment altered reporter expression from this 230-bp construct (Fig. 5, B and C).
ICSBP Interacts with a PRDI-like Sequence in the Proximal NF1 Promoter—Our experiments indicate that truncation of the PRDI-like sequence from the NF1 promoter abolishes ICSBP-induced transcriptional activation. To determine whether ICSBP interacts with this NF1 promoter region, EMSAs were performed with an oligonucleotide probe representing –310 to –337 bp of the NF1 promoter and nuclear proteins from U937 cells, with and without IFN differentiation. In EMSA with nuclear proteins from differentiated U937 cells, we identified a protein complex, not present in experiments with nuclear proteins from undifferentiated U937 cells (Fig. 5D). To determine whether this complex binds the PRDI-like sequence, EMSAs were performed with double-stranded oligonucleotide competitors. We found that either homologous oligonucleotide or an oligonucleotide representing the PRDI consensus competed for binding of this IFN-induced complex.
We next investigated whether this complex includes proteins cross-immunoreactive with ICSBP. EMSA were performed with the –310 to –337 bp NF1 promoter oligonucleotide probe and nuclear proteins from U937 cells (with and without IFN differentiation) in the presence of ICSBP antibody or irrelevant control antibody. We found that the PRDI-cross-competitive complex was cross-immunoreactive with ICSBP, but not control antibody, in EMSA with nuclear proteins from IFN-differentiated U937 cells (Fig. 5E). In contrast, ICSBP antibody did not recognize complexes in EMSA with undifferentiated U937 nuclear proteins.
Nuclear proteins used in these experiments were tested for integrity in preliminary EMSA with control probes. Each experiment was repeated at least three times, using different batches of nuclear proteins, and representative results are shown. Autoradiogram exposures are shown, which emphasize the increase in binding of the ICSBP-cross-immunoreactive complex with IFN differentiation.
Examination of the murine NF1 5'-flank indicates the –334 to –312 bp sequence is 100% identical to the –337 to –315 bp sequence in the human promoter (22). Therefore, we determined whether ICSBP interacts in vivo with this murine promoter sequence by chromatin immunoprecipitation (Fig. 1B). As discussed above, we found that the region of the murine promoter, homologous to the ICSBP-binding sequence in the human promoter, co-precipitates from lysates of murine bone marrow-derived monocytes with ICSBP antibody. These results further suggest the potential significance of ICSBP interaction with the proximal NF1 promoter.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The Ras-GAP function of Nf1 has previously been shown to antagonize hematopoietic cell proliferation (19), and Nf1-deficient cells are hypersensitive to various cytokines (15). In previous studies, colony formation from ICSBP–/– (14) and Nf1-deficient (15) bone marrow cells was greater than normal at low GM-CSF doses (0.01–1.0 ng/ml) but not at higher doses (10 ng/ml) (14, 15). This GM-CSF dose response is consistent with our studies of Ras and ERK activation in freshly isolated bone marrow myeloid cells. In comparison with wild type cells, we found only a minor increase in the abundance of active Ras in ICSBP–/– bone marrow progenitor cells in 10 ng/ml GM-CSF. Also, there was no consistent difference in ERK phosphorylation between wild type and ICSBP-deficient cells at this GM-CSF dose. However, our reconstitution experiments with ICSBP–/– myeloid cells document an impact of ICSBP on Nf1-expression, Ras activity, and ERK phosphorylation in GM-CSF-treated cells.
In ICSBP-deficient cells, we found expression of ICSBP or the Nf1 GAP-related domain reverses hypersensitivity to proliferation at GM-CSF doses from 0.01 to 10.0 ng/ml. Consistent with this, ICSBP expression in ICSBP–/– murine myeloid cells also increases Nf1 expression, decreases activated Ras, and decreases phosphorylated ERK in cells treated with 10 ng/ml GM-CSF. The lack of GM-CSF dose dependence in these reconstitution experiments is probably due to the impact of ICSBP protein abundance on Nf1 expression. We found that ICSBP is overexpressed in ICSBP-transduced ICSBP–/– cells in comparison with endogenous ICSBP expression in wild type cells treated with 10 ng/ml GM-CSF. Indeed, ICSBP expression in transduced ICSBP–/– cells is similar to endogenous ICSBP expression in M-CSF-treated wild type cells. Therefore, our results correlate increased ICSBP-expression, due to either ICSBP transduction or M-CSF stimulation, with increased Nf1 expression. In either case, increased Nf1 is associated with decreased Ras activation, ERK phosphorylation, and proliferation.
In normal myeloid progenitors, M-CSF activates Ras and ERK, leading to proliferation (33). Ras and ERK activation are maximal in the first hour and sustained over 24 h (33). After 48 h, differentiation occurs, accompanied by proliferation arrest (33). Most prior studies of M-CSF focus on the early response to ligand binding. In contrast, down-regulation of M-CSF-induced proliferation, in differentiating myeloid cells, has not been extensively investigated. To understand the mechanism by which ICSBP deficiency contributes to the myeloproliferative phenotype, we focused on sustained cytokine exposure, since this mimics the in vivo situation. We found that ICSBP deficiency is associated with sustained M-CSF-induced proliferation, accompanied by increasing Ras and ERK activation, over 72 h. This aspect of the ICSBP-deficient phenotype is rescued by ICSBP or Nf1-GRD expression, thereby identifying a role for these proteins in down-regulation of M-CSF-induced proliferation. Consistent with this, we found that M-CSF increases Nf1 mRNA, protein, and NF1 promoter activity. Therefore, our studies identify a role for Nf1 during normal myelopoiesis. These results are complementary to previous investigations, which defined a role for Nf1 in myeloid malignancy. Additionally, we identify a mechanism which regulates Nf1 expression and therefore Ras activity in hematopoietic cells.
Our studies specifically investigate the proliferative response to M-CSF, during 24–72 h of differentiation. Previously, other investigators found that cultures of ICSBP–/– bone marrow contained fewer cells, in comparison with wild type cultures, after 8 days in M-CSF (14, 34). Cell numbers in such experiments are a function of both proliferation and apoptosis. Therefore, these results suggest that ICSBP deficiency may have more impact on M-CSF-induced proliferation than on survival. This differential effect may be more pronounced with prolonged incubation in M-CSF as a sole cytokine. Investigations to address this are ongoing in the laboratory. These investigators also found less M-CSF receptor protein and M-CSF-induced Ras activation in ICSBP–/– macrophages than in wild type cells (34). This is probably due to the differentiation stage of ICSBP–/– cells in comparison with wild type in these studies. Blocked terminal differentiation is often difficult to detect by flow cytometry, using the usual myeloid markers (35). However, gp91PHOX (not shown) and granule protease (34) expression is decreased in ICSBP–/– macrophages, consistent with late differentiation block. Decreased granule protease expression in ICSBP–/– macrophages has been shown to increase Cbl protein, consequently decreasing M-CSF receptor abundance (34). This suggests an additional mechanism of differentiation block in ICSBP-deficient macrophages.
To avoid difficulties in comparing differentiation-blocked ICSBP-deficient cells with mature wild type cells, we focused on the impact of ICSBP deficiency on the differentiation process. To do this, we used murine bone marrow-derived cells and long term cultured ICSBP–/– myeloid cells as differentiation models. Short term culture of ScaI+ bone marrow cells, in GM-CSF + IL-3, induces differentiation to committed myeloid progenitors (29). To study the effect of M-CSF-induced differentiation, we compared cells cultured for 72 h in M-CSF (monocytes) versus GM-CSF + IL-3 (myeloid progenitors). We chose 72 h based on previous studies of cultured bone marrow cells (30) and flow cytometry of wild type cells. We treated some cells with IFN to promote monocyte differentiation (36) and increase ICSBP expression in mature cells (30).
Similarly, we studied NF1 promoter activity in IFN-treated U937 cells as a model of monocyte differentiation (23). In U937 transfection experiments, we found that overexpressed ICSBP activates transcription from the NF1 promoter, and IFN increases the effect of ICSBP without increasing ICSBP expression (6). One possible explanation is that IFN induces expression of another factor involved in NF1 transcription. Alternatively, IFN-induced phosphorylation may increase ICSBP affinity for essential protein-protein interactions, as we found for activation of the CYBB and NCF2 genes (6) and as other investigators found for activation of an artificial promoter construct with the PRDI consensus (7). Consistent with this latter possibility, our studies indicate ICSBP interacts with a sequence in the NF1 promoter, similar to the PRDI consensus. Therefore, our result suggest ICSBP may interact with this PRDI-like NF1 cis element as part of a multi-IRF complex, dependant on IRF protein tyrosine phosphorylation. Although beyond the scope of the present investigations, these possibilities are currently under investigation in the laboratory.
Although mice with ICSBP and Nf1 deficiency both develop a myeloproliferative disorder, only ICSBP–/– mice develop blast crisis. This suggests that Nf1 deficiency is not the only explanation for the ICSBP anti-leukemia effect, although it contributes to the phenotype. Progression to clonal myeloid malignancy involves not only disregulated proliferation but also deranged apoptosis and chromosomal stability. This is consistent with the observation that CML blast crisis is often associated with accumulation of additional chromosomal translocations. Using chromatin immunoprecipitation, we have identified additional putative ICSBP target genes that address these potential mechanisms. These will be the focus of our future investigations.
|
FOOTNOTES |
|---|
These authors contributed equally to this work.
To whom correspondence should be addressed: Feinberg School of Medicine, 710 North Fairbanks Court, Olson Pavilion Rm. 8524, Chicago, IL 60611. Tel.: 312-503-4625; Fax: 312-908-5717; E-mail: e-eklund{at}northwestern.edu.
1 The abbreviations used are: ICSBP, interferon consensus sequence-binding protein; N
