Published online before print July 18, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
in differentiating myeloid cells results in neutrophil expansion and induces a distinct, antiapoptotic and pro-oncogenic gene expression pattern
,1
Departments of
* Pediatrics, Section of Hematology-Oncology, and
Medicine, Section of Infectious Diseases, and
Breast Center Biostatistics and Informatics Core, Baylor College of Medicine, Houston, Texas, USA
1 Correspondence: Section of Infectious Diseases, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030, USA. E-mail: dtweardy{at}bcm.tmc.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and Stat3ß, are expressed in myeloid cells. To understand the contribution of each Stat3 isoform to myelopoiesis, we conditionally overexpressed Stat3 or Stat3ß in the murine myeloid cell line 32Dcl3 (32D) and examined the consequences of overexpression on cell survival and differentiation. 32D cells induced to overexpress Stat3, but not Stat3ß, generated a markedly higher number of neutrophils in response to G-CSF. This effect was a result of decreased apoptosis but not of increased proliferation. Comparison of gene expression profiles of G-CSF-stimulated, Stat3-overexpressing 32D cells with those of cells with normal Stat3 expression revealed novel Stat3 gene targets, which may contribute to neutrophil expansion and improved survival, most notably Slc28a2, a purine nucleoside transporter, which is critical for maintenance of intracellular nucleotide levels and prevention of apoptosis, and Gpr65, an acid-sensing, G protein-coupled receptor with pro-oncogenic and antiapoptotic functions.
Key Words: transcription factor • apoptosis • hematopoiesis
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. A shorter, 83-kD Stat3ß variant is generated by alternate splicing near the carboxy terminus [1
, 5
6
7
], resulting in replacement of the 55 amino acid transactivation domain with seven unique amino acids. The ratio of Stat3:Stat3ß protein varies within myeloid cells, and the ratio is reported to decrease with cell maturation and activation [8
, 9
].
Stat3ß was shown initially to transcriptionally antagonize Stat3 and was viewed as a dominant-negative isoform [6
]. However, in other reports, Stat3ß has been demonstrated to enhance Stat3-dependent gene regulation, particularly genes encoding anti-inflammatory proteins [10
11
12
]. In addition, Stat3ß can substitute for Stat3 during embryonic development [13
]. Consistent with the view that Stat3ß is a dominant-negative isoform are findings that overexpression of Stat3ß in malignant cells results in apoptosis and tumor regression [14
15
16
]. However, the effect of Stat3ß on apoptosis in normal cells, including those within the myeloid lineage, has not been reported.
Most in vitro studies suggest that Stat3 activation in myeloid cell lines promotes differentiation of neutrophils and macrophages [17
18
19
20
21
22
], although overexpression of Stat3 partially impaired macrophage differentiation of M1 cells [18
]. However, data from transgenic mouse studies are conflicting. Transgenic expression of a G-CSFR mutant, unable to activate Stat3, resulted in profound neutropenia, and G-CSF-dependent myelopoiesis was rescued by transduction of mutant bone marrow progenitor cells with constitutively active Stat3 (Stat3-C) [23
]. Transgenic mice deficient for suppressor of cytokine signaling 3 (Socs3), an important feedback inhibitor of Stat3 activity, demonstrated persistent Stat3 activity and neutrophilia [24
25
26
]. Paradoxically, however, several groups found that mice with selective ablation of the Stat3 gene in hematopoietic cells also developed neutrophilia [27
28
29
30
]. Of note, none of these studies took into account the existence of two distinct Stat3 isoforms. Clearly, more investigations are required to resolve these disparities and determine the precise role of Stat3 isoforms in normal myeloid development.
Dysregulated Stat3 activity is strongly implicated in oncogenesis. Fibroblasts expressing Stat3-C developed malignant properties in culture and formed tumors in nude mice [31
]. In humans, Stat3-C has been demonstrated in many different types of cancers including leukemias [32
33
34
35
] and in some cases, has been associated with poor outcome [16
, 36
, 37
]. One of the primary means by which Stat3-C contributes to oncogenesis is by conferring a survival advantage to transformed cells by inhibition of apoptosis. Stat3 also promotes tumor growth by weakening anti-tumor immune responses [38
]. Thus, understanding the functions of Stat3 in myeloid cells, including the differential roles of Stat3 and Stat3ß, will provide insight into processes such as oncogenesis and immune surveillance, as they relate to myeloid leukemias and many other diseases.
We developed stably transfected lines of 32Dcl3 (32D) cells, which conditionally overexpress Stat3 or Stat3ß. Conditional overexpression of Stat3, but not Stat3ß, markedly increased the proportion and absolute number of neutrophils after G-CSF treatment, compared with cells expressing normal levels of Stat3 isoforms. Neutrophil expansion was a result of improved survival of neutrophil progenitor cells. Comparison of the gene expression profiles of Stat3-overexpressing cells with control cells suggested that Stat3 overexpression inhibited apoptosis by up-regulating antiapoptosis genes dramatically, most notably the purine nucleoside transporter Slc28a2, which is critical for maintenance of intracellular nucleotide levels and prevention of apoptosis, and Gpr65, an acid-sensing, G protein-coupled receptor with pro-oncogenic and antiapoptotic functions.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Generation of conditional Stat3 isoform-overexpressing cells
32D cells were first transfected with 20 µg linearized plasmid pUbiq.irtTa-VP16-GR*-IRESeGFP, a gift of A. Francis Stewart (Max-Planck-Institut für Molekulare Zellbiologie und Genetik, Dresden, Germany) [39
], by electroporation (Bio-Rad Gene Pulser, Bio-Rad, Hercules, CA, USA). This plasmid encodes the glucocorticoid/tetracycline-inducible transactivator and GFP with an internal ribosomal entry sequence (IRES). Stable transfectants were selected in G-418 (800 µg/ml) and then sorted by flow cytometry for GFP expression into 96-well plates (Beckman-Coulter Altra, Baylor Flow Cytometry Core Facility, Baylor College of Medicine, Houston, TX, USA). Individual clonal lines were evaluated for inducible gene expression by luciferase assay (luciferase assay system, Promega, Madison, WI, USA). Cells were transfected with 2 µg pTRE2-Hyg-Luc (Clontech, Palo Alto, CA, USA) and treated with dexamethasone (100 nM) and doxycycline (1 µg/ml; dex/dox). After 48 h, luciferase activity was measured and normalized to luciferase activity in HeLa cells expressing a tetracycline-dependent transactivator (Clontech) and the reporter.
Two clones (3A4 and 2F8) with strong inducible and low baseline luciferase activity were selected for subsequent transfection, and the second vector encoded the Stat3 or Stat3ß cDNA sequence downstream of a tetracycline response element. Human Stat3 and Stat3ß cDNAs were a gift of Dr. Rolf P. de Groot (University Hospital Utrecht, Utrecht, The Netherlands) [6
]. The BamHI/KpnI fragment containing the cDNA was removed from pSG513 and subcloned into pSP72; the BamHI/ClaI fragment was then subcloned into pRevTRE (Clontech) to construct pRevTRE/Stat3 and -/Statß. Linearized plasmid (20 µg) was introduced by electroporation, and stable transfectants were selected in 1 mg/ml hygromycin. Clonal lines were screened for inducible isoform overexpression by immunoblot. The Stat3-overexpressing clones 3A4/Stat35 (5) and 3A4/Stat36 (6) and the Stat3ß-overexpressing clones 2F8/Stat3ß15 (ß15) and 3A4/Stat3ß1 (ß1) were selected for further study. Isoform overexpression was induced by treating cells for 48 h with dex/dox, and overexpression was maintained by adding dex/dox to the media every 2 days.
Immunoblotting
For most experiments, total cellular protein was extracted by resuspending frozen cell pellets in ice-cold, high-salt buffer [20 mM HEPES, pH 7.9, 420 mM NaCl, 20 mM NaF, 1 mM Na3VO4(ortho), 1 mM Na4P2O7, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 20% glycerol], containing protease inhibitors (Protease Inhibitor Cocktail Set III, Calbiochem, San Diego, CA, USA; 1:100 dilution and PMSF 1 mM), and lysing the cells by repeated freeze-thaw cycles. In some cases, cells were incubated in 1 mM PMSF for 1 h at 37°C to inhibit intracellular proteases, followed by extraction of total cellular protein in ice-cold lysis buffer containing detergent and protease inhibitors [50 mM PIPES, pH 7.0, 0.5% Triton X-100, 50 mM KCl, 10 mM EGTA, 2 mM MgCl2, 1 mM Na3VO4(ortho), 5 mM NaF, 1 mM PMSF, 10 µg/ml leupeptin, and 100 µg/ml aprotinin], as described [40
]. Immunoblotting was done with the following antibodies: antiphospho-Y705-Stat3 (1:1000, Cell Signaling Technology, Beverly, MA, USA), total Stat3 (1:2000, N-terminal epitope, BD Transduction Laboratories, Franklin Lakes, NJ, USA), ß-actin (1:10,000, Abcam, Cambridge, MA, USA), and HRP-conjugated, goat anti-mouse antiserum (1:2000–1:20,000, BD PharMingen, San Diego, CA, USA). Bands were visualized by ECL (GE Biosciences, UK). Densitometric values were obtained using Image J (National Institutes of Health, Bethesda, MD, USA), and Stat3 isoform band densities were normalized by dividing by the corresponding ß-actin band densities.
Evaluation of differentiation
Cytospin cell preparations were stained with Wright-Giemsa for morphologic evaluation. Cells were classified as undifferentiated (blasts), partially differentiated (intermediate), or fully differentiated (bands/neutrophils), without prior knowledge of cell line or treatment group. Surface expression of the myeloid marker CD11b was measured at Days 0, 3, and 7 by flow cytometry (FACScan) using PE-conjugated anti-CD11b antibodies and PE-conjugated isotype control (BD PharMingen), according to the manufacturer's instructions. Gates were set to eliminate nonviable [7-amino-actinomycin (7-AAD)-positive] cells from analysis.
Apoptosis assays
Apoptosis assays were performed using the Annexin V-PE apoptosis detection kit (BD PharMingen). Cells were labeled with Annexin V-PE and 7-AAD, according to the manufacturer's instructions. The percentage of Annexin V-labeled cells was determined by flow cytometry (Beckman-Coulter Epics XL). The baseline percentage of Annexin V-positive cells prior to IL-3 removal (Day 0) was subtracted from subsequent measurements (Days 1–3).
Analysis of cell proliferation
Proliferation rates were analyzed with the Cell Census System (Sigma Chemical Co.), as described [41
]. Cells were labeled with the fluorescent membrane dye PKH-26, according to the manufacturer's instructions, and analyzed at Days 0, 1, 2, and 3 by flow cytometry (FL-2, Beckman Coulter Epics XL). Listmode data files were analyzed with ModFit, which resolves the fluorescence distribution into distinct, generational peaks and calculates the proliferation index (total number of cells/calculated number of cells in parental generation).
Microarray studies and analysis
Stat3-transfected (5) and control (3A4) cells were treated with dex/dox for 2 days (induced) or not (uninduced). Four induced/uninduced pairs of cultures for each cell line were prepared (batches "a"–"d"). Cells were stimulated with G-CSF (100 ng/ml) for 4 h. Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA, USA) and then purified further with the RNeasy cleanup kit (Qiagen, Valencia, CA, USA), according to manufacturers' instructions. Sixteen Affymetrix GeneChip® Mouse Expression Set 430 2.0 microarrays were used. Generation of cRNA, labeling, hybridization, laser scanning, and quality assurance was done by the Baylor Microarray Core Facility, according to Affymetrix protocols. Probe set-intensity values were normalized, modeled by the perfect match-only model, and log-transformed in dChip [42
, 43
]. Subsequent analyses were performed using Biometric Research Branch array tools, developed by Dr. Richard Simon and Amy Peng Lam (http://linus.nci.nih.gov/BRB-ArrayTools.html). Data were filtered to exclude probe sets with <25% (four of 16) present calls. Next, two-way ANOVA (groupxbatch) was performed. We included two groups in this analysis: group "5 Induced" and group "3A4 Uninduced + 3A4 Induced + 5 Uninduced" (combined control group). Random variance model modification of ANOVA, developed for experiments with small sample size [44
], was used. Genes with a false discovery rate (FDR) <0.05 were considered to be significantly, differentially expressed. The difference (d) in the mean log-transformed expression values between 5 Induced and the combined control group was calculated, and fold change was 2d.
Quantitative real-time PCR
Quantitative RT-PCR (qRT-PCR) was performed on an ABI Prism 7700 sequence detector system (Perkin Elmer-Applied Biosystems, Wellesley, MA, USA). RNA samples of 500 ng were reverse-transcribed using TaqMan® RT reagents (Perkin Elmer-Applied Biosystems) with oligo-dT(16) to prime first-strand synthesis. For selected Stat3 target genes, Taqman® primer/probe kits (Perkin Elmer-Applied Biosystems) were used, and three or more RNA samples from each cell group (5 Induced, 5 Uninduced, ß15 Induced, ß15 Uninduced, 3A4 Induced, 3A4 Uninduced) were assayed for each of the selected genes. Parallel RNA samples were diluted 1:100 for 18S rRNA amplification with 18S primer/probe kit (Perkin Elmer-Applied Biosystems). The cycle threshold (Ct) for amplification of a given mRNA transcript was normalized by subtracting the Ct for the corresponding 18S amplification, yielding the 
Ct, and the Ct values were then calibrated by subtracting the uninduced Ct from the corresponding induced Ct (Ct), and the fold change was calculated as 2–Ct, as described [45
]. For each gene and cell group (5, ß15, or 3A4), paired t-test was performed between induced and uninduced Ct values, and P < 0.05 was considered statistically significant.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
or Stat3ß or Stat3ß cDNA downstream of a tetracycline response element (pRevTRE/Stat3 or -/Stat3ß). Two double stably transfected clones were isolated for each isoform and expanded for further use: 5 and 6 and ß1 and ß15.
Inducible isoform overexpression occurred in double stably transfected clones after 48 h in dex/dox (Fig. 1
). There was a 5.3-fold and 3.7-fold increase in Stat3ß immunoblot band density in ß1 and ß15 cells, respectively (Fig. 1B)
, and a 1.6- and 1.3-fold increase in Stat3 immunoblot band density in the 5 and 6 cells, respectively (Fig. 1B)
. Neither 32D cells stably expressing only the transactivator (Clone 3A4), nor the original, untransfected 32D cells showed any difference in Stat3 isoform levels after dex/dox treatment (Fig. 1
and data not shown). Also, there was no difference in Stat3 isoform expression after treatment with either drug alone (data not shown). The lower fold induction of Stat3 in 5 and 6 cells is accompanied by an increase in the intensity of the Stat3ß band intensities. It has been demonstrated previously that limited proteolysis of Stat3 occurs in myeloid cells [47
], resulting in a truncated form of Stat3 with identical SDS-PAGE mobility as Stat3ß; it is likely that this process accounts for the combination of muted induction of Stat3 and the increased levels of the Stat3ß bands observed in these cells. Proteolysis of Stat3 likely occurred within intact cells, as immunoblot results were not altered when cells were lysed by an alternative method [40
], reported to inhibit neutrophil proteases effectively during protein extraction (data not shown).
 View larger version (36K): [in a new window] |
Figure 1. Overexpression of Stat3 isoforms was inducible in double stably transfected 32D cells. Double stably transfected Stat3 cells (5, 6), Stat3ß cells (ß1, ß15), and parental cells expressing the dex/dox-responsive transactivator only (3A4) were extracted prior to dex/dox treatment (–; solid bars) and after 48 h of dex/dox induction (+; shaded bars). (A) Extracted proteins were separated by SDS-PAGE and immunoblotted with antibody to Stat3 (upper panel) or ß-actin (lower panel). (B) Densitometry was performed with Image J according to the user's manual (National Institutes of Health). Values for Stat3 and Stat3ß bands were normalized by dividing by the corresponding ß-actin band density and are expressed as mean ± SEM of n = 3 (3A4), n = 8 (5), or n = 4 (6, ß1, ß15); *, P < 0.05, for Induced versus Uninduced by paired t-test.
|
and total Stat3 levels were increased statistically significantly in the induced 5 cells over 7 days of G-CSF treatment compared with simultaneously cultured, uninduced cells; phospho-Stat3ß and total Stat3ß levels were likewise increased statistically significantly in the induced ß15 cells (Fig. 2B)
. Moreover, when Stat3 was overexpressed, phospho-Stat3 levels exceeded phospho-Stat3ß levels at Days 3 and 7; conversely, phospho-Stat3ß levels exceeded phospho-Stat3 levels in Stat3ß-overexpressing cells. Of note, the presence of tyrosine-phosphorylated Stat3ß prior to G-CSF treatment was observed frequently; constitutive Stat3ß phosphorylation has been reported in several other studies [6
, 48
].
 View larger version (45K): [in a new window] |
Figure 2. G-CSF-mediated, tyrosine phosphorylation of Stat3 isoforms was increased and sustained in double stably transfected 32D cells when induced with dex/dox compared with uninduced cells. Double stably transfected Stat3 cells (5; left) and Stat3ß cells (ß15; right) were induced (left panels; shaded bars) with dex/dox on Day –2 (dotted bar) and then removed from IL-3 and stimulated with G-CSF on Day 0 (hatched bars) or were not induced (right panels; solid bars) but were otherwise treated identically and simultaneously. Cells were removed on the days indicated for protein extraction. The Day –2 samples were taken prior to dividing the cells into induced and uninduced parallel cultures. The Day 0 samples were taken prior to the addition of G-CSF. (A) Extracted proteins were separated by SDS-PAGE and immunoblotted with antibody to phosphotyrosine-Stat3 (pStat3/ß; top panels), Stat3 N terminus (middle panels), or ß-actin (bottom panels). (B) Density values for phospho-Stat3 and total Stat3 isoforms were normalized by dividing by the corresponding ß-actin band density and are expressed as fold change from baseline Day –2 values. Data shown are mean ± SEM of n = 4 for 5 and ß15 cells. ANOVA for repeated measures was performed and showed a significant difference (P<0.05) between Induced and Uninduced groups (i.e., between-subjects effect) for phospho-Stat3 and total Stat3 values from 5 cells and for phospho-Stat3ß and total Stat3ß values from ß15 cells.
|
, but not Stat3ß, generate more neutrophils-transfected clones 5 and 6 showed a 2.3- and 2.4-fold increase, respectively, in the percentage of neutrophils on Day 7 (Fig. 3A
) compared with uninduced cells (P<0.05). Morphologic differentiation was accompanied by an increase in CD11b expression at Day 7, as measured by flow cytometry (data not shown). There was no difference between the percentages of mature cells on Day 7 in the induced versus uninduced Stat3ß-transfected clones or 3A4 control cells (Fig. 3A)
. We also noted a consistently increased percentage of neutrophils in the uninduced, Stat3-overexpressing cells compared with the Stat3ß-overexpressing cells and control cells (Fig. 3A)
. This finding may be a result of clonal differences or leakiness of exogenous Stat3 expression in the absence of dex/dox, which was not apparent by immunoblot but nevertheless functionally detectable.
 View larger version (25K): [in a new window] |
Figure 3. Overexpression of Stat3 resulted in more mature neutrophils after culture in G-CSF. Cells were treated with dex/dox at Day –2 (Induced; shaded bars) or not (Uninduced; solid bars) and then removed from IL-3 and cultured in G-CSF from Days 0–7. (A) Greater percentages of mature neutrophils were found in cultures of induced 5 and 6 cells, compared with uninduced 5 and 6 cells and compared with ß15, ß1, and 3A4 cells. Values are percentages of neutrophils among cells counted at Day 7 of G-CSF treatment. Data shown are the mean ± SEM for n = 5 (5) or n = 4 (6, ß15, ß1, 3A4); *, significant differences (P<0.05) by Student's t-test, Induced versus Uninduced. (B) Higher overall cell numbers were seen in Stat3-overexpressing cultures. Viable cell number was determined by Trypan blue dye exclusion and is expressed as the percent of cells plated on Day 0. Data are mean ± SEM for n = 5 (5), n = 7 (6), or n = 4 (ß15 and ß1). *, P < 0.05, Induced versus Uninduced by ANOVA. (C) Overexpression of Stat3 resulted in expansion of the absolute number of mature neutrophils upon G-CSF stimulation. Data from clone 5 are shown. The absolute number of mature neutrophils (left), the absolute number of intermediate morphology cells (middle), and the absolute number of immature blasts (right) on each day were calculated by multiplying the percent of cells in the category times the total number of cells. Values are mean ± SEM for n = 5. *, P < 0.05, by ANOVA.
|
-inducible clones, the number of viable cells 2–3 days after transfer to G-CSF was increased two- to fivefold in induced cells compared with uninduced cells (P<0.05; Fig. 3B
), and for the 5 clone, this increase in viable cell number persisted out to Day 7. No difference in viable cell number was seen in 32D clones overexpressing Stat3ß (Fig. 3B)
nor in 3A4 control cells, with and without dex/dox treatment (data not shown), indicating that the observed difference was not simply a result of dex/dox treatment. Calculation of the absolute number of cells in each morphologic category revealed that overexpression of Stat3 led to a significant expansion of cells of all categories, but the effect was most pronounced within the neutrophil population (5, Fig. 3C
; 6, data not shown). The numbers of neutrophils at Days 3, 5, and 7 were increased 5.1-, 4.7-, and 15.5-fold, respectively, in the induced 5 cells compared with uninduced 5 cells (P<0.05).
Stat3 overexpression does not increase proliferation but reduces apoptosis
The increased cell number in the clones overexpressing Stat3 could have been a result of increased proliferation, decreased apoptosis, or both. To test whether Stat3 overexpression caused increased proliferation, we used the Cell Census System (Sigma Chemical Co.), which uses a fluorescent dye to label cell membranes. As cells divide, each daughter cell has half of the fluorescence intensity of its parent. We found no difference in the proliferation rates between induced and uninduced Stat3-transfected cells after transfer to G-CSF (Fig. 4A
) nor between 3A4 cells, with and without dex/dox treatment (data not shown).
 View larger version (12K): [in a new window] |
Figure 4. Stat3 overexpression did not affect proliferation but did inhibit apoptosis. Stat3 clone 5 cells were treated with dex/dox at Day –2 (Induced; solid gray line) or not (Uninduced; dashed black line) and then removed from IL-3 and stimulated with G-CSF at Day 0 (A), where cells were labeled with the fluorescent membrane dye PKH-26. Fluorescence was measured at the indicated days. ModFit was used to resolve generational peaks and calculate the proliferation index (total number of cells/number of cells in the parental generation). Values are mean ± SEM for n = 4. (B) Cells were removed at the indicated days for labeling with Annexin V-PE. Values are mean ± SEM for n = 5, corrected for the baseline percent of Annexin V-positive cells on Day 0. *, P < 0.05, by ANOVA.
|
overexpression resulted in improved survival, induced and uninduced cells were switched from IL-3 to G-CSF, and apoptotic cells were quantified at various time-points by Annexin V-PE (BD PharMingen) staining. Stat3-overexpressing cells had 20–50% as many Annexin V-positive cells as the uninduced cells at Days 1–3, when apoptosis as a result of IL-3 withdrawal is greatest (P<0.05, Fig. 4B
). No difference in the percentage of Annexin V-positive cells was seen between induced and uninduced Stat3ß-transfected cells nor between 3A4 control cells, with and without dex/dox treatment (data not shown). Taken together, these data show that Stat3 overexpression in 32D cells resulted in a dramatic expansion of neutrophil numbers through resistance of progenitors to apoptosis.
Stat3 overexpression differentially regulates gene expression
To identify the gene expression profile associated with increased neutrophil production and apoptosis resistance, we performed microarray analyses comparing cells with increased Stat3 expression with cells with normal Stat3 expression. Stat3-transfected cells (5) were induced with dex/dox for 48 h (Induced) or not (Uninduced). To control for genes which were affected by dex/dox treatment itself, we also assayed 3A4 cells, with and without dex/dox treatment. After the induction period, cells were removed from IL-3 and stimulated with G-CSF (100 ng/ml) for 4 h. This brief time period was chosen to capture genes which were regulated directly by Stat3. Four induced/uninduced pairs of cultures ("a"–"d") for the 5 and 3A4 cell lines were prepared for analysis on Affymetrix GeneChip® Mouse Expression Set 430 2.0 arrays.
After appropriate normalization, modeling, and filtering, ANOVA analysis identified 148 gene probe sets, which were differentially expressed in cells with increased Stat3 expression (5 Induced) compared with cells with normal Stat3 expression (5 Uninduced+3A4 Induced+3A4 Uninduced), using as a statistical threshold a FDR <0.05 (full list included as supplemental material). After excluding redundant probe sets and unnamed sequences, 84 known genes were differentially expressed between the two groups, and 45 were up-regulated, and 39 were down-regulated. The 84 genes were categorized according to function based on Gene Ontology and PubMed databases (Table 1
) and illustrated by hierarchical clustering (Fig. 5
). As the analysis was designed to identify differences in gene regulation between cells with overexpression versus normal expression of Stat3, the fold change values were generally 1.5–3. Nevertheless, a few genes showed increases over tenfold, notably two linked to apoptosis resistance—Slc28a2, which encodes a purine nucleoside transporter, and Gpr65, which encodes a member of the acid-sensing, G protein-coupled receptor family.
|
View this table: [in a new window] |
Table 1. Genes Expressed Significantly Differentially in Stat3-Overexpressing Cells
|
 View larger version (23K): [in a new window] |
Figure 5. Microarray analysis identified 84 known genes, which were expressed differentially using a limit of FDR <0.05. High expression is indicated in red, and low expression is indicated in green (5 I, 5 Induced; 5 U, 5 Uninduced; 3A4 I, 3A4 Induced; 3A4 U, 3A4 Uninduced).
|
5-induced versus uninduced cells (Table 1)
. For six of eight genes examined, there was a statistically significant difference in mRNA levels [i.e., normalized Ct (Ct)] between 5-induced and 5-uninduced cells, which confirmed the results of the microarray analysis in terms of direction and relative magnitude of expression changes. Two genes, Casp6 and Hip1, were down-regulated by microarray analysis and were also down-regulated by qRT-PCR, but the difference between induced and uninduced Cts did not reach statistical significance. Socs3 was not regulated differentially by microarray, qRT-PCR, or Western blot analyses (data not shown). qRT-PCR also was performed for these nine genes using RNA from dex/dox-induced and uninduced 3A4; only one of the selected genes, ecotropic virus integration site-1 (Evi1), showed a significant change with treatment: 1.5-fold increase in 3A4 cells versus 4.0-fold increase in 5 cells. Thus, the overall rate of agreement between mRNA changes detected by microarray versus qRT-PCR was 15 out of 18 or 83%.
qRT-PCR performed using RNA from dex/dox-induced and uninduced ß15 cells revealed that there was no difference in mRNA levels between induced and uninduced cells for five of nine genes examined. In four of nine genes (Evi1, Pdcd1, Tde1, and Slc28a2), the levels of mRNA were increased in induced versus uninduced cells, but the fold increase was smaller than that seen in the 5-induced versus uninduced pair (data not shown). Thus, the changes in mRNA levels observed in dex/dox-treated 32D cells were selective for Stat3 overexpression.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and Stat3ß in normal myeloid cell survival, proliferation, and differentiation. We developed a system for conditionally enforcing increased expression of Stat3 or Stat3ß in 32D cells to better understand the distinct functions of each isoform in G-CSF-stimulated myelopoiesis. Cell counts, Annexin V-binding analyses, and cell division assays indicated that Stat3 overexpression did not promote proliferation but rather inhibited apoptosis, resulting in markedly higher myeloid cell numbers at all stages of differentiation but especially neutrophils. Proliferation, apoptosis, and differentiation were not affected when Stat3ß overexpression caused phospho-Stat3ß levels to exceed phospho-Stat3, in agreement with previous work [49
]. Experiments with 3A4 control cells showed no significant effect of dex/dox treatment itself.
Activation of Stat3 has been shown to promote differentiation of myeloid cell lines [17
18
19
20
21
22
]. However, Hevehan et al. [8
] reported that Stat3 levels are highest in immature myeloid cells and decrease as the cells differentiate, suggesting that persistent elevation of Stat3 could inhibit differentiation. Our data do not support this hypothesis, as cells with increased Stat3 expression had increased proportions of differentiated cells in response to G-CSF stimulation, as measured by morphology and by CD11b expression. Our data do not provide concrete evidence that Stat3 activity promotes myeloid differentiation directly either. Rather, our results suggest that Stat3 activity plays a permissive role in this system, allowing more cells to survive and fully differentiate.
Data from transgenic mice with Stat3 deletions targeted to the hematopoietic system make it clear that Stat3 activity is not absolutely required for neutrophil production in vivo but is required for neutrophil homeostasis [27
28
29
30
], a complex process that may involve feedback inhibitors such as Socs3 as well as Stat3 signals generated in peripheral tissues in response to increased neutrophil numbers [50
]. In our system, Socs3 levels were the same in cells with normal and increased Stat3 activity, and any peripheral regulatory signals would be absent. Thus, the antiapoptosis effects of increased Stat3 prevailed to promote neutrophil expansion. Our data are consistent with the results from the Socs3 knockout mice, in which Stat3 activity was increased, and the animals developed severe neutrophilia [24
25
26
]. Indeed, these investigators also described reduced apoptosis of cells within the myeloid lineage. Our findings are also consistent with a more recent study in the Stat3 knockout model, which demonstrated that Stat3 is required for G-CSF-dependent, Socs3-independent emergency granulopoiesis [51
].
Stat3 overexpression generated a distinct gene expression pattern, in which 84 known genes were significantly differentially modulated. Prominent among these were genes involved in apoptosis. Our microarray experiments were designed to identify genes, which may contribute to the neutrophil expansion observed in Stat3-overexpressing cells upon G-CSF stimulation. Previously, proapoptotic proteins (e.g., Bax, Bad) and antiapoptotic proteins (e.g., Bcl-2, Bcl-XL, Mcl-1) have been reported to be down-regulated and up-regulated, respectively, by Stat3 activation in neoplastic cells [14
, 52
53
54
]. Our microarray studies identified a new set of Stat3-responsive genes involved in apoptosis resistance in 32D cells, which have been used over the past 20 years as an in vitro model of normal neutrophil differentiation. It is notable that the purine nucleoside transporter CNT2 (Slc28a2) was up-regulated 13.5-fold in the microarray analysis and nearly 600-fold by qRT-PCR. The Slc28a2 gene product is essential for maintaining intracellular purine nucleotide levels and is particularly noteworthy in light of a recent report describing inhibition of the intrinsic apoptosis pathway by maintenance or elevation of intracellular nucleotide concentrations [55
]. Two genes with antiapoptotic and oncogenic effects, Evi1 and methionine aminopeptidase-2 (Metap2), were also up-regulated. Evi1 encodes a transcription factor implicated in acute myeloid leukemias (AMLs) and some epithelial cancers. It has been shown to prevent apoptosis by inhibiting JNK activity [56
] and by increasing PI-3K/Akt activity [57
]. Metap2 is garnering attention as a prosurvival and proangiogenesis factor, which is overexpressed in tumors [58
, 59
]. In addition, genes encoding caspase 6 and the caspase 8 activator Hip1 were modestly down-regulated in cells with increased Stat3 expression, and mRNAs encoding tyrosine phosphatases Phlpp and Inpp5a were reduced significantly, which may allow increased signaling of the prosurvival PI-3K-Akt pathway.
These data contribute to our understanding of the role of Stat3 as an oncogene. Stat3 is overexpressed and/or constitutively activated in a large number of different cancers. We have identified several genes implicated in oncogenesis, which were not known to be regulated by Stat3 previously. In particular, the acid-sensing, G protein-coupled receptor Gpr65 was the most strongly induced gene in our microarray analysis (16-fold). Exogenous expression of Gpr65 protein in a mouse epithelial cell line enabled the cells to form tumors in nude mice, and Gpr65 mRNA levels are increased in many tumors [60 ].
Finally, recent evidence has described Stat3 as a key mediator of tumor immune tolerance [38
]. Our finding that programmed death-1 (PD-1) was induced over eightfold by microarray (
30-fold by qRT-PCR) in Stat3-overexpressing cells is highly significant in this regard. PD-1 is an inhibitory immunoreceptor expressed on activated B cells, T cells, and myeloid cells. Activation of the receptor by one of its ligands, PD-L1 or PD-L2, results in recruitment of the tyrosine phosphatase Shp2, which can inactivate JAK-Stat signaling, among other effects [61
]. However, as PD-L1 and PD-L2 are cell surface ligands, PD-1 is not likely to be active in 32D suspension cultures. In contrast, PD-1-dependent feedback could be important in knockout mouse models and may partly explain the disparity between the results of Stat3 modulation studies in cell lines versus whole animals. It is interesting that PD-1-deficient mice have splenomegaly and increased neutrophils in the spleen [62
], similar to the phenotype seen in Socs3 mutant mice [24
25
26
], consistent with the idea that both proteins mediate a negative regulatory role in neutrophil homeostasis.
A recent review of leukemic stem cell biology has put forth the idea that transcription factor concentration, rather than simply presence or absence, has a dramatic influence on hematopoiesis and leukemogenesis [63
]. Regulation of Stat3 and Stat3ß concentrations is likely to be essential for maintaining normal myelopoiesis. In this regard, the reported decrease in the Stat3:Stat3ß ratio [8
] may represent a myeloid cell-intrinsic means of limiting expansion of the neutrophil lineage. In contrast, inappropriately maintained or increased levels of Stat3 would have deleterious effects. For example, AML cells with increased or constitutively active Stat3 would be expected to be resistant to apoptosis-inducing stimuli, possibly including chemotherapy. A variety of approaches aimed at inactivating Stat3 signaling in cancer cells has proven successful in preclinical models (reviewed in refs. [64
, 65
]). As the field of molecularly targeted therapeutics develops, agents targeting Stat3 are expected to be a valuable option in the treatment of patients with AML and other diseases.
Received December 27, 2006; revised June 13, 2007; accepted June 18, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and Stat3ß in immature normal and leukemic human myeloid cells Blood 88,2442-2449 and STAT3ß messenger RNA splicing Blood 98,3853-3856 and Stat3ß Mol. Cell. Biol. 17,5307-5316 and ß have unique and specific functions Nat. Immunol. 5,401-409[CrossRef][Medline] in granulocyte colony-stimulating factor signaling pathway J. Biol. Chem. 280,12621-12629 to STAT3
in human neutrophils J. Biol. Chem. 279,31076-31080
B, interleukin 6, and granulocyte-colony stimulating factor Shock 21,519-525[CrossRef][Medline]This article has been cited by other articles:
 |
|
 |
|
  K.-H. Chuang, S. Altuwaijri, G. Li, J.-J. Lai, C.-Y. Chu, K.-P. Lai, H.-Y. Lin, J.-W. Hsu, P. Keng, M.-C. Wu, et al. Neutropenia with impaired host defense against microbial infection in mice lacking androgen receptor J. Exp. Med., May 11, 2009; 206(5): 1181 - 1199. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 B. J. van Raam, A. Drewniak, V. Groenewold, T. K. van den Berg, and T. W. Kuijpers Granulocyte colony-stimulating factor delays neutrophil apoptosis by inhibition of calpains upstream of caspase-3 Blood, September 1, 2008; 112(5): 2046 - 2054. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Jiang, N. Hein, K. Eckert, J. Luscher-Firzlaff, and B. Luscher Regulation of the MAD1 promoter by G-CSF Nucleic Acids Res., March 1, 2008; 36(5): 1517 - 1531. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
