of Granulopoiesis but Is Not Required
for G-CSF-Dependent Differentiation
sites on receptor cytoplasmic tails. Activation of the PI3 kinase pathway through recruitment of adaptor proteins such as GAB2, SHP2, or Grb2 contributes to cell survival (Hibi and Hirano, 2000), while the Ras-MAP kinase cas-cade is activated during cell proliferation (Ihle, 1996). Chien-kuo Lee,1,5,6Regina Raz,1,5
Ramon Gimeno,1,5Rachel Gertner,1 Birte Wistinghausen,2Kenichi Takeshita,2 Ronald A. DePinho,3and David E. Levy1,4 1Department of Pathology and
A largely uncharacterized signaling pathway leads to 2Department of Medicine
c-myc gene induction, also required for cell cycle pro-Kaplan Comprehensive Cancer Center
gression. Finally, activation of STAT transcription factors New York University School of Medicine
by specific tyrosine phosphorylation is a common fea-New York, fea-New York 10016
ture of cytokine signaling and adds potential for consid-3Department of Adult Oncology
erable specificity given the presence of seven family Dana Farber Cancer Institute
members that are differentially activated by distinct re-Department of Medicine and
ceptors (Ihle, 2001). Each of these signaling pathways Department of Genetics
requires receptor docking of JAK enzymes; distinct dis-Harvard Medical School
tal receptor cytoplasmic regions mediate downstream Boston, Massachusetts 02115
signals.
Selective mutagenesis of receptor signaling motifs coupled with ectopic expression of dominant-negative Summary
signaling molecules has been used to dissect the indi-vidual requirements of distinct pathways in cell re-STAT3 has been described as an essential component
sponses. However, results of such studies have been of G-CSF-driven cell proliferation and granulopoiesis.
contradictory and controversial. A case in point is analy-This notion was tested by conditional gene ablation
sis of the role of STAT3 during response of cells to G-CSF, in transgenic mice. Contrary to expectation,
granulo-a growth fgranulo-actor criticgranulo-al for grgranulo-anulopoiesis (Smithggranulo-all et cytes developed from STAT3 null bone marrow
pro-al., 2000). Early work using G-CSF-responsive cell lines genitors, and STAT3 null neutrophils displayed mature
demonstrated potent activation of STAT3 as well as effector functions. Rather than a deficit in
granulo-weaker activation of STAT1 and STAT5 (Tidow and poiesis, mice lacking STAT3 in their hematopoietic
Welte, 1997). However, activation of STAT3 was linked progenitors developed neutrophilia, and bone marrow
to both proliferation and terminal differentiation and ar-cells were hyperresponsive to G-CSF stimulation.
rest of granulocytes, possibly reflecting cell type-spe-These studies provide direct evidence for
STAT3-inde-cific differences among the cell lines tested (Coffer et pendent granulopoiesis and suggest that STAT3
di-al., 2000; Takeda and Akira, 2000). Analysis in transgenic rects a negative feedback loop necessary for
control-mice held out the prospect of precise definition of the ling neutrophil numbers, possibly through induced
role of individual STAT proteins in cytokine responses. expression of the signaling inhibitor, SOCS3.
Unlike other STAT proteins, none of which is required for viability, ablation of STAT3 produced early
embry-Introduction onic lethality (Takeda et al., 1997). In fact, STAT3 is
required for maintenance of embryonic stem cells, at Mammalian hematopoiesis follows an orchestrated pro- least in vitro (Boeuf et al., 1997; Matsuda et al., 1999; gram of cell proliferation, differentiation, and apoptosis, Niwa et al., 1998; Raz et al., 1999), suggesting that it resulting in expansion of progenitor stem cells that be- might be generally required for cell proliferation or sur-come progressively restricted along distinct develop- vival. G-CSF plays a critical role in granulopoiesis in mental lineages. The signals that control these steps vivo, as demonstrated by severe neutropenia in animals remain poorly understood, but cytokines have been im- lacking the gene for G-CSF or for G-CSFR (Lieschke et plicated in many aspects of this process, including as al., 1994; Liu et al., 1996). Moreover, it was recently mitogens, survival factors, inducers of differentiation, reported that ablation of the STAT3 recruitment site on and stimulators of cell migration and effector functions. the G-CSFR or ectopic expression of DN-STAT3 led to Cytokines function through specific cell surface recep- a similar neutropenic phenotype, suggesting STAT3 is tors that stimulate multiple intracellular signaling path- a major target for G-CSFR signaling (McLemore et al., ways, largely through activation of receptor-associated 2001). However, selective ablation of STAT3 expression and cytoplasmic protein kinases (Ihle et al., 1995). in committed myeloid cells through Cre recombinase-At least four signaling pathways contribute to cytokine dependent gene deletion directed by the macrophage lysozyme promoter did not prevent normal production action. All of these pathways require activation of protein
of neutrophils (Takeda et al., 1999), limiting any STAT3 tyrosine kinases of the JAK family and depend at least
requirement to an earlier, progenitor stage. in part on phosphorylation of specific tyrosine docking
We have examined the role of STAT3 in hematopoiesis by selective ablation of the STAT3 gene in hematopoietic
4Correspondence: levyd01@med.nyu.edu
progenitor cells. In contrast to an expected neutropenic
5These authors contributed equally to this work.
phenotype, these mice produced excess neutrophils in
6Present address: Graduate Institute of Immunology, National
pro-assay of total bone marrow cells of poly(I:C)-treated mice showed efficient loss of the STAT3-flox allele and creation of a novel, ablated STAT3 null allele (STAT3⌬, lanes 1 and 2). Similar treatment of STAT3f/fmice lacking the Mx-Cre transgene showed no change in the struc-ture of the STAT3 gene (lanes 3 and 4). Only trace amounts of STAT3-flox were detectable following Mx-Cre induction. Quantitation of the comparative levels of STAT3 null to STAT3-flox indicated that the efficiency of STAT3 deletion after poly(I:C) treatment was⬎95% in the marrow (data not shown). Deletion of STAT3 by poly(I:C) was also observed in other organs, with highest efficiency in liver (ⵑ100%), moderate efficiency in kid-ney, spleen, lung, and heart (40%–60%), and lower effi-ciency in skeletal muscle and testis (10%–20%) (data not shown). For experiments in this paper, comparisons were made between littermate STAT3f/f mice with or without the Mx-Cre transgene, treated equivalently with poly(I:C). STAT3⫹/fheterozygous mice containing Mx-Cre responded to poly(I:C) in a manner similar to that of Cre-negative mice (data not shown), demonstrating that the phenotypes observed were not due to expres-sion of Cre or presence of the transgene but rather to loss of STAT3. While adverse effects of Cre expression have been noted in experiments with fibroblasts (Silver and Livingston, 2001), hematopoietic cells do not appear to be affected (de Alboran et al., 2001).
To further confirm a functional deletion of STAT3 in Figure 1. STAT3 Is Deleted in Bone Marrow
bone marrow from Mx-Cre animals, nuclear extracts pre-(A) Mx-Cre:STAT3f/f(Cre) or STAT3f/fmice were sacrificed 10 days
after poly(I:C) treatment. DNA from BM was analyzed for STAT3 by pared from poly(I:C)-treated, G-CSF-stimulated bone mar-PCR (upper left panel). BM cells were stimulated with G-CSF (10 row cells were subjected to gel mobility shift analysis ng/ml), and protein extracts were analyzed by EMSA (lower left) and in the absence or presence of anti-STAT antibodies. As by immunoblotting with antibodies against phospho-STAT3, STAT3,
shown in Figure 1A (lower left panel), a G-CSF-inducible and STAT1, as indicated (right panel). ns, nonspecific.
protein-DNA complex, which was absent in cells from (B) Increased late-stage neutrophils in BM lacking STAT3. Neutrophil
Mx-Cre animals treated with poly(I:C), was detected in lineage cells were enumerated from control and STAT3KO marrow
extracts from control cells. This complex contained following Wright-Giemsa staining.
STAT3 as indicated by reactivity with STAT3 anti-bodies but not anti-STAT1 antianti-bodies, consistent with genitors differentiated normally in response to G-CSF
previous reports that STAT3 is the principal STAT acti-and produced neutrophils that were morphologically
vated in response to G-CSF. Significantly, no STAT3-and functionally mature. These data demonstrate that DNA complex was detected in extracts from G-CSF-STAT3 is not required for G-CSF-dependent
prolifera-treated cells isolated from Mx-Cre mice, confirming loss tion, differentiation, or survival of neutrophils. Instead, of STAT3. A similar conclusion was supported by West-STAT3 is necessary for a negative feedback loop that ern blotting. STAT3 protein and tyrosine-phosphory-controls neutrophil numbers. lated STAT3, a signature for activation, were readily de-tected in STAT3f/fcells, but no signal was detected in
Results cells from Mx-Cre:STAT3f/f animals (Figure 1A, right
panel). Antibodies that recognized either N-terminal or Generation of Mice Lacking STAT3 C-terminal regions of STAT3 that were not deleted in in Hematopoietic Progenitors the mutant allele failed to detect STAT3 in mutant cells, To overcome the STAT3 requirement during embryogen- showing that no cryptic product was expressed from esis, we created conditionally mutant mice by using the the targeted allele. In particular, no immunoreactive
pro-loxP-Cre recombinase system (Zou et al., 1994). Mice tein of a size consistent with the targeted deletion (ⵑ57
were created carrying a conditional STAT3 allele with kDa) was detected using antibodies directed against exons 16–21 flanked by loxP sites (STAT3-flox), thus nondeleted regions. STAT1␣ and STAT1 proteins were targeting critical portions of the STAT3 DNA binding and detected at equal levels in cells of both genotypes, dem-SH2 domains for Cre-mediated excision to functionally onstrating the specificity of the disruption of STAT3. ablate both STAT3␣ and STAT3 (Raz et al., 1999). These These results demonstrated that deletion of STAT3 DNA mice were mated with a transgenic line bearing Cre in Mx-Cre animals effectively ablated STAT3 protein in recombinase driven by the IFN-inducible Mx promoter the marrow.
(Kuhn et al., 1995). Induction of Cre expression following
injection of the IFN-inducer poly(I:C) or of recombinant Granulopoiesis in the Absence of STAT3
IFN␣ led to effective deletion of STAT3 gene segments To examine the consequence of STAT3 loss in Mx-Cre:STAT3f/fmice, peripheral white blood cells were pre-in bone marrow (Figure 1A). A PCR-based genotyppre-ing
Table 1. WBC Differential Count and Hematological Parameters
WBC Differential Count Control STAT3KO
Neutrophils (%) 11⫾ 4 69⫾ 12 Lymphocytes (%) 85⫾ 4 28⫾ 12 Macrophages (%) N.D. 2.7⫾ 1 Eosinophils (%) 4⫾ 0 N.D. Hematological Parameters Hematocrit (%) 54⫾ 0 54⫾ 2 Hemoglobin (g/dl) 15.2⫾ 0.7 15⫾ 0.8 Total WBC (⫻ 10⫺3/l) 6.5⫾ 0.7 7.2⫾ 0.7 Platelets (⫻ 10⫺6/l) 1.2⫾ 0 1.6⫾ 0.2
Peripheral blood from STAT3f/f(control) and Mx-Cre:STAT3f/fmice
(STAT3KO) treated with poly(I:C) was analyzed by Wright-Giemsa staining and enumerated for granulocyte cell types and various he-matological parameters, by standard methods. N.D., not detected.
pared for Wright-Giemsa staining 3 weeks after induced deletion of STAT3. Morphologically identifiable neutro-phils, lymphocytes, and myeloid cells were detected in both mutant and control animals (Table 1). Normal numbers of mature red blood cells were also detected (data not shown). In contrast to animals lacking G-CSF or G-CSFR that displayed a decrease in peripheral neu-trophils (Lieschke et al., 1994; Liu et al., 1996), an in-creased percentage of neutrophils and a corresponding decreased percentage of lymphocytes were detected in STAT3 mutant mice relative to controls. Increased neutrophils were not likely due to an overt inflammatory
Figure 2. Peripheral Neutrophils Lacking STAT3 Are Functional response because the total white blood cell count was
(A) In vitro differentiated granulocytes were stained with Giemsa comparable between control and mutant mice and (upper panels) or for myeloperoxidase (lower panels).
would be expected to increase during infection (Table (B) Peritoneal exudate cells from thioglycollate-treated mice were 1). Moreover, animals were maintained on antibiotic analyzed for oxidation of dihydrorhodamine by FACS with or without treatment with PMA. Mean fluorescence intensities and stimulation throughout these experiments to suppress infections
index (S.I.) are indicated. that could complicate the analysis. Neutrophils were
(C) STAT3-deficient neutrophils are phagocytic. CFSE-labeled bac-also detected in spleen where a similar phenomenon of
teria were incubated with peritoneal neutrophils at 37⬚C for 30 min a 4-fold increase in neutrophil numbers was detected (T⫽ 0), and gentamycin was added for another 30 min (T ⫽ 30 min). in mutant mice relative to controls (data not shown). Cells washed free of extracellular bacteria at T0or T30were analyzed
by FACS for engulfed bacteria.
(D) STAT3-deficient neutrophils are bactericidal. Bacteria engulfed STAT3-Deficient Neutrophils Mature Normally
by neutrophils were enumerated by serial dilution. and Are Functional
Granulopoiesis in the absence of G-CSF or G-CSFR, in
the presence of a mutant G-CSFR unable to activate phil differentiation, and cell surface expression of these proteins was enhanced by stimulation with PMA (data STAT3, or in the presence of dominant-negative STAT3
resulted in a profound neutropenia accompanied by ac- not shown).
Functional competence of STAT3-deficient neutro-cumulation of immature precursors (Lieschke et al., 1994;
Liu et al., 1996; McLemore et al., 2001). In contrast, an phils was confirmed by four different assays: differentia-tion marker staining, respiratory burst, phagocytosis, excess of morphologically mature neutrophils was
pres-ent in the periphery of STAT3 null animals. Also in con- and bacterial killing. Bone marrow of control or mutant mice was differentiated in vitro with G-CSF for 1 week trast to the results observed in G-CSF or G-CSFR mutant
animals, no significant increase in precursors or imma- followed by staining with Giemsa and myeloperoxidase; control and mutant neutrophils were comparable (Figure ture granulocytes was detected in STAT3 null bone
mar-row. Rather, increased numbers of relatively mature band 2A). When neutrophils are exposed to appropriate stim-uli, NADPH oxidase is activated, resulting in the con-and segmented stage PMN were observed (Figure 1B).
To further examine functional parameters of STAT3 sumption of oxygen and its conversion to superoxide in a respiratory burst. Production of reactive oxygen mutant neutrophils, the ability of these cells to emigrate
to the peritoneal cavity was measured in response to intermediates was measured in peritoneal and periph-eral blood neutrophils by oxidation of dihydrorhodamine thioglycollate (TGA). STAT3 mutant neutrophils
traf-ficked normally to the peritoneum, with 4- to 8-fold to fluorescent rhodamine following stimulation with PMA. Neutrophils from both control and STAT3 null mice greater numbers accumulating within 4 hr compared to
control animals. TGA-elicited neutrophils stained nor- displayed an equivalent oxidative burst (Figure 2B). A similar equivalence of respiratory burst potential was mally for Gr-1 and CD11b, late-stage markers of
neutro-observed using wild-type and mutant peripheral blood (data not shown).
Neutrophils are phagocytic and bactericidal. To ex-amine these parameters, carboxyfluorescein succinimi-dyl ester (CFSE)-labeled bacteria were incubated with TGA-elicited peritoneal granulocytes, and bacterial take was monitored by FACS (Figure 2C). Bacterial up-take was monitored before (T⫽ 0) and after a further incubation with antibiotic to kill noninternalized bacteria (T⫽ 30). Both wild-type and STAT3 null cells were capa-ble of engulfing bacteria, although the STAT3-deficient cells displayed a 30%–50% reduced phagocytic activity. Bacterial killing was measured by determining the viabil-ity of engulfed bacteria following lysis of bacteria-loaded neutrophils (Figure 2D). Despite the reduction in phago-cytic activity observed in STAT3 null neutrophils, bacte-ria that were engulfed were effectively killed during a 30 min incubation. Taken together, the results show that STAT3 null neutrophils are functional.
Normal Granulocyte Precursor Frequency in the Absence of STAT3
Myeloid cells are derived from bone marrow progenitors, and granulocytic differentiation is potentiated by G-CSF. Granulocyte precursor frequencies were assessed in the absence of STAT3 by colony formation assays. Bone marrow from poly(I:C) treated control or Mx-Cre mice was seeded in semisolid medium containing a cocktail of IL-3, SCF, IL-6, and EPO (complete medium), support-ing growth of a panoply of myeloid progenitors. Alterna-tively, cells were cultured with G-CSF as the only sup-portive cytokine to enumerate granulocyte precursors.
A comparable number of colonies formed from bone Figure 3. Normal Numbers of Granulocyte Precursors in the Ab-marrow cells of both control and STAT3 mutant mice in sence of STAT3
complete medium (Figure 3A, left panel) and produced Control and STAT3KO BM were analyzed for myeloid precursors by colony formation in methylcellulose.
a mixture of mature cell types (right panel), including
(A) Colonies developed in complete medium (MethoCult M3434) morphologically mature neutrophils (arrows). Similarly,
containing SCF, IL-3, IL-6, and EPO. Example colonies of CFU-medium containing only G-CSF supported increased
GEMM are shown, and mature neutrophils are indicated by arrows. colony numbers in a dose-dependent manner from bone (B) Colonies developed in the presence of G-CSF as the only sup-marrow of both genotypes (Figure 3B, left panel), dem- porting cytokine.
onstrating equal numbers of general precursors and of (C) Individual colonies developed in complete medium (left panel) or in response to G-CSF (right panel) were genotyped for STAT3. granulocyte precursors despite loss of STAT3. Similar
A heterozygous flox/⫺ sample was included for comparison. ns, to the result with cells matured in response to complete
nonspecific. medium, cells with typical mature neutrophil
morphol-ogy developed in response to G-CSF, regardless of
STAT3 status (right panel). creased abundance of neutrophils in the periphery could Since poly(I:C)-induced deletion may not be 100%, we be due to increased proliferation. G-CSF is known to tested whether cytokine-dependent colony formation increase production and release of neutrophils from favored selective outgrowth of residual wild-type cells bone marrow, and it is widely used to treat neutropenia from mutant marrow. PCR assays performed on isolated in clinical settings. To directly investigate possible roles individual colonies verified the presence of cells with for STAT3 in G-CSF action in vivo, an experimental auto-the expected genotypes (Figure 3C). Colonies devel- reconstitution assay was used to monitor newly formed oped from control marrow carried the STAT3-flox allele bone marrow granulocytes and release of peripheral while Mx-Cre:STAT3-flox marrow produced only STAT3 neutrophils. Mice were deleted for STAT3 by two doses null colonies. These data confirm that G-CSF-dependent of poly(I:C) treatment, and their marrow was ablated by granulopoiesis is not impaired in the absence of STAT3, treatment with 5-fluorouracil. Subsequent hematopoi-but the neutrophilia observed in STAT3 null animals can- etic repopulation was followed in mice with or without not be explained by increased progenitor frequency. treatment with G-CSF. Peripheral blood Gr-1⫹ leuko-cytes from autoreconstituted mice were assayed for in-tracellular STAT3 and quantified by FACS. Figure 4A STAT3 Is Not Required for Response to G-CSF In Vivo,
but STAT3 Null Granulocytes Are Hyperproliferative shows results from a representative experiment in which 9% of peripheral blood cells were Gr-1 positive in a The similar numbers of granulocytic progenitors
row could be due to the cell-autonomous loss of STAT3 protein in progenitor cells or to alterations in the bone marrow microenvironment, such as from altered cytokine production. The cell-autonomous contribution of STAT3 to granulopoiesis was investigated by bone marrow trans-plantation. Control and STAT3 null marrow were injected intravenously into lethally irradiated wild-type recipi-ents. Five weeks after transplantation, bone marrow, peritoneal exudate, and peripheral organs were ana-lyzed for granulocytes by FACS (Figure 4B). Gr-1⫹cells developed from both wild-type and STAT3 null precur-sors and were detected in both the marrow and the periphery of recipients. The great majority of neutrophils developing from STAT3 null donors lacked detectable STAT3 protein, with the small number of STAT3-positive cells being derived either from residual recipient-derived cells due to incomplete irradiation or from the small number of donor cells that did not undergo recombina-tion. Consistent with the neutrophilia observed in Mx-Cre mice, an increased percentage of granulocytes ac-cumulated in the marrow of recipients reconstituted with STAT3 null cells, and peripheral cells were morphologi-cally mature (Figure 4C). The absence of STAT3 in the reconstituted cells derived from mutant donors was con-firmed by immunoblotting (Figure 4D). Increased neutro-phils were also observed in spleen and peripheral blood of mice transplanted with STAT3 null cells (data not shown). These data conclusively demonstrate that STAT3 is not required for the development of mature neutrophils and that loss of STAT3 in hematopoietic progenitors rather than in stromal tissue or other organs contributes to increased granulopoiesis.
Enhanced Proliferation, Prolonged Signaling, Figure 4. Enhanced Granulopoiesis In Vivo in the Absence of STAT3 and Reduced Induction of SOCS3 by G-CSF (A) Hematopoiesis in the absence of STAT3. 5-Fluorouracil-treated in the Absence of STAT3
control and STAT3KO mice were treated with G-CSF or BSA (10
Proliferation of control and STAT3 null bone marrow g/kg) daily for 7 days. Peripheral blood from control (upper panel) cells cultured in vitro was measured in response to in-or STAT3KO mice (lower panel) was analyzed by FACS fin-or Gr-1 and
creasing amounts of G-CSF (Figure 5A, left panel) or of STAT3.
IL-3 (right panel). Proliferation of control bone marrow (B) BM transplantation. Donor BM from control or STAT3KO mice
was analyzed by FACS for Gr-1 and STAT3 prior to transplantation cells plateaued at approximately 1 ng/ml G-CSF while (left panels). Five weeks after transplantation, peritoneal exudate proliferation of STAT3 null cells continued to increase cells (PEC, middle panels) and BM cells (BM, right panels) were
with increasing G-CSF. In contrast, control and STAT3 analyzed by FACS.
null cells responded similarly to IL-3. The increased pro-(C) PEC from recipient mice were stained with Wright-Giemsa.
liferation of mutant cells could not be accounted for by (D) Cell extracts from PEC and BM of recipient mice were
immu-decreased apoptosis, since no significant difference in noblotted for STAT3 and STAT1.
TUNEL-positive or Annexin V-positive cell numbers was observed between control and STAT3 null cultures (data not shown). Importantly, no decrease in cell proliferation approximately 3-fold following G-CSF treatment. In
con-trast, the percentage of neutrophils in the periphery of was observed in the absence of STAT3 as would be expected if it contributed an essential component to STAT3-deficient mice was high even in the absence
of cytokine treatment and did not increase further in G-CSF-driven cell growth.
It must be noted that bone marrow cells from wild-response to exogenous G-CSF. The cells repopulating
mutant mice were not due to inefficient deletion of type and STAT3-ablated mice are not identical popula-tions due to enhanced mutant granulopoiesis in vivo. STAT3 in response to poly(I:C), as indicated by loss
of intracellular STAT3 in mutant cells. The response to To determine whether enhanced response to G-CSF was cell autonomous or reflected different starting cell G-CSF averaged 3-fold in control mice while no increase
in peripheral or bone marrow neutrophils was observed populations, we isolated immature granulocytes from mice shortly after gene ablation in response to poly(I:C). in G-CSF-treated STAT3 null animals, although these
animals displayed a 3- to 4-fold higher granulocyte Following enrichment by density gradient for immature cells, equivalent numbers of Gr-1/CD11b⫹cells (Figure abundance despite whether or not they were treated
with G-CSF. 5B) were analyzed for proliferation. Similar to
unfraction-ated bone marrow, STAT3 null cells showed enhanced Enhanced granulopoiesis from STAT3 null bone
mar-Figure 6. G-CSF Stimulates Prolonged STAT1 Phosphorylation but Impaired Gene Induction in STAT3 Null Cells
(A) BM cells were untreated or stimulated with 10 ng/ml G-CSF for 15 min, 1 hr, or 2 hr, or with 10 ng/ml IL-3 for 15 min and analyzed for phospho-STATs and STAT1 levels, as indicated.
(B) Percoll-purified immature BM cells were stimulated with 10 ng/ ml G-CSF and analyzed for SOCS3, JunB, and-actin expression by RT-PCR.
absence of STAT3, activation of different STAT proteins was measured in response to G-CSF. Control cells dis-Figure 5. Enhanced G-CSF-Dependent Proliferation in the Absence
played transient and high level activation of phospho-of STAT3
STAT3, with lower levels of phospho-STAT1 and barely (A) Enhanced proliferation of STAT3 null BM in response to G-CSF
(left panel) but not to IL-3 (right panel). detectable levels of phospho-STAT5, a protein that was (B) Comparable numbers of immature granulocytes were fraction- much more sensitive to IL-3 treatment (Figure 6A). ated from BM of control and STAT3 null mice 4 days after poly(I:C) STAT3 null cells displayed essentially no phospho-treatment.
STAT3, while both STAT1 and STAT5 were still phos-(C) Enhanced proliferation of STAT3 null immature granulocytes in
phorylated. Interestingly, significant phospho-STAT1 response to G-CSF, as measured by3H-thymidine incorporation.
was observed even under basal conditions, and its (D) Dose-dependent enhancement of STAT3 null cell proliferation,
measured by BrdU incorporation into Gr-1⫹immature granulocytes. phosphorylation was more sustained, remaining signifi-cantly elevated 2 hr after G-CSF stimulation. Altered phosphorylation kinetics was not due to changes in STAT1 abundance, which remained approximately equal proliferation in response to increasing G-CSF treatment
by both3H-thymidine incorporation (Figure 5C) and by between control and STAT3 null cells (Figure 6A). Phos-phorylation levels of STAT5 were comparable, regard-BrdU incorporation into Gr-1⫹cells (Figure 5D).
We considered what might underlie the increased less of genotype.
Prolonged activation of STAT1 might reflect sustained granulocyte abundance in STAT3 null bone marrow and
the G-CSF-dependent proliferation of STAT3 null granu- signaling in response to G-CSF. Important feedback in-hibitors of JAK catalytic activity in response to cytokines locytes. One possibility might be increased expression
of G-CSF or G-CSF receptors. However, no significant are the SOCS proteins, and expression of SOCS3 is significantly induced in response to G-CSF (Endo et differences in G-CSF or G-CSFR mRNA levels were
de-tected (data not shown). Alternatively, enhanced prolif- al., 1997; Naka et al., 1997; Starr et al., 1997). Once expressed, SOCS3 is recruited to phosphorylated cyto-eration could be due to aberrant mitogenic signaling.
G-CSF is known to activate STAT1 and STAT5 in addition kine receptors and inhibits JAK catalytic activity (Krebs and Hilton, 2001). As shown in Figure 6B, SOCS3 was to STAT3, each of which might be involved in regulating
con-trast, induction of SOCS3 was impaired in the absence showed complete loss of STAT3 protein, and the bulk of autoreconstituted or transplant-derived neutrophils of STAT3. Similarly, induction of other STAT3-inducible
genes, such as JunB (Figure 6B) and Fc␥RI (data not showed loss of STAT3 (Figure 4A). Finally, Western blot analysis of elicited peritoneal exudate cells, which con-shown), was decreased in response to G-CSF,
consis-tent with the loss of STAT3. These results demonstrate sisted of highly enriched mature neutrophils, showed no STAT3 (Figure 4D).
that STAT3 is required for induction of SOCS3 in
re-sponse to G-CSF and suggest that absence of SOCS3 Another consideration is the distinct differences be-tween previous studies implying a requirement for STAT3 negative feedback may allow prolonged JAK catalytic
activity, resulting in enhanced signaling and heightened and the present gene-targeting approach. The mutant G-CSF receptor unable to activate STAT3 (d715F) is also proliferation.
severely truncated in its cytoplasmic domain, possibly complicating the phenotype of this mouse. In fact, the Discussion
deleted G-CSFR induces a heightened proliferative re-sponse to G-CSF, resulting in prolonged activation of STAT3 Is Dispensable for Granulocyte Proliferation
STAT1, STAT3, and STAT5 and impaired ligand-depen-and Differentiation
dent receptor downregulation (Hermans et al., 1998). Normal development of the granulocyte lineage requires
Similarly skewed responses have been observed with G-CSF signaling (Lieschke et al., 1994; Liu et al., 1997),
other mutant cytokine receptors. Forced expression of and a major target of the G-CSFR is STAT3 (Figures
mutant gp130 proteins that contain only one of the criti-1A and 6A). Indeed, transgenic mice with a targeted
cal tyrosine residues normally found in its cytoplasmic mutation of G-CSF receptor (d715F) that abrogates
tail showed that different tyrosine residues contribute STAT3 activation display severe neutropenia with an
distinct signals to IL-6-mediated responses. Moreover, accumulation of immature myeloid precursors in their
point mutant receptors behave differently from trun-bone marrow. Expression of a constitutively active
ver-cated mutant receptors (Schmitz et al., 2000). In the sion of STAT3 rescues this phenotype, a condition that
case of G-CSFR, the d715 receptor caused increased can also be mimicked by expression of a
dominant-proliferation in response to G-CSF. Mutation of its single negative form of STAT3 in wild-type mice. Such data
remaining tyrosine residue resulted in a loss of STAT5 have led to the conclusion that STAT3 activation is
re-phosphorylation, in addition to STAT3 (McLemore et al., quired for normal G-CSF-dependent proliferation and
2001). While STAT5 is not required for granulopoiesis granulocyte differentiation (McLemore et al., 2001). In
or G-CSF-dependent proliferation (Teglund et al., 1998), this study, we have directly tested this notion through
loss of STAT5 phosphorylation by the d715F receptor targeted deletion of STAT3 in hematopoietic progenitor
indicates that this mutation cannot be considered spe-cells in the bone marrow of transgenic mice. In contrast
cific for STAT3. Consistent with this notion, it is probable to previously reported results, we detected no
impair-that impaired granulopoiesis observed in mice express-ment in granulocyte proliferation and terminal
differenti-ing a dominant-negative STAT3 could reflect crossinhi-ation in the absence of STAT3. Rather, evidence was
bition of other receptor-mediated signals beyond STAT3 obtained that STAT3 is a negative regulator for
G-CSF-activation that depend on the same recruitment site. mediated proliferation and homeostasis of
granulo-Similar discrepancies between the effects of dominant-cytes. In the absence of STAT3, enhanced
G-CSF-negative STAT proteins and the corresponding null phe-dependent proliferation was observed in responsive
notypes have been reported (Ihle, 2001). progenitor cells, and mice lacking STAT3 in their
hema-topoietic precursors developed neutrophilia, as did
wild-type mice transplanted with STAT3 null progenitors. STAT3 as a Negative Regulator of G-CSF-Mediated Granulopoiesis Loss of STAT3 affected neither the frequency of
granulo-cytic precursors nor their ability to differentiate into ma- The enhanced production of neutrophils in the absence of STAT3 is reminiscent of the phenotype derived from ture, functional effector cells. STAT3 null granulocytes
displayed a respiratory oxidative burst, increased ex- the G-CSFR mutant found in human severe congenital neutropenia (Hermans et al., 1998; McLemore et al., pression of surface markers, and the ability to
phagocy-tose and kill bacteria. Therefore, STAT3 is dispensable 1998). The hypergranulopoiesis of mutant mice bearing this receptor correlated with prolonged activation of for granulocyte proliferation, differentiation, and
func-tion but is required for homeostasis. STAT1, STAT3, and STAT5 in response to G-CSF treat-ment (Hermans et al., 1999). Interestingly, prolonged Several possibilities may account for the discrepancy
between functional impairments of STAT3 studied pre- activation of STAT1 though not STAT5 was also ob-served in the absence of STAT3 following G-CSF stimu-viously and the absence of STAT3 studied here. A trivial
explanation could be that our mice express a low but lation. However, a requirement for STAT1 in G-CSF-mediated granulopoiesis is unlikely, given the normal sufficient level of STAT3. Although the deletion of STAT3
is not absolute in Mx-Cre mice, several lines of evidence proliferation, accumulation, and function of granulo-cytes in STAT1-deficient mice (Durbin et al., 1996; Meraz preclude the possibility that mature neutrophils are
de-rived from residual STAT3-containing precursors. First, et al., 1996). Nonetheless, a critical interaction between STAT1 and STAT3 in G-CSF-dependent granulopoiesis little residual wild-type STAT3 allele was detected in
mutant bone marrow (Figure 1A), and individual G-CSF- cannot be excluded.
More likely, however, is that enhanced proliferation derived colonies from mutant progenitors demonstrated
complete absence of STAT3 (Figure 3C). Similarly, pro- of granulocytes in response to G-CSF and neutrophilia observed in mutant mice resulted from loss of negative tein immunoblots and DNA binding assays consistently
regulatory signals and that prolonged STAT1 activation devoid of intracellular tyrosines (Ward et al., 1999). In contrast, activation of STAT3 by the truncated d715 is a symptom of that impaired negative feedback. The
severe impairment of SOCS3 induction is consistent receptor is completely blocked by loss of the single remaining tyrosine, even at high G-CSF concentrations. with this notion, since SOCS3 as an inducible
suppres-sor of cytokine signaling is normally responsible for inhi- These results suggest that emergency granulopoiesis in response to high levels of G-CSF may be accomplished bition of receptor-associated JAK catalytic function. The
promoter of SOCS3 contains a functional STAT3 binding through an independent signaling pathway mediated by a non-phosphotyrosine-dependent mechanism involv-element that is critical for LIF-mediated induction
(Auernhammer et al., 1999). SOCS3 also negatively regu- ing the distal region of the G-CSFR. While STAT3 activa-tion is a marker for both receptor tyrosine-dependent lates IL-6 signaling, and antagonism of SOCS3 function
can lead to development of colitis, due at least in part (low dose G-CSF) and receptor tyrosine-independent (high dose G-CSF) signaling pathways, it appears to be to hyperactivity of cytokine signals that target STAT3
(Suzuki et al., 2001). dispensable for both mechanisms of granulopoiesis. In fact, mice lacking STAT3 in the bone marrow develop It is interesting that a constitutively active form of
STAT3 partially rescued the ability of progenitor cells severe neutrophilia in response to bacterial infection (data not shown).
from d715F mice to differentiate into neutrophils in
re-sponse to G-CSF (McLemore et al., 2001). These data Emergency granulopoiesis can also be independent of G-CSF, presumably through the action of other hema-could be interpreted as showing that STAT3 is capable
of driving G-CSF-dependent differentiation, even though topoietic cytokines. For instance, G-CSF-deficient mice develop neutrophilia during Candida infections (Basu et STAT3 deletion showed that it is not necessary. It must
be borne in mind, however, that the constitutively active al., 2000). Although IL-6 is an independent regulator of granulopoiesis and is increased in the serum of Can-STAT3 protein also displays oncogenic properties
(Bromberg et al., 1999) and therefore cannot be consid- dida-infected, G-CSF-deficient mice, this cytokine was
also dispensable for the observed neutrophilia since ered a faithful mimic of normal STAT3 function. In fact,
expression of constitutively active STAT3 in bone mar- neutrophilia developed in infected mice deficient for both IL-6 and G-CSF (Basu et al., 2000). Although both row progenitors led to increased cell numbers as well
as increased differentiation (McLemore et al., 2001). In IL-6 and G-CSF are efficient activators of STAT3, it is intriguing that neutrophilia can develop in the absence a similar fashion, constitutively active STAT5 is capable
of causing myeloproliferative disease (Schwaller et al., of all three of these molecules. It would be interesting if emergency, G-CSF-independent granulopoiesis was 2000), even though it is not required for normal
myelo-proliferation or differentiation (Teglund et al., 1998). It is somehow caused by antagonism of the normally inhibi-tory function of STAT3, such as inhibition of SOCS pro-possible that STAT3 will be found to contribute to
my-eloid oncogenesis in spite of its dispensability for normal tein expression. granulopoiesis.
Experimental Procedures Due to alternative splicing, STAT3 is expressed as
two isoforms, STAT3␣ and STAT3 (Schaefer et al.,
Animals
1997), both of which were functionally ablated by our Generation of a conditional STAT3 allele in ES cells has been pre-targeting strategy. STAT3 lacks a transactivation do- viously described (Raz et al., 1999). Mice were generated by injection main and antagonizes the activity of STAT3␣. Indeed, of heterozygous targeted ES cells into blastocysts at the transgenic mouse facility of Albert Einstein College of Medicine by standard recent gene targeting of the STAT3 locus documented
methods. Interbreeding with transgenic Mx-Cre mice (Kuhn et al., a critical role for this isoform in recovery from endotoxic
1995) generated littermate mice that were homozygous for the con-shock (Yoo et al., 2002), demonstrating its role as a
ditional STAT3 allele (STAT3f/f) with or without the Mx-Cre transgene.
negative regulator of gene expression. Therefore, it is Induction of Cre and subsequent deletion of STAT3 was accom-possible that lack of STAT3 function contributes to the plished by two successive intraperitoneal injections separated by loss of negative regulation responsible for the observed 7 days with 100g poly(I:C) per mouse (Radtke et al., 1999) unless otherwise specified and comparisons were made between Mx-Cre-neutrophilia in STAT3 null mice. If so, STAT3 would be
positive and Mx-Cre-negative STAT3f/flittermates that had been
a more general negative regulator than originally
sug-similarly treated with poly(I:C). Neomycin (2 mg/ml) was added to gested, inhibiting gene expression beyond inhibition of
drinking water to eliminate bacterial infection and intestinal flora.
STAT3␣. All work with animals conformed to guidelines approved by the
Institutional Animal Care and Use Committee of the New York Uni-versity School of Medicine.
STAT3 and Emergency Granulopoiesis
Under certain stressful situations, such as response to
Hematology pathogen infection, rapid expansion of neutrophil
popu-Peripheral blood or in vitro differentiated bone marrow (BM) was lations can be achieved through a process known as
prepared by blood smear or cytospin centrifugation and stained emergency granulopoiesis that is accompanied by high with Wright-Giemsa. Differential cell counts were scored visually on levels of G-CSF (Basu et al., 2000; Lieschke et al., 1994; coded samples. For in vitro differentiation, BM collected from femurs was cultured in RPMI-1640 medium containing 20% fetal bovine Ward et al., 1999). Emergency granulopoiesis
demon-serum and 10 ng/ml recombinant murine G-CSF (PeproTech). strates some of the complexity of molecular responses
to G-CSF. For instance, STAT3 activation by G-CSFR
Autoreconstitution and Colony Formation can be achieved through both receptor
tyrosine-depen-Age 6- to 8-week-old mice (five mice per group) were injected with dent and -independent mechanisms, depending on 100g poly(I:C) on days 1 and 7. 5-Fluorouracil (300 mg/kg) was G-CSF concentration. At saturating concentrations (100 injected on day 10 followed by subcutaneous injection of BSA or G-CSF (10g/kg) daily for 7 days. BM cells were prepared 4 hr after ng/ml), STAT3 can be activated by a full-length G-CSFR
the last injection. For colony formation assays, 4⫻ 104 BM cells Phagocytosis and Bactericidal Assays
Peritoneal granulocytes were harvested by lavage from thioglycol-were incubated in complete medium (MethoCult M3434, Stem Cell
Technology) containing 0.9% methylcellulose in Iscove’s modified late-treated mice, washed, and resuspended in medium containing 5% FCS and lacking antibiotics. Listeria monocytogenes labeled Dulbecco’s media, 1% BSA, 10⫺4 M 2-mercaptoethanol, 2 mM
L-glutamine, 15% fetal bovine serum, 10g/ml bovine pancreatic with 5M CFSE was mixed with 2 ⫻ 106peritoneal exudate cells
and incubated at 37⬚C for 30 min. Subsequently, extracellular bacte-insulin, 200g/ml human transferrin, 10 ng/ml recombinant murine
IL-3, 10 ng/ml recombinant human IL-6, 50 ng/ml recombinant mu- ria were killed by incubation with 50g/ml gentamycin. Phagocytic activity was assessed by FACS on paraformaldehyde-fixed cells. rine stem cell factor, and 3 U/ml recombinant human EPO. For
granulocyte-specific colonies, 2⫻ 105BM cells were incubated in To measure bactericidal activity, engulfed bacteria were released
by lysing eukaryotic cells in water, and surviving bacteria were enu-medium (MethoCult M3234) containing the same components
ex-cept with 1, 10, or 100 ng/ml of recombinant murine G-CSF as the merated by serial dilution. only cytokine, followed by incubation for 7–10 days.
Acknowledgments
Thioglycollate Challenge and Respiratory Burst Assay
We thank Ross Basch for invaluable discussions and advice on CFU Age-matched STAT3f/fand Mx-Cre:STAT3f/fmice were treated with
assays, Giorgio Inghirami, Alan Frey, Joellen Shaw, and Jack Hessler poly(I:C) followed by i.p. injection of 2 ml 3% TGA. Peritoneal
exu-for helpful discussions, and John Hirst exu-for FACS analysis. We are date cells were harvested after 4 hr by PBS lavage. Granulocytes
grateful to Michel Aguet for the gift of Mx-Cre mice. R.A.D. is an were enumerated following staining with CD11b-FITC and
anti-American Cancer Society Research Professor. This work was sup-Gr-1-PE (Caltag) by FACS analysis. Dihydrorhodamine 123 (DHR)
ported by grants from the NIH and the AHA. oxidation assay was performed as described (Vowells et al., 1995).
In brief, 1⫻ 106cells were incubated with 100M DHR and 1000
Received: December 6, 2001 U/ml catalase in 100l of medium at 37⬚C for 5 min. One hundred
Revised: June 7, 2002 microliters of medium containing 200 ng PMA was added and
incu-bated for another 20 min. Cells were washed with cold PBS twice,
References resuspended in PBS containing 1% paraformaldehyde, and
ana-lyzed by FACS.
Auernhammer, C.J., Bousquet, C., and Melmed, S. (1999). Autoregu-lation of pituitary corticotroph SOCS-3 expression: characterization of the murine SOCS-3 promoter. Proc. Natl. Acad. Sci. USA 96, BM Transplantation
6964–6969. 4⫻ 106STAT3f/fBM cells were injected intravenously into lethally
irradiated (950 rad) wild-type recipients. Five weeks later, peritoneal Basu, S., Hodgson, G., Zhang, H.H., Katz, M., Quilici, C., and Dunn, exudate and BM cells were prepared for FACS analysis, Wright- A.R. (2000). “Emergency” granulopoiesis in G-CSF-deficient mice Giemsa staining, genomic PCR, and Western blotting. in response to Candida albicans infection. Blood 95, 3725–3733.
Boeuf, H., Hauss, C., Graeve, F.D., Baran, N., and Kedinger, C. (1997). Leukemia inhibitory factor-dependent transcriptional activa-PCR and RT-activa-PCR Assays
tion in embryonic stem cells. J. Cell Biol. 138, 1207–1217. Mice, tissues, and cells were genotyped by PCR, using primer sets
that distinguished STAT3, STAT3-flox, and STAT3-deleted alleles, Bromberg, J.F., Wrzeszczynska, M.H., Devgan, G., Zhao, Y., Pestell, as previously described (Raz et al., 1999). To genotype single hema- R.G., Albanese, C., and Darnell, J.E., Jr. (1999). Stat3 as an onco-topoietic colonies grown in methylcellulose, colonies were resus- gene. Cell 98, 295–303.
pended in 5l of buffer containing 50 mM KCl, 10 mM Tris-HCL Coffer, P.J., Koenderman, L., and de Groot, R.P. (2000). The role (pH 8.3), 2.5 mM MgCl2, 0.1 mg/ml gelatin, 0.45% NP-40, 0.45% of STATs in myeloid differentiation and leukemia. Oncogene 19, Tween 20, and 200g/ml proteinase K; incubated at 55⬚C for 2 hr 2511–2522.
and then at 95⬚C for 10 min; and subjected to 40 cycles of PCR.
de Alboran, I.M., O’Hagan, R.C., Gartner, F., Malynn, B., Davidson, RT-PCR was performed as previously described (Lee et al., 2000)
L., Rickert, R., Rajewsky, K., DePinho, R.A., and Alt, F.W. (2001). using total RNA prepared from four to six million cells. Cycle number
Analysis of C-MYC function in normal cells via conditional gene-and template dilutions were empirically chosen to insure that
reac-targeted mutation. Immunity 14, 45–55. tions remained within a linear range.
Durbin, J.E., Hackenmiller, R., Simon, M.C., and Levy, D.E. (1996). Sequences of PCR primers are available on request.
Targeted disruption of the mouse Stat1 gene results in compromised innate immunity to viral disease. Cell 84, 443–450.
Proliferation Assay
Endo, T.A., Masuhara, M., Yokouchi, M., Suzuki, R., Sakamoto, H., BM cells were seeded at 2⫻ 105cells per well and incubated in
Mitsui, K., Matsumoto, A., Tanimura, S., Ohtsubo, M., Misawa, H., the absence or presence of 1, 10, and 100 ng/ml recombinant murine
et al. (1997). A new protein containing an SH2 domain that inhibits G-CSF or 1 and 10 ng/ml recombinant murine IL-3. 0.5Ci3
H-JAK kinases. Nature 387, 921–924. thymidine was added after 48 hr and incubated for a further 16 hr
Hermans, M.H., Ward, A.C., Antonissen, C., Karis, A., Lowenberg, prior to harvest. To enrich immature granulocyte precursors, BM
B., and Touw, I.P. (1998). Perturbed granulopoiesis in mice with was fractionated by Percoll step-gradient ( ⫽ 1.065 and ⫽ 1.08),
a targeted mutation in the granulocyte colony-stimulating factor and cells located at the 1.065 interface were recovered for FACS
receptor gene associated with severe chronic neutropenia. Blood and proliferation analysis by3H-thymidine and BrdU incorporation.
92, 32–39.
For BrdU incorporation, cells treated with G-CSF for 24 hr were
Hermans, M.H., Antonissen, C., Ward, A.C., Mayen, A.E., Ploe-incubated with 10M BrdU for 1 hr, followed by FACS staining for
macher, R.E., and Touw, I.P. (1999). Sustained receptor activation Gr-1 and BrdU. Stimulation index was calculated as the ratio of
and hyperproliferation in response to granulocyte colony-stimulat-Gr-1⫹/BrdU⫹cells with and without G-CSF.
ing factor (G-CSF) in mice with a severe congenital neutropenia/ acute myeloid leukemia-derived mutation in the G-CSF receptor Electrophoretic Mobility Shift Assay (EMSA) and Western Blots gene. J. Exp. Med. 189, 683–692.
EMSA and Western blots were performed as described (Lee et al.,
Hibi, M., and Hirano, T. (2000). Gab-family adapter molecules in 1999). Nuclear extracts or total cell extracts were prepared from five
signal transduction of cytokine and growth factor receptors, and T million BM cells following treatment with 10 ng/ml G-CSF. Nuclear
and B cell antigen receptors. Leuk. Lymphoma 37, 299–307. extracts were tested by EMSA using a radiolabeled GAS probe in
Ihle, J.N. (1996). STATs and MAPKs: obligate or opportunistic part-the absence or presence of anti-STAT1 or anti-STAT3 antibodies
ners in signaling. Bioessays 18, 95–98. (Zymed). Total cell extracts were examined by Western blot using
anti-STAT and anti-phospho-STAT antibodies (Zymed and Cell Sig- Ihle, J.N. (2001). The Stat family in cytokine signaling. Curr. Opin. Cell Biol. 13, 211–217.
Ihle, J.N., Witthuhn, B.A., Quelle, F.W., Yamamoto, K., and Silven- Stat5 is essential for the myelo- and lymphoproliferative disease induced by TEL/JAK2. Mol. Cell 6, 693–704.
noinen, O. (1995). Signaling through the hematopoietic cytokine
re-ceptors. Annu. Rev. Immunol. 13, 369–398. Silver, D.P., and Livingston, D.M. (2001). Self-excising retroviral vec-tors encoding the Cre recombinase overcome Cre-mediated cellular Krebs, D.L., and Hilton, D.J. (2001). SOCS proteins: negative
regula-toxicity. Mol. Cell 8, 233–243. tors of cytokine signaling. Stem Cells 19, 378–387.
Smithgall, T.E., Briggs, S.D., Schreiner, S., Lerner, E.C., Cheng, H., Kuhn, R., Schwenk, F., Aguet, M., and Rajewsky, K. (1995). Inducible
and Wilson, M.B. (2000). Control of myeloid differentiation and sur-gene targeting in mice. Science 269, 1427–1429.
vival by Stats. Oncogene 19, 2612–2618. Lee, C.K., Gimeno, R., and Levy, D.E. (1999). Differential regulation
Starr, R., Willson, T.A., Viney, E.M., Murray, L.J., Rayner, J.R., Jen-of constitutive major histocompatibility complex class I expression
kins, B.J., Gonda, T.J., Alexander, W.S., Metcalf, D., Nicola, N.A., in T and B lymphocytes. J. Exp. Med. 190, 1451–1464.
and Hilton, D.J. (1997). A family of cytokine-inducible inhibitors of Lee, C.K., Rao, D.T., Gertner, R., Gimeno, R., Frey, A.B., and Levy,
signalling. Nature 387, 917–921. D.E. (2000). Distinct requirements for IFNs and STAT1 in NK cell
Suzuki, A., Hanada, T., Mitsuyama, K., Yoshida, T., Kamizono, S., function. J. Immunol. 165, 3571–3577.
Hoshino, T., Kubo, M., Yamashita, A., Okabe, M., Takeda, K., et al. Lieschke, G.J., Grail, D., Hodgson, G., Metcalf, D., Stanley, E.,
(2001). CIS3/SOCS3/SSI3 plays a negative regulatory role in STAT3 Cheers, C., Fowler, K.J., Basu, S., Zhan, Y.F., and Dunn, A.R. (1994).
activation and intestinal inflammation. J. Exp. Med. 193, 471–481. Mice lacking granulocyte colony-stimulating factor have chronic
Takeda, K., and Akira, S. (2000). STAT family of transcription factors neutropenia, granulocyte and macrophage progenitor cell
defi-in cytokdefi-ine-mediated biological responses. Cytokdefi-ine Growth Factor ciency, and impaired neutrophil mobilization. Blood 84, 1737–1746.
Rev. 11, 199–207. Liu, F., Wu, H.Y., Wesselschmidt, R., Kornaga, T., and Link, D.C.
Takeda, K., Noguchi, K., Shi, W., Tanaka, T., Matsumoto, M., Yoshi-(1996). Impaired production and increased apoptosis of neutrophils
da, N., Kishimoto, T., and Akira, S. (1997). Targeted disruption of in granulocyte colony-stimulating factor receptor-deficient mice.
Im-the mouse Stat3 gene leads to early embryonic lethality. Proc. Natl. munity 5, 491–501.
Acad. Sci. USA 94, 3801–3804. Liu, F., Poursine-Laurent, J., Wu, H.Y., and Link, D.C. (1997).
Interleu-Takeda, K., Clausen, B.E., Kaisho, T., Tsujimura, T., Terada, N., kin-6 and the granulocyte colony-stimulating factor receptor are
Forster, I., and Akira, S. (1999). Enhanced Th1 activity and develop-major independent regulators of granulopoiesis in vivo but are not
ment of chronic enterocolitis in mice devoid of Stat3 in macrophages required for lineage commitment or terminal differentiation. Blood
and neutrophils. Immunity 10, 39–49.
90, 2583–2590.
Teglund, S., McKay, C., Schuetz, E., van Deursen, J.M., Stravopodis, Matsuda, T., Nakamura, T., Nakao, K., Arai, T., Katsuki, M., Heike,
D., Wang, D., Brown, M., Bodner, S., Grosveld, G., and Ihle, J.N. T., and Yokota, T. (1999). STAT3 activation is sufficient to maintain
(1998). Stat5a and Stat5b proteins have essential and nonessential, an undifferentiated state of mouse embryonic stem cells. EMBO J.
or redundant, roles in cytokine responses. Cell 93, 841–850.
18, 4261–4269.
Tidow, N., and Welte, K. (1997). Advances in understanding postre-McLemore, M.L., Poursine-Laurent, J., and Link, D.C. (1998).
In-ceptor signaling in response to granulocyte colony-stimulating fac-creased granulocyte colony-stimulating factor responsiveness but
tor. Curr. Opin. Hematol. 4, 171–175. normal resting granulopoiesis in mice carrying a targeted
granulo-Vowells, S.J., Sekhsaria, S., Malech, H.L., Shalit, M., and Fleisher, cyte colony-stimulating factor receptor mutation derived from a
pa-T.A. (1995). Flow cytometric analysis of the granulocyte respiratory tient with severe congenital neutropenia. J. Clin. Invest. 102,
burst: a comparison study of fluorescent probes. J. Immunol. Meth-483–492.
ods 178, 89–97. McLemore, M.L., Grewal, S., Liu, F., Archambault, A.,
Poursine-Ward, A.C., Hermans, M.H., Smith, L., van Aesch, Y.M., Schelen, A.M., Laurent, J., Haug, J., and Link, D.C. (2001). STAT-3 activation is
Antonissen, C., and Touw, I.P. (1999). Tyrosine-dependent and required for normal G-CSF-dependent proliferation and granulocytic
-independent mechanisms of STAT3 activation by the human granu-differentiation. Immunity 14, 193–204.
locyte colony-stimulating factor (G-CSF) receptor are differentially Meraz, M.A., White, M.J., Sheehan, K.C.F., Bach, E.A., Rodig, S.J.,
utilized depending on G-CSF concentration. Blood 93, 113–124. Dighe, A.S., Kaplan, D.H., Riely, J.K., Greenlund, A.C., Campbell,
Yoo, J.Y., Huso, D.L., Nathans, D., and Desiderio, S. (2002). Specific D., et al. (1996). Targeted disruption of the Stat1 gene in mice reveals
ablation of Stat3beta distorts the pattern of Stat3-responsive gene unexpected physiologic specificity in the JAK-STAT signaling
path-expression and impairs recovery from endotoxic shock. Cell 108, way. Cell 84, 431–442.
331–344. Naka, T., Narazaki, M., Hirata, M., Matsumoto, T., Minamoto, S.,
Zou, Y.R., Muller, W., Gu, H., and Rajewsky, K. (1994). Cre-loxP-Aono, A., Nishimoto, N., Kajita, T., Taga, T., Yoshizaki, K., et al.
mediated gene replacement: a mouse strain producing humanized (1997). Structure and function of a new STAT-induced STAT
inhibi-antibodies. Curr. Biol. 4, 1099–1103. tor. Nature 387, 924–929.
Niwa, H., Burdon, T., Chambers, I., and Smith, A. (1998). Self-renewal of pluripotent embryonic stem cells is mediated via activation of STAT3. Genes Dev. 12, 2048–2060.
Radtke, F., Wilson, A., Stark, G., Bauer, M., van Meerwijk, J., Mac-Donald, H.R., and Aguet, M. (1999). Deficient T cell fate specification in mice with an induced inactivation of Notch1. Immunity 10, 547–558.
Raz, R., Lee, C.K., Cannizzaro, L.A., d’Eustachio, P., and Levy, D.E. (1999). Essential role of STAT3 for embryonic stem cell pluripotency. Proc. Natl. Acad. Sci. USA 96, 2846–2851.
Schaefer, T.S., Sanders, L.K., Park, O.K., and Nathans, D. (1997). Functional differences between Stat3alpha and Stat3beta. Mol. Cell. Biol. 17, 5307–5316.
Schmitz, J., Dahmen, H., Grimm, C., Gendo, C., Muller-Newen, G., Heinrich, P.C., and Schaper, F. (2000). The cytoplasmic tyrosine motifs in full-length glycoprotein 130 have different roles in IL-6 signal transduction. J. Immunol. 164, 848–854.
Schwaller, J., Parganas, E., Wang, D., Cain, D., Aster, J.C., Williams, I.R., Lee, C.K., Gerthner, R., Kitamura, T., Frantsve, J., et al. (2000).