Increased Myeloproliferation in Glutathione S-Transferase π-deficient Mice Is Associated with a Deregulation of JNK and Janus Kinase/STAT Pathways*

It has been shown that glutathione S-transferase π (GSTπ) interacts with and suppresses the activity of c-Jun NH2-terminal kinase (JNK). GST-deficient mice (GSTπ–/–) have higher levels of circulating white blood cells, with similar proportions of lymphocytes, monocytes, and granulocytes. Interestingly, a selective expansion of splenic B lymphocytes was observed in GSTπ–/– animals but no change in T lymphocytes or natural killer cells. A peptidomimetic inhibitor of GSTπ that disrupts the interaction between GSTπ and JNK mimics in wild type mice the increased myeloproliferation observed in GSTπ–/– animals. Until now, the molecular basis for this effect has not been defined. In an in vitro hematopoiesis assay, interleukin-3, granulocyte colony-stimulating factor, and granulocyte/macrophage colony-stimulating factor were more effective at stimulating proliferation of hematopoietic cells in GSTπ–/– mice than in wild type. The JNK inhibitor SP600125 which caused little inhibition of cytokine-induced myeloproliferation in wild type mice, decreased the number of colonies in GSTπ–/– animals. A more sustained phosphorylation of the STAT family of proteins was also observed in GSTπ–/– bone marrow-derived mast cells exposed to interleukin-3. This was associated with an increased proliferation and a down-regulation of expression of negative regulators of the Janus kinase-STAT pathway SHP, Src homology 2 domain-containing tyrosine phosphatase-1 and -2. The increased activation of JNK and STATs in GSTπ-deficient mice provides a viable mechanism for the increased myeloproliferation in these animals. These data also confirm the important role that GSTπ plays in the regulation of cell signaling pathways in a myeloproliferative setting.

Glutathione S-transferases (GSTs) 1 are expressed ubiquitously in plants and animals and have diverse roles in the conjugation of glutathione to electrophilic species. In mammals, six different isoforms ␣, , , , , and have been identified (1). GST overexpression has been observed in many tumors compared with the surrounding normal tissues and in various cancer cell lines resistant to anticancer agents (2). However, the precise role that this enzyme plays in resistance to anticancer drugs remains ill defined, particularly because GST transfection does not always confer resistance to chemotherapeutic agents (3). More recently, it has been shown that GST acts as a regulator of mitogen-activated protein (MAP) kinases. GST is an endogenous inhibitor of c-Jun NH 2 -terminal kinase (JNK), mediated by interactions with the NH 2terminal region of the kinase (4,5). After oxidative stress, GST oligomerizes and disassociates from JNK, which then becomes phosphorylated (4). GST can also modulate the activation of p38 and extracellular signal-regulated kinase (ERK). For example, in NIH 3T3 cells, the forced expression of GST inhibits JNK activity and activates ERK and p38 kinase (6). In addition, in GST-deficient mouse embryo fibroblasts, the basal activities of JNK and ERK are higher than in their wild type counterparts (4,7).
The principle that a therapeutic advantage might be gained by inhibiting GST led to the design and synthesis of a peptidomimetic inhibitor ␥-glutamyl-S-(benzyl)cysteinyl-R(-)-phenylglycine diethyl ester (TLK199) (8). Although the drug was active in modulating resistance associated with GST overexpression both in vitro and in vivo (9), a distinct pharmacological property emerged, namely, GST-dependent regulation of myeloproliferation (7). The intraperitoneal injection of this drug increased the number of peripheral white blood cells in wild type mice but not in GST-null animals. Furthermore, GSTdeficient mice presented a higher number of circulating leukocytes than wild type animals. The production of fully differentiated peripheral blood cells from pluripotent cells is tightly regulated by cytokines produced by the surrounding stroma and by the hematopoietic cells themselves (10). The signal transduction pathways mediated by cytokines have been studied extensively. Although each cytokine has its own mechanism of action, canonical pathways have been described (11). The binding of a cytokine to its receptor leads to its dimerization and to the recruitment of Janus kinases (JAKs). These kinases phosphorylate each other and subsequently phosphorylate the receptor on tyrosine residues, allowing the binding of various proteins containing an SH2 domain. The binding of signal transducer and activator of transcription (STAT) molecules to the receptor induces their phosphorylation and leads to dimers that are translocated into the nucleus where they bind to their specific DNA binding site. Src protein family members also bind to the cytokine receptor, leading to their phosphorylation and subsequent activation of Ras, which then activates JNK, ERK, and p38 MAP kinase pathways. The activation of the JNK pathway by cytokines such as interleukin-3 (IL-3; 12, 13), granulocyte/macrophage colony-stimulating factor (GM-CSF; 12), and granulocyte colony-stimulating factor (G-CSF; 13,14) has been described previously. However, the precise role(s) of these kinases in the proliferation of hematopoietic cells is still unclear. Because GST is a natural inhibitor of JNK and a modulator of MAP kinase activities and the lack of GST expression is associated with an increase in myeloproliferation, we set out to define a mechanistic link between myeloproliferation and GST inhibition.
Cell Lines-Bone marrow-derived mast cells (BMMC) were obtained from the bone marrow of C57BL/6-129/Sv wild type and GST Ϫ/Ϫ mice and grown in 50% WEHI-3 cell (ATCC) conditioned medium and 50% RPMI 1640 culture medium supplemented with 10% fetal bovine serum, 50 M 2-mercaptoethanol, 2 mM L-glutamine, 10 g/ml streptomycin, and 10 units/ml penicillin. Cells in suspension were resuspended in fresh medium twice a week. After 4 weeks in culture, more than 95% of cells were BMMC. BMMC purity was determined by toluidine blue staining and was confirmed by flow cytometry by analyzing the expression of c-Kit (SCF-R) and IgE receptor.
Hematology Studies-All animal experiments were approved by the Institutional Animal Care and Use Committee of Fox Chase Cancer Center. Wild type and GST-deficient mice, kindly provided by Drs. C. J. Henderson and R. Wolf (Cancer Research UK, Molecular Pharmacology Unit, Dundee, UK), were housed in conventional conditions in the Fox Chase Cancer Center laboratory animal facility. 6-12-week-old wild type and GST Ϫ/Ϫ male mice were sacrificed by CO 2 asphyxiation. Blood was collected by intracardial puncture in heparin tubes. Platelets and red and white blood cells were counted using a hemocytometer. Leukocyte composition was determined by staining blood smears with a KWIK DIFF TM staining kit from Shandon (Pittsburgh, PA). Spleen and thymus were collected, and cells were extracted and labeled with specific antibodies against mouse surface markers B220, IgM, CD3⑀, DX5, CD8, and CD4 (BD Biosciences). Bone marrow cells were flushed out of the femoral bone, and red blood cells were lysed with 0.8% ammonium chloride. Nucleated cells were counted with a hemocytometer and labeled with specific antibodies against mouse surface markers B220, CD43, CD11b (BD Biosciences), Ter119, and Gr1 (E-Bioscience). Surface marker expression was determined by flow cytometry using a FacsVantage or a Facscan flow cytometer (BD Biosciences). Results were analyzed using FlowJo software.
In Vitro Hematopoiesis Assay-Bone marrow cells were collected as described previously (7). Briefly wild type and GST Ϫ/Ϫ mice 6 -12 weeks old were sacrificed with CO 2 , the pelt clipped, and hind limbs exposed. Both femurs and tibiae were removed and cells collected in Iscove's modified Dulbecco's medium containing 2% fetal bovine serum. Red blood cells were lysed with 0.8% NH 4 Cl. Cells (5 ϫ 10 4 cells/ml) were plated in Methocult TM medium containing 1% methylcellulose in Iscove's modified Dulbecco's medium, 20% fetal bovine serum, 1% bovine serum albumin (Stemcell Technologies, Vancouver, BC), 2 mM glutamine into 35-mm culture dishes, treated with cytokines (10 ng/ml G-CSF, GM-CSF, or IL-3) in combination with TLK199, MAP kinase inhibitors or vehicle, and incubated at 37°C in 5% CO 2 . Colony-forming units were scored (colony Ͼ50 cells) 7 days after plating and expressed as a percentage of control. For colony composition and cell counts, cells were washed extensively with ice-cold phosphate-buffered saline to remove methyl cellulose, slides were prepared by cytospin, and cells were stained using a KWIK DIFF TM staining kit.
Proliferation Assay-2 ϫ 10 5 BMMC/ml were seeded in complete medium without IL-3 for 12 h. IL-3 was then added at various concentrations for 24 h. 1 Ci/ml [methyl-3 H]thymidine (ICN, Irvine, CA) was added to the cells for an additional 12 h. Cells were collected and washed three times with an ice-cold 5% trichloroacetic acid solution, and final pellets were lysed with 0.1 N NaOH. Radiolabeled thymidine incorporation was quantified with a scintillation counter (Beckman LS6500).
Immunoblot-After treatments, cells were collected and washed twice in ice-cold phosphate-buffered saline. The final pellets were resuspended in lysis buffer (20 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM ␤-glycerophosphate, 1 mM Na 2 VO 4 ) and incubated for 30 min on ice. Lysates were sonicated 3 ϫ 10 s and centrifuged for 30 min at 10,000 ϫ g. The resulting supernatants were considered whole cell extracts. Protein concentrations were determined using Bio-Rad DC protein assay. Equal amounts of protein were separated on SDS-polyacrylamide gels and transferred overnight onto polyvinylidene difluoride membranes (PerkinElmer Life Sciences). Protein expression was determined using specific primary and secondary antibodies. Briefly, after transfer, membranes were incubated in TBS-Tween containing 5% skimmed milk, washed with TBS-Tween, and incubated with primary antibody in TBS-Tween containing 5% skimmed milk for 1 h. Membranes were washed twice with TBS-Tween and incubated with horseradish peroxidase-conjugated secondary antibody in TBS-Tween containing 5% skimmed milk for 1 h and finally washed three times with TBS-Tween. When phospho-specific antibodies were used, 5% skimmed milk was replaced by 10% IgG-free bovine serum albumin (Jackson Immunoresearch). Specific proteins were revealed by chemiluminescence using ECL or ECL Plus Western blotting reagents from Amersham Biosciences.
Reverse Transcription and Real Time Quantitative PCR-BMMC were incubated for 4 h in the absence of IL-3 and then exposed to this cytokine (100 ng/ml). At various time points, cells were collected and washed twice with ice-cold phosphate-buffered saline and total RNA extracted with RNeasy mini kit (Qiagen, Carlsbad, CA) as described by the manufacturer. RNA was reverse transcribed as follows. 1 g of RNA was mixed with 1 l of 0.5 g/l random hexamer and incubated for 10 min at 70°C. The sample was mixed with 4 l of first strand buffer, 2 l of 0.1 M dithiothreitol, 1 l of dNTP mix, and 1 l of Superscript II (Invitrogen) and incubated 1 h at 42°C and 15 min at 70°C. 50 ng of cDNA was used for PCR amplification, and gene expression was determined using a Cepheid Smartcycler and a Quantitect SYBR Green kit (Qiagen). Relative amounts of each gene were determined by the comparative C T method using the expression of each gene in the untreated wild type sample as the calibrator (15). The quantification of expression for each sample was calculated from the threshold cycle (C T ) value, which is the number of cycles for which an increase in PCR product is first detected at a statistically significant level. The relative expression value of each gene is obtained by evaluating the C T values for the unknown reaction using the equation 2 ⌬⌬CT . The ⌬C T values for the calibrator and samples in each gene expression assay were obtained by subtracting the C T value of the actin from the C T value for the gene. ⌬⌬C T was calculated by subtracting the average ⌬C T (calibrator) values from the ⌬C T (sample). The relative quantification was calculated by 2 ⌬⌬CT . The mRNA quantity of the calibrator (untreated wild type) is expressed as 1, and all other quantities are expressed as a -fold difference relative to the calibrator. The primers used for this experiment were: 5Ј-GTCGCCAACGGAACTGCT-3Ј and 5Ј-GGAAGGGGAAGGAA-CTCA-3Ј for SOCS-1; 5Ј-GCTGGACCGACTAACCTG-3Ј and 5Ј-CTAC-ATCTTCACACCTTT-3Ј for SOCS-2; 5Ј-TTTGCCACCCACGGAACC-3Ј and 5Ј-CCCCCTCTGACCCTTTTG-3Ј for SOCS-3; 5Ј-TCCAGGCAGAG-AATGAAC-3Ј and 5Ј-GATAGGCAGCACCGACTC-3Ј for CIS-1; 5Ј-TA-GAAGAATGCTCACAGA-3Ј and 5Ј-GTTTAGATACAGGCTCAG-3Ј for CD45; 5Ј-GTGGGCCGCTCTTAGGCACC-3Ј and 5Ј-CTCTTTGATGTC-ACGCACGAT-3Ј for actin.

Increased Number of White Blood Cells in GST Ϫ/Ϫ Mice-
We observed that GST Ϫ/Ϫ mice had about twice the number of circulating white blood cells compared with wild type animals (16.1 Ϯ 2.0 versus 9.6 Ϯ 1.3 ϫ 10 6 /ml, respectively) ( Table I). The percentage of lymphocytes, monocytes, and granulocytes in the peripheral blood of both mouse strains was similar. In contrast, the numbers of red blood cells and platelets were not increased significantly in GST Ϫ/Ϫ mice. Spleens from GST Ϫ/Ϫ mice also contained significantly more cells than wild type (74.7 Ϯ 5.3 versus 46.1 Ϯ 9.4 ϫ 10 6 cells/spleen, respectively) (Table II). Both the number and percentage of splenic B lymphocytes (B220 ϩ /IgM ϩ ) were significantly higher in knockout animals compared with wild type. In contrast, the percentage of T lymphocytes (CD3⑀ ϩ ) was lower in GST Ϫ/Ϫ mice, the cell counts were similar in both animals. In addition, there was no difference in number or proportion of splenic natural killer cells (DX5 ϩ ). Analysis of thymus cells showed that cell counts and thymocyte subsets were similar in both wild type and GST Ϫ/Ϫ mice (Table II), which is consistent with the lack of an impact on mature peripheral T cells. Therefore, our analysis indicates a selective expansion of B cells within the lymphocyte population in GST Ϫ/Ϫ mice.
Both the total number of nucleated bone marrow cells and the percentages of erythroid cells (Ter119 ϩ ) were similar in each mouse strain (Fig. 1, A-C). However, the percentages of granulocytic cells (CD11b ϩ /Gr1 hi ) and monocytic cells (CD11b ϩ /Gr1 lo ) were lower in the GST-deficient animals than in the wild type. In contrast, the percentage of cells committed to the B lymphocyte lineage (B220 ϩ ) was higher in the GST Ϫ/Ϫ mice (Fig. 1).

Modulation of Myeloproliferation by MAP Kinase Inhibitors-In vitro hematopoiesis experiments
showed that all three cytokines tested (G-CSF, GM-CSF, and IL-3) stimulated the formation of colony-forming units more effectively in GSTdeficient than in wild type animals. In addition, as reported previously (7), TLK199 stimulated cytokine-induced myeloproliferation in wild type mice but not in GST Ϫ/Ϫ animals (Fig.  2). Therefore, myeloproliferative responses to cytokine stimulation are enhanced in the absence of GST.
Because GST is known to be a regulator of MAP kinase pathways, we assessed the effect of MAP kinase inhibitors on the myelostimulant properties of TLK199. As shown in Fig. 2, A-C, MEK1/2 inhibitors PD98059 and U0126 decreased myeloproliferation induced by cytokines but failed to inhibit the myelostimulant effect of TLK199 completely. These molecules decreased the number of colonies produced after exposure to cytokines by about 2-3-fold in both wild type and GST-defi-cient mice. However, in combination with TLK199, a greater number of colonies were produced in wild type, but not in GST-deficient animals. In addition, it appeared that U0126 was a more potent inhibitor of proliferation than PD98059. This could be explained in part by the more favorable inhibitory constant of U0126 toward MEK1/2 compared with PD98059. In contrast, the p38 inhibitor SB202190 significantly stimulated cytokine-induced myeloproliferation independent of GST expression. TLK199 myelostimulant properties were not detected when used in combination with this p38 inhibitor. Similarly, differential myeloproliferation observed between wild type and GST Ϫ/Ϫ cells was abrogated in the presence of SB202190. The JNK inhibitor SP600125 had a slight inhibitory effect on the number of colonies produced in wild type animals but significantly decreased their numbers in GST Ϫ/Ϫ mice. In addition, SP600125 inhibited the myelostimulant effects of TLK199 in wild type mice. The cell composition and number of colonies were assessed in wild type and GST-deficient mice. As expected, the number of cells/colony was higher in GST-deficient samples than in wild type. However, the results presented in Table III show that there was no significant difference in the colony cell composition. Therefore, our data suggest that the JNK pathway could play an important role in the stimulation of myeloproliferation in GST Ϫ/Ϫ cells or in wild type cells exposed to TLK199.
Influence of GST Expression on IL-3-mediated Signal Transduction in Bone Marrow Cells-Because of the increased myeloproliferation observed in GST-deficient cells in response to cytokines, we assessed the influence of GST expression on IL-3-mediated signal transduction in bone marrow cells collected from wild type and GST Ϫ/Ϫ animals. Treatment of these cells with IL-3 induced a rapid and transient phosphorylation of STAT5 (between 10 to 30 min) in wild type cells (Fig. 3). In contrast, in GST-deficient cells, this cytokine induced a rapid but more sustained activation of STAT5 (up to 120 min). In addition, IL-3 induced a stronger activation of STAT3 without affecting the duration of stimulation and of JNK in GST-deficient cells compared with wild type. The activation of the other proteins was similar in both mouse strains (Fig. 3). In addition, the phos-  phorylation of STAT1 was not detected in our model. Increased DNA Synthesis and Sustained STAT Activation in GST Ϫ/Ϫ BMMC in Response to IL-3-Because the bone marrow is a heterogeneous cell system that does not allow accurate study of the effect of cytokines on signal transduction path-ways, we decided to use homogeneous cell populations such as BMMC to define better the role of GST in the cellular response to cytokines. GST Ϫ/Ϫ BMMC had a higher DNA synthesis rate than wild type in response to IL-3 stimulation (Fig.  4). Stimulation of BMMC proliferation by IL-3 was also ob-  (n ϭ 4). *, statistically different from wild type (Student's t test, p Ͻ 0.01). C, total nucleated cells/femur. served at a lower concentration of cytokine in GST-deficient cells than in the wild type mast cells. Studies of cell signaling pathways involved in IL-3-mediated signal transduction showed that this cytokine induced a stronger and more sustained phosphorylation of STAT1 (Tyr 701 ), STAT3 (Tyr 705 and Ser 727 ), and STAT5 (Tyr 694 ) in GST-deficient cells than in wild type (Fig. 5A). Such an effect was also observed to a lesser degree with JAK2 phosphorylation. In contrast, levels of AKT phosphorylation were lower in knock-out than in wild type cells. Although levels of phosphorylated JNK were higher in wild type cells, the ratio of JNK phosphorylation to JNK total protein was higher in the BMMC from knock-out animals (Fig.  5B). Activation of p38 and ERK was similar in both cell types.
Decreased Expression of Negative Regulators of the JAK-STAT Pathway in GST Ϫ/Ϫ BMMC-The basal and IL-3-induced mRNA expression of negative regulators of the JAK-STAT pathway in BMMC was assessed by quantitative PCR and immunoblot. The basal mRNA levels for CIS-1, SOCS-1, SOCS-2 were similar, whereas that of SOCS-3 was 2-fold lower in GST-deficient BMMC than wild type (Fig. 6A). IL-3 stimulated the expression of CIS-1, SOCS-1, SOCS-3, and to a lesser extent SOCS-2 in both cell types; however, this was more pronounced in GST-deficient cells. Control expression of the regulatory phosphatase CD45 was similar in both cell lines, but its expression was repressed by IL-3 treatment. By immunoblot, we observed that the expression of the phosphatases SHP-1 and SHP-2 was also much lower in GST Ϫ/Ϫ cells than in wild type but was not significantly affected by IL-3 exposure (Fig. 6B).

Influence of Phosphatase Inhibitor Orthovanadate on IL-3induced BMMC Proliferation and JAK-STAT Activation-Pretreatment of wild type and GST-deficient mast cells for 4 h with 10
M orthovanadate leads to a more sustained phosphorylation of JAK2, STAT-3, -5, ERK, and p38 proteins (up to 4 h) in response to IL-3 (Fig. 7B). In the presence of the tyrosine phosphatase inhibitor, the time course of activation of those proteins was similar in wild type and knock-out cells, whereas in the absence of orthovanadate the activation of the JAK-STAT pathway was more sustained in GST Ϫ/Ϫ (Fig. 5A). In addition, pretreatment with orthovanadate increased wild type and GST-deficient BMMC proliferation in response to IL-3 (Fig. 7A). When exposed to orthovanadate, the levels of thymidine incorporation in response to the cytokine were similar in both cell lines, whereas in the absence of this inhibitor, GST Ϫ/Ϫ cells accumulated more thymidine than wild type. DISCUSSION Increased myeloproliferation in GST-deficient mice compared with wild type animals has been reported previously (7). However, the molecular basis of this effect has, until now, not  2. In vitro hematopoietic assay in wild type and GST ؊/؊ mice. Bone marrow cells were exposed to 10 ng/ml IL-3 (A), 10 ng/ml G-CSF (B), or 10 ng/ml GM-CSF (C) in the presence or absence of 10 M TLK199, 10 M PD98059, 10 M U0126, 10 M SP600125, or 10 M SB202190. Colonies (Ͼ50 cells) were counted 7 days after treatment. Results are expressed as a percentage of the wild type control (cytokine alone). Statistical analysis was performed by repeated measures of one-way analysis of variance and Tukey-Kramer post-test using Graph-Pad software. *, statistically different from the control (p Ͻ 0.01); **, statistically different from GST Ϫ/Ϫ "cytokine alone" (p Ͻ 0.01); #, statistically different from MEK1/2 inhibitor alone (p Ͻ 0.05). been addressed. We observed an increase in white blood cell counts in GST Ϫ/Ϫ mice, but no modification in leukocyte composition was detected. Spleen cell counts were higher in knockout animals than in wild type, and this was associated with a 2-fold increase in B lymphocytes, whereas T lymphocytes and natural killer cell counts were similar in both animal strains. In contrast, no difference in thymocyte counts and thymus subset composition was observed. Red blood cell and platelet counts were also similar in both mouse strains. Despite the increased number of circulating white blood cells, no difference in the total number of nucleated bone marrow cells was ob-served. Interestingly, the percentage of B lymphoid cells was increased in the GST-deficient animals, whereas the percentages of myeloid cells were lower. This could be explained by the fact that blood cell count is used to assess the production of mature cells, whereas bone marrow cell analysis is used to determined commitment of cells to the production of mature circulating blood cells. The results obtained here are not necessarily contradictory and may suggest that bone marrow from GST-deficient animals produces more mature cells by increasing the proliferation and/or differentiation of committed progenitors into fully differentiated circulating cells. B cells represent about 70% of the total circulating white blood cells and can be found in other organs such as spleen and lymph nodes. Their proportional abundance within the total white blood cell count could explain why GST-deficient animals have a higher percentage of B220 ϩ than wild type animals. Perhaps because of limited space for bone marrow expansion, lymphoid progenitors could compete with myeloid counterparts with the outcome of an increased percentage of lymphoid cells and a decrease of myeloid progenitors. In addition, in in vitro hematopoiesis experiments we observed that proliferation and differentiation of myeloid cells in response to various cytokines were higher in GST-deficient animals, suggesting that de-

FIG. 3. Effect of GST expression on IL-3-mediated signal transduction pathways in bone marrow cells.
Bone marrow (B.M.) cells were collected from wild type (WT) and GST Ϫ/Ϫ mice and incubated for 2 h in Iscove medium containing 2% bovine serum albumin. Cells were exposed to 100 ng/ml IL-3 and collected at various time points. Protein expression and phosphorylation were analyzed by immunoblotting.

FIG. 4. Effect of GST expression on BMMC proliferation.
Wild type (WT) and GST-deficient BMMC were incubated overnight without IL-3. Increasing concentrations of IL-3 were added to the cells for 24 h. Radiolabeled thymidine was then added for an extra 12 h. Thymidine incorporation was assessed using a scintillation counter.

FIG. 5.
A, effect of GST expression on IL-3-mediated signal transduction pathways in BMMC. Wild type and GST Ϫ/Ϫ were IL-3 starved for 4 h. 100 ng/ml IL-3 was then added, and cells were collected at various time points. Protein expression and phosphorylation were analyzed by immunoblot. B, ratio of phosphorylated JNK to total JNK in wild type and GST Ϫ/Ϫ cells after IL-3 treatment. Band intensities were determined by densitometry using NIH Image software. Results are normalized to the untreated wild type control. spite their lower number in the knock-out animals they are able to produce higher numbers of mature circulating cells. Taken together, these data infer that the absence of GST expression potentiates hematopoiesis by influencing the proliferation and/or differentiation of hematopoietic cells.
In in vitro hematopoiesis experiments, IL-3, GM-CSF, or G-CSF induced more colonies in GST Ϫ/Ϫ cells than in wild type. TLK199, a specific GST inhibitor, was able to stimulate colony formation in wild type but not in knock-out animals. The JNK inhibitor SP600125 decreased the number of colonies produced by cytokine treatment of GST-deficient animals. In addition, we observed here that JNK phosphorylation was increased in GST Ϫ/Ϫ bone marrow cells. These data are consistent with GST being a physiological inhibitor of JNK (4) and TLK199 disassociating GST from JNK, allowing kinase phosphorylation and subsequent activation of the kinase cascade (4,7). SP600125 as an inhibitor of JNK (16) abrogates the increased phosphorylation of this kinase observed in the presence of TLK199 and consequently reduces the myeloproliferative effect of this agent. Because the cell compositions of the colonies in both mouse strains were similar after exposure to cytokines, the increase is a function of proliferation, rather than differentiation of hematopoietic cells. Taken together, these data suggest that JNK plays an integral role in the elevated myeloproliferation observed in GST-deficient mice and the myelostimulant properties of TLK199. This is consistent with recent reports showing a possible role for JNK in proliferation. Yang and colleagues (17) have shown that the inhibition of JNK1 or JNK2 expression in human prostate carcinoma was associated with a decrease in cell proliferation. Similarly, it has been observed that JNK activity was required for rat liver regeneration by increasing cyclin D1 expression and allowing G 0 to G 1 transition (18). In addition, JNK phosphorylates and induces the transactivation of the transcription factors c-Jun (19), JunB, and JunD (20). Overexpression or activation of Jun and JunD has been linked with cell proliferation and transformation (21,22). JNK activation has also been associated with induction of apoptosis (23); however, more recent data have suggested that after UV exposure, JNK1 was more likely proapoptotic, whereas JNK2 was associated with survival (24). The discrimination between the survival and apoptotic functions of JNK also seems to correlate with the level and duration of the enzyme activation. A strong and sustained activation is associated with apoptosis, whereas a weaker and transient phosphorylation is correlated with proliferation (25). For example, in mouse hematopoietic BaF3 cells, JNK activity was three times lower when cells were exposed to mitogenic concentrations of IL-3 than exposed to cytotoxic concentrations of anisomycin (13). Thus, in the context of our work, the increased JNK activation observed in GST Ϫ/Ϫ bone marrow cells could contribute to the increased proliferation of these cells. In addition, we observed that JNK expression was lower in GST-deficient BMMC than in wild type, but the level of activation of this kinase (pJNK versus total JNK) was higher in the knock-out cells. Because of the proapoptotic properties of JNK, lower levels of the kinase may serve to maintain a higher level of viability in BMMC from GST Ϫ/Ϫ animals. However, lower expression of JNK was observed only in mast cells, and its exact significance remains unclear.
The MEK1/2 inhibitors PD98059 and U0126 decreased myeloproliferation in both wild type and GST-deficient animals. However, the latter was more potent than PD98059, an observation that could be explained in part by the higher inhibitory activity of U0126 toward MEK1/2. ERK pathway activation has been linked with proliferation and differentiation in various hematopoietic cell lines, and its inhibition has been associated with growth inhibition (26 -28). However, because MEK1/2 inhibitors failed to suppress myelostimulation caused by TLK199, our results would suggest that ERK is not directly involved in the myeloproliferative properties of this agent.
The p38 inhibitor SB202190 stimulated myeloproliferation in both wild type and GST-deficient mice, data consistent with recent studies showing that p38 controls the G 2 /M checkpoint by phosphorylating two key cell cycle regulators, cdc25B and cdc25C. Inhibition of p38 activity blocks the initiation of this checkpoint and leads to increased proliferation (29). In addition, p38 has been demonstrated to phosphorylate and stabilize p53 and consequently to induce either cell cycle arrest or apoptosis (30,31). Our results suggest that the modulation of p38 kinase pathway could also be an interesting therapeutic target to stimulate myeloproliferation in patients with impaired hematopoiesis. We also observed that in cotreatment with this p38 inhibitor, TLK199 did not increase myeloproliferation already stimulated by SB202190. Because of its potent effects on myeloproliferation, SB202190 could mask the myelostimulant properties of this GST inhibitor. Perhaps for the same reason, this kinase inhibitor abrogated the differential myeloproliferation observed in wild type and GST Ϫ/Ϫ animals. In addition, recent reports have suggested the presence of cross-talk among the various MAP kinase pathways (32). Such findings could also explain why SB202190 myelostimulant properties are not affected by TLK199, which affects JNK activity directly and may through this kinase alter p38 activity.
The increased proliferation of GST Ϫ/Ϫ bone marrow and mast cells was associated with a sustained activation of STAT proteins. These molecules have been linked with the proliferation and differentiation of hematopoietic cells (33). For example, STAT5-deficient animals present a decreased proliferation of hematopoietic cells, which is correlated with a defect in response to various cytokines including IL-3, GM-CSF, and G-CSF (34). Constitutive activation of STAT5 has been observed in various leukemias (35)(36)(37). In addition, forced expression of a constitutively active form of STAT5 allows IL-3-dependent BaF3 cells to grow in medium without cytokines, whereas mock-transfected cells died by apoptosis (38). Similarly, constitutive activation of STAT3 has been observed in various tumors (39,40). Because of the crucial role played by the JAK-STAT pathway in hematopoiesis, our data suggest that the increased STAT activation in the GST-deficient cells could be a critical component of the accelerated proliferation of these cells. This observation could also explain the higher number of circulating white blood cells in the GST Ϫ/Ϫ animals.
In BMMC, elevated STAT protein activation was associated with a down-regulation of the levels of mRNA for various negative regulators of JAK-STAT pathways. CIS-1 and SOCS-1, -2, and -3 have been shown to inhibit the phosphorylation of JAK and STAT proteins (41). Their expression is usually induced by cytokine treatment. Mechanistically, SOCS-1 and SOCS-3 inhibit JAK catalytic activity, whereas CIS-1 blocks STAT binding to the receptor; the mechanism of action of SOCS-2 is still unclear (41). In our studies, we observed that the basal expression of the SOCS family members was similar or slightly decreased (especially SOCS-3) in the GST-deficient mast cells. However, upon IL-3 treatment, their expression was more stimulated in the knock-out than in wild type cells. This could be a consequence of the stimulation of SOCS expression by STAT proteins. Because activation of the latter is more sustained in GST-deficient BMMC, their overall transcriptional activity is also increased. SHP-1 and SHP-2 are phosphatases that can bind to the ␤-chain of the IL-3 receptor (42) and can inhibit IL-3-induced phosphorylation of STAT5 (43). In addition, it has been shown recently that SHP-2 could bind to STAT5 and was able to dephosphorylate this protein (44). Negative regulation of these phosphatases has been observed in various lymphomas and associated with increased activation of the JAK-STAT pathway (45). Similarly, transmembrane protein tyrosine phosphatase CD45 inhibits cytokine signaling by dephosphorylating JAK (46). Taken together, our present results suggest that increased activation of STAT proteins in GST Ϫ/Ϫ cells might be a consequence of the decreased expression of these endogenous inhibitors of their phosphorylation.
To demonstrate the role of phosphatases in the more sustained activation of JAK-STAT pathway, we cotreated cells with the tyrosine phosphatase inhibitor orthovanadate, which has been shown previously to potentiate the activation of those proteins in response to interferon or leptin (47,48). We observed that the IL-3-induced proliferation of wild type and GST-deficient mast cells was potentiated by orthovanadate. This was associated with a more sustained activation of JAK-STAT pathways in both cell lines. These results suggest that negative regulation of phosphatase expression in GST-deficient cells was, at least in part, responsible for the more sustained activation of the JAK-STAT pathway and consequently the increased proliferation of the GST-deficient cells.
In conclusion, the elevated activations of JNK and STAT in GST -deficient mice are contributory factors to the increased myeloproliferation observed in these animals. These data also confirm the important role that GST plays in the regulation of cell signaling pathways through its interaction with JNK.