STAT5 Activation by BCR-Abl Contributes to Transformation of K562 Leukemia Cells

U937 and K562 cells were maintained in RPMI 1640 supplemented with 8% fetal calf serum (FCS) (Hyclone, Greiner, Logan, UT). NIH 3T3 cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 8% FCS (Life Technologies, Breda, The Netherlands). Focus assays on NIH-3T3 cells were performed as described previously.

The phosphotyrosine monoclonal antibody (MoAb) 4G10 was obtained from UBI (Lake Placid, NY). The MoAb anti-STAT1 (S21120, directed against amino acids [aa] 592-731) was obtained from Transduction laboratory (Lexington, KY). The polyclonal antibodies (pAb) directed against STAT3 (C-20, aa 750-769), STAT5A (L-20, aa 774-793), STAT5B (C-17, aa 711-727), and STAT5 (N-20, aa 5-24) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

The following oligonucleotide was used in this study (only the upper strands are shown); for bandshift assays: β-casein (5′-AGCTTAGATTTCTAGGAATTCAA ATCA-3′). The expression plasmids pMXmSTAT5A, pMXmSTAT5A, Δ75037 (a kind gift of Dr Fabrice Gouilleux, Tumor Biology Center, Freiburg, Germany), pSG5-STAT3, pSG5-STAT3β, and pMvSrc were described previously.22,38 Full-length cDNAs of the different STATs were excised from the plasmids described above and cloned into the LNCX expression vector (Clontech, Palo Alto, CA), which contains a neomycin resistance gene. The reporter constructs 4xIREtkCAT, 4xGAStkCAT, 4xSIEtkCAT, and 4xβCaseintkCAT were previously described.38,39 

For transient assays, K562 cells were electroporated with 2 μg of reporter plasmid and 10 μg of bluescript SK− carrier DNA. For some experiments, dominant negative expression vectors for STAT3 or STAT5 were cotransfected instead of bluescript SK−. Cells were harvested for CAT assays 48 hours after transfection. Chloramphenicol acetyltransferase (CAT) assays were performed as described previously.38 

For soft agar assays, 24 hours posttransfection K562 cells were treated with 1 mg/mL G418 for 10 days. Cells were then plated in soft agar and scored as was described previously,40 except that RPMI 1640 with 10% FCS was used as medium.

Nuclear extracts were prepared following a previously described procedure.41 Oligonucleotides were labeled by filling in the cohesive ends with [a-³²P]deoxycytidine triphosphate (dCTP) using Klenow fragment of DNA polymerase I. Gel retardation assays were performed according to published procedures.38 Supershift analysis was performed by preincubating 10 μg of nuclear extract with 1 μg of anti-STAT antibody for 30 minutes on ice before addition of the binding buffer and ³²P-labeled probe.

K562 cells (10.10⁶ per sample) were washed with ice-cold phosphate-buffered saline and subsequently lysed in 750 μL radioimmunoprecipitation assay (RIPA) buffer (150 mmol/L NaCl, 20 mmol/L Tris pH 7.4, 5 mmol/L EDTA, 1% Nonidet P-40, 0.1% sodium dodecyl sulfate [SDS], 0.5% sodium deoxycholate, 1 mmol/L phenylmethylsulphonylfluoride [PMSF], 2 μg/mL leupeptin, 4 μg/mL aprotinin, 1.5 μg/mL pepstatin, 1 μg/mL trypsin inhibitor, and 50 mmol/L NaF) for 10 minutes on ice. Cell lysates were centrifuged to remove the insoluble material, and supernatants were precleared with protein A-Sepharose beads for 15 minutes at 4°C. Thereafter, supernatants were incubated with STAT antibodies for 1.5 hours at 4°C. Protein A-Sepharose was added to the reaction mixture and the incubation was continued for 45 minutes. Beads were collected by centrifugation and washed eight times with cold RIPA buffer, resuspended in Laemmli sample buffer, and boiled for 3 minutes. Protein samples were separated on 8% SDS-polyacrylamide gels, and electrotransferred to Immobilon-P membranes (Millipore, Bedford MA). Membranes were blocked in TBST-buffer (150 mmol/L NaCl, 10 mmol/L Tris pH 8.0, 0.3% Tween 20) containing 5% bovine serum albumin (BSA) for 30 minutes and probed with either an antiphosphotyrosine MoAb 4G10 or the anti-STAT antibodies described above for 1 hour. After three washes with TBST, the membranes were incubated for 1 hour with either peroxidase-conjugated rabbit antimouse antibodies (after MoAb) or peroxidase-conjugated swine antirabbit antibodies (DAKO, Glostrup, Denmark) (after pAb), followed by five washes with TBST. Proteins were visualized with enhanced chemiluminescence (ECL; Amersham, Buckinghamshire, UK).

Previously, it was demonstrated that BCR/Abl activates STAT5, although activation of STAT1 was also reported in some cell types.7-14 To determine which STAT family members were activated in BCR/Abl-expressing K562 cells, we performed immunoprecipitation/Western blotting experiments using antiserum specific for STAT1, 3, 5A, and 5B. Figure1A shows that both STAT5A and 5B are phosphorylated on tyrosine residues in these cells. The second band observed in the STAT5B immunoprecipitation is likely to be caused by serine phosphorylation of STAT5B (data not shown). By contrast, activation of STAT1 or 3 could not be detected. These results were confirmed by gel-shift analysis. Figure 1B shows that STAT5A and 5B constitutively bind to the STAT binding site from the β-casein promoter and from the FcγRI promoter (not shown), while DNA binding of STAT1 and STAT3 could not be observed. These results clearly show that STAT5A and 5B are the major targets for BCR/Abl in K562 cells.

Although tyrosine phosphorylation and DNA binding of STAT5 was previously reported in BCR/Abl-expressing cells, transcriptional activation of STAT5-dependent promoters was not shown. We therefore transfected various STAT-dependent reporter constructs into K562 cells. Figure 2A shows that constructs containing STAT binding sites from the β-casein promoter (which can bind STAT1 and STAT5) and from the FcγRI promoter (GAS, which can bind STATs 1, 3, and 5) are more active than the TK-CAT control reporter. By contrast, the IRE-CAT and SIE-CAT reporters, which cannot bind STAT5, are comparable to the TK-CAT control reporter. These results suggest that STAT5 is indeed transcriptionally active in K562 cells. To further extend these results, we transfected K562 cells with the β-casein-CAT reporter together with either dominant-negative STAT5 (STAT5δ750) or STAT3β (Fig 2B). While STAT3β did not significantly alter the activity of the β-casein-CAT reporter, STAT5δ750 caused a strong repression of CAT activity, further suggesting that STAT5 is transcriptionally active in these cells.

In contrast with most nontransformed hematopoietic cells, one of the transformed properties of K562 cells is their ability to grow in semisolid media (soft agar). To determine whether activation of STAT5 is involved in this property of K562 cells, we used STAT5δ750. K562 cells were transfected with STAT5δ750, STAT3β, or the empty expression vector (LNCX), after which the transfected cells were selected in G418-containing media for 10 days. Thereafter, cells were plated in dishes containing soft agar. Figure 3 shows that K562 cells transfected with an empty expression vector (LNCX) grow efficiently in soft agar (Fig 3A and D). By contrast, K562 cells transfected with STAT5δ750 form fewer and smaller colonies in soft agar (Fig 3B and D), suggesting that STAT5 is indeed involved in this aspect of cellular transformation by BCR/Abl. As a control, cells transfected with STAT3β grew as efficiently in soft agar as vector controls, further showing that STAT3 is not involved in transformation by BCR/Abl (Fig 3C and D).

We next wanted to investigate whether STAT5 also contributes to transformation by other oncogenes. Recently, we and others have shown that STAT3 is involved in cellular transformation by v-Src.18,22 Figure4A shows that besides STAT3, STAT5A and 5B are also tyrosine-phosphorylated in v-Src–transformed NIH-3T3 cells, but not in the parental NIH-3T3 cells. However, in contrast to K562 cells, we failed to detect substantial STAT5 DNA binding activity in v-Src–transformed cells (data not shown). To investigate whether STAT5 is causally involved in v-Src–induced transformation, we performed focus assays in NIH-3T3 cells. Transfection of NIH-3T3 cells with v-Src efficiently induces focus formation in these cells (Fig 4B). As we have previously shown, this can be strongly repressed by cotransfecting STAT3β. Interestingly, cotransfection of STAT5δ750 also represses v-Src–dependent focus formation in NIH-3T3 cells, albeit much less efficiently than STAT3β. The combination of STAT3β and STAT5δ750 was somewhat more potent in repression focus formation than either plasmid alone. In addition, the repression observed with STAT5δ750 could be overcome by cotransfection of wild-type STAT5, further suggesting that STAT5 indeed plays a role in transformation by v-Src. To rule out potential a-specific effects of STAT5δ750 on STAT3-dependent signaling, we performed cotransfection of a STAT3-dependent reporter construct (SIE-CAT) together with v-Src, STAT3β, and STAT5δ750. Figure 4C shows that STAT3β efficiently blocks v-Src–induced SIE-CAT activity. By contrast, STAT5δ750 only slightly reduced v-Src–induced SIE-CAT activity, suggesting that STAT5δ750 does not repress STAT3 function in these cells. These results show that although STAT3 is the major STAT involved in transformation by the v-Src oncogene, STAT5 is also likely to play a minor role.

Taken together, we have shown that besides STAT3, STAT5 is also involved in transformation mediated by at least two oncoproteins. The mechanism by which active STAT5 contributes to cellular transformation remains to be determined, but is likely to involve constitutive activation of STAT5 target genes, which are somehow involved in the control of proliferation (eg, c-fos). In this respect, it is noteworthy that blocking STAT5 function in mouse BaF3 cells results in a significant decrease in IL-3–dependent proliferation,35while a constitutively active form of STAT5 renders IL-3–dependent cells partially IL-3–independent.36 On the other hand, because STAT5 is involved in the regulation of the antiapoptotic Bcl2 homologue A1,42 enhanced survival and escape from apoptosis of cells containing active STAT5 might also contribute to cellular transformation. The availability of constitutively active STAT5 variants36,43 will undoubtedly give new insights into the mechanism by which STAT5 contributes to cellular transformation.