Introduction
As mediators of cytokine-induced and growth factor-induced gene expression, signal transducers and activators of transcription (STATs) are involved in cellular differentiation, proliferation, and survival. Upon cytokine or growth factor binding to its receptor, the latent cytoplasmic STAT proteins are recruited to the receptor complex resulting in STAT activation by either receptor tyrosine kinases or nonreceptor tyrosine kinases such as Janus kinases or c-Src. Activation of STAT proteins requires phosphorylation on a conserved tyrosine residue located in the carboxy terminus. Phosphorylation of this tyrosine leads to phosphotyrosine–Src homology domain 2-mediated reciprocal dimerization. The activated STAT dimer then translocates to the nucleus and binds to a STAT consensus DNA element, resulting in gene transcription. The STAT family consists of seven members that can be divided into two categories: those that respond to cytokine signals (STAT2, STAT4, STAT6), and those that respond to cytokine and growth factor signals (STAT1, STAT3, STAT5a, STAT5b) [
1‐
3].
Although STAT5a and STAT5b play a fundamental role in normal growth and development of the mammary gland, both proteins are overexpressed or constitutively activated in cancers, including some breast cancer tumors [
4‐
9]. Owing to their ability to regulate the expression of genes involved in cell-cycle regulation (cyclin D
1, c-
myc, and p21) and cellular survival (Bcl-X
L), STAT5a and STAT5b have emerged as possible targets for cancer therapeutics [
10]. Recent evidence indicates that STAT5b, but not STAT5a, has a proproliferative role in breast cancer, head and neck cancer, and prostate cancer [
11‐
14]. Since STAT5b mediates breast cancer proliferation, identification of kinases that increase STAT5b activity is critical to identifying potential therapeutic targets.
Breast tumor kinase (Brk) is a nonreceptor tyrosine kinase originally isolated from an involved axillary node of a patient with metastatic breast cancer, and is expressed in more than 60% of breast cancers [
15,
16]. With 46% amino acid identity to c-Src, Brk is distantly related to the Src family of tyrosine kinases [
17,
18]. Although normally expressed in the gastrointestinal tract, expression of Brk is not detected in the normal mammary gland [
16,
19]. Stable transfection of Brk in the immortalized nontransformed human mammary cell lines HB4a and MCF10A, however, leads to sensitization to epidermal growth factor and results in a partially transformed phenotype [
20]. Brk also enhances epidermal growth factor-induced ErbB3 and Akt phosphorylation in the HB4a cells [
21]. Furthermore, knockdown of the Brk protein decreases proliferation of breast cancer cell lines [
22]. Given its role in breast cancer cell proliferation, survival, and tumorigenesis, identification of the substrates of this tyrosine kinase is of utmost importance.
Although STAT5b is involved in cancer proliferation, mutations of STAT5b to account for this increased biological activity have not been identified. Alternatively, increased STAT5b activation results from the overexpression and/or the constitutive activation of tyrosine kinases, such as the epidermal growth factor receptor (EGFR), c-Src, and the fusion protein Bcr/Abl [
4,
6]. Since all identified Brk substrates are also substrates for c-Src, we examined the ability of Brk to mediate STAT5b phosphorylation, a known c-Src substrate.
In the studies presented here, the ability of Brk to phosphorylate and activate STAT5b and the biological significance of this activation were investigated. Exogenous expression of Brk and STAT5b demonstrated that Brk mediates the phosphorylation of the activating tyrosine, Y699. Furthermore, an in vitro kinase assay determined that Brk can directly phosphorylate STAT5b on Y699. Subsequently, this Brk-mediated STAT5b phosphorylation leads to STAT5b transcriptional activity, and this activity is further increased by kinase active c-Src. The results of siRNA experiments suggest that Brk and STAT5b are in the same signaling pathway, which ultimately leads to the proliferation of breast cancer cells. These studies further support targeting STAT5b as a potential breast cancer therapeutic.
Materials and methods
Cell lines and transient transfections
The human breast cancer cell lines SKBr3, BT-20, BT-549, MDA-MB-468, and T47D were obtained from the American Type Culture Collection (Manassas, VA, USA). Cells were maintained in DMEM plus 10% FCS and were passaged twice per week. Mouse embryo fibroblasts (MEF5
-/-) from STAT5a/b knockout mice (provided by Dr J Ihle, St Jude Children's Hospital, Memphis, TN, USA) were passaged twice per week and maintained in DMEM plus 10% FCS. Cells were transfected with STAT constructs [
23], Brk constructs (generous gift from Dr C Lange, University of Minnesota, Minneapolis, MN, USA), and c-Src constructs as previously described [
24], using LipofectAMINE and PLUS reagent according to the manufacturer's instructions (Invitrogen, Gaithersburg, MD, USA).
Reagents
The polyclonal STAT5a-specific and STAT5b-specific antibodies were developed in our laboratory, as previously described [
23]. Polyclonal anti-STAT3 antibodies, polyclonal anti-Brk antibodies, and monoclonal antiphosphotyrosine antibodies (PY-99) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The monoclonal anti-β-actin antibody was from Sigma (St Louis, MO, USA). The antiphospho-STAT5a/b (Y694/Y699) antibody was developed in conjunction with Aves Labs (Tigand, OR, USA) as described elsewhere (EM Fox, T Bernaciak, J Wen, A Weaver, M Shupnik, CM Silva - unpublished data). The protease inhibitor cocktail was from Calbiochem (San Diego, CA, USA). The acrylamide was obtained from Bio-Rad (Hercules, CA, USA), and the prestained molecular weight marker was from Invitrogen. Except as noted, other reagents were of either reagent grade or molecular biological grade from Sigma.
Immunoprecipitations and immunoblotting
Cells were lysed in 150 mM NaCl, 5 mM ethylenediamine tetraacetic acid, 1% Triton X-100, 1% deoxycholate, 50 mM Tris (pH 7.4), containing protease inhibitors and phosphatase inhibitors. Lysates were incubated with the indicated antibody overnight at 4°C, and protein A-agarose (Santa Cruz Biotechnology) was added for an additional 1 hour at 4°C. Agarose pellets were washed three times in detergent lysis buffer, and the bound proteins were removed by boiling in 1 × Laemmli buffer. Proteins were separated on polyacrylamide gels and were analyzed as previously described [
12].
Luciferase assay
MEF5
-/- and BT-549 cells were transfected with the Spi2.1-containing luciferase reporter plasmid. Forty-eight hours post transfection, lysates were prepared and luciferase activity was measured [
24]. The luciferase values (arbitrary units), as measured by a Berthold Luminometer (Berthold, Oak Ridge, TN, USA), were normalized to total protein.
Recombinant protein
STAT5b or Y699F cDNA was cloned into the pGEX4T-1 plasmid (Amersham, Piscataway, NJ, USA) using the EcoRI and NotI restriction sites. The GST, GST-STAT5b, and GST-Y699F plasmids were generated and transformed into Escherichia coli BL21. Protein expression was induced by 1 mM isopropyl-beta-D-thiogalactopyranoside (IPTG in Luria broth (LB) broth at 18°C for 18 hours. Bacterial cells were lysed in 1 × PBS containing 1 mM ethylenediamine tetraacetic acid, 5 mM dithiothreitol, 1.5% sarkosyl, 1% Triton, 2 mM phenylmethylsulfonyl fluoride (PMSF), and 1 μg/ml pepstatin followed by sonication. GST, GST-wtSTAT5b, and GST-Y699F were isolated using glutathione agarose beads (Sigma) following the manufacturer's instructions. Following elution with excess glutathione (50 mM Tris–HCl/10 mM glutathione, pH 8.0), the recombinant proteins were dialyzed in 1 × Tris-buffered saline/10% glycerol. Protein concentration was determined by the Bio-Rad Protein Assay (Bio-Rad, Hercules, CA, USA).
In vitrokinase assay
Purified recombinant Brk (Upstate, Billerica, MA, USA) and purified recombinant GST, GST-STAT5b, or GST-Y699F were incubated with 20 nM ATP in reaction buffer (100 mM Tris–HCl, pH 7.4, 125 mM MgCl2, 25 mM MnCl2, 2 mM ethylene glycol tetraacetic acid (EGTA), 0.25 mM NaVO4, 2 mM dithiothreitol) at 30°C for 30 minutes. An equal volume of 2 × Laemmli was added to end the reaction. Phosphorylation of STAT5b was analyzed by immunoblotting with our specific phospho-Y694/Y699 STAT5a/b antibody.
RNA isolation and reverse transcriptase polymerase chain reaction
Total RNA was isolated from the breast cancer cell lines using the RNeasy mini kit (Qiagen, Valencia, CA, USA) following the manufacturer's instructions. The cDNA was generated using iScript cDNA synthesis (Bio-Rad). The cDNA was amplified using primers for Brk or β-actin as described by Kasprzycka and colleagues [
25].
Small interfering RNA methodology
Knockdown of Brk and/or STAT5b was performed using the siGenome SMARTpool duplex (Dharmacon, Lafayette, CO, USA) transfected with Oligofectamine (Invitrogen) according to the manufacturer's instructions.
DNA synthesis assay
Following 48 hours of transfection with siRNA, SKBr3 cells or BT-20 cells were serum starved for an additional 18 hours and were then incubated with 100 μM bromodeoxyuridine (BrdU) for 6 hours. Cells were fixed and permeabilized as previously described [
12]. Cells were blocked in 20% goat serum/PBS for 20 minutes at 37°C, and then were incubated with anti-BrdU-Alexa-Fluor 594 (Molecular Probes, Carlsbad, CA, USA) for 1 hour at 37°C. BrdU incorporation was visualized using a Leica DM RBE Fluorescence microscope (model RS232C; Leica Microsystems, Bannockburn, IL, USA).
Discussion
Although traditionally known for its role in growth hormone signaling and mammary gland development, STAT5b has emerged as a therapeutic target due to its pivotal role in cancer. Inhibition of STAT5b, but not of STAT5a, in xenograft models of head and neck carcinomas via antisense oligonucleotides repressed tumor growth and hindered expression of the STAT5-regulated genes cyclin D
1 and Bcl-X
L [
13,
14]. STAT5b is more abundantly expressed than STAT5a in prostate cancer and breast cancer cell lines [
12,
31]. Inhibition of STAT5b via dominant-negative constructs and siRNA technology decreases DNA synthesis (Figure
6) [
32], while exogenous expression of a basally active STAT5b mutant (Y740/743F) increases DNA synthesis of breast cancer cells [
12]. Together, these data establish a fundamental role for STAT5b in the process of breast cancer tumorigenesis.
Since activating mutations in STAT5b have not been found in breast cancer, it is important to look upstream to identify the kinases that regulate STAT5b, thus potentially leading to its increased activity in breast cancer cells. Both STAT5b and STAT3 are activated by several kinases overexpressed in breast cancer, including the EGFR, HER2, and c-Src. STAT3 was recently shown to also be phosphorylated by the nonreceptor tyrosine kinase Brk [
1]. Since Brk is expressed in more than 60% of breast tumors but not in normal mammary tissue, it has been suggested to be a potential therapeutic target for breast cancer. Since evidence supports the involvement of both Brk and STAT5b in breast cancer proliferation, we investigated the ability of Brk to phosphorylate STAT5b and the biological significance of this activation.
Using synthetic peptides containing consensus motifs for the EGFR, insulin receptor, c-Src, or Abl tyrosine kinases, it was determined that the c-Src-preferred synthetic peptide is the best Brk substrate [
17]. Given these data, it is not surprising that all of the identified Brk substrates (Sam68, SLM-1, SLM-2, STAP-2, paxillin, STAT3) are also c-Src substrates [
1,
30,
33‐
35]. We have identified two more Brk substrates, STAT5a and STAT5b, which are also c-Src substrates (Figure
1). c-Src can mediate the phosphorylation of Y699 and Y725, however, Brk mediates Y699 phosphorylation but not Y725 phosphorylation (Figure
2b) [
12,
27]. A previous study demonstrated that Brk mediated phosphorylation of STAT3, but not of STAT1, STAT2, STAT5, or STAT6 [
1]. These studies were performed in COS1 cells, however, not in breast cancer cells, and
in vitro kinase assays were not performed for STATs other than STAT3. Using our previously characterized STAT5b-specific antibodies, we demonstrated here that Brk can also directly phosphorylate Y699 on STAT5b in breast cancer cell lines and in an
in vitro kinase assay.
Exogenous expression of Brk in the Brk-negative breast cancer cell line BT-549 increased endogenous STAT5b transcriptional activity. Interestingly, the catalytically inactive K219M Brk mutant also significantly increased STAT5b transcriptional activity compared with vector alone, although not to the extent seen with wildtype Brk or the constitutively active (Y447F). In fact, Harvey and Crompton have previously reported that the kinase-inactive Brk mutant (K219M) could increase the proliferation of T47D cells compared with vector [
22]. Since the K219M mutation disrupts the ATP-binding motif, but not the Src homology domain 2 or the Src homology domain 3, these data suggest that Brk has a role in intracellular signaling that does not require its kinase activity. In these cases, Brk may function as an adaptor protein. Finally, although the constitutively active Y447F Brk mutant was able to increase STAT5b transcriptional activity, it was not significantly higher than that seen with wildtype Brk. This mutation is at the presumptive autoinhibitory tyrosine phosphorylation site of Brk (Y447), equivalent to that identified in c-Src (Y527). Although the Y447F Brk increases phosphorylation of a synthetic peptide, the autophosphorylation of the activating tyrosine in Brk (Y342) is comparable with that seen with wildtype Brk [
17,
36]. Nevertheless, the Y447F Brk mutant has decreased transforming potential when compared with wildtype Brk in NIH3T3 cells [
20]. Together, these data as well as the results we have presented here suggest that the regulation of Brk is more complex than originally thought, and probably involves its role as a kinase and an adaptor protein depending on the cell context.
Since both c-Src and Brk tyrosine kinases are frequently overexpressed in breast cancer and since they both mediate Y699 phosphorylation of STAT5b, there is potential for these kinases to either substitute for one another or work together to activate STAT5b. As shown in Figure
5b, exogenous overexpression of c-Src, unlike Brk, did not enhance STAT5b transcriptional activity in the BT-549 cells, although we have reported that it does in other cell lines [
12]. Exogenous expression of c-Src along with Brk, however, enhanced STAT5b transcriptional activity to a level greater than that with Brk alone. As the kinase-inactive c-Src did not enhance Brk-mediated STAT5b transcriptional activity, c-Src kinase activity may play a role in increasing the phosphorylation and functional activation of Brk. Together, these results demonstrate that c-Src and Brk do not merely substitute for one another in mediating STAT5b transcriptional activation. Rather, Brk can function independently of c-Src, or these two kinases can work together to enhance STAT5b activity.
While Brk is not expressed in normal mammary epithelial cells, it is expressed in 60% of breast tumors – suggesting that Brk expression is regulated at the transcriptional level in breast cancer cells [
16]. Furthermore, the DNA sequence of Brk isolated from gastrointestinal epithelial cells and that of Brk isolated from breast tumor cells are identical, suggesting that activating mutations in Brk are not accountable for Brk activity in breast cancer [
18]. Located within the minimal functional promoter of Brk are NF-κB, Sp1, and STAT consensus binding sites, all known to play a role in tumorigenesis. Only NF-κB and Sp1, however, have thus far been shown to bind the Brk promoter [
37]. Reports in human breast cancer samples have varied, showing that Brk correlates with HER2 and HER4 overexpression, as well as with estrogen receptor positivity [
38‐
40]. There is also conflicting evidence showing a strong correlation between Brk staining by immunohistochemistry and tumor grade in one report [
29], and another correlating Brk expression with long-term survival [
39].
Our mRNA and protein analysis across a panel of human breast cancer cell lines showed no correlation of Brk expression with estrogen receptor status or EGFR/HER2 overexpression (Figure
4). We have, however, shown that knockdown of Brk significantly decreased DNA synthesis in the EGFR-overexpressing and c-Src-overexpressing breast cancer cell lines, SKBr3 and BT-20 (Figure
6). The effect of Brk knockdown on inhibiting basal DNA synthesis was comparable with that seen with STAT5b knockdown in both cell lines. Furthermore, simultaneous knockdown of both Brk and STAT5b had the same effect on decreasing DNA synthesis as their individual knockdowns, suggesting that this kinase and substrate converge upon a pathway ultimately leading to proliferation. Ostander and colleagues have shown that Brk knockdown decreases cell growth as well as epidermal growth factor-induced (or heregulin-induced) migration of breast cancer cells [
29]. Additionally, knockdown of Brk decreases epidermal growth factor-induced (or heregulin-induced) cyclin D
1 expression in breast cancer cells [
29]. As STAT5b is known to regulate cyclin D
1 through consensus sites in the promoter, this pathway may be one mechanism by which Brk and STAT5b together regulate increases in DNA synthesis. Given the evidence that we have provided for the functionally relevant regulation of STAT5b by Brk, further pursuing the role of this Brk–STAT5b pathway in breast cancer may provide important novel therapeutic targets.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
AMW performed all the experiments. Both AMW and CMS conceived the experiments, analyzed the data, and drafted the manuscript. Both authors read and approved the final manuscript.