Introduction
Breast cancer is the second most common cancer in American women. Despite improvements in detection and the development of new treatment strategies, the American Cancer Society estimates that more than 180,000 new cases of breast cancer will be diagnosed, and more than 40,000 women will die of breast cancer this year alone. Because many cancers arise from dysregulation of signaling pathways found in normal cells, one of the difficulties in treating cancers is identifying cancer-specific therapeutic targets. Current targeted therapies have not been as successful as anticipated. This lack of success is due in part to the ability of cancer cells to upregulate alternative signaling pathways to promote growth and tumor progression. Many tumorigenic signaling pathways converge on common nuclear transcription factors, and therefore, targeting these downstream proteins may be more effective [
1].
One such group of transcription factors is the signal transducer and activator of transcription (STAT) family. STATs are a family of transcription factors activated by cytokines or growth factors or both. Seven members of the STAT family are known: STAT 1, 2, 3, 4, 5a, 5b, and 6. STAT proteins are latent in the cytoplasm and require phosphorylation of a conserved C-terminal tyrosine residue for activation. This allows dimerization to occur between the phosphorylated tyrosine of one STAT and the Src homology 2 (SH2) domain of another. Active dimers are translocated to the nucleus, where they bind DNA and regulate gene transcription. STAT proteins regulate transcription of genes involved in a variety of biologic processes, including proliferation, survival, and angiogenesis, all of which are involved in cancer development and progression. Thus, it is not surprising that in the last several years, a role for STATs in tumorigenesis has emerged. Activation of STAT5a and STAT5b occurs in a variety of cancers including both hematopoietic cancers and solid tumors, such as those of the breast, prostate, lung, head and neck, and brain [
2,
3]. STAT5a and STAT5b regulate the transcription of the pro-proliferative genes
c-myc and
cyclin D1 and the anti-apoptotic genes
Bcl-xL and
Pim-1, to stimulate tumor growth and survival [
4‐
8]. In addition, STAT5b has been implicated in prostate cancer cell invasion [
9].
To date, most of the work examining STAT5b in breast cancer has focused on its pro-proliferative function, and its role in breast cancer cell migration has not been examined. Importantly, a recent study investigating the effects of STAT5a on breast cancer cell migration and invasion showed that prolactin (Prl)-induced activation of STAT5a inhibited migration and invasion of BT-20 and T-47D human breast cancer cells [
10]. STAT5a and STAT5b, although highly homologous, are encoded by two separate genes and function independently in mammary gland development. STAT5a is necessary for lobuloalveolar outgrowth and lactation mediated by Prl signaling, whereas STAT5b is vital for establishing growth hormone (GH)-directed sexual dimorphism [
11,
12].
Given this background, we sought to investigate the potential role of STAT5b, specifically, in the migration of two highly aggressive, highly migratory breast cancer cell lines. We found that STAT5b knockdown inhibited serum- and fibronectin-stimulated migration of both BT-549 and MDA-MB-231 human breast cancer cell lines in a transwell assay. This inhibition was rescued by co-expression of wild-type, Y699F-, and dominant-negative STAT5b but not STAT5b containing a mutation in the SH2 domain. With real-time imaging, we showed that knockdown of STAT5b resulted in decreased directional migration and the formation of multiple protrusions, giving rise to an overall reduction in motility. These results establish, for the first time, an important SH2-dependent function of STAT5b in aggressive breast cancer, further defining its role in tumorigenesis and supporting its potential as a therapeutic target for the treatment of breast cancer.
Materials and methods
Cell culture
BT-549 and MDA-MB-231 human breast cancer cell lines were obtained from the American Type Culture Collection (Manassas, VA). Cells were passaged twice per week and maintained in Dulbecco's Modified Eagle Medium (DMEM) supplemented with 10% fetal bovine serum (FBS). All tissue-culture reagents were purchased from Invitrogen (Gaithersburg, MD).
siRNA Transfection
BT-549 and MDA-MB-231 cells were transfected with siGENOME SMARTpool siRNA targeting human STAT5b or individual custom oligonucleotides specific for STAT5a or STAT5b (siGENOME STAT5b SMARTpool duplex #3), or luciferase duplex control, all purchased from Dharmacon (Lafayette, CO). Transfections were performed by using either Oligofectamine (Invitrogen) or Amaxa nucleofection (Amaxa/Lonza, Walkersville, MD), by using solution T and program X-013 (MDA-MB-231 cells) or A-023 (BT-549 cells) as per manufacturers' instructions. For knockdown-rescue experiments, cells were transfected simultaneously with siSTAT5b SMARTpool duplex #3 and HA-tagged wild-type-, Y699F-, dominant-negative-, or R618K-STAT5b engineered to be immune to knockdown by introduction of four silent point mutations in the siRNA target sequence. These point mutations were introduced by using QuikChange site-directed mutagenesis (Stratagene, La Jolla, CA), and constructs were sequenced to verify mutations.
Immunoblotting
Cells were lysed in RIPA buffer (150 mmol/L NaCl, 50 mmol/L Tris, pH 7.4, 1% deoxycholate, 1% Triton X-100, 5 mmol/L EDTA) containing protease inhibitor cocktail (Calbiochem, San Diego, CA) and sodium orthovanadate (Sigma, St. Louis, MO), and boiled in 2× Laemmli buffer containing β-mercaptoethanol or 20 mmol/L dithiothreitol for 5 minutes at 100°C. Protein lysates were separated on 7.5% or 12.5% polyacrylamide gels and transferred to nitrocellulose (Pall Corporation, Pensacola, FL). Membranes were blocked and incubated with primary antibodies in TBST (150 mmol/L NaCl, 0.1% Tween 20, 50 mmol/L Tris, pH 8.0) containing 5% nonfat dry milk or 3% BSA. STAT5a- and STAT5b-specific polyclonal antibodies were developed by our laboratory, as previously described [
13]. Monoclonal anti-β-actin antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). HA monoclonal antibody was obtained from the University of Virginia hybridoma facility. Secondary antibodies were applied in TBST and were HRP-conjugated sheep anti-mouse or donkey anti-rabbit (GE Healthcare, Piscataway, NJ). The enhanced chemiluminescence detection kit (GE Healthcare) was used to detect antibody binding. Acrylamide was from Bio-Rad (Hercules, CA), prestained molecular weight standards were from Sigma, and all other reagents were of reagent or molecular biologic grade from Sigma.
Transwell migration assays
BT-549 and MDA-MB-231 cells were transfected with siRNA, as described earlier. Seventy-two hours after transfection, 5 × 104 BT-549 cells or 1 × 105 MDA-MB-231 cells were plated in serum-free (DMEM/0.1% BSA) media into the upper chambers of BD BioCoat Control Chambers (BD Biosciences, San Jose, CA), and DMEM containing 0 to 10% FBS was placed in the lower chamber. For β1-integrin blocking experiments, cells were pretreated for 1 hour with DMSO (vehicle control) or 10 μg/ml monoclonal anti-human β1-integrin antibody (R&D Systems, Minneapolis, MN), and treatment was left on for the duration of the assay. For migration to extracellular matrix components, the undersides of filters were coated with 0 to10 μg/ml human plasma fibronectin (FN) (BD Biosciences) or recombinant human vitronectin (VN) (R&D Systems) overnight at 4°C, as indicated, and DMEM/0.1% BSA was used in both the upper and lower chambers. Plates were incubated at 37°C, and migration was allowed to proceed for 3 to 6 hours. After this time, nonmigratory cells in the upper chambers were removed with cotton swabs, and the remaining cells were stained with 0.1% crystal violet (Sigma) in 20% ethanol. Cells were counted by using a Zeiss Invertoskop light microscope. Four fields were counted on each of two filters. Results are expressed as average cells per field or relative migration (compared with control).
Wound healing
Cell were plated on FN-coated (5 μg/ml) 35-mm Bioptechs delta T dishes (Fisher Scientific, Pittsburgh, PA). Confluent cell monolayers were wounded by using a 20-μl pipette tip, and cells migrating into the wound were filmed at 37°C with time-lapse microscopy by using a Nikon TE200 inverted microscope with a 20× differential interference contrast objective and a Bioptechs heated stage. Images were taken with a Hamamatsu Orca camera every 5 minutes for 6 hours and collected with Openlab software (Improvision, Lexington, MA). Movies were analyzed by using Image J Manual Tracking software (National Institutes of Health, Bethesda, MD). Single cell nuclei were tracked over time, and migratory speed was calculated by dividing the length of the migration path by the total movie time. Directional persistence was determined as the net displacement divided by the total length of the migration path.
Total internal reflection fluorescence (TIRF) microscopy
Cells were transfected with siRNA oligonucleotides, mKO-paxillin, and GFP-speckle-actin by using Amaxa nucleofection, as described earlier. Seventy-two hours after transfection, cells were plated on FN (2 μg/ml), and after 20 to 30 minutes of spreading, TIRF images were acquired by using an inverted microscope, 60× objective (Olympus, model IX70). GFP was excited by using the 488-nm laser line of an Ar ion laser, and RFP was excited by using the 543-nm laser line of an He-Ne laser (Mells Griot, Carlsbad, CA). Time-lapse images were taken every 3 seconds for 5 minutes with a charge-coupled device camera (QImaging, Surrey, BC, Canada) and analyzed by using MetaMorph (MDS Analytical Technologies, Mississauga, Ontario, Canada). Protrusion rates were determined by using ImageJ Kymograph software (National Institutes of Health) and calculated as the length of the protrusion divided by the total time of the movie.
Discussion
The work presented here establishes an important, previously undiscovered role for STAT5b in the migration of highly aggressive breast cancer cells. Knockdown of STAT5b inhibits migration of BT-549 and MDA-MB-231 human breast cancer cells to serum and fibronectin (Figures
1 and
3). This inhibition can be rescued by expression of wt-, Y699F-, or dn-STAT5b, but not with the SH2 domain mutant R618K-STAT5b (Figure
2). Upon attachment to fibronectin, STAT5b-knockdown cells form multiple, contractile protrusions resulting in loss of directionality and inefficient migration (Figures
4 and
5).
Knockdown of STAT5b inhibits migration of both BT-549 and MDA-MB-231 breast cancer cell lines to serum, whereas knockdown of STAT5a has no effect (Figure
1). These results may seem contradictory to those published by Sultan and colleagues [
10], which reported a suppressive effect of STAT5a on breast cancer cell migration. However, in those studies, STAT5a was overexpressed in the moderately migratory BT-20 and T-47D breast cancer cell lines, which contain little to no endogenous STAT5a. The STAT5a and STAT5b expression pattern in those cell lines is very different from that seen in MDA-MB-231 cells, and consequently, the signaling may differ. Moreover, the role of endogenous STAT5b was not investigated in the studies of Sultan and associates. We and others found that STAT5b is the predominant STAT5a/b protein expressed in breast cancer cell lines and tissues, and that STAT5b, not STAT5a, mediates proliferation of breast cancer cells [
10,
14,
16]. Differential expression or activity (or both) of STAT5a and STAT5b has been reported in other cancer model systems [
9,
24,
25]. STAT5b levels and phosphorylation are elevated in SCCHN tumors compared with control mucosa, whereas STAT5a levels to do not change [
24]. Consistent with these findings, antisense inhibition of STAT5b, but not STAT5a, inhibits
in vivo growth of SCCHN xenografts [
24]. In prostate cancer, differential STAT5a and STAT5b protein expression can be correlated with metastatic potential. STAT5a is expressed in nonmetastatic C1D mouse prostate cancer cells, but not in their metastatic C2H counterparts, whereas STAT5b is expressed in both [
9]. Additionally, STAT5a is expressed in LNCaP human prostate cancer cells but not the more highly migratory PC-3 prostate cancer cell line, but STAT5b levels are comparable [
9]. For these reasons, it is imperative to determine the individual contributions of STAT5a and STAT5b, as they may have distinct functions in tumorigenesis.
After establishing the necessity of STAT5b for maximal migration of breast cancer cells to serum, we investigated the mechanism by which this occurs. Interestingly, the migratory function of STAT5b does not require phosphorylation at tyrosine 699 (Y699) or the C-terminal transactivation domain (TAD), evidenced by equivalent rescue of migration with either wt-, Y699F-, or dn-STAT5b (Figure
2). Phosphorylation of Y699 is a hallmark of STAT5b transcriptional activation. If this residue is mutated to a phenylalanine, such that it cannot be phosphorylated (Y699F), the resulting STAT5b mutant is transcriptionally inactive [
17,
20]. Dn-STAT5b is also transcriptionally inactive because of the loss of the TAD and the ability to bind transcriptional cofactors [
14]. Although STAT proteins are predominantly thought of as transcription factors that function in the nucleus, recent work identified a non-transcriptional, cytoplasmic role for unphosphorylated STAT3 in regulating tubulin dynamics [
26]. Unphosphorylated STAT dimers take on an antiparallel configuration through interactions between the DNA-binding domains [
27,
28]. In this conformation, the SH2 domains are on opposite sides of the structure and are free to interact with phosphorylated tyrosines of other proteins. The necessity of the SH2 domain of STAT5b in migration is consistent with a cytoplasmic function of unphosphorylated STAT5b. We propose that STAT5b uses its SH2 domain to act as a scaffolding protein, bringing together signaling molecules necessary for efficient, directional migration. Although migration is inhibited to similar levels after STAT5b knockdown or blocking of β
1-integrin (Figure
3), it remains to be determined whether STAT5b interacts directly with β
1-integrin or if this effect is indirect. There are many tyrosine-phosphorylated proteins in the cytoplasm involved in migratory signaling with which STAT5b may interact, and future studies are aimed at uncovering these associations.
Loss of STAT5b leads to a polarity defect that impedes directional movement (Figure
4). During spreading on fibronectin, STAT5b-knockdown cells take on a remarkable phenotype distinguished by the formation of multiple, unstable protrusions (Figure
5). No defect is found in initial attachment, and once attached, protrusions extend rapidly. However, these protrusions are highly dynamic, and over time, they contract back into the cell. This phenotype is indicative of disrupted equilibrium between Rho family GTPases. Rac is localized predominantly at the leading edge, with Rho in the tail. Normal ratios of Rho and Rac lead to Rac-mediated formation and spreading of broad lamellipodia in the front of the cell, followed by Rho-mediated tail retraction, overall resulting in directional migration toward a stimulus [
29]. Increased Rho activity in the front of the cell would disrupt this equilibrium and could account for the multiple protrusions and increased contraction at the front of the cell seen with STAT5b knockdown. Based on the data presented here, we postulate that unphosphorylated STAT5b mediates migration of breast cancer cells through regulation of cytoplasmic Rho GTPase family signaling.
In summary, these studies are the first to report a role for STAT5b in the migration of breast cancer cells. It is well established that STAT5b positively regulates breast cancer cell proliferation and survival, two processes important for initial tumor formation and growth. These data implicate STAT5b in the later stages of tumorigenesis also, such as migration. Future studies will further elucidate the mechanism by which STAT5b exerts its effect on migration, thereby broadening our understanding of how STAT5b promotes tumorigenesis and possibly metastasis. This will facilitate the long-term goal of defining conditions whereby STAT5b would be an effective therapeutic target for the treatment of breast cancer.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
TMB, CMS, and JTP designed the studies, and TMB performed the experiments. JZ assisted with the TIRF microscopy. All authors analyzed the data, and TMB drafted and revised the manuscript. All authors read and approved the final manuscript.