Role of tyrosine kinase Jak2 in prolactin-induced differentiation and growth of mammary epithelial cells.

Genetic studies in mice have established a critical role for prolactin receptors and transcription factor Stat5 in mammary gland differentiation. However, the enzymatic coupling between prolactin receptors and Stat5 in this process has not been established. In addition to Jak2, several other tyrosine kinases reportedly also are associated with prolactin receptors and may phosphorylate Stat5. Because Jak2 null mice die in utero, we targeted Jak2 in an ex vivo model of prolactin-induced mammary epithelial cell differentiation to determine the role of Jak2 in regulation of cell differentiation and growth. Two independent targeting strategies were used to suppress Jak2 in immortalized HC11 mouse mammary epithelial cells: 1) stable expression of a specific Jak2 antisense construct and 2) adenoviral delivery of a dominant-negative Jak2 gene. We now demonstrate that Jak2 is essential for prolactin-induced differentiation and activation of Stat5 in normal mouse mammary epithelial cells. Furthermore, suppression of Jak2 in HC11 cells was associated with constitutive activation of oncoprotein Stat3 and a hyperproliferative phenotype characterized by increased mitotic rate, reduced apoptosis, and reduced contact inhibition. Collectively, our data suggest that Jak2 is differentiation-inducing and growth-inhibitory in normal mammary epithelial cells, observations that may shed new light on the role of the Jak2-Stat5 pathway in breast cancer.

Prolactin is a principal differentiation factor for human and mouse mammary epithelial cells and is required for milk production (1). However, prolactin may also stimulate mammary epithelial cell growth and act as a mammary tumor promoter (1,2). Identifying the roles of individual signaling molecules and pathways activated by prolactin in normal mammary epithelial cells is therefore needed to better understand the role of prolactin in breast cancer.
Prolactin activates tyrosine kinase Jak2 1 (3) and transcription factor Stat5 (4) in target cells, including Nb2 lymphocytes, ovarian cells, and mammary cells. Specifically, the Jak2-Stat5 pathway is expected to mediate prolactin-induced mammary epithelial cell differentiation (5). Although genetic studies have established a critical role for Stat5 in mouse mammary gland differentiation (6,7), corresponding genetic evidence is not yet available for Jak2, because Jak2 null mice die in utero (8,9) and conditional Jak2 null mice have not been established. Although Jak2 has been regarded as the principal tyrosine kinase activated by prolactin (3), the picture has been complicated by evidence that prolactin also can activate other tyrosine kinases, including Src family kinases (10,11), focal adhesion kinase (12), Tec kinase (13), and the ErbB-2 receptor tyrosine kinase (14). Experimental testing of the importance of Jak2 for prolactin-induced differentiation of mammary epithelial cells is therefore warranted. Furthermore, because Jak2 is oncogenic in hematopoietic cells (15), it is also critical to establish the role of Jak2 in regulating growth of normal mammary epithelial cells.
In this study, we targeted Jak2 in an ex vivo model of prolactin-induced mammary epithelial cell differentiation to determine the role of Jak2 in cell differentiation and growth. Two distinct targeting strategies were used to suppress Jak2 in immortalized HC11 mouse mammary epithelial cells. First, an effective Jak2 antisense construct was generated and stably introduced into HC11 cells. Second, a functional dominantnegative Jak2 mutant was generated and introduced into HC11 cells by adenoviral delivery. We now report that Jak2 is essential for prolactin-induced differentiation and activation of transcription factor Stat5 in normal HC11 mouse mammary epithelial cells. Importantly, suppression of Jak2 in HC11 cells was associated with a hyperproliferative phenotype characterized by increased mitotic rate, reduced apoptosis, and reduced contact inhibition. In addition, constitutive activation of Stat3 was associated with suppression of Jak2 in HC11 cells. Collectively, our data suggest that Jak2 mediates growth-suppressive and differentiation-inducing effects on normal mouse mammary epithelial cells. These observations may provide important new insight into the role of the prolactin-activated Jak2-Stat5 pathway in breast cancer.
HC11 Cell ex Vivo Model of Mammary Epithelial Cell Differentia-tion-The mouse mammary epithelial cell line HC11 (16) was grown to confluence in RPMI 1640 medium (Biofluids, Rockville, MD) supplemented with 10% heat-inactivated fetal calf serum (Atlanta Biologicals, Norcross, GA), insulin (5 g/ml), and EGF (10 ng/ml). Prior to hormone treatment, HC11 cells were starved for 48 h in medium lacking EGF and containing only 2% fetal calf serum. For studies of prolactin-induced differentiation, HC11 cells were then incubated with RPMI 1640 medium supplemented with 10% fetal calf serum, dexamethasone (0.1 M), and insulin (5 g/ml) in the presence or absence of ovine prolactin (10 nM) as indicated.
Expression Vectors-Expression vector for murine Stat5a (pXM-Stat5a) was kindly provided by Xiuwen Liu and Lothar Hennighausen (National Institutes of Health, Bethesda, MD). Plasmid p3PRLR contains a 2.7-kb human prolactin receptor cDNA (kindly provided by Dr. Paul A. Kelly, INSERM, Paris, France) and subsequently cloned into the EcoRI site of pcDNA3 expression vector (Invitrogen, Carlsbad, CA) as described previously (17). Rat Jak2 cDNA was originally cloned from an Nb2-SP cell cDNA library (18), and a 3.7-kb open reading frame cDNA fragment was subcloned into the NotI and ApaI sites of pcDNA3 for expression studies.
Construction of V5/His Epitope-tagged Wild-type (Wt) and Dominant-negative (Dn) Jak2 Expression Vectors-To generate expression constructs encoding Wt-Jak2 and a kinase-deleted Dn-Jak2 with Cterminal V5/His epitope tags that could be further subcloned into an adenoviral vector with limited selection of cloning sites, a two-step strategy was used. First, a Wt-Jak2 cDNA was cloned into the pcDNA3.1/V5-His vector (Invitrogen, Carlsbad, CA) 5Ј to the V5/His sequence. Second, irrelevant intervening sequence, including Jak2 untranslated repeat was deleted to generate Wt-Jak2-V5-His, and further deletion of the entire JH1 kinase domain was carried out to generate Dn-Jak2-V5-His. Specifically, rat Jak2 cDNA containing the open reading frame was released from the pBK cloning vector (18) by digestion with ApaI and NotI. This fragment was subcloned into the ApaI and NotI sites of the pcDNA3.1/V5-His vector. To remove intervening sequence to generate V5-His-tagged Wt-Jak2, and in the case of Dn-Jak2 also to remove the JH1 domain, a modified pCR2.1 vector (Invitrogen) was used from which we had removed the KpnI site. The KpnI site in pCR2.1 vector was deleted by KpnI digestion, blunt-ending by T4 DNA polymerase, relegation with T4 DNA Ligase, and confirmation by KpnI re-digestion. The Jak2-V5-His pcDNA3.1 construct was digested with NotI and PmeI, and the resulting fragment was subcloned into the NotI-PmeI site of the modified pCR2.1 vector.
High fidelity PCR was used to 1) generate a short fragment A spanning from upstream of the AvrII site within the JH1 domain and containing a 3Ј SacII restriction site after the last Jak2 codon and 2) generate a second short fragment B spanning from upstream of the KpnI site within the JH2 domain to introduce a 3Ј SacII restriction site within the hinge region located between the JH1 and the JH2 domains. To generate a contiguous Wt-Jak2-V5-His construct, irrelevant sequence was released from the original Jak2-pcDNA 3.1/V5-His vector by AvrII and SacII digestion and replaced by correspondingly digested high fidelity PCR fragment A. For the Dn-Jak2 construct, the original Jak2-pcDNA 3.1/V5-His vector was digested with KpnI and SacII, and the released fragment was replaced by the high fidelity PCR fragment B. Finally, the Wt-Jak2-V5-His and the Dn-Jak2-V5-His genes were released by NotI and SpeI digestion, blunted by T4 DNA polymerase, and cloned into blunt-ended NotI and XbaI restriction sites of the pcDNA3 vector. All cloning was verified by DNA sequencing.
Antisense Jak2 Construct and Generation of Stably Expressing HC11 Clones-After testing several alternatives, an effective and specific Jak2 antisense oligonucleotide was determined as 5Ј-TGTCTTCAAAA-GCACCAGAAAATCCTAGGGCACCTATTCTCATGTTGGGTA-3Ј. This targeting sequence was verified to be unique to Jak2 by searching the NCBI GenBank TM data base. The sequence targets a region of Jak2 mRNA encoding the amino acid sequence PNMRIGALGFSGAFEDR within the hinge region between the JH1 and JH2 domains (18). The sense control nucleotide sequence is 5Ј-TACCCAACATGAGAATAGGT-GCCCTAGGATTTTCTGGTGCTTTTGAAGACA-3Ј. Both the antisense and sense Jak2 DNA were generated by PCR, and a 5Ј EcoR V flanking site was introduced for orientation identification and subsequent cloning into pCR2.1. The sense and antisense Jak2 DNA were released by HindIII and NotI digestion and were subcloned into the pcDNA3 vector at the multiple cloning site. The purified antisense and sense Jak2 cDNA-pcDNA3 constructs were transfected into HC11 cells using LipofectAMINE 2000 (Invitrogen, Carlsbad, CA), and stable clones were isolated after 10 days of treatment with 300 g/ml G418 (Invitrogen). The clones were validated with neomycin resistance gene PCR product analysis.
Cell Culture and Transient Transfections-Construct expression and functional tests were performed by transient transfection of COS-7 cells (ATCC, Manassas, VA). COS-7 cells were grown in RPMI 1640 medium containing 10% fetal calf serum, 2 mM L-glutamine, and penicillinstreptomycin (50 IU/ml and 50 g/ml, respectively) at 37°C with 5% CO 2 . Sub-confluent COS-7 cells in 6-well plates were transfected using LipofectAMINE 2000 according the manufacturer's protocol and were kept without fetal calf serum for 24 -48 h followed by stimulation of 10 nM prolactin for 30 min. The harvested cell pellets were frozen on dry ice and stored at Ϫ80°C.
Protein Solubilization, Immunoblotting, and Immunoprecipitation-For protein solubilization, the cell pellets were solubilized in lysis buffer (10 mM Tris-HCl, pH 7.6, 5 mM EDTA, 50 mM NaCl, 30 mM sodium pyrophosphate, 50 mM sodium fluoride, 1 mM sodium orthovanadate, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 5 g/ml aprotinin, 1 g/ml pepstatin A, and 2 g/ml leupeptin). Cell lysates were rotated end-over-end at 4°C for 60 min, and insoluble material was pelleted at 12,000 ϫ g for 30 min at 4°C. For immunoprecipitations, the protein concentrations of clarified tissue lysates were determined by simplified Bradford method (Bio-Rad Laboratories, Hercules, CA). Clarified lysates corresponding to 3.5 mg of total protein were immunoprecipitated by rotation for 2 h at 4°C with the appropriate antibodies. Antibodies were captured by incubation for 60 min with protein A-Sepharose beads (Amersham Biosciences, Inc., Piscataway, NJ), and washed three times in 1 ml of lysis buffer. Immunoprecipitated proteins were dissolved in 1.1ϫ loading buffer containing reducing agent. The proteins were analyzed by SDS-polyacrylamide gel electrophoresis and immunoblotting as described previously using polyvinylidene difluoride membranes (Millipore, Bedford, MA) and horseradish peroxidase-conjugated secondary antibodies in conjunction with enhanced chemiluminescence substrate mixture (Amersham Biosciences, Inc.) and exposed to x-ray film.
Dominant-negative Jak2 Recombinant Adenovirus-A replicationdefective human adenovirus (Ad5) carrying Dn-Jak2 was generated using the AdEasy Vector system (Qbigene, Carlsbad, CA). Briefly, the sequence-verified and expression-confirmed Dn-Jak2-V5/His expression cassette was released by ApaI and NotI digestion. Before NotI digestion, the ApaI digested ends were blunt-ended by T4 DNA polymerase. The fragment was then subcloned into the NotI and EcoRV sites of Adv-shuttle vector. After creating a recombinant Dn-Jak2-V5/His Adv gene using pAdEasy, the virus was packaged in QBI-293A cells and subsequent plaques were isolated. Expression of Dn-Jak2 was verified by Western blotting using anti-V5 and anti-Jak2 antibodies. The selected recombinant virus was expanded, purified, and titered in QBI-293A cells as per the manufacturer's recommendation. Two additional recombinant adenoviruses were used as controls: Adv-LacZ, which specifies a nuclear-localized form of ␤-galactosidase, and Adv-Control, which does not express a protein.
Flow Cytometry-HC11 cells were washed once in PBS, trypsinized, pelleted at 1000 ϫ g, and washed once in 5 ml of cold PBS. After a second centrifugation, cells were resuspended in 0.5 ml of cold PBS and fixed by dropwise addition of 1.5 ml of cold 100% ethanol, while slowly vortexing the cell suspension. After having been fixed for 1 h at 4°C, cells were stained with 100 g/ml propidium iodide (Roche Molecular Biochemicals, Indianapolis, IN) and treated with 100 ng/ml RNase A (Invitrogen) for 30 min at 37°C. The cells were measured for DNA content by flow cytometry using a Coulter EPICS XL cell analyzer (Beckman-Coulter, Brea, CA).
TUNEL Assay-In situ detection of apoptotic cells was performed using terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling (TUNEL). Cells were air-dried on glass slides and fixed in 4% paraformaldehyde for 20 min at room temperature and permeabilized with 0.1% Triton X-100 and 0.1% sodium citrate for 1 min on ice. The slides were rinsed with PBS several times, and the samples were then processed for TUNEL labeling using the fluorescein-based In Situ Cell Death Detection kit (Roche Molecular Biochemicals, Indianapolis, IN) according to the manufacturer's instructions. Samples were rinsed three times with PBS, mounted, and analyzed under a fluorescence microscope.
Anchorage-independent Survival Analysis-Confluent HC11 cells were trypsinized into a single cell suspension. A total of 700,000 cells per group were plated in 150-mm culture dishes that had previously been coated with 0.8% agarose. Cells were collected at various time points and washed in PBS, and cell aggregates were dispersed by trypsinization. Parallel samples were analyzed for apoptosis by TUNEL staining and flow cytometry for hypodiploid cells.
The cells were gently detached by a cell scraper in PBS. The detached cells were stretched into monolayer sheets in warm PBS and adhered to poly-lysine pretreated glass slides. Before immunocytochemical staining, sample slides were pretreated with an antigen-unmasking procedure by boiling in an antigen-retrieval solution for 10 min. The slides were incubated at 4°C overnight by using a 1:2000 dilution of the primary anti-phosphoTyr-Stat5 monoclonal antibody AX1 (Advantex Bioreagents, Conroe, TX). For secondary detection the Histomouse kit (Zymed Laboratories Inc., South San Francisco, CA) was used, and active Stat5 was visualized with aminoethyl carbazole and counterstaining with hematoxylin. The prolactin-treated COS-7 cells cotransfected with Stat5a and prolactin receptor expression constructs were used as positive control, and subtype-specific mouse IgG and PBS were used for negative controls (not shown).

Prolactin-induced Differentiation of HC11 Mouse Mammary Epithelial Cells Correlates with Activation of Tyrosine Kinase
Jak2-Confluent, growth-arrested HC11 mouse mammary epithelial cells can be induced to differentiate in vitro by prolactin in medium supplemented with glucocorticoids and insulin (16). This differentiation process leads to formation of mammospheres, which are acinar-like structures that have been shown to express milk proteins (19). HC11 cells have been widely used as an ex vivo model of mammary gland epithelial cell differentiation (20,21). We took advantage of this model to determine whether Jak2 was critical for prolactin-induced differentiation.
The time-dependent differentiation of HC11 cells induced by prolactin as measured by the appearance of mammospheres is presented in Fig. 1A. Mammospheres were detectable within 1 day of prolactin treatment, and additional mammospheres continued to form over a period of 4 to 5 days of culture, reaching a plateau at a density of ϳ15 mammospheres per cm 2 . Although glucocorticoids and insulin are required supplements in the differentiation medium, mammosphere formation critically requires prolactin as demonstrated by a concentration-dependent effect of prolactin ranging from 0 to 20 nM (Fig. 1B).
To determine expression and activation patterns of Jak tyrosine kinases at the initiation of differentiation treatment, HC11 cells were treated with or without prolactin for 30 min and harvested. Individual Jak kinases were immunoprecipitated from cell lysates and immunoblotted for phosphotyrosine and reprobed for Jak protein levels. These analyzes showed that prolactin treatment of HC11 cells correlated with selective activation of Jak2 and not of other members of the Jak tyrosine kinase family (Fig. 1C). Specifically, phosphotyrosine immunoblotting of immunoprecipitated Jak proteins established that only Jak2 became detectably tyrosine-phosphorylated in response to prolactin. Furthermore, of the four Jak kinases, Jak2 was the only Jak family member expressed at significant levels in HC11 cells. Control experiments verified that the antibodies used for immunoprecipitation and immunoblotting of the various Jak tyrosine kinases were effective against mouse isoforms (data not shown). We therefore conclude that prolactininduced differentiation of HC11 mammary epithelial cells correlated with selective activation of Jak2 tyrosine kinase.
Jak2 Antisense Blocks Prolactin-induced Differentiation of Stably Transfected HC11 Clones-As one strategy to suppress Jak2 function and test the importance of Jak2 in prolactininduced differentiation of HC11 cells, we generated an antisense construct to inhibit Jak2 protein expression. The genetic engineering of this Jak2 antisense construct, which targets a region of 51 bp unique to the Jak2 transcript, is described in detail under "Materials and Methods." This region is located within the hinge region between the JH2 pseudokinase and JH1 kinase domains of Jak2. Functional testing of the antisense construct was first carried out using COS-7 cells and cotransfection experiments with a V5-His tagged Wt-Jak2 construct. We determined that the Jak2 antisense construct effec-tively blocked Wt-Jak2 expression in a dose-dependent manner under conditions where total amounts of DNA transfected were kept constant ( Fig. 2A). Equivalent amounts of cell lysates in the various lanes were verified by parallel blotting for ␤-actin. Furthermore, the control sense construct did not affect Jak2 levels in parallel experiments (data not shown).
To directly test whether Jak2 is required for prolactin-induced mammary epithelial cell differentiation, we then generated stable HC11 clones expressing the Jak2 antisense con-FIG. 1. Prolactin-induced differentiation of HC11 mammary epithelial cells correlates with activation of Jak2, and not other Jaks. A, time-dependent stimulation of HC11 cell differentiation by prolactin (Prl). HC11 cells were stimulated with prolactin (10 nM) for up to 7 days, and density of mammospheres was recorded by manual counting under phase-contrast microscopy. Data presented represent three independent experiments carried out in duplicate (error bars represent S.E.). See "Materials and Methods" for detailed culture conditions. B, concentration-dependent stimulation of HC11 cell differentiation by prolactin. HC11 cells were incubated with concentrations of prolactin ranging from 0 to 20 nM for 4 days, and density of mammospheres was recorded. Data represent three independent experiments carried out in duplicate (error bars represent S.E.). C, prolactin selectively activated tyrosine kinase Jak2, and not other Jaks, in HC11 cells. HC11 cells at day 0 of differentiation treatment were exposed to prolactin (10 nM) for 30 min. Jak1, Jak2, Jak3, and Tyk2 were individually immunoprecipitated (IP) with specific antibodies and first subjected collectively to anti-pTyr immunoblotting and subsequently reprobed individually for corresponding Jak protein levels. struct. A vector sense-control clone (HC11-Vs) was also selected for use as a second control cell line in addition to the parental line. Stable clones A and B were selected for further study from several positive clones, and showed markedly reduced Jak2 protein levels by immunoprecipitation and Western blot analysis (Fig. 2B, lower panel). Correspondingly, basal and prolactin-induced Jak2 tyrosine phosphorylation was lost in clones A and B (Fig. 2B, upper panel). When examined in differentiation assays, suppression of Jak2 levels in clones A and B was associated with complete disruption of prolactin-induced mammosphere differentiation, as illustrated by representative images of cultures on day 4 of treatment (Fig. 2C). In contrast, mammosphere formation remained intact in parental HC11 and vector-sense control cells. Mammosphere formation was quantified by counting, and the data from three independent experiments were expressed as number of mammospheres per cm 2 of culture surface (Fig. 2D). The inhibitory effect of Jak2 suppression on cell differentiation was equally pronounced after extended treatment with prolactin for up to 7 days (data not shown). We therefore conclude that suppression of Jak2 levels by stable expression of a Jak2 antisense construct blocked prolactin-induced differentiation of HC11 cells.
Construction and Functional Testing of a Dominant-negative Jak2 Mutant-To further test by an independent strategy whether Jak2 was essential for prolactin-induced mammary differentiation, we generated a dominant-negative Jak2 protein by deletion of the C-terminal kinase domain as described under "Materials and Methods." As a functional test of this construct, transient transfection assays in COS-7 cells were used to examine the ability of this kinase-deleted mutant Jak2 to block prolactin-induced activation of Stat5 by Wt-Jak2. COS-7 cells were cotransfected with plasmids encoding PrlR, Stat5a, and Jak2 forms as indicated and stimulated with or without prolactin for 30 min (Fig. 3A). In whole cell lysates of mock transfected, negative control cells, immunoblotting revealed no detectable endogenous Stat5, and therefore no response to prolactin stimulation (Fig. 3A, lanes a and b). When Stat5a and prolactin receptor were cotransfected into COS-7 cells, modest but detectable prolactin-stimulated Stat5a phosphorylation was observed, presumably mediated by low levels of endogenous Jak2 (Fig. 3A, lanes c and d).
This inducible Stat5a activation was inhibited by cotransfection with Dn-Jak2, which migrated with the expected size of 90 kDa in SDS-PAGE (Fig. 3, lanes e and f; top and bottom panels, respectively). Cotransfection of Wt-Jak2 with Stat5a led to basal tyrosine phosphorylation of Stat5a that was markedly enhanced by prolactin treatment (lanes g and h). However, further cotransfection of Dn-Jak2 with Wt-Jak2 showed complete inhibition of both basal and prolactin-induced Stat5a activation (lanes i and j). This effect was not due to reduced levels of Stat5 or Wt-Jak2 as demonstrated by reprobing with anti-Stat5 or anti-V5 antibodies, respectively (Fig. 3A, middle  and lower panels). Furthermore, these and other immunoblotting experiments showed that Dn-Jak2 effectively inhibited Wt-Jak2 function at equivalent protein levels, providing direct evidence for dominant rather than a simple competitive inhibitory effect of the kinase-deleted Jak2 mutant. Thus, we conclude that the engineered Dn-Jak2 functioned as predicted.
Adenoviral Delivery of Dn-Jak2 into HC11 Cells Blocks Prolactin-induced Differentiation-As the second approach to effectively inhibit the Jak2 tyrosine kinase in HC11 cells, we then generated a replication-defective adenovirus for high efficiency gene delivery of Dn-Jak2 into HC11 cells. Detailed description of this construct is presented under "Materials and Methods." Functional testing of Adv-Dn-Jak2 was first carried out in HC11 cells using prolactin-stimulated Jak2 tyrosine phosphorylation as a readout. Cells were mock infected or infected with virus carrying no insert (Adv-control), or Adv-Dn-Jak2 at increasing multiplicity of infection (m.o.i.). Twentyfour hours later, cells were incubated with or without prolactin for 30 min, and Jak2 phosphotyrosine levels were examined. In HC11 cells, both basal and prolactin-activated Jak2 tyrosine phosphorylation was inhibited by Dn-Jak2 at m.o.i. values of 5 and 25, whereas infection with control virus had no effect. Reblotting of samples for V5-tagged Dn-Jak2 protein verified specific and dose-dependent expression of Dn-Jak2 in HC11 cells (Fig. 3B) and that Adv-Dn-Jak2 was functional.
To determine whether Dn-Jak2 would block prolactin-induced differentiation of HC11 cells, cells were infected with or without Adv-DN-Jak2 as described and mammosphere formation in response to prolactin treatment was monitored. While prolactin-induced differentiation remained intact in mock and Adv-control infected cells, Dn-Jak2 effectively disrupted prolactin-induced mammosphere formation in HC11 mammary cells (Fig. 3C). The effect of Dn-Jak2 was dose-dependent and could not be attributed to general protein overexpression, because infection with Adv-LacZ did not disrupt mammosphere formation (Fig. 3D). Therefore, based on two independent approaches that involved either Dn-Jak2 or antisense to inactivate Jak2, we conclude that Jak2 is required for terminal differentiation of mammary epithelial cells. To our knowledge, these data provide the first direct evidence that Jak2 tyrosine kinase activity is critical for prolactin-induced differentiation of mammary epithelial cells.

Disruption of Jak2 Activity Is Associated with Inhibition of Stat5a Tyrosine Phosphorylation in HC11 Cells-Transcription
factor Stat5 is critical for terminal differentiation of mammary cells and for lactogenesis as determined from genetic studies in mice (6,7). Furthermore, Stat5 is recognized to be a substrate of Jak2 in the context of the prolactin receptor complex (22). To experimentally determine whether inhibition of Jak2 would block prolactin-induced Stat5 activation in HC11 cells, we tested the effect of Adv-Dn-Jak2 on prolactin-induced Stat5 tyrosine phosphorylation. Stat5 activation was measured by protein immunoblotting of samples from whole cell lysates using a monoclonal anti-phosphotyrosine-Stat5 antibody (Fig.  4A). Whereas prolactin-induced Stat5 activation was readily detectable in mock-infected cells or cells infected with Advcontrol (lanes a-d), prolactin-induced Stat5 activation was inhibited in a dose-dependent manner in cells infected with Adv-Dn-Jak2 (Fig. 4A, lanes e-h, upper panel). Furthermore, inhibition of Stat5 activation correlated with Dn-Jak2 levels as detected by anti-V5 immunoblotting and was not due to reduction in Stat5 protein levels (Fig. 4A, middle and bottom panels,  respectively).
Inhibition of prolactin-induced Stat5 tyrosine phosphorylation by Jak2 suppression was verified at the subcellular level by anti-phosphoTyr-Stat5 immunocytochemistry. HC11 cells infected with Adv-control and treated with prolactin showed marked nuclear tyrosine phosphorylation of Stat5, whereas infection of cells with Adv-Dn-Jak2 markedly inhibited Stat5 phosphorylation (Fig. 4B, panels 1 and 2). Likewise, HC11 cell clones stably expressing antisense-Jak2 displayed only minor levels of prolactin-stimulated Stat5 phosphorylation as detected by anti-phosphoTyr-Stat5 immunocytochemistry (Fig.  4B, panel 3; only clone A shown). Finally, infection of clone A with Adv-Dn-Jak2 led to an even more pronounced inhibition of prolactin-induced Stat5 tyrosine phosphorylation (Fig. 4, panel  4). Therefore, both molecular approaches to inhibit Jak2 activity also inhibited prolactin-induced Stat5 activation. Collectively, these observations support the concept that the Jak2-Stat5 pathway is a differentiation-inducing axis in mammary epithelial cells.
Targeted Inactivation of Jak2 in HC11 Cells Resulted in a Hyperproliferative Phenotype-Terminal differentiation of cells is associated with exit from the cell cycle and inhibition of cell proliferation. To determine the effect of Jak2 on growth characteristics of HC11 cells, we compared the growth rates of HC11 clones A and B to those of parental HC11 cells and the sense-control clone. As shown in Fig. 5A, HC11 clones A and B exhibited significantly higher growth rates than parental HC11 or the sense-control clone. In fact, the growth rates of clones A and B stably expressing Jak2 antisense were approximately double that of parental or sense-control HC11 cells.
The increased growth rates of Jak2-suppressed clones A and B were also correlated with increased rates of mitosis during exponential growth, as visualized by propidium iodide staining of cells (Fig. 5B). Flow cytometry of cells during exponential growth verified a markedly increased proportion of cycling cells with a corresponding reduction in cells in the G 0 /G 1 phase in clones A and B when compared with control cells (Fig. 5C). Specifically, whereas almost 50% of parental or vector-sense control cells were in G 0 /G 1 , less than 25% of clones A and B were in G 0 /G 1 . For these cell cycle experiments, cells were harvested and measured at ϳ50% confluency. The data suggest that suppression of Jak2 is associated with increased cycling of HC11 cells and higher growth rate during exponential growth.
We then examined the cell cycle characteristics of superconfluent cultures. Intriguingly, Jak2-deficient clones A and B consistently grew to a density nearly 3-fold greater than that of parental or vector sense-control HC11 cells (Fig. 6A). Furthermore, during superconfluency, Jak2-deficient clones displayed reduced growth suppression and retained a markedly elevated S phase population compared with parental HC11 cells and vector sense-control cells (Fig. 6B). Specifically, 15-21% of parental or vector sense-control cells were in S-phase during superconfluent conditions, whereas ϳ30% of antisense Jak2 clones A and B remained in S-phase.
We also introduced Dn-Jak2 into HC11 cells by adenoviral gene transfer and assessed its effect on cell cycling. Superconfluent cells that had been serum-deprived for 48 h were cycling only to a very low extent (Ͻ4% in S or G 2 /M) in mock infected or control virus-infected cells (Fig. 6C, left two panels). However, a dose-dependent increase in the fraction of cycling cells was observed in superconfluent, serum-deprived cells overexpressing Dn-Jak2 (Fig. 6B, right two panels). These observations are consistent with a general growth-suppressive effect of the Jak2-Stat5 pathway in HC11 cells. The data are also consistent with reduced contact inhibition following suppression of Jak2.
Inhibition of Jak2 Suppresses Apoptosis of HC11 Cells Induced by Anchorage-independent Culture Conditions-The HC11 cell is a nontransformed mammary epithelial cell line that does not survive under anchorage-independent conditions. To investigate whether inhibition of Jak2 would affect the rate of apoptosis induced by culture under anchorage-independent culture conditions, we first examined apoptosis rates in HC11 cells stably expressing antisense Jak2. Parental HC11 cells, Vs-control cells, and clones A and B were cultured on 0.8% agar in normal growth medium, collected after 12 and 36 h, and assayed for apoptosis by flow cytometry.
The hypodiploid fraction of HC11 cells at 12 h was markedly lower in Jak2-suppressed clones A and B than in parental or vector control cells (Fig. 7A), and although the number of apoptotic cells increased in both control cells and Jak2-suppressed cells over the next 24 h, Jak2-suppressed cells showed consistently reduced rates of apoptosis. Examination of apoptosis in parallel samples using TUNEL staining of fragmented DNA at 12 h verified reduced number of apoptotic cells in Jak2-suppressed clones, whereas phase-contrast and DAPI staining verified equal cell numbers in the selected fields (Fig.  7B). However, extended cultures revealed that suppression of Jak2 levels in HC11 cells did not confer long term survival under anchorage-independent culture conditions (data not shown).
Parallel studies of HC11 cells grown under anchorage-dependent conditions also provided evidence for reduced apoptosis following delivery of Dn-Jak2 into confluent cultures, as demonstrated by TUNEL staining of cells undergoing DNA fragmentation (Fig. 7C). Representative fields from cell cultures presented by phase-contrast, TUNEL, and DAPI staining, showed that a larger number of HC11 cells undergo apoptosis when infected with Adv-control compared with Adv-Dn-Jak2 (Fig. 7C). We conclude from these experiments that suppression of Jak2 in HC11 cells inhibits cellular apoptosis under several culture conditions. Thus, the hyperproliferative phenotype of HC11 cells associated with suppression of Jak2 also involved anti-apoptotic elements.
The Hyperproliferative Phenotype Resulting from Jak2 Sup- Twenty-four hours later all cells were exposed to prolactin for 30 min, followed by fixation and immunocytochemistry for activated Stat5 using anti-Stat5 pTyr antibodies.
FIG. 5. Suppression of Jak2 tyrosine kinase in Jak2 antisense expressing HC11 cell clones is associated with a hyperproliferative phenotype. A, increased growth rate in Jak2 antisense expressing HC11 clones A and B. The growth rates of parental HC11 cells, vector-sense (Vs) control expressing control clone, and Jak2 antisense-expressing clones A and B were compared by plating cells at the same low density and following cell numbers over 72 h. Cell numbers were counted manually in a hemacytometer, and the data represent means of three independent experiments (S.E. indicated by bars). B, increased growth rate in Jak2-suppressed HC11 cells was associated with increased number of mitotic figures. Exponentially growing cultures of parental HC11 cells, vector-sense (Vs) control expressing control clone, and Jak2 antisense expressing clones A and B were stained with propidium iodide to better visualize dividing cells (indicated by arrows). C, increased growth rate in Jak2-suppressed HC11 cells was associated with increased number of cycling cells by flow cytometry. Exponentially growing cultures of parental HC11 cells, vector-sense (Vs) control expressing control clone, and Jak2 antisense-expressing clones A and B were stained with propidium iodide and analyzed by flow cytometry to determine fraction of cells in the various stages of cell cycle. pression in HC11 Cells Is Associated with Constitutive Activation of Stat3-The observed hyperproliferative and undifferentiated phenotype of HC11 mammary epithelial cells with disrupted Jak2-Stat5 signaling is intriguing in light of the general loss of differentiation associated with progressing breast cancer cells. Because Stat3 has been shown to be an oncogene (23) and to be constitutively activated in human breast cancer (24), we examined the effect of Jak2 suppression on basal Stat3 activation in HC11 cells. Interestingly, Western blot analysis of whole cell lysates from subconfluent HC11 cells indicated that Stat3 was constitutively active in hyperprolif-erative clones A and B, as determined by increased Stat3 phosphotyrosine levels in the absence of increased Stat3 levels (Fig. 8A).
Similar results were obtained with introduction of Dn-Jak2 into HC11 cells by adenoviral transfer. Dn-Jak2 also led to constitutive activation of Stat3 as measured by Stat3 phosphotyrosine and Stat3 protein immunoblotting (Fig. 8B). The effect was dose-dependent and was not induced by infection with control adenovirus. Based on two independent molecular approaches, we conclude that the hyperproliferative phenotype of Jak2-suppressed HC11 cells correlated with constitutive activation of Stat3. Thus, the possibility exists that constitutive activation of Stat3 is involved in the hyperproliferative phenotype of the cells associated with Jak2 suppression. Further investigations into the role of Stat3 in proliferation of HC11 cells, and the relationship between Stat3 activation and suppression of the Jak2-Stat5 pathway, are underway. DISCUSSION The present study used adenoviral delivery of dominantnegative Jak2 and stable expression of Jak2 antisense mRNA to identify Jak2 as a critical mediator of prolactin-induced differentiation of nontransformed HC11 mammary epithelial cells. The associated disruption of prolactin-induced Stat5 activation most likely represents the key molecular mechanism responsible for disrupting prolactin-induced differentiation. Furthermore, targeted suppression of Jak2 in HC11 cells led to a hyperproliferative phenotype, suggesting that Jak2 exerts a growth-inhibitory influence on normal mammary epithelial cells. Whereas suppression of Jak2 did not lead to transformation of HC11 cells, Jak2 suppression did, however, reduce contact inhibition and extend cell survival under anchorage-independent growth conditions. Stat3 was constitutively activated in hyperproliferative, Jak2-suppressed HC11 mammary epithelial cells, and this activation of Stat3, a known oncogene (23), may contribute to the hyperproliferative phenotype.
Jak2 as a Mediator of Prolactin-induced Differentiation-We originally identified Jak2 as the prolactin-associated tyrosine kinase in Nb2 lymphoma cells (3,25). Although other Jak tyrosine kinases reportedly are not activated by prolactin receptors, numerous studies indicate that prolactin may activate other, non-Jak tyrosine kinases. These include tyrosine kinases Fyn (10), Src (11), Tec (13), and focal adhesion kinase (12). In most cases, temporal activation profiles suggest that prolactin activates Jak2 upstream of the non-Jak tyrosine kinases, but at least one report indicated that prolactin can activate Src independent of Jak2 (26). Consequently, the notion that all prolactin-induced effects are mediated by Jak2 may not be correct. Therefore, it was of particular importance to determine whether Jak2 mediates prolactin-induced differentiation of mammary epithelial cells.
The present study now provides direct evidence that Jak2, in fact, is critical for mammary epithelial cell differentiation. After establishing that prolactin activated Jak2, and not other Jaks, in HC11 cells, and that concentration-dependent induction of cell differentiation by prolactin correlated with Jak2 activation, two independent strategies were used in vitro for targeted suppression of Jak2. These included construction of two sets of novel reagents, a vector for stable or transient expression of a novel and specific Jak2 antisense-mRNA, and a replication-defective adenovirus for efficient gene delivery of Dn-Jak2. Both sets of molecular tools were validated and determined to be effective in independent experiments, and both strategies independently demonstrated that Jak2 is critical for prolactin-induced differentiation by selectively disrupting mammosphere formation. Disruption of prolactin-induced differentiation was not associated with any general cytotoxic effects, because suppression of Jak2 by either strategy led to increased proliferation rates and reduced cellular apoptosis. Furthermore, the described molecular tools may be applied to determine whether Jak2 is critical for biological effects of other cytokine receptors in a variety of cell types.
Evidence That Jak2 Phosphorylates and Activates Stat5 in HC11 Cells-Transcription factor Stat5, and especially the Stat5a isoform, is critical for mammary gland differentiation (6). Stat5 is phosphorylated on a single tyrosine residue following prolactin receptor activation, and this modification causes Stat5 to dimerize, which in turn is needed for DNA binding and transcriptional regulation (4). Jak2 is presumed to mediate prolactin-induced tyrosine phosphorylation of Stat5 (22), although Stat5 can also be phosphorylated by the Src tyrosine kinase (27). In the present study, we determined that suppression of Jak2 activity blocked prolactin-stimulated Stat5 tyrosine phosphorylation in HC11 cells, both by immunoblotting and by immunocytochemistry using an antiphosphotyrosine-Stat5 antibody. Specifically, in Jak2suppressed HC11 clones, Stat5 activation was significantly down-regulated. Furthermore, Stat5 activation was inhibited in a dose-dependent manner by Dn-Jak2 delivered by adenoviral infection. These findings support a view of the Jak2-Stat5 pathway as a differentiation-inducing axis in mammary epithelial cells.
Targeted Inactivation of Jak2 Resulted in Hyperproliferative Phenotype of HC11 Cells-The Jak2-deficient HC11 clones showed significantly increased growth rates, loss of contact inhibition, and prolonged survival under anchorage-indepen-dent culture conditions. Although suppression of Jak2 in HC11 cells was associated with loss of Stat5 activation, Stat3 became constitutively tyrosine-phosphorylated. Constitutive activation of Stat3 was observed both in HC11 clones stably expressing the Jak2 antisense construct and in cells overexpressing Dn-Jak2 by adenoviral delivery, raising the possibility that the Jak2-Stat5 pathway normally inhibits Stat3 activation in HC11 cells. Interestingly, a similar mutual exclusion of Stat5 activation and Stat3 activation has been observed in mammary epithelial cells within the physiological setting of mammary gland involution (28). Weaning, or artificially induced milk stasis, rapidly shuts off Stat5 activation in lactating mammary epithelial cells, and Stat3 becomes activated instead (29). Although the importance of Stat3 activation in hyperproliferative HC11 cells remains to be determined, this observation is particularly intriguing in light of the established tumor-promoting role of Stat3 (23).
In summary, the present work demonstrated that Jak2 is critical for prolactin-induced differentiation of HC11 mouse mammary epithelial cells. Equally important, the data are also consistent with an overall growth-inhibitory role of the Jak2-Stat5 pathway in mammary epithelial cells. This notion is of direct relevance to mammary tumorigenesis, because the data suggest a tumor-suppressive role of the Jak2-Stat5 pathway in the mammary gland. In general, cancer cells are characterized by enhanced growth and reduced levels of differentiation. Future work will address the possibility that progression of mammary and breast tumors involve a gradual loss of Jak2-Stat5 signaling, with a consequent increase in Stat3 activation.