Differential interleukin-6/Stat3 signaling as a function of cellular context mediates Ras-induced transformation

The pBabe-H-RasV12 construct was a gift from P. Sicinski (Dana-Farber Cancer Institute) [21]. Stat3shRNA lentiviral and scrambled control shRNA constructs were previously described [22]. The pSuper-Il-6 shRNA-GFP retroviral construct was generated by substituting PKG-puro with CMV-GFP. pSuper IL-6 shRNA was a gift from C. Counter [23]. Nuclear and cytoplasmic extracts were prepared as previously described [24]. Radioimmunoprecipitation assay (RIPA) buffer extracts were utilized in the protein extraction of all tissues and Western blots were carried out as previously described [25]. Protein concentrations were determined using the Bradford assay (BioRad, Hercules, CA, USA). EMSAs were carried out as previously described [26] by using a radiolabeled high-affinity m67 DNA binding probe and an anti-Stat3 antibody for supershifting (Stat3 (K-15): sc-483 X, Santa Cruz Biotechnology, Santa Cruz, CA, USA). RNA was isolated using the RNeasy kit (QIAGEN, Valencia, CA, USA). Two micrograms of total RNA was used for IL-6 and β-actin RT-PCR using an iScript RT-PCR kit (Bio-Rad) according to the manufacturer's instructions. Sequences of primers for amplification of the IL-6 gene were as follows: forward primer, 5'-TAGCCGCCCCACACAGACAG-3'; reverse primer, 5'-GGCTGGCATTTGTGGTTGGG-3'. The β-actin primers were as follows: forward primer, 5'-CGTGCGTGACATTAAGGAGA-3'; reverse primer, 5'-TGATCCACATCTGCTGGAAG-3'. For quantitative PCR 1 ug of total RNA was reverse-transcribed using the Thermoscript RT-PCR system (Invitrogen) at 52°C for one hour. 20 ng of resultant cDNA was used in a Q-PCR reaction using an iCycler (Biorad) and pre-designed TaqMan ABI Gene expression Assays (Hs00985639_m1 for IL-6). Amplification was carried for 40 cycles (95°C for 15 seconds, 60°C for 1 minute). To calculate the efficiency of the PCR reaction, and to assess the sensitivity of the assay, we also performed a seven-point standard curve (5, 1.7, 0.56, 0.19, 0.062, 0.021, and 0.0069 ng). To obtain normalized qPCR values for IL-6, triplicate cycle threshold values were averaged, amounts of target were interpolated from the standard curves and normalized to HPRT (hypoxanthine guanine phosphoribosyl transferase, assay Hs99999909_m1). Antibodies used were: anti-tubulin monoclonal antibody (1:2500) (Sigma, St. Louis, MO, USA); Anti-H-Ras polyclonal antibody (1:1000) (Santa Cruz Biotechnology); Anti-Stat 3 and anti-Tyr 705 Stat3 polyclonal antibodies (1:1000) (Cell Signaling, Danvers, MA, USA) and Anti-E-cadherin monoclonal antibody (1:1000) (Zymed Laboratories, San Francisco, CA, USA).

MCF10A and 293T cells were obtained from the American Type Culture Collection. MCF10A cells were cultured as previously described [27]. 293T cells were grown in Dulbecco's modified Eagle medium (DMEM) containing 10% Cosmic Calf serum (Hyclone, Logan, UT, USA). Cells were plated on plastic coated Matrigel, Fibronectin, Laminin, CollagenI, CollagenV (BD Biosciences). Human IL-6 was used at 10 ng/ml (R&D Systems, Minneapolis, MN, USA). To measure cell growth, 2 × 10⁴ cells per well (2 × 10³/cm²) were plated into six-well dishes. Cells were counted at Days 1, 2, 3, 4, 5 and 6. Each data point represents the mean value from triplicate wells. All transfections were carried out using Superfect (QIAGEN). Retroviral infections were carried out using the RetroMax Retroviral Expression System pCL-Ampho (Imgenex, San Diego, CA, USA) according to the supplier. Clonal selection was carried out using puromycin (2 ug/ml) (Sigma). Lentiviral infections were performed using lentivirus-based vectors encoding shRNA IRES eGFP targeted to either Stat3 or IL-6 transcripts, and were generated by transient co-transfection of 293T cells with a three-plasmid combination, as described previously [28]. EGFP expressing cell populations positive for each shRNA were sorted by FACS.

Migration/invasion experiments were performed as described previously [29, 30]. Briefly, cells (5 × 10⁴) were starved for 16 hours in media lacking EGF and were subsequently applied to an 8-μm pore cell culture inserts or in Matrigel Invasion Chambers (Falcon/BD Labware, Franklin Lakes, NJ, USA). MCF10A media containing 2% horse serum and EGF was used as the chemoattractant with or without IL-6 (10 ng/ml). After 16 h of incubation, the cells were stained with crystal violet and counted. Each condition was assayed in triplicate, experiments were performed independently at least three times, and the results were expressed as the number of cells per field.

Morphogenesis assays were carried out as previously described [27]. Growth Factor Reduced Matrigel was obtained from BD Biosciences (BD No.354230). For inhibitor incubation, the following concentrations of inhibitor or antibody were used: 1 μM P6 (Calbiochem, San Diego, CA, USA), 10 μg/mL BR3 (Diaclone, Stamford, CT, USA) and 2 μg/mL anti-IL-6 522 (Cell Sciences, Conton, MA, USA). Antibodies for staining of acini were as follows: anti-phospho Stat3 # 9135, 1:50 (Cell Signaling); anti-E-Cadherin 1:200 (Zymed Laboratories). Examination of the staining of cells was carried out using a Leica inverted confocal microscope.

Anchorage-independent growth in triplicate was assessed as previously described [17]. Cells (4 × 104) in 4 ml of a 0.35% agar-MCF10A media solution were plated in triplicate on 35-mm-diameter dishes containing a 0.7% agar plug (BiTek, DIFCO Laboratories, Detroit, MI, USA). Colonies were stained with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (Sigma) and counted after three weeks. Cells (107) were harvested and mixed with an equal volume of Matrigel (Becton Dickinson Labware, Bedford, MA, USA), and 200-μl doses were injected into the flanks of six- to eight-week-old male NCr athymic nude mice (NCI, Frederick, MD, USA). Tumor sizes were measured after four weeks and tumor volume was calculated using the formula length × width × depth × 1/2. Each cell line was injected in a minimum of five animals.

Mammary tumors from MMTV-KRas mice and xenograft MCF10A-Ras tumors were dissected [31]. Half of the tumors were frozen in liquid nitrogen, and ground to a powder for RNA and protein analysis. The remaining tumors were mechanically dissociated with scalpels and enzymatically digested in a) DMEM: F-12 with 1 mM glutamine, 5 μg/ml insulin, 500 ng/ml hydrocortisone, 10 ng/ml epidermal growth factor, 20 ng/ml cholera toxin, 5% horse serum, 300 U/ml collagenase and 100 U/ml hyaluronidase (Sigma) for one hour at 370C followed by 0.1 mg/mL DNase treatment (Worthington, Lakewood, NJ, USA) for one minute subsequently washed in DMEM with 10% FBS and resuspended and plated in MEGM media (Cambrex, Walkersville, MD, USA) (for MMTV-KRas tumors) b) 0.25% trypsin for 5' at 37°C followed by washing in DMEM with 10% FBS and plated in MEGM media (for MCF10A Ras tumors). Cells were plated and subcultured daily. Supernatants were saved for IL-6 ELISA's, and RIPA extracts were made for protein analysis. ELISA analysis for human and murine IL-6 was performed using the manufacturer's instructions (Cell Sciences).

Tumors were fixed in 4% paraformaldehyde and embedded in paraffin. Tissue sections were stained for IL-6 using anti-IL-6 (1:200; Abcam, Cambridge, MA, USA) using previously described methods [32]. Imaging was carried out on a Leica Inverted Confocal Microscope.

Data are expressed as means ± standard deviation (SD). The statistical significance of differences was evaluated using an unpaired, non-parametric Student's t-test. Significant differences between experimental groups were * P < 0.05 or ** P < 0.01.

We examined the role of Stat3 in H-RasV12 mediated cell migration, invasion and cellular transformation using H-RasV12 transformed mammary epithelial cells (MCF10A-Ras). MCF10A cells are a spontaneously immortalized human breast epithelial cell line, mutant in the cdk inhibitor p16, yet has many of the characteristics of normal breast epithelium, do not form tumors in nude mice nor form colonies in soft agar, but undergo transformation upon the introduction of Ha-Ras [33‐35]. The H-RasV12 oncogene was introduced into MCF10A cells by retroviral gene transfer, and Ras expressing cells were selected in puromycin containing media. A Stat3sh or scrambled control shRNA-GFP constructs were introduced into H-RasV12 transformed MCF10A cells by lentiviral infection and sorted for GFP expression (Stat3sh and control lines respectively). Tyrosine phosphorylated Stat3 was undetectable in the MCF10A-Ras cells (Figure 1a). Stat3sh expressing cells displayed lower levels of total Stat3 protein by Western blot analysis and Ras protein levels were constant (Figure 1a). The morphology and growth rates of the Stat3sh cells were similar to control cells, as were their requirements for defined media components including EGF (Figure 1b and data not shown).

Ras transformed cells have increased invasive and migratory potential over control non-transformed cells [36]. We investigated whether Stat3 may be required for this phenotype and compared Stat3sh to control cells (Figure 1c). The Stat3sh expressing cells displayed an approximately four-fold decrease in migration in a Boyden chamber assay, compared to control cells (Figure 1c). Additionally, we observed a nearly 10-fold reduction in invasion through a matrigel coated insert in the Stat3sh cells compared with controls. Hence, Stat3 is a modulator of the invasive and migratory potential of Ras transformed mammary epithelial cells.

The finding that Stat3 is required for invasion and migration in Ras transformed MCF10A cells led us to determine if Stat3 might also be required for tumorigenesis. Anchorage-independent-growth is a measure of a cells capacity to grow in three dimensions, without contacting a basement membrane. We next determined whether Stat3 expression affects anchorage independent growth of MCF10A-Ras cells. Control cells displayed robust colony formation while Stat3sh cells formed very few colonies (approximately 10-fold less) (Figure 1d). Tumorigenesis was determined by injecting both control and Stat3sh cells into nude (nu/nu athymic) mice. Mice injected with Stat3sh cells formed tiny acellular tumors (0.007 cm3) relative to control cells (0.412 cm3) (Figure 1d). Taken together, these results indicate that Stat3 is required for Ras mediated transformation of MCF10A cells.

We next examined the levels of pStat3 in these tumors by immunohistochemical and Western blot analyses. Surprisingly, we observed high pStat3 levels in the control tumors (both tumor cells and adjacent stromal cells), while the Stat3sh tumors had very low levels of pStat3 (Figure 2a). The cells expressing pStat3 and total Stat3 within the Stat3sh tumors were principally non-tumor cells (Figure 2a). We recently showed that a principal mechanism of Stat3 activation in breast and lung cancers is through autocrine production of IL-6 [6, 32]. Furthermore, it was shown that a number of Ras transformed cells express high levels of IL-6 which promotes angiogenesis and tumorigenesis [23]. We therefore analyzed these tumor samples for IL-6 expression by immunohistochemistry and determined that control tumors expressed high levels of IL-6 (Figure 2a). To demonstrate that our observations were not specific to the MCF10A-Ras cell line, we examined mice expressing the K-Ras oncogene within the mammary gland (MMTV-K-Ras). These tumors also expressed high levels of pStat3 (by immunohistochemistry and Western blot analysis) and IL-6 (by immunofluoresence) (Figure 2b).

It was recently determined that exogenous sources of IL-6 could enhance autocrine production of IL-6 in models of breast cancer where IL-6/pStat3 levels were very low [37]. Furthermore, IL-6 has been shown to promote an epithelial-mesenchymal transition in breast cancer which correlated with enhanced invasion [38]. Given our observation that MCF10A-Ras cells express IL-6 and pStat3 in 3-D, we wished to determine whether exogenous IL-6 could lead to Stat3 phosphoryaltion and inducible expression of endogenous IL-6 in MCF10A-Ras cells grown in 2-D. We treated cells with IL-6 and observed a marked increase in pStat3 levels by Western blot analysis, Stat3 DNA binding activity by electrophorectic mobility shift assay (EMSA) as well as IL-6 mRNA levels (Figure 3a, b, c). We also examined the effect of exogenous IL-6 on MCF10A-Ras cell migration and determined that IL-6 enhanced MCF10A-Ras cell migration in a Stat3 dependent manner as IL-6 could not promote migration in RasS3Sh cells (Figure 3d). Thus, paracrine or exogenous sources of IL-6 enhances pStat3 levels, Stat3 binding activity and cell migration in a Stat3 dependent manner.

To further characterize the requirement for Stat3 in MCF10A-Ras cells, we utilized a Matrigel assay to examine growth in three-dimensions (3-D). The culture of MCF10A mammary epithelial cells on a defined basement membrane (Matrigel) results in the formation of polarized, hollow acini which recapitulates several aspects of glandular architecture in vivo [27]. Furthermore, oncogenes introduced into MCF10A cells disrupt this ordered process and elicit distinct morphological phenotypes [27]. MCF10A-Ras cells grew as amorphous structures which were not hollow and expressed high levels of pStat3 as determined by immunofluorescence (Figure 4a). MCF10A cells were also plated in matrigel revealing hollow acini which were negative for pStat3 by immunocytochemistry (Figure 4a). Inhibition of IL-6 signaling using an anti-IL6 blocking antibody or using a pan-Jak inhibitor (P6) led to a reduction in pStat3 levels (Figure 4a). Matrigel is a mixture of extracellular matrix (ECM) proteins composed primarily of laminin and collagen. We tested the role of matrigel and its components for the ability to enhance Stat3 phosphorylation and determined that matrigel and laminin were capable of inducing pStat3 (Supplemental figure S1 in Additional file 1). Thus, the growth of MCF10A-Ras on defined ECM proteins can enhance Stat3 activation. A role for Stat3 in MCF10A cell growth and acini formation was also examined. A reduction in Stat3 had no effect on the morphology of the acini nor on growth in 2-D (Supplemental figure S2 in Additional file 2). MCF10A-Ras cells lack E-cadherin expression which marks organized cell-cell contacts. Stat3sh cells continued to grow as filled acini but interestingly E-Cadherin expression was restored (Figure 4b). Similarly, treatment of MCF10A-Ras cells with inhibitors of IL-6/Jak/pStat3 signaling led to the expression of E-cadherin. Treatment of MCF10A-Ras 3-D structures with a pan-Jak inhibitor did not lead to a hollowing out of the structures despite the reappearance of E-Cadherin (Supplemental figure S3 in Additional file 3). IL-6/Stat3 signaling has been shown to inhibit E-Cadherin expression in models of prostate and breast cancer [38, 39]. This data suggest that phosphorylated Stat3 has a role in regulating Cadherin expression and the loss of a requirement for cell-cell contacts in transformed cells.

To investigate the relationship between cell transformation and IL-6 signaling, we introduced an shRNA construct targeting the IL-6 mRNA transcript into Ras transformed MCF10A cells. IL-6 shRNA functionality was confirmed using semi-quantitative RT-PCR analysis on cells stimulated with TNF-α, a ligand known to induce IL-6 mRNA production (Figure 5a). IL-6 transcript levels were significantly reduced in the IL-6 shRNA expressing cells. Total Stat3, pStat3 and Ras levels were unchanged by the loss of IL-6 (Figure 5b). However, the addition of exogenous IL-6 led to robust Stat3 tyrosine phosphorylation (data not shown). MCF10A-Ras cells expressing either control shRNA or the IL-6 shRNA grew at similar rates when grown on plastic (2-D) (Figure 5c). Furthermore, the chronic administration of IL-6 to MCF10A-Ras cells did not increase their proliferation in 2-D (Figure 5c). We examined the loss of IL-6 on cell migration and determined that MCF10A-Ras IL-6Sh cells (IL6Sh) did not migrate as well as control cells but the addition of IL-6 could restore cell migration (Figure 5d). Tumorgenicity was also assessed. MCF10A-Ras (C) cells were injected into the flanks of nu/nu athymic mice and tumors grew as expected (>0.4 cm3), while IL-6 shRNA expressing cells failed to form tumors, further implicating IL-6 signaling as a key pathway in Ras-mediated transformation (Figure 5e).

Our data indicate that MCF10A-Ras cells when grown in two dimensions (2-D) do not express detectable levels of pStat3 (Figure 1a). However, when these same cells were grown in three dimensions (3-D), either in a Matrigel assay or as tumors in nude mice, we observed high levels of pStat3 (Figures 2 and 4). We then asked whether the presence of pStat3 observed in 3-D growth was reversible upon culturing in 2-D. MCF10A-Ras tumors were surgically removed, and the epithelial cell population was serially cultured over four days. Protein extracts and RNA was isolated from the initial tumor growth, as well as from each passage of these cells. Additionally, the supernatant from the cultured cells of each passage was obtained in order to measure IL-6 levels. MCF10A-Ras tumors exhibited high levels of pStat3, however the passaging of these cells over time resulted in decreased pStat3 levels (Figure 6a). The decrease in pStat3 by Western blot correlated directly with a decrease in IL-6 protein levels in the cultured cell supernatants as determined by ELISA (Figure 6b) and of IL-6 mRNA levels by real-time PCR (Figure 6c). Conversely, we observed an increase in E-Cadherin levels as pStat3 and IL-6 levels were decreasing (Figure 6a). Similar results were obtained from MMTV-K-Ras tumors cultured as described above, whereby pStat3 and IL-6 levels were markedly decreased upon two passages (data not shown). Hence, the growth environment (growth in 2-D versus 3-D), markedly affects the IL-6/Jak/Stat3 signaling pathway in Ras transformed mammary epithelial cells.