LIP expression is regulated by IGF-1R signaling and participates in suppression of anoikis

To determine whether IGF-1 regulates C/EBPβ-LIP expression in mammary epithelial cells, MCF10A cells were serum starved for 24 hours and then stimulated with IGF-1 (2.6 nM, 20 ng/ml) for 4 or 16 hours prior to harvesting. Western blot analysis of whole cell extracts demonstrated that treatment with IGF-1 led to an increase in the LIP isoform (Figure. 1A). The LIP isoform was more significantly elevated as compared to the LAP isoforms, resulting in a statistically significant increase in the LIP/LAP ratio of 3.5 fold (p < 0.05) after 16 hrs of treatment as compared to LIP/LAP levels observed in serum starved, non-treated cells (Figure 1A). Similar increases in LIP expression and the LIP/LAP ratio were observed in MCF 7 cells treated with 2.6 nM IGF-1 for 16 hours (Figure 1B). Treatment of cells with insulin (10 nM) also led to increases in LIP protein expression (Figure 1A). The identification and sizes of the human LAP 1 and LAP 2 isoforms were confirmed in our previous study [32].

An IGF-1 concentration of 2.6 nM was chosen for this study because it is within the Kd of the IGF-1 receptor, and will not result in activation of the insulin receptor [35, 36]. In some experiments the IGF-1 concentration was increased 15× to 39 nM in order to generate a maximal LIP induction due to activation of IGF-1R, hybrid receptors and the insulin receptor. Likewise an insulin concentration of 10 nM activates insulin receptors but not IGF-1 receptors [35, 36]. Because a strong induction in LIP expression was commonly observed 16 hr after IGF-1 treatment, this time point was selected for all consequent analyses in this study.

To determine whether the increase in LIP expression might be the result of transcriptional increases in C/EBPβ mRNA, RNA was purified from IGF-1 (2.6 nM) treated MCF10A and MCF7 cells and C/EBPβ mRNA expression levels were analyzed by real-time qPCR. No statistically significant changes were observed in the levels of C/EBPβ mRNA in response to a 16 hour treatment of cells with 2.6 nM IGF-1 (Figure 1C, D). These data suggest that IGF-1R signaling does not increase C/EBPβ-LIP expression via an increase in C/EBPβ mRNA transcription, but rather via post-transcriptional mechanisms.

It was next important to determine whether the increased expression of LIP and the elevations observed in the LIP/LAP ratio in response to IGF-1 treatment were biologically active. To serve as a control, we first validated the activity of the individual LIP and LAP2 constructs on a C/EBPβ responsive promoter as shown in Figure 2A. C/EBPβ null mammary epithelial cells were transfected with either LIP, or LAP2 individually or together with a C/EBP responsive, firefly luciferase construct and renilla luciferase construct as control. As expected, LAP2 expression led to an increase in C/EBP responsive luciferase activity while LIP alone reduced promoter activity (Figure 2A). In combination with LAP2, LIP expression antagonized and decreased LAP2 induced promoter activity and led to a decrease in luciferase activity. To test for IGF-1 induced, endogenous C/EBPβ activity, MCF10A cells were transfected with a C/EBP responsive, luciferase construct before stimulation with IGF-1. To maximize LIP expression for a significant increase the LIP/LAP ratio, cells were stimulated for 16 hrs with 39 nM IGF-1. This led to an expected decrease in C/EBP responsive luciferase activity due to the antagonistic effects of increased LIP expression (Figure 2B). These data demonstrate that IGF-1R induced increases in the LIP/LAP ratio are biologically active.

Because IGF-1R signaling has been observed to cross-talk with EGFR signaling, it was necessary to determine whether the IGF-1R induced expression of LIP was, in part, mediated by EGFR signaling. We therefore investigated whether treatment of MCF10A and MCF7 cells with IGF-1 leads to phosphorylation of EGFR. As determined by Western blot analysis, neither IGF-1 nor insulin stimulation led to a significant increase in EGFR phosphorylation as assessed in whole cell protein extracts 10 minutes (Figure 3A) after addition of ligand. Additionally, neither a 10× increase in IGF-1 nor insulin activated the EGF receptor (Figure 3A). However, immunoprecipitation followed by immunoblot analysis did show a modest increase in phosphorylated EGFR after 10 minutes of IGF-1 stimulation (Figure 3B).

In addition to IGF-1 and insulin receptors, mammary epithelial cells can also express insulin/IGF-1 hybrid receptors [36‐38]. Hybrid receptors have been detected in most tissues that express both insulin receptor and IGF-1 receptor. An IGF-1 concentration of 2.6 nM will not activate the insulin receptor, but could potentially lead to the activation of the insulin/IGF 1 hybrid receptors. Data presented in Figure 3C supports this hypothesis and suggests that IGF-1 (2.6 nM) signaling has led to the formation of insulin/IGF-1 hybrid receptors. Functional studies with hybrid receptors demonstrate that they behave more like IGF-1 receptors rather than insulin receptors because they bind IGF-1 with a much better affinity than insulin [37, 38]. As expected, we did not observe activation of the hybrid receptor with 10 nM insulin (Figure 3C). Although the significance of the hybrid receptors in mammary epithelial cells in unclear, we hypothesize that the insulin/IGF-1 hybrids may be more abundant in MCF10A cells than otherwise expected and this hypothesis is supported by reports that insulin and hybrid insulin/IGF -1 receptors are important regulators of breast cancer cells [36, 38]. Throughout this study, we will refer to the IGF-1R mediated induction in LIP for simplicity, but the reader should understand that hybrid receptors may also be involved in regulation of LIP/LAP.

Because LIP expression is analyzed 16 hr after addition of ligand, we also checked p-EGFR expression at this later time point. EGFR was not phosphorylated in MCF10A cells or MCF 7 cells 16 hr after addition of IGF-1 (Figure 3D, E) To confirm that IGF-1 was indeed activating the IGF-1R signaling cascade, we analyzed p-IGF-1R and p-Akt expression at 20 min and 16 hr (Figure 3D, E).

To further assess the possibility that EGFR activity may play a role in the IGF-1R stimulated increase in LIP expression, we tested the sensitivity of IGF-1 treated MCF10A cells to the selective EGFR kinase inhibitor, AG1478. Pretreatment of cells for 30 minutes with 0.1, 1 or 5 μM AG1478 prior to addition of 2.6 nM IGF-1 for 16 hr did not inhibit or reduce the IGF-1 mediated increases in LIP expression and did not inhibit the increase in the LIP/LAP ratio (Figure 4A left portion & 4B). As a control, 5 μM AG1478 did lead to the expected decrease in p-EGFR (Figure 4A, lower panel), decreases in EGF mediated LIP expression and the LIP/LAP ratio, and lesser reductions with 0.1 and 1 μM (Figure 4A right portion). Treatment of cells with 0.1, and 1.0 μM AG1478 effectively reduced IGF-1-induced Erk1/2 phosphorylation (Figure 4C left panel) and as expected EGF-induced Erk1/2 phosphorylation (Figure 4C right panel). These data demonstrate that inhibition of EGFR kinase activity reduces IGF-1R mediated Erk1/2 activity and suggest that IGF-1R and EGFR signaling crosstalk in MCF10As to regulate Erk1/2 activity (as suggested in Figure 3B). Our data also demonstrate that inhibition of EGFR signaling with AG1478 does not inhibit IGF-1R induced Akt activity but does block EGF induced Akt activity (Figure 4C). These data are in agreement with published results and demonstrate that IGF-1R mediated Akt activity is not regulated by EGFR signaling, and that IGF-1R mediated Erk1/2 activity is ErbB dependent [26]. IGF-1R mediated Akt activity thus appears to be an important regulator of IGF-1R induced-LIP expression and may also be important for EGF mediated LIP expression.

To validate that IGF-1R induced LIP expression is EGFR independent, we tested an additional EGFR inhibitor. IGF-1R induced LIP expression was not reduced by treatment of MCF10A cells with the EGFR specific, monoclonal antibody, mAb528, which blocks the ligand epitope binding site of EGFR. Although this antibody blockade had no affect on IGF-1R induced LIP expression or the LIP/LAP ratio, it did reduce EGF-induced LIP expression, and the LIP/LAP ratio as expected (Figure 4D). Taken together, these data suggest that although EGFR signaling can crosstalk with IGF-1R signaling, the crosstalk is not required for the IGF-1R mediated regulation of LIP expression in MCF10A cells.

To better understand the importance of p44/42 MAPK (Erk1/2) and phosphatidylinositol 3-kinase (PI3K)/serine-threonine protein kinase B (Akt) in the regulation of IGF-1R induced LIP expression, cells were pre-treated with a Mek1/2 inhibitor, (U0126), or an Akt inhibitor, (SH-6), 30 minutes prior to stimulation with 2.6 nM IGF-1. As anticipated, 5 and 10 μM U0126 effectively inhibited the IGF-1R induced phosphorylation of Erk1/2 but did not inhibit Akt phosphorylation or the increase observed in LIP expression and the LIP/LAP ratio (Figure 5A). Treatment of MCF10A and MCF7 cells with SH-6, which acts to prevent membrane localization of Akt by competing with Inositol (3, 4, 5) phosphate binding to the Akt pleckstrin homology domain [39], effectively reduced p-Akt expression and LIP expression in IGF-1 treated cells and led to a reduction in the LIP/LAP ratio (Figure 5B, C). Taken together, these results suggest that Akt activity is an important regulator of IGF-1R induced LIP expression.

To better understand the biological significance of C/EBPβ expression in response to IGF-1R signaling, we investigated how knock down of C/EBPβ expression affects the well established, anti-apoptotic role of IGF-1R in cell survival. Anoikis, which is an induction of apoptosis that occurs upon loss of cellular adhesion [22], was induced in MCF10A cells via forced suspension culture on low adherence plates for up to 96 hrs, and apoptosis was analyzed as a sub-G1 fraction or Annexin V staining by flowcytometry (Figure 6A, B, C).

Treatment of cells that were serum starved for 24 hrs prior to anoikis, with 39 nM IGF-1, led to an expected increase in cell survival as shown by a significant decrease in apoptosis and reduction in the percent of vector control cells in sub-G₁ from 2.5% to 1.5% at 48 hr (bars 3 & 4) and from 9% to 6% at 96 hr (bars 7 & 8) of treatment (Figure 6A). Treatment of cells with 2.6 nM IGF-1 led to similar results (data not shown). It is important to note, that before placing IGF-1 treated, vector control cells into the anoikis assay, we checked duplicate plates of cells to validate IGF-1R induced LIP expression. Because the C/EBPβ isoforms are translated from a single mRNA, it is not possible to selectively knock down the individual LIP and LAP isoforms; however successful knockdown of total C/EBPβ expression with shRNA (Figure 6E) led to decreases in cell survival. Increased apoptosis, as observed by the increased number of cells in sub-G₁ as compared to vector control rose from 2.5% to 5.1% at 48 hr (bars 3 & 5) and 9% to 22% at 96 hr (bars 7 & 9) in the cells with knocked down C/EBPβ expression (Figure 6A). Moreover, in the presence of knocked-down C/EBPβ expression, IGF-1 treatment only moderately increased survival, with decreases in apoptosis from 5.1% to 4% at 48 hr (bars 5 & 6) and 22% to 16% at 96 hr (bars 9 & 10). These decreases in apoptosis were not statistically significant.

Because we have demonstrated in this study that IGF-1R signaling increases LIP expression and the ratio of LIP/LAP, we sought to test the effects of LIP overexpression on survival from anoikis, in a manner similar to that described in Figure 6A. Overexpression of LIP in MCF10A cells was accomplished using a pEIZ (HIV-Zsgreen) lentiviral construct driven by the EF-alpha 1 promoter [40]. Overexpression of LIP (Figure 6F) led to decreases in apoptosis as evidenced by the number of Annexin V positive cells (Figure 6B) and the accumulation of cells in sub-G1 at both 48 hr and 96 hr of anoikis (Figure 6C). These data suggest that the LIP isoform has an anti-apoptotic action and plays a role in cellular survival of anoikis. Thus the biological consequence of IGF-1R mediated increases in LIP expression may include the actions of LIP to participate in the regulation of cell survival. Our data demonstrate that treatment of cells with IGF-1 or overexpression of LIP leads to decreases in the percentage of cells in sub-G1, and decreases in the number of cells positive for Annexin V, thus representing a decrease in apoptosis (Figure 6A, B, C).

Taken together, the data in Figure 6 demonstrate that C/EBPβ knockdown leads to increased cell death and an accumulation of cells in sub-G₁ and suggest that C/EBPβ expression is important for survival and resistance to anoikis. Furthermore, we showed that IGF-1R treatment can partially rescue control cells from anoikis; however, cells with reduced C/EBPβ expression, are not successfully rescued from anoikis. This is most clearly observed in clonogenic outgrowth assays of C/EBPβ knock-down cells (Figure 6D). Suspension culture of vector control and C/EBPβ knock-down cells, in the presence of IGF-1 for 24 hr, followed by harvest and subsequent plating for adherent growth revealed a dramatic reduction in the survival and clonogenic activity of cells with knocked down C/EBPβ expression (Figure 6D). Similarly, overexpression of LIP reduced anoikis, as evidenced by the decreased number of Annexin V positive cells and the decreased number of sub G1 cells. In summary, C/EBPβ expression appears to play an important role in protection from anoikis and may be an integral downstream mediator of the protective effects of IGF-1R signaling.

In summary, our data demonstrate that IGF-1 stimulation of mammary epithelial cells leads to increased expression of LIP and an elevation in the LIP/LAP ratio. We additionally demonstrate that IGF-1R induced-LIP expression is biologically active as determined on a C/EBP responsive promoter construct. Although IGF-1R signaling can crosstalk with EGFR signaling to regulate Erk1/2 activity in our study, IGF-1R induced LIP expression is independent of EGFR signaling. We demonstrate that Akt activity is a critical determinant in the regulation of IGF-1R induced LIP expression and that EGFR-dependent, Erk1/2 activity is not necessary for IGF-1R induced LIP expression. Lastly we show that LIP plays a role to increase the survival of cells from anoikis and may participate in IGF-1R mediated suppression of anoikis.

Cultured mammary epithelial cells, MCF10A, were grown in Dulbecco modified Eagle medium (DMEM)-F12 (Invitrogen, USA) supplemented with 5% donor horse serum (Invitrogen, USA), 20 ng/ml of recombinant human EGF (Invitrogen, USA), 10 μg/ml of bovine pancreatic insulin (Sigma, USA), 100 ng/ml of cholera toxin (Sigma, USA), 0.5 μg/ml of hydrocortisone (Sigma, USA), and 5 μg/ml of gentamycin sulfate (Invitrogen, USA). MCF7 cells were grown in Eagle's Minimum Essential Medium (MEM) (Mediatech, USA) supplemented with 0.01 mg/ml bovine insulin and 10% fetal bovine serum (Hyclone, USA). C/EBPβ null cells were culture in Hepes buffered, Dulbecco modified Eagle medium (DMEM)-F12 (Invitrogen, USA) supplemented with 2% adult bovine serum (Invitrogen, USA), 5 ng/ml of recombinant human EGF (Invitrogen, USA), 10 μg/ml of bovine pancreatic insulin (Sigma, USA) and 5 μg/ml gentamycin sulfate.

To knock down C/EBPβ expression, C/EBPβ (Tet-On) and control TRIPZ™lentiviral shRNAmir constructs (Open Biosystem, USA) were stably transduced into MCF-10A cells by infection and puromycin selection. Prior to suspension culture, the cells were treated with Doxycycline (1 μg/ml) for 2 days to activate shRNA expression, followed by one more day of Dox treatment in serum free conditions to synchronize the cells and to generate a maximal knockdown of C/EBPβ expression. To prevent adherence, cells were transferred to Costar 6 well ultra low attachment plates (20,000 cells/well) or to 1% agar coated plates (500,000 cells/10 cm dish) for 24, 48 and 96 hrs in the presence or absence of IGF-1 (2.6 nM and 39 nM). After 24 hrs, suspended cells were transferred to standard 6 well cell culture plates and permitted to adhere to analyze survival via clonogenic outgrowth for two weeks followed by staining with crystal violet. Flow cytometry was conducted on cells collected at 48 and 96 hrs of suspension culture. Briefly, suspended cells were collected by centrifuge at 1000 rpm for 5 min. To prevent clustering, cells were digested in 1× trypsin at 37°C for 5 min, followed by washing with HBSS. Cells were then resuspended in (0.6% NP-40, 3.7% Formalin, 11% Hoechst 33258 in PBS) for Flow cytometry. Cell death was analyzed by measuring the sub-G₁ cell cycle fraction. LIP was overexpressed in MCF-10A cells using a pEIZ (HIV-Zsgreen) lentiviral construct driven by the EF-alpha 1 promoter (kindly provided by Dr. Zena Werb [40]) and cells were sorted. Annexin V-PE Apoptosis detection kit was purchased from BD Biosciences and performed according to manufacturer's instructions.

MCF10A and MCF7cells were plated at a density of 1.7 × 10⁶/100 mm and upon reaching 75 to 80% confluency, the growth medium was removed and replaced with a serum-free, defined medium containing DMEM-F12, 100 ng/ml cholera toxin, 0.5 μg/ml of hydrocortisone, and 5 μg/ml of gentamycin sulfate for MCF10A, and MEM for MCF7. Cells were maintained in defined medium for 24 hour prior to the addition of ligand: human EGF (Invitrogen, USA), IGF-1 (Sigma, USA), insulin (Sigma, USA) and harvested at 10-20 min or 16 hr after the addition of ligand. The MEK inhibitor, U0126 (Calbiochem, USA), the Akt inhibitor, SH-6 (Axxora Platform, USA), the EGFR inhibitor, AG1478 (Sigma, USA), and the blocking antibody EGFR-mAb528 (Santa Cruz Biotechnology, USA) were added 30-60 min before addition of ligand. Cells harvested at 16 hr were sonicated in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl [pH 7.4], 1% NP-40, 0.25% deoxycholate, 150 mM NaCl, 10 mM EGTA, 0.2% sodium dodecyl sulfate [SDS]) containing a protease inhibitor cocktail (Sigma, USA) and a phosphatase inhibitor I and II mixture (Sigma, USA). Aliquots of the lysates containing 100-200 μg of protein were boiled at 100°C for 10 min, electrophoresed on denaturing SDS-7% or 12% polyacrylamide minigels, and then transferred to polyvinylidene difluoride membranes (PVDF, Millipore, Bedford, Mass. USA). Blots were blocked 1-2 hr in TBST (20 mM Tris [pH 7.5], 150 mM NaCl, 0.5% Tween 20) containing 5% Carnation dry milk and then incubated with primary antibody for 1-2 hr (or overnight for antibodies directed against phospho-proteins) in TBST-1-5% carnation milk. Primary antibodies used were monoclonal and polyclonal anti-C/EBPβ (1:250, Santa Cruz, USA), polyclonal anti-GAPDH (1:5000, Trevigen, Gaithersburg, MD, USA), polyclonal β-actin (1:1000 Santa Cruz, USA), monoclonal anti-phospho-p44/42 (1:2000, Cell Signaling, Beverly, MA, USA), polyclonal anti-p44/42 (1:2000, Cell Signaling, USA), monoclonal anti-phospho Akt, polyclonal Akt (1:1000, Cell Signaling, USA), polyclonal anti-EGFR (1:1000 Santa Cruz, USA), monoclonal anti-phospho-EGFR (Tyr 845, 1:1000, Cell Signaling, USA). Blots were washed with TBST three times for 5 to 10 min each with agitation and then incubated for 1 hr with either goat anti-mouse-horseradish peroxidase (HRP) conjugate (Santa Cruz. USA) or goat-anti-rabbit-HRP (Bio-Rad, Hercules, CA, USA) in TBST-1-5% carnation. Proteins were visualized by either DURA or FEMTO chemiluminescence (Super Signal; Pierce, Rockford, Ill. USA) and HyBlot CL film (Denville Scientific, Metuchen, NJ, USA). Blots were stripped in Re-blot Plus Mild Solution (Chemicon, Temecula, CA, USA) for reprobing.

Proteins were electrophoresed and transferred to PVDF membranes as described above. Membranes were blocked for 1 hr in Odyssey blocking buffer. Primary antibodies (1:250, monoclonal anti-C/EBPβ, Santa Cruz, USA), polyclonal anti-GAPDH (1:5000, Trevigen, Gaithersburg, MD, USA) and secondary antibodies (1:5000, goat anti-mouse IR Dye 800 CW, LI-COR Biosciences and 1:5000, goat anti-mouse IR Dye 680 DX, LI-COR, USA) were diluted in blocking buffer with 0.1% Tween-20 and incubated with the blot for 1 hr at room temperature. After washing, the membranes were scanned using Li-COR's Odyssey infrared imaging system and quantitated using Odyssey 3 software.

MCF10A and MCF7cells were plated at a density of 75 to 80% confluency, the growth medium was removed and replaced with a serum-free, defined mediums as described. Cells were maintained in defined medium for 24 hour prior to the addition of human IGF-1 (Sigma, USA) and harvested at 16 hr after the addition of ligand by adding 1 ml Trizol (Invitrogen, USA). Total RNA was extracted according to the manufacturer's instruction. First-strand cDNA was prepared with 5 μg total RNA, random primers and reverse transcriptase (SuperScript II RNase H, Invitrogen, USA) according to the manufacturer's instruction. Quantitative PCR was performed by using real-time PCR iCycler (Bio-Rad, USA). PCR reaction and C/EBPβ primers were: sense 5' AACTCTCTGCTTCTCCCTCTG 3'; antisense 5'AAGCCCGTAGGAACATCTTT 3'. Ct values were converted to relative expression using the delta-delta Ct method, allowing normalization to both 18S and untreated control. The primer sequences for 18S were sense 5' GTAACCCGTTGAACCCCATTC 3'; antisense: CCATCCAATCGGTAGTAGCG 3'.

To validate the activity of individual LIP and LAP2 constructs, a C/EBP consensus luciferase construct (500 ng) and a Renilla construct (20 ng) as internal control were cotransfected with LAP2 and LIP individually or together at different ratios into C/EBPβ null cells to a total of 2500 ng plasmid DNA. Control vector serves as both a control for basal activity and to match the quantity of plasmid DNA. Luciferase and Renilla activities were recorded at 48 hrs. For the IGF experiment, MCF-10A cells were cultured in Falcon 24-well plates and at 70% confluency, were transfected with a C/EBP consensus Luciferase construct (500 ng) and a Renilla construct (20 ng) as internal control. Transfection was conducted using Fugene reagent (Roche, Switzerland, according to manufacturer instructions) and cells were maintained in serum free medium for 24 hrs. The cells were then treated with 2.6 nM IGF-1 for 16 hrs in serum free medium and luciferase activity was analyzed at the end of treatment. The relative luciferase activity was calculated as Luciferase value/Renilla value. n = 5

MCF10A cells incubated with ligand for 10 min were extracted in RIPA buffer without SDS, and sonicated. Protein extracts (1 mg) were pre-cleared for 1 hr at 4°C with protein G-PLUS agarose (Santa Cruz, USA), then immunoprecipitated overnight at 4°C with anti-EGFR (1:1000, Santa Cruz, USA) or 4G10-conjugated agarose beads (Upstate Biotechnology, Waltham, MA, USA) to immunoprecipitate IGF-1R/IR. The beads were rinsed 3 times with RIPA (without SDS), sample buffer was added, the mixture boiled for 10 minutes followed by electrophoresis through SDS-7% polyacrylamide minigels, and transfer to PVDF. Immuno-blots were performed as above using anti-phospho-EGFR (1:1000, Cell Signaling, USA), anti-IR (1:1000 Upstate Biotechnology, USA) or anti-IGF-1R (1:1000 Upstate Biotechnology, USA).