Transcriptional and posttranslational regulation of insulin-like growth factor binding protein-3 by Akt3

Abstract

Insulin-like growth factor (IGF)-dependent and -independent antitumor activities of insulin-like growth factor binding protein-3 (IGFBP-3) have been proposed in human non-small cell lung cancer (NSCLC) cells. However, the mechanism underlying regulation of IGFBP-3 expression in NSCLC cells is not well understood. In this study, we show that activation of Akt, especially Akt3, plays a major role in the mRNA expression and protein stability of IGFBP-3 and thus antitumor activities of IGFBP-3 in NSCLC cells. When Akt was activated by genomic or pharmacologic approaches, IGFBP-3 transcription and protein stability were decreased. Conversely, suppression of Akt increased IGFBP-3 mRNA levels and protein stability in NSCLC cell lines. Characterization of the effects of constitutively active form of each Akt subtype (HA-Akt-DD) on IGFBP-3 expression in NSCLC cells and a xenograft model indicated that Akt3 plays a major role in the Akt-mediated regulation of IGFBP-3 expression and thus suppression of Akt effectively enhances the antitumor activities of IGFBP-3 in NSCLC cells with Akt3 overactivation. Collectively, these data suggest a novel function of Akt3 as a negative regulator of IGFBP-3, indicating the possible benefit of a combined inhibition of IGFBP-3 and Akt3 for the treatment of patients with NSCLC.

Introduction

Insulin-like growth factor binding protein-3 (IGFBP-3), the most abundant IGFBP in human serum (1), regulates the activation of the insulin-like growth factor (IGF)-1R pathway by sequestering free IGF-I and thus modulating IGF-I bioavailability (2). Beyond its direct role in modulating the action of IGF, IGFBP-3 also plays a role in an IGF-independent manner, in which it induces G₁ cell cycle arrest and apopotosis in several human cancer cells (3–6). Several factors regulate the expression and stability of IGFBP-3. For instance, growth hormone and insulin are considered as inducers of IGFBP-3 (7). Expression of IGFBP-3 is also mediated by stimulation with a variety of proapoptotic and growth-inhibitory factors, such as transforming growth factor-β, retinoic acid, tumor necrosis factor-α, vitamin D, antiestrogens, antiandrogens and tumor suppressors (4,7). Several proteases have been involved in the non-responsiveness of cancer cells to IGFBP-3, including matrix metalloproteinases, cathepsins, neutrophil elastase and other serine proteases; these proteases represent a potential hurdle for the use of IGFBP-3 in lung cancer therapy (8–10). However, most of the studies involving these proteases were focused on the role of IGFBP-3 as a reservoir of IGF-I and little is known about the mechanisms underlying regulation of cellular IGFBP-3.

We have previously demonstrated that treatment with the farnesyltransferase inhibitor SCH66336, a pharmacologic approach to inhibit Ras activation, decreases Akt activity in H1299 non-small cell lung cancer (NSCLC) cells (11). Recent reports have suggested that Akt, a serine/threonine protein kinase that serves as a key player in the control of cell transformation, proliferation, survival and metabolism (12), has an effect on the stability of several proteins, including BRCA1 (13) and the L-type subunits of Ca²⁺ channels (14). Based on these previous findings, we hypothesized that Akt may counteract IGFBP-3’s antitumor actions through regulating the expression and/or stability of IGFBP-3 in NSCLC cells. This study was performed to investigate the role of Akt in the growth-inhibitory function of IGFBP-3 and the detailed mechanisms responsible for the effects of Akt on IGFBP-3 function. Here we show that Akt, especially Akt3, regulates cellular IGFBP-3 function by modulating its transcription and protein stability. Our data demonstrate that the antiproliferative and proapoptotic effects of IGFBP-3 are enhanced by inactivation of Akt, implying that one way to enhance the therapeutic potential of IGFBP-3 in NSCLC cells is to inhibit Akt activity. Our findings indicate a potential benefit to using Akt inhibitors in combined treatments with IGFBP-3 or other drugs that induce IGFBP-3 expression.

Materials and methods

Reagents

Phosphate-buffered saline and cell culture media were purchased from Invitrogen (Carlsbad, CA). Fetal bovine serum was purchased from Gemini Bio-Products (West Sacramento, CA). Penicillin-streptomycin and trypsin-ethylenediaminetetraacetic acid were purchased from Invitrogen (Carlsbad, CA). Hygromycin B was purchased from Roche Applied Science (Indianapolis, IN). The adenoviral constructs expressing kinase-inactive Akt (Ad-Akt-KM), phosphatase and tensin homolog (PTEN) (Ad-PTEN) and empty vector (Ad-EV) were amplified as described previously (15). HA-Akt1, HA-Akt2 and HA-Akt3 (T308D/S473D) expression vectors (HA-Akt1DD, HA-Akt2DD and HA-Akt3DD) were kindly provided by Dr Gordon Mills (University of Texas M. D. Anderson Cancer Center, Houston, TX). IGF was purchased from R&D Systems (Minneapolis, MN). Perifosine was purchased from Selleckchem (Houston, TX) or LC Laboratories (Woburn, MA). Recombinant human IGFBP-3 (rBP3) was obtained from R&D Systems. LY294002 was purchased from EMD Chemicals (Gibbstown, NJ). Reagents unless otherwise indicated were purchased from Sigma–Aldrich (St Louis, MO).

Cell culture

The human NSCLC lines (A549, H460, H226B, H1299, H226Br, H322, H358 and H292) were purchased from the American Type Culture Collection or kindly provided by Dr Jack A. Roth (MD Anderson Cancer Center, Houston, TX). They were cultured in RPMI-1640 medium supplemented with 10% fetal bovine serum and antibiotics. TSC2-knockout (TSC2^–/–) mouse embryonic fibroblast immortalized by p53 knockout (kindly provided by Dr D.J.Kwiatkowski at Brigham and Women’s Hospital, Boston, MA) were maintained in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum. TSC2^–/– and H226B cells expressing constitutively active Akt1, Akt2 or Akt3 were established by introduction of the pBABE retrovirus expressing hemagglutinin (HA)-tagged constitutively active Akts (HA-Akt1DD, HA-Akt2DD or HA-Akt3DD) (16) and selection by hygromycin B (50 μg/ml). To analyze the effects of Akt activity on IGFBP-3 expression, H322 cells infected with Ad-PTEN, Ad-Akt (KM), or Ad-EV and H226B cells stably transfected with HA-Akt3DD were treated with all-trans-retinoic acid (ATRA) prior to cycloheximide treatment (10 μg/ml) or with recombinant IGFBP-3 (rBP3). Cell lines used in this study were authenticated and validated prior to performing experiments. These cell lines were tested for authentication at Genetic Resources Core Facility of Johns Hopkins University in 2010 or at the Korean Cell Line Bank using AmplFLSTR identifiler PCR Amplification kit (Applied Biosystems, Foster, CA; cat. No. 4322288) in 2013.

Immunohistochemistry and immunofluorescence

Immunohistochemistry was performed using standard procedures (ABC-Elite, Vector Laboratories, Burlingame, CA). Formalin-fixed xenograft samples were paraffin embedded and cut into 4 μm sections. The sections were deparaffinized and dehydrated, and then treated with methanol containing 0.3% H₂O₂ to inhibit endogenous peroxidase. The slides were incubated with anti-IGFBP-3 and anti-phosphorylated S6 antibodies (Santa Cruz Biotechnology) at 4ºC overnight, followed by incubation with a biotinylated secondary antibody (Vector Laboratories) for 1 h. Solutions A and B (ABC-Elite) were added simultaneously for 30 min, and the signals were detected using Diaminobenzidine Substrate kit (Vector Laboratories). Counterstaining was performed with hematoxylin.

For immunofluorescence, cells were fixed with 4% paraformaldehyde for 30min at room temperature, incubated with the blocking buffer [5% bovine serum albumin in TBST (TBS containing 0.05% Tween-20)], and then incubated with anti-IGFBP-3 or anti-HA primary antibodies diluted in TBST containing 1% bovine serum albumin . After incubating with the corresponding fluorochrome-conjugated (FITC or rhodamine) secondary antibodies, coverslips were mounted with a mounting solution containing 4′,6-diamidino-2-phenylindole. The fluorescence was observed under the fluorescent microscope.

Metabolic labeling

Cells were incubated in a methionine- and cysteine-free RPMI medium (Sigma) for 1h and pulse-labeled with trans-³⁵S label (0.5 mCi; ICN Radiochemicals, Irvine, CA) for the indicated time periods. For the pulse-chase experiment, cells pulse labeled for 1 h were chased in fresh RPMI containing methionine (150mg/l) and cysteine (150mg/l) for the indicated time periods. Equal amounts of proteins from the total cell lysates were immunoprecipitated using an antibody against Flag or IGFBP-3 and then analyzed as described previously (11,17). Two independent experiments were performed with similar results; representative results from one experiment are presented.

Western blot analysis

Whole cell extracts were prepared as described previously (11,15). The antibodies used in this study include IGFBP-3 (Diagnostic Systems Laboratories, Webster, TX); Akt, pAkt (Ser473), pAkt (Thr308), PTEN, TSC2, pS6 ribosomal protein, S6 ribosomal protein, cleaved caspase-3 (Cell Signaling Technology, Danvers, MA); and Akt1, Akt2, Akt3, α-tubulin, β-actin, glyceraldehyde 3-phosphate dehydrogenase (GAPDH), HA- and FLAG tags (Santa Cruz Biotechnology, Santa Cruz, CA).

Reverse transcription-polymerase chain reaction

Total RNA was reverse-transcribed using the Superscript® first strand cDNA system (Invitrogen) and further analyzed by quantitative real-time polymerase chain reaction (RT–PCR) (7500 ABI) using the SYBR® Green PCR Master Mix kit (Applied Biosystems, Foster, CA) and Applied Biosystems 7500 Real-Time PCR System (Applied Biosystems). The primer sequences used for the quantitative RT–PCR are as follows: Mouse IGFBP-3 forward, 5′-CCA GGA AAC ATC AGT GAG TCC-3′; mouse IGFBP-3 reverse, 5′-GGA TGG AAC TTG GAA TCG GTC A-3′; mouse GAPDH forward, 5′-TGC ACC ACC AAC TGC TTA GC-3′; mouse GAPDH reverse, 5′-GGC ATG GAC TGT GGT CAT GAG-3′; human IGFBP-3 forward, 5′-TCT GCG TCA ACG CTA GTG C-3′; human IGFBP-3 reverse, 5′-GCT CTG AGA CTC GTA GTC AAC T-3′; human β-actin forward, 5′-GCG AGA AGA TGA CCC AGA TC-3′; human β-actin reverse, 5′-GGA TAG CAC AGC CTG GAT AG-3′. Quantification of mRNA expression was performed by the comparative cycle threshold method and normalized to the amount of GAPDH or β-actin mRNA.

Cell viability and colony formation assays (anchorage-independent and -dependent)

For the analysis of cell viability, NSCLC cells (1−2 × 10³ cells/well) seeded in 96-well plates were treated with rhIGFBP-3 and/or LY294002 and incubated for 3−5 days. The 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay was performed as described previously (11). For the colony formation analysis, Akt3DD was overexpressed in H226B cells through the transfection with recombinant pcDNA3-Akt3DD plasmid. For the anchorage-independent colony forming analysis, cells were suspended in a 0.35% agar/RPMI mixture (1 × 10³/ml) and plated in six-well plates precoated with 0.6% bottom agar. Cells were then allowed to propagate for 14 days. Colonies <0.2mm in diameter were counted. For the anchorage-dependent colony forming analysis, transfected cells were cultured on 12-well plates for colony formation, followed by the addition of IGFBP-3 and/or perifosine to the medium. After 2 weeks, the remaining colonies were washed twice with phosphate-buffered saline and counted on crystal violet-stained plates. All experiments for observing colony formation were independently repeated at least three times. The combination index for the determination of effects of drug combination was calculated as described previously (18).

Caspase-3 assay

Caspase-3 activity in H1299 cells treated with LY294002, IGFBP-3 or both for 48h was measured using a caspase-3/CPP32 colorimetric assay kit (BioVision, Mountain View, CA) according to the manufacturer’s suggested protocol. Briefly, 150 μg of protein was incubated in 50 μl of cell lysis buffer with an equal volume of 2× reaction buffer and 5 μl of DEVD-pNA substrate at 37°C for 2h. The amount of pNA released from the substrate was measured at 405nm using a 96-well plate reader. Relative caspase-3 activity was calculated by comparing the absorbance of the treated cells with that of the control cells.

RNA interference

Small interference RNAs (siRNAs) targeting Akt3 and negative control siRNA (si-NC) were used in the functional experiment assays. The sequences of Akt3 siRNA duplexes are as follows: Akt3-siRNA #1, sense, 5′-GGA CCG CAC ACG UUU CUA U-3′; antisense, 5′-A UAG AAA CGU GUG CGG UCC-3′; Akt3-siRNA #2, sense, 5′-GGA CCG CAC ACG UUU CUA UUU-3′, antisense, 5′-AAA UAG AAA CGU GUG CGG UCC-3′. The siRNA for negative control was purchased from Dharmacon (Waltham, MA). Cells were transfected with siRNAs using Lipofectamine RNAiMAX (Invitrogen) according to the manufacturer’s recommended protocol. Knockdown of Akt3 expression was determined by RT−PCR using the following primers: human Akt3 forward, 5′-ATG AGC GAT GTT ACC ATT GT-3′; and human Akt3 reverse, 5′-CAG TCCT GTC TGC TAC AGC CTG GAT A-3′.

In vivo xenograft model

All animal procedures were performed according to a protocol approved by the Seoul National University Institutional Animal Care and Use Committees (approval nos. SNU-121207-2 and SNU-130130-1). H226B cells stably expressing either EV or Akt3DD were subcutaneously injected into the flank of nude mice. Tumor growth was monitored by measuring tumor size every 3 days. In additional experiment, NOD/SCID mice bearing H226B-EV xenograft tumors were subcutaneously administered 3mg/kg of rBP3 twice a week. Mice bearing H226B-Akt3DD xenograft tumors were given rBP3 (subcutaneously, twice a week) alone or in combination with 19.6875mg/kg of perifosine (p.o., four times a week) for the indicated times. Tumor growth was determined by measuring tumor size every days, and body weight was also measured to check toxicity in each treatment group. Tumor volume was calculated using the following formula: 1/2 × width² × length.

Statistics

Statistical significance was analyzed by two-sided Student’s t-test (Microsoft Excel 2013, Microsoft, Redmond, WA) or one-way analysis of variance (GraphPad Prism 6, GraphPad Software, La Jolla, CA). The difference was considered to be statistically significant when P < 0.05.

Results

Inhibition of PI3K/Akt activity enhances antiproliferative and proapoptotic effects of IGFBP-3 in NSCLC cells

Previously, we have reported that IGFBP-3 has the regulatory effect over phosphorylated Akt in NSCLC cells (19,20) and combined treatment with the PI3K inhibitor LY294002 enhances antiproliferative and proapoptotic activities of adenoviral IGFBP-3 compared with single agent treatment (11). It was possible that overexpression of adenoviral infection caused an artifact that would not be observed with a pharmacological inducer or a recombinant protein of IGFBP-3. We have shown that exogenously added IGFBP-3 localize in the cells via as yet unknown mechanisms (11) and inhibits the growth of NSCLC cells and angiogenesis through IGF-independent mechanism (20,21). Hence, in the current study, we determined whether blockade of the PI3K/Akt pathway would enhance the antitumor activities of recombinant IGFBP-3 protein (rBP3) in H1299 and A549 NSCLC cells which expressed all three Akt isoforms (Supplementary Figure 1, available at Carcinogenesis Online) (22). In line with the previous results from adenoviral IGFBP-3 treatment (11), rBP3 treatment showed notably increased inhibitory effects on H1299 and A549 NSCLC cell viability, in which Akt activity was suppressed by treatment with LY294002 (10 μM, 3 or 5 days) (Figure 1A). Because drug response of cancer cells grown in 2D culture conditions might differ from that of tumors under the 3D condition, we further analyzed the response of these cell lines grown in a 3D culture condition (soft agar). As shown in Figure 1B, the H1299 and A549 NSCLC cells experienced significantly enhanced inhibition of anchorage-independent colony forming abilities after the combined treatment with rBP3 and LY294002 compared with those treated with single drug treatment. Similarly, addition of LY294002 enhanced the apoptotic activities of rBP3 in H1299 cells as shown by an increase in the expression of cleaved caspase-3 (Figure 1C) and by the upregulation of caspase-3 enzymatic activity (Figure 1D). These findings indicate that PI3K/Akt inhibition potentiates the antiproliferative and proapoptotic activities of IGFBP-3 in NSCLC cells.

Fig. 1.

Enhancement of the antiproliferative and proapoptotic effects of IGFBP-3 by disruption of Akt activation. (A) H1299 and A549 cells were treated with the PI3K inhibitor LY294002 (LY), recombinant IGFBP-3 (rBP3) or their combination for 3 or 5 days. Cell proliferation was assessed by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assay. Cell viability values are expressed as relative to non-treated cells for 3 days, normalized to 100%. Error bar represents standard deviation; **P < 0.01; ***P < 0.001. (B) H1299 and A549 cells seeded in soft agar were treated with LY294002 (LY), IGFBP-3 (rBP3) or their combination. Growth media were changed every 2 days. Colony formation was analyzed 14 days later. Left, quantitation of colonies. Right, representative images of colonies in soft agar. Error bar represents standard deviation; ***P < 0.001. (C and D) H1299 cells were treated with LY294002 (LY, 25 μM), recombinant IGFBP-3 (rBP3, 5 μg/ml) or their combination for 48h. (C) Levels of phosphorylated Akt (pAkt), Akt and cleaved caspase-3 were analyzed by western blot analysis. (D) Caspase-3 activity was measured with the caspase-3 activity assay kit under the manufacturer’s recommended procedure. The results are expressed as relative caspase-3 activity compared with vehicle-treated control cells, normalized to 100%. Error bar represents standard deviation; ***P < 0.001.

Akt activity regulates IGFBP-3 mRNA expression and protein stability in NSCLC cells

We assessed the impact of Akt on IGFBP-3 expression in NSCLC cells. We observed a considerable increase in IGFBP-3 protein expression in H460 cells that had been infected with adenoviruses carrying PTEN (Ad-PTEN), a negative regulator of Akt (23) (Figure 2A). Similarly, H460 cells, in which Akt was inactivated by the infection with adenoviruses carrying dominant negative Akt [Ad-Akt (K179M)], showed increased levels of IGFBP-3 expression (Figure 2B).

Fig. 2.

IGFBP-3 mRNA expression and protein stability by regulation of Akt activation. (A and B) H460 cells were infected with the indicated doses of adenoviruses [Ad-EV (A, B), Ad-PTEN (A) or Ad-Akt (K179M) (B)] and subjected to western blot analysis for the indicated protein expressions. (C) H226Br cells were transfected with a PTEN-specific (siPTEN) or a control (siSCR) siRNA. Cells were subsequently treated with recombinant IGFBP-3 (rBP3) (5 μg) for 48h and subjected to western blot analysis for the indicated protein expressions. (D) H1299 and H226Br were treated with LY294002 (LY, 25μM) or with recombinant IGFBP-3 (rBP3) (5 μg), or their combination for 2 days. IGFBP-3 expression was analyzed by western blot analysis. (E) H226Br were treated with perifosine (1, 2 μM) or with recombinant IGFBP-3 (rBP3) (5 μg) for 2 days. The protein level of IGFBP-3 was analyzed by western blot analysis. (F) H322 cells were treated with LY294002 (LY, 25 μM) for 2 days. RT–PCR analysis was performed for analyzing IGFBP-3 mRNA level. (G) H661 and H292 cells were treated with the Akt inhibitor perifosine for 2 days. RT–PCR analysis was performed for analyzing IGFBP-3 mRNA level. Error bar represents standard deviation; *P < 0.05, **P < 0.01. (H) H292 cells were treated with perifosine (2 μM) for days, followed by actinomycin D for indicated times. RT–PCR analysis was performed for IGFBP-3 mRNA level and quantification analysis showed the half-life of IGFBP-3 mRNA.

To investigate the mechanisms by which Akt activity regulated IGFBP-3 expression, we first explored whether Akt activity was implicated in stability of intracellular IGFBP-3 protein. Because recombinant IGFBP-3 protein (rBP3) is internalized in NSCLC cells, probably through its membrane receptor (20,21), we assessed protein levels of IGFBP-3 in NSCLC cells after treatment with exogenous rBP3 in the presence or absence of various PI3K/Akt regulators. In H226Br cells, upregulation of Akt activity by transfection with a siRNA against PTEN (siPTEN) decreased the level of IGFBP-3 (Figure 2C). In contrast, suppression of PI3K/Akt activity by treatment with LY294002 (Figure 2D) resulted in an increased level of IGFBP-3 in H1299 and H226Br cells. Because both PTEN and LY294002 impact cellular level of PIP₃, which can regulate a variety of pathways (24–27), we further employed treatment with perifosine, a pharmacological inhibitor of Akt, and found that IGFBP-3 protein level was increased by the treatment (Figure 2E). We also observed a considerable increase in IGFBP-3 mRNA expression in H322 (Figure 2F), H661 and H292 (Figure 2G) cells had been treated either with LY294002 or perifosine. We then analyzed whether Akt inhibition affected the mRNA stability of IGFBP-3 or not. When treated with actinomycin D, the half-life values of IGFBP-3 mRNA in vehicle-treated and perifosine-treated H292 cells were 16.49 and 16.84 h, respectively, suggesting that Akt inhibition did not influence the mRNA stability of IGFBP-3 (Figure 2H). These results support our hypothesis that Akt has suppressive effects on the mRNA expression and protein stability of IGFBP-3 in NSCLC cells.

Akt3 inhibits IGFBP-3 mRNA expression and protein stability in NSCLC cells

We sought to determine which Akt isoform affects IGFBP-3 expression. We introduced HA-conjugated constitutively active Akt isoforms (HA-Akt1DD, HA-Akt2DD or HA-Akt3DD) to mouse embryonic fibroblasts derived from a TSC2-knockout mouse line (TSC2^{– /–}) by stable transfection. These cells express residual levels of phosphorylated Akt (pAkt) due to the activation of the mTOR/p70S6K-mediated feedback inhibition of the IRS/PI3K/Akt pathway (25) but have markedly increased levels of IGFBP-3 protein (Supplementary Figure 2A, available at Carcinogenesis Online). Surprisingly, downregulation of IGFBP-3 protein (Supplementary Figure 2B, available at Carcinogenesis Online) and mRNA (Supplementary Figure 2C, available at Carcinogenesis Online) expression was observed in TSC^{– /–} cells transfected with Akt3DD (TSC^{– /–}/HA-Akt3DD).

We then assessed the predominant role of Akt3 over the two other isoforms in the regulation of IGFBP-3 expression in NSCLC cells. Similar to the findings in TSC2^–/– cells, constitutive activation of Akt3 by transient transfection with HA-Akt3DD induced downregulation of IGFBP-3 protein and mRNA expressions in H226B cells (Figure 3A), which expressed low levels of endogenous Akt3 (Supplementary Figure 1A, available at Carcinogenesis Online). With regards to the stability of IGFBP-3, IGFBP-3 levels resulting from rBP3 treatment in H226B cells with stable transfection with HA-Akt3DD (H226B/HA-Akt3DD) decreased faster than that in H226B/HA-EV cells (Figure 3B). By performing metabolic labeling analysis using [³⁵S] Met-Cys, we confirmed Akt3-mediated regulation of IGFBP-3 expression in NSCLC cells. Pulse labeling of the cells revealed that de novo synthesized IGFBP-3 protein level in H226B/HA-Akt3DD cells was markedly lower than that in any other H226B-derived cells (Figure 3C). Furthermore, chase of the pulse-labeled protein revealed that H226B/HA-Akt3DD cells had the fastest degradation of the newly synthesized IGFBP-3 protein. H226B/HA-Akt3DD also showed markedly decreased levels of IGFBP-3 mRNA expression (Figure 3D). Conversely, loss of Akt3 expression by siRNA transfection led to increases in IGFBP-3 mRNA level in H661 and H358 cells with high levels of Akt3 (Figure 3E and Supplemental Supplementary Figure 1, available at Carcinogenesis Online).

Fig. 3.

Modulation of IGFBP-3 mRNA expression and protein stability by Akt3 activation. (A) The protein and mRNA expressions of indicated protein and IGFBP-3 in H226B expressing constitutively active Akt mutant constructs (HA-Akt1DD, HA-Akt2DD or HA-Akt3DD) were analyzed by western blot (top) and RT–PCR analysis (bottom), respectively. (B) Serum-starved H226B/HA-EV and H226B/HAAkt3DD cells were treated with rBP3 for the indicated times. Protein expression was analyzed by western blot analysis. Expression levels of protein were represented as relative to that of control and normalized to 1. (C) H226B cells derivatives expressing control (HA-EV) or constitutively active Akt constructs (HA-Akt1DD, HA-Akt2DD or HA-Akt3DD) were pulse labeled with Trans-³⁵S label for up to 9h (top). H226B cells derivatives were pulse labeled with Trans-³⁵S for 1h and then chased for 3 or 6h (bottom). Cells were subjected to immunoprecipitation for [³⁵S]-labeled IGFBP-3 expression. Expression levels of protein were represented as relative to that of control and normalized to 1. (D) RT–PCR analysis was performed to analyze IGFBP-3 in H226B cells expressing control (HA-EV) and constitutively active Akt3 mutant (HA-Akt3DD) (top). Error bar represents standard deviation; ***P < 0.001. RT–PCR analysis was performed to analyze IGFBP-3 in H292 cells expressing control (HA-EV) and constitutively active Akt3 mutant (HA-Akt3DD) (bottom). (E) H661 and H358 cells were transfected with an Akt3-specific (Akt3) or a control (Scr) siRNA. Cells were subjected to RT–PCR (top) and RT–PCR (bottom) for the IGFBP-3 and Akt3, respectively. (F) Western blot analysis was performed to analyze IGFBP-3 and indicated proteins in H226B cells expressing control (HA-EV) and constitutively active Akt3 mutant (HA-Akt3DD).

Given our previous finding showing IGFBP-3-mediated downregulation of bFGF (21), we analyzed bFGF expression in cell lysates and conditioned media from H226B/HA-EV and H226B/HA-Akt3DD cells. We found that H226/HA-Akt3DD had increased cellular expression and secretion of bFGF protein along with decreased IGFBP-3 protein levels and increased pS6 levels compared with those transfected with control vector (HA-EV) (Figure 3F). Together, these results clearly indicate that transcriptional and posttranslational suppressions of IGFBP-3 expression may be under the control of Akt3 in NSCLC cells.

Akt3 activation leads to a decrease in IGFBP-3 expression in the nucleus

Akt has an important role in nuclear localization of several proteins, including EGFR (28), FOXO3a (29) and IKKα (30). In addition, previous reports suggest the antitumor and proapoptotic effects of nuclear IGFBP-3 (31), suggesting Akt might regulate cellular distribution of IGFBP-3 by destabilizing IGFBP-3 in the nucleus, thereby abrogating its antitumor actions. Hence, we investigated whether Akt activation affects the subcellular localization of IGFBP-3 in NSCLC cells. Immunofluorescence analysis showed that IGFBP-3 localized mainly in the nucleus of H226B/HA-EV, H226B/HA-Akt1DD and H226B/HA-Akt2DD cells. In contrast, nuclear IGFBP-3 staining was dramatically decreased in H226/HA-Akt3DD cells (Figure 4A). Comparing IGFBP-3 expression between cytosol and nucleus with cellular fractionation also indicates that Akt3 activation leads to substantial decrease of nuclear IGFBP-3 (Figure 4B).

Fig. 4.

Subcellular expression of IGFBP-3 regulated by Akt3 activation in H226B NSCLC cells. (A and B) H226B cells were transfected with control (HA-EV) or constitutively active Akt constructs (HA-Akt1DD, HA-Akt2DD or HA-Akt3DD). (A) H226B cells were immunostained using antibodies against an IGFBP-3 (red), followed by an Alexa 546 conjugated antibody. Cells were also counterstained with 4′,6-diamidino-2-phenylindole to visualize nuclei (blue). Scale bars are 20 μm (left). Quantification of relative intensity of IGFBP-3. Values were normalized to 100% at the IGFBP-3 in control (HA-EV) cells. Error bar represents standard deviation; **P < 0.01 (right). (B) Cytosolic extract or nuclear extract were isolated from indicated cells. IGFBP-3 expression was determined by western blot analysis. Expression levels of protein were represented as relative to that of control and normalized to 1.

Based on the previous report showing induction of IGFBP-3 expression by ATRA (32), we further analyzed ATRA-induced IGFBP-3 expression in H226B/HA-EV and H226B/HA-Akt3DD cells. As shown in Figure 5A, IGFBP-3 expression was increased in both H226B/HA-EV and H226B/HA-Akt3DD cells. However, IGFBP-3 protein expression increased in H226B/HA-Akt3DD cells in a slower rate than in H226B/HA-EV cells. We further analyzed the stability of ATRA-induced IGFBP-3 protein by exposing ATRA-treated cells to cycloheximide, a widely used protein synthesis inhibitor (33). The ATRA-induced IGFBP-3 protein level in H226B/HA-Akt3DD cells decreased more rapidly than in H226B-EV cells (Figure 5B). We also determined ATRA-induced IGFBP-3 in the nucleus in the cells. As shown by the results from immunoblotting of cytosolic extracts and nuclear extracts, both cytosolic and nuclear IGFBP-3 protein levels were markedly lower in H226B/HA-Akt3DD cells than in H226B/HA-EV cells (Figure 5C). These results suggest that Akt3-mediated suppression of IGFBP-3 expression results in decrease in nuclear IGFBP-3 in NSCLC cells.

Fig. 5.

Nuclear IGFBP-3 downregulation by Akt3 activation in H226B NSCLC cells. (A) H226B/HA-EV and H226B/HA-Akt3DD cells were treated with ATRA (1 μM) for indicated time period. IGFBP-3 protein expressions were determined by RT–PCR. Expression levels of protein were represented as relative to that of control and normalized to 1. (B) H226B/HA-EV and H226B/HA-Akt3DD cells were treated with ATRA (1 μM) for 24h prior to exposure to cycloheximide. IGFBP-3 expression was determined by western blot analysis. LE, long exposure; SE, short exposure. Expression levels of protein were represented as relative to that of ATRA-treated group and normalized to 1. (C) H226B/HA-EV and H226B/HA-Akt3DD cells were treated with ATRA (1 μM) for 24h and cytosolic extract or nuclear extract were isolated. IGFBP-3 expression was determined by western blot analysis. Expression levels of protein were represented as relative to that of control and normalized to 1.

Akt3 is a main player in the regulation of IGFBP-3 expression in vivo

To further correlate the constitutive activation of Akt3 with IGFBP-3 expression in NSCLC cells in vivo, nude mice were injected with H226B/HA-EV or H226B/HA-Akt3DD cells. Mice injected with these cells started to form similarly sized tumors ~10–14 days after the injection. However, xenograft tumors of HA-Akt3DD-expressing H226B cells showed significantly faster growth compared with those of control tumors (Figure 6A). Representative nude mice from these groups are shown. IGFBP-3 mRNA and protein levels in H226B/HA-Akt3DD xenograft tumors were obviously downregulated compared with those in H226B/HA-EV xenograft tumors (Figure 6B). Immunohistochemical analysis was performed to evaluate IGFBP-3 and pS6 expression in the H226B xenograft tissues. Compared with the control group, xenograft tumors of H226B/HA-Akt3DD cells showed markedly greater pS6 along with obviously weaker IGFBP-3 staining (Figure 6C). These results indicate increase of tumor growth by Akt3 activation, which is accompanied by IGFBP-3 decrease, in lung cancer.

Fig. 6.

IGFBP-3 downregulation by Akt3 activation in vivo. (A– C) H226B cells (5×10⁶ cells) stably transfected with the control vector (HA-EV) or expression vector carrying active Akt3 construct (HA-Akt3DD) were subcutaneously injected into nude mice. The tumors were measured every 3 days, and the results were expressed as the mean (±SD) tumor volume (calculated from 10 mice) (A). Representative mice from each group are shown. IGFBP-3 expression was analyzed by RT–PCR, western blot assay (B) and immunohistochemical analysis (C) in H226B (HA-EV and HA-Akt3DD) xenograft tumor. Error bar represents standard deviation; *P < 0.05.

Inhibition of Akt enhances the efficacy of IGFBP-3 in NSCLC cells with Akt3 hyperactivation

We performed in vitro clonogenic assays to test whether inactivation of Akt would enhance antitumor activities of IGFBP-3 in NSCLC cells with overactivation of Akt3. H226B/HA-EV and H226B/HA-Akt3DD cells were treated with rBP3 (5 µg/ml), perifosine, or their combination. The inhibitory effects of perifosine or rBP3 alone on colony formation were greater in H226/HA-EV cells than in H226B/HA-Akt3DD cells (Figure 7A). However, the inhibitory actions of the combination of the drugs were greater in H226B/HA-Akt3DD than in H226/HA-EV cells (Figure 7A). We also confirmed that the Akt inhibition by perifosine enhanced the BP3-induced increase in IGFBP-3 level and decrease in pAkt level in H226B/HA-Akt3DD cells more effectively than in H226/HA-EV cells. The combination regulated expressions of bFGF in both H226B/HA-Akt3DD and H226/HA-EV cells (Figure 7B). To confirm the role of Akt3 in the regulation of IGFBP-3, the antitumor effects of IGFBP-3, with or without Akt inhibition, were investigated and compared in vivo using H226B/HA-EV or H226B/HA-Akt3DD cells. NOD/SCID mice bearing H226B/HA-EV and H226B/HA-Akt3DD xenograft tumors were treated with rBP3 (3mg/kg), perifosine (19.6875mg/kg) or their combination. The significant inhibitory effect of rBP3 on tumor growth was seen in H226/HA-EV xenografts but not in H226B/HA-Akt3DD xenografts, supporting negative regulation of IGFBP-3 by Akt3 activation (Figure 7C). In addition, treatment with IGFBP-3 in combination with an Akt inhibitor perifosine statistically significantly suppressed the growth of H226B/HA-Akt3DD xenograft tumors compared with that of tumors treated with rBP3 only (Figure 7D). Collectively, these results suggest that Akt negatively regulates IGFBP-3 and thus Akt inhibition may be a strategy to restore antitumor effect of IGFBP-3 in NSCLC cells harboring active Akt, especially Akt3.

Fig. 7.

Enhanced antitumor activities of IGFBP-3 in NSCLC cells with overactivation of Akt3 mediated by Akt inactivation. (A) In clonogenic assay, H226B cells expressing control (HA-EV) or constitutively active Akt3 (HA-Akt3DD) were treated with IGFBP-3 (5 μg) or with perifosine (2 μM), or their combination for 7 days. Colony formation was analyzed 7 days later. Error bar represents standard deviation; ***P < 0.001. (B) H226B cells expressing control (HA-EV) or constitutively active Akt3 (HA-Akt3DD) were treated with IGFBP-3 perifosine (2 μM) or with recombinant IGFBP-3 (rBP3) (5 μg), or their combination for 2 days. Indicated proteins and IGFBP-3 expression were analyzed by western blot analysis. (C and D) H226B cells (4×10⁶ cells) stably transfected with the control vector (HA-EV) or expression vector carrying active Akt3 construct (HA-Akt3DD) were subcutaneously injected into nude mice. (C) Mice were subcutaneously injected with 3mg/kg of rBP3 twice a week. Data were expressed as the mean (±SD) tumor volume (calculated from six mice). Error bar represents standard deviation; **P < 0.01. (D) Mice were subcutaneously injected with either 3mg/kg of rBP3 twice a week or 19.6875mg/kg of perifosine 4 times a week, or their combination. Data were expressed as the mean (±SD) tumor volume (calculated from six mice). Error bar represents standard deviation; *P < 0.05. (E) Schematic model of Akt3-mediated decrease in IGFBP-3 expression and its involvement in NSCLC proliferation and survival.

Discussion

This study elucidates a novel function of Akt3 as a negative regulator of IGFBP-3 in NSCLC cells. Our study provides in vitro and in vivo evidences that activated Akt, especially Akt3, decreases IGFBP-3 levels through mechanisms that involve transcriptional and posttranslational regulations of IGFBP-3 expression. We also show that Akt3 interferes with the antitumor activities of IGFBP-3 in NSCLC cells. These results show that the mechanisms responsible for the Akt3-mediated increase in NSCLC cell proliferation is at least in part through regulating IGFBP-3 expression.

A growing number of epidemiologic studies suggest that increased serum levels of IGFs, altered levels of IGFBP-3 or both are associated with an increased risk for developing several malignancies, including NSCLC (34,35). We have demonstrated that the loss of IGFBP-3 expression is a common event in patients with stage I NSCLC and correlates closely with poor prognosis (22,36). We have also shown that IGFBP-3 has antitumor activities in NSCLC cells in vitro and in vivo by inducing apoptosis and by inhibiting angiogenic and metastatic activities (11,19,21). Recent studies revealed that the loss of IGFBP-3 expression is involved in the development of resistance to conventional therapies and molecularly targeted agents, such as cisplatin and gefitinib, in NSCLC cell lines (8,37,38). These data suggest that IGFBP-3 may be a useful drug for the treatment of lung cancer.

One potential concern about the use of IGFBP-3 in lung cancer therapy is the presence of various mechanisms that can abolish the antitumor actions of IGFBP-3. Posttranslational modifications, including proteolysis, phosphorylation and glycosylation, have been proposed to antagonize IGFBP-3. IGFBP-3-proteases, such as kallikrein-like serine proteases, cathepsins and matrix metalloproteinases, can cleave IGFBP-3, releasing IGFs from IGFBP-3/IGF complexes and thus mediating IGF signaling-dependent cell proliferation and survival (7,9,10). Phosphorylation has been proposed to influence ligand binding and antiproliferative and proapoptotic effects of IGFBP-3 (39). Mechanisms involving proteasomes and lysosomes have also been proposed to regulate intracellular IGFBP-3, and IGFBP-3 phosphorylation tends to occur prior to ubiquitination and subsequent degradation by the proteasome system (40). Potential phosphorylation sites have been identified for several kinases, including casein kinase II, cyclic adenosine monophosphate-dependent protein kinase (PKA), calcium/phospholipid-dependent protein kinase, mitogen-activated protein kinase, DNA-PK and other unidentified kinases (39), suggesting a potential role of these kinases in IGFBP-3 function.

We have previously reported that overexpression of IGFBP-3 by adenoviral infection has regulatory effects on Akt activation in NSCLC cell lines (11,19). Consistently, in the current study we confirmed the rBP3-mediated regulatory effects on pAkt levels in H226Br and H1299 cells. Given the fact that IGFBP-3 binds to free IGF in the extracellular milieu with high affinity and specificity, thus reducing its bioavailability, IGFBP-3 could suppress Akt activity through IGF-dependent mechanisms (Figure 7E, blue line). IGFBP-3 has been demonstrated to contain nuclear localization sequences and nuclear IGFBP-3 induces apoptosis through binding with RXR/Nur77 (7,31,41). Others and we have shown that IGFBP-3 localize in NSCLC cells via as yet unknown mechanisms and induces antitumor activities through IGF-independent mechanisms (Figure 7E, green line). Results from our current study show that approaches to inhibit Akt activity potentiate antiproliferative and proapoptotic activities of IGFBP-3 in NSCLC cells. These findings point to activation of Akt as a potential regulator of cellular IGFBP-3 in NSCLC cells. Our subsequent studies using derivatives of TSC2^{– /–} mouse embryonic fibroblast and H226B NSCLC cells, in which each Akt subfamilies were engineered to be constitutively active, indicate that Akt activities, especially those mediated by Akt3, play a prominent role in transcriptional and posttranslational regulation of IGFBP-3. Because Akt is constitutively activated in NSCLC cell lines, probably through various mechanisms, such as activating mutations of EGFR and RAS and overexpression of the EGFR and transforming growth factor α (42) in addition to the IGF-1R pathway, Akt may have a critical role in regulating the ability of cellular IGFBP-3 to act as an IGF-independent growth modulator (Figure 7E, red line). Consistent with the notion, we have shown previously that SCH66336, a pharmacologic farnesyltransferase inhibitor for Ras inactivation, enhances antitumor actions of IGFBP-3 in NSCLC cells by suppressing Akt expression (11).

Our findings also show that nuclear IGFBP-3 expression is decreased by Akt3 activity in NSCLC cells. A previous report demonstrated that the distinct cellular localization of each Akt subtype; Akt3 was predominantly found in the nucleus (43). In addition, Akt moves into nucleus where it interacts with target molecules such as the FOXO family of transcription factors and the transcriptional coactivator p300 (29,44). The phosphorylated form of Akt in the nucleus has been also reported in lung, breast and prostate, as well as in thyroid cancers (45,46). Because several factors, including insulin, nerve growth factor and IGF-1, have been known to induce Akt nuclear translocation (47–49), it is possible that Akt3 regulates IGFBP-3 in the nucleus, serving as a cellular regulator of IGF-independent antitumor activities of IGFBP-3. We hypothesized that Akt3 may induce phosphorylation of IGFBP-3, leading to its degradation. However, we were unable to locate an Akt consensus sequence in IGFBP-3. It is possible that Akt recruits proteases of IGFBP-3, such as matrix metalloproteinases (50), or other mediators involved in extracellular and/or intracellular proteolysis of IGFBP-3 protein. Given the previous results indicating loss of IGFBP-3 expression through promoter hypermethylation (22,36), epigenetic regulation of IGFBP-3 expression could have been involved in Akt3-mediated transcriptional regulation of IGFBP-3 in NSCLC cells. One of the main players to maintain DNA methylation is DNA methyltransferase I (DNMT1) (51). Previous reports demonstrate that Akt stabilizes DNMT1 by decreasing glycogen synthase kinase 3β-mediated phosphorylation and subsequent ubiquitin-proteasomal degradation and by direct phosphorylation at Ser143 of DNMT1 (52,53). Akt also regulates DNMT3B expression at transcriptional and posttranscriptional levels in hepatocellular carcinoma cells (54). Interestingly, we observed the increase of IGFBP-3 mRNA expression by Akt blockade using small molecule inhibitors such as LY294002 or perifosine (Figure 2G and F) and using siRNA transfection (Figure 3E) in H322 and H661 cells, which have been reported to express no detectable IGFBP-3 mRNA due to promoter methylation (22). Therefore, it is possible that Akt might play a role in regulating methylation status of IGFBP-3 promoter by modulating DNA methyltransferase activity and/or expression including DNMT1 and DNMT3B, thereby downregulating IGFBP-3 expression in NSCLC cells. However, H226B cells expressing constitutively active Akt isoform did not show detectable difference in the methylation status of IGFBP-3 promoter as well as the DNMT1 expression levels (data not shown). Further studies are warranted to understand the mechanisms by which Akt, in particular each Akt isoform, regulates IGFBP-3 expression in relation to both genetic and epigenetic modulation including promoter methylation.

In conclusion, our study identified Akt, especially Akt3, as a key regulator of the antitumor activities of IGFBP-3 in NSCLC cells through its regulation of IGFBP-3 mRNA expression and protein stability. Our results indeed provide in vitro evidence for the significantly increased therapeutic efficacy of IGFBP-3 when combined with genomic or pharmacologic inhibitors of Akt in NSCLC cells. Our findings have a therapeutic implication for the treatment of NSCLC patients with Akt activation. Our findings may have broader application for the clinical management of patients with a variety of cancers that exhibit highly activated Akt, such as breast and prostate cancers (55). The potential therapeutic benefit of this new combinatorial regimen requires clinical investigation in patients with NSCLC. Further investigation of the detailed mechanism through which Akt3 regulates IGFBP-3 expression is also warranted.

Supplementary material

Supplementary Figures 1 and 2 can be found at http://carcin.oxfordjournals.org/

Funding

National Research Foundation of Korea (NRF), the Ministry of Science, ICT and Future Planning (MSIP), Republic of Korea (2011-0017639 and 2011-0030001); National Institutes of Health (R01 CA100816).

Conflict of Interest Statement: None declared.

Glossary

Abbreviations:

ATRA

all-trans-retinoic acid

DNMT1

DNA methyltransferase I

EV

empty vector

GAPDH

glyceraldehyde 3-phosphate dehydrogenase

HA

hemagglutinin

IGF

insulin-like growth factor

IGFBP-3

insulin-like growth factor binding protein-3

NSCLC

non-small cell lung cancer

PTEN

phosphatase and tensin homolog

RT–PCR

reverse transcription-polymerase chain reaction

TGF-β

transforming growth factor-β.