Both IGF1R and INSR Knockdown Exert Antitumorigenic Effects in Prostate Cancer In Vitro and In Vivo

Abstract

The IGF network with its main receptors IGF receptor 1 (IGF1R) and insulin receptor (INSR) is of major importance for cancer initiation and progression. To date, clinical studies targeting this network were disappointing and call for thorough analysis of the IGF network in cancer models. We highlight the oncogenic effects controlled by IGF1R and INSR in prostate cancer cells and show similarities as well as differences after receptor knockdown (KD). In PC3 prostate cancer cells stably transduced with inducible short hairpin RNAs, targeting IGF1R or INSR attenuated cell growth and proliferation ultimately driving cells into apoptosis. IGF1R KD triggered rapid and strong antiproliferative and proapoptotic responses, whereas these effects were less pronounced and delayed after INSR KD. Down-regulation of the antiapoptotic proteins myeloid cell leukemia-1 and survivin was observed in both KDs, whereas IGF1R KD also attenuated expression of prosurvival proteins B cell lymphoma-2 and B cell lymphoma-xL. Receptor KD induced cell death involved autophagy in particular upon IGF1R KD; however, no difference in mitochondrial energy metabolism was observed. In a mouse xenograft model, induction of IGF1R or INSR KD after tumor establishment eradicated most of the tumors. After 20 days of receptor KD, tumor cells were found only in 1/14 IGF1R and 3/14 INSR KD tumor remnants. Collectively, our data underline the oncogenic functions of IGF1R and INSR in prostate cancer namely growth, proliferation, and survival in vitro as well as in vivo and identify myeloid cell leukemia-1 and survivin as important mediators of inhibitory and apoptotic effects.

Among all tumor entities, prostate cancer represents the most commonly diagnosed cancer and the second leading cause of cancer death among men in Western countries (1). Initially, advanced prostate cancer is mostly hormone sensitive and therefore treated by hormonal therapies. Despite a high rate of initial responsiveness to androgen ablation, prostate tumors progress to a castration resistant, metastatic stage usually after 24–36 months (2).

An important feature of advanced stages of cancer is the inability to undergo apoptosis. In this regard, cytokines and growth factors play an important role as stimulators of survival (3, 4). A crucial player in this context is the IGF network. It consists of 3 transmembrane receptors, IGF receptor 1 (IGF1R), IGF2R, and insulin receptor (INSR) with its 2 subtypes INSR A and B, the growth factors IGF-1, IGF-2, and insulin that bind and activate the receptors with different affinities, and 6 IGF-binding proteins participating in the IGF network via interaction with IGF-1 and IGF-2. IGF1R and INSR are tyrosine kinase receptors that share a high degree of homology and thus can form homo- as well as heterodimers with each other in every constellation (5). In cancer tissues IGF1R and INSR are often elevated and their activation is associated with enhanced growth, proliferation, angiogenesis and survival (6, 7). This crucial role of the IGF network makes it an interesting and promising target for molecular therapeutic approaches, although care should be taken in view of intervening with such a highly complex signaling network.

The IGF axis can be targeted at different levels, including neutralizing or blocking monoclonal antibodies, small drug inhibitors of the receptor tyrosine kinase activities, receptor knockdown (KD) using antisense oligonucleotides and increasing IGF-binding proteins to sequester and reduce IGF bioavailability (8). Numerous clinical trials using mostly IGF1R targeting monoclonal antibodies or receptor tyrosine kinase inhibitors have been conducted or are underway for different tumor entities, including prostate cancer (9 and https://clinicaltrials.gov/). Unfortunately, the results have not met the high expectations so far. Some studies were even prematurely terminated due to lack of efficacy and side effects (10).

Most therapeutic strategies focus on the IGF1R, an approach that might be insufficient for a redundant multireceptor and multicomponent system. In our previous studies we demonstrated oncogenic effects of IGF1R and INSR in stimulating proliferation, migration, colony formation, and angiogenesis (11, 12). Moreover, we demonstrated a role of INSR similar to that of IGF1R and essential differences between benign prostate epithelial and cancer cells. In a clinical therapeutic approach, long-term treatment effects are critical for therapeutic efficacy. Therefore, the aim of this study was to demonstrate the long-term inhibitory effects of inducible IGF1R/INSR KD and confirm tumor inhibition in an in vivo xenograft mouse model. In addition, we analyzed the inhibitory pathways involved and the relation between apoptosis, autophagy and mitochondrial activity.

Materials and Methods

Cell culture and reagents

PC3, LNCaP, and DuCaP cells obtained from the American Type Culture Collection were maintained in RPMI 1640 with 10% fetal bovine serum and 2mM glutamax (Gibco, Life Technologies) at 37°C and 5% CO₂. PC3 cells transduced with short hairpin RNA (shRNA) constructs were grown in the presence of 2.5-μg/mL puromycin as selection marker. All experiments were performed in the presence of fetal bovine serum to ensure receptor activation by IGFs and INS (Supplemental Figure 1A) (12). Reagents were from Sigma-Aldrich unless otherwise specified.

KD experiments

PC3 cells were stably transduced with inducible shRNA targeting IGF1R (shIGF1R), shRNA targeting INSR (shINSR), or shRNA targeting luciferase (shLUC) (control) via lentiviral transduction using a doxycycline (dox)-inducible Tet-On shRNA system as described previously (13). Three shRNAs for each target were tested. For receptor down-regulation, cells were treated with dox (1 μg/mL). Omission of dox served as an additional control. For transient down-regulation of IGF1R, INSR, or myeloid cell leukemia 1 (Mcl-1), cells were transfected with small interfering RNAs (siRNA) using Lipofectamine 2000 (Invitrogen) as described (12). All shRNA and siRNAs used in this study are listed in Supplemental Table 1.

RNA isolation and quantitative real-time PCR (qRT-PCR)

RNA isolation (RNeasy Mini kit; QIAGEN), cDNA synthesis (iScript Select kit; Bio-Rad), and qRT-PCR (ABI Prism 7500 PCR-System; Applied Biosystems, Life Technologies) were described elsewhere in detail (14). Primer and TaqMan probes for IGF1R, INSR, and TATA-binding protein as endogenous control were designed according to the NCBI database. Primer/probe sequences are summarized in Supplemental Table 2.

Analysis of cell growth

Cells were seeded into 24-well plates at a density of 10 000 cells per well. On day 2, cells were transfected with appropriate siRNAs and 72 hours thereafter trypsinized and counted using the CASY Cell Counter and Analyzer System TTC (Schärfe System). For stably transduced shPC3 cells, shRNA expression was induced by addition of dox (1 μg/mL). Medium and dox were replaced twice per week, and cells without dox treatment were analyzed in parallel. Cells were trypsinized and counted 4 and 12 days after dox supplementation.

Mouse xenograft model

Four-week-old male BALB/c nu/nu mice (Charles Rivers) were sc injected with 1 × 10⁶ shPC3 cells (stably transduced with shLUC, shIGF1R, or shINSR) into the right and left flanks (7 mice per group). After 1 week (tumor volume, ∼50 mm³), dox (final concentration of 1 g/L) was added to the drinking water, which was changed twice a week. Tumor size was measured twice per week using a caliper. After 20 days, the mice were killed, and the remaining tumors isolated and weighed, and volume was calculated; volume = (length × width²)/2. Tumors were fixed in formalin, embedded in paraffin, sectioned, and stained with hematoxylin/eosin. The animal experiment was approved by the local animal experiments ethics committee and the Austrian Ministry of Health (BMWF-66.011/0119-II/3b/2013).

Cell apoptosis analysis

For apoptosis analysis, 180 000 shPC3 cells were seeded into 6-well plates. The next day, dox (1 μg/mL) was added. After 4 and 12 days, cells were scraped off and divided for measurement of caspase 3/7 activity (Caspase-Glo 3/7 Assay; Promega) or propidium iodine staining as described (15). Sub-G₁ cell fraction was measured by flow cytometry (Calibur; BD) to assess percentage of apoptotic cells. For Mcl-1 and survivin overexpression rescue experiments, shPC3 cells were transfected with pCMV6-Mcl-1, pCMV6-GFP-survivin, or empty vector and treated with dox for 72 hours before Western blot and apoptotic analysis. Effects of transient receptor KD in PC3, LNCaP, and DuCaP cells or treatment with the survivin inhibitor YM155 and the Mcl-1 inhibitor UMI-77 (Selleckchem) were analyzed 72 hours after or start of KD or inhibitor treatment, respectively.

Proliferation and viability assay

Cells were seeded in quintuplicates into 96-well plates. On day 2, cells were treated with dox (1 μg/mL) or transfected with siRNA. After 4 days, 1-μCi/well [methyl-³H]-thymidine (Moravek Biochemicals) was added. After 24 hours, cell viability was measured via water-soluble tetrazolium (WST-1) assay (Roche) and cell proliferation was determined via detection of 3H-thymidine incorporation (16).

Western blotting

Cells were harvested by scraping, washed with PBS and pelleted by centrifugation. For autophagic analysis, cells were additionally treated with 25μM chloroquine (CQ), 5mM 3-methyladenine (3-MA), or 10mM metformin (Mf) for 48 hours before harvesting. Cell pellets were lysed with radio-immunoprecipitation assay buffer containing 1% Triton X-100, 1mM phenylmethylsulfonylfluoride, 5mM sodium fluoride, phosphatase (1:100; Sigma-Aldrich), and protease inhibitor cocktails (1:200) (VWR International). Western blotting was performed as described (17). The next antibodies from Cell Signaling were used at a dilution of 1:1000: anti-IGF1Rβ (111A9), anti-B cell lymphoma 2 (Bcl-2) (50E3), anti-Bcl-xL, anti-light chain 3 (LC3)A/B, anti-autophagy-related gene 5 (Atg5) (D5F5U), anti-Atg7 (D12B11), anti-protein-62 (p62)/sequestosome 1, anti-Beclin1, anti-phospho-mammalian target of rapamycin (pmTOR) (Ser2481), anti-mTOR, anti-phospho-Protein kinase B (pAKT) (Ser473), anti-AKT, and anti-pIGF1R/INSR (Tyr1135/1136, 19H7). The remaining antibodies were anti-INSRβ C-19 (1:500; Santa Cruz Biotechnology, Inc), anti-poly ADP-ribose polymerase (PARP) p85 fragment (1:1000; Promega), anti-Mcl-1 (S-19) (1:500; Santa Cruz Biotechnology, Inc), anti-survivin (1:1000; Novus Biologicals), and anti-glyceraldehyde-3-phosphate-dehydrogenase (GAPDH) (1:10 000; Millipore). Antibodies were incubated overnight at 4°C. The next day, membranes were washed and incubated with secondary infrared-dye-labeled antibodies for 1 hour at room temperature and scanned using an Odyssey Infrared Imaging System (LI-COR Biosciences). For Mcl-1 quantification, both bands representing 2 splice variants (18) were included.

Immunofluorescence (IF)

IF staining was performed as described (15). Briefly, cells seeded onto glass coverslips were treated for 48 hours with dox, 25μM CQ, 5mM 3-MA, 10mM Mf, or vehicle equivalent. After fixation in 100% methanol for 15 minutes, cells were blocked and incubated with primary antibodies against LC3A/B (1:50). For identification, cells were costained with cytokeratin 8/18 (1:500). Coverslips were mounted in Vectashield Hard Set mounting medium containing 4′,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories). For imaging, a Zeiss Axio Imager M1 fluorescent microscope was used.

High-resolution respirometry

To test the impact of IGF network inhibition on energy metabolism, oxidative phosphorylation (OXPHOS) was measured via high-resolution respirometry in an Oroboros Oxygraph-2k (Oroboros Instruments). For each experiment, 2 × 10⁶ PC3 cells (shLUC/shIGF1R/shINSR) were seeded into 75-cm² culture flasks and incubated ± dox for 4 days. After treatment, cells were trypsinized, centrifuged, and resuspended in mitochondrial respiration medium (MiR05) (110mM sucrose, 20mM HEPES, 10mM KH₂PO₄, 20mM taurine, 60mM K-lactobionate, 3mM MgCl₂, 0.5mM EGTA, 1 g L⁻¹ BSA, 20mM creatine, and 280 U mL⁻¹ catalase; pH adjusted to 7.1 with KOH at 30°C) to a concentration of 2.5 × 10⁵ cells per ml. 2.3 mL of the cell suspension was added to each glass chamber of the Oxygraph-2k (19). Oxygen consumption was measured at 37°C under normoxic conditions and expressed as pmol O₂ per second per million cells (pmol s⁻¹/10⁻⁶). Real-time data acquisition and analysis were performed using DatLab v5 software (Orobos Instruments) (20).

First, basal respiration was measured with intact cells. The combined complex I and II linked respiration (coupled respiration) was measured after permeabilization with digitonin (10 μg/10⁶ cells) and successive addition of substrates (2mM malate, 200μM octanoylcarnitine, 10mM glutamate, 5mM pyruvate, and 10mM succinate) (20). The maximal electron transfer system (ETS) capacity was examined by addition of the uncoupler carbonyl cyanide m-chlorophenyl hydrazone that annihilates the proton gradient, thus enabling OXPHOS complexes to work at their maximal capacity.

Statistical analysis

Data analyses were performed using GraphPad Prism 5. ANOVA with Bonferroni post hoc test was applied to test significances between treatment groups unless otherwise stated. Significances between dox treated shLUC/shIGF1R/shINSR cells were calculated and shown in the figures. All differences between cells without dox treatment were statistically insignificant. P ≤ .05 was considered a significant difference. Bars and errors represent mean ± SD of at least 4 independent experiments unless otherwise stated.

Results

Inducible IGF1R and INSR KD for analysis of long-term treatment

For investigation of the therapeutic effect of long-term receptor inhibition, it was essential to generate cancer cell lines exhibiting a uniform, inducible long-term KD of IGF1R and INSR. For this purpose, we selected PC3 cells, which originated from a bone metastasis and represent a late stage of prostate cancer, the primary target of an anti-IGF-axis therapy. The cells were stably transduced with inducible shRNA constructs targeting IGF1R (shIGF1R), INSR (shINSR), and luciferase (shLUC, control) using lentiviruses. Of 3 tested shRNA sequences, shIGF1R sequence 2 and shINSR sequence 1, which targets both INSR isoforms, were selected for further analysis according to their ability to efficiently down-regulate IGF1R and INSR at the mRNA and protein level after induction with dox (Figure 1, A and B, and Supplemental Figure 2, A and B).

Figure 1.

IGF1R and INSR KD attenuated growth of prostate cancer cells. Efficiency of IGF1R and INSR KD by dox induction of shRNAs for 4 days was confirmed at mRNA (A) and at protein (B) level by qRT-PCR and Western blot analysis in shPC3 cells (shLUC control, shIGF1R, and shINSR). C, Representative images of shPC3 cells ± dox incubation for 4 and 12 days. D, Growth of shLUC control, shIGF1R, and shINSR cells ± dox stimulation for 4 and 12 days. E, Analysis of DNA synthesis assessed by measurement of [methyl-³H]-thymidine incorporation ± dox for 4 and 12 days. All data represent mean ± SD from 4 independent experiments (*, P < .05; **, P < .01; ***, P < .001, ANOVA).

IGF1R and INSR are both essential for prostate cancer cell growth

Receptor KD was induced with dox for 4 or 12 days in PC3 cells. Cell numbers were significantly reduced at both time points in shIGF1R and shINSR cells when compared with the shLUC control. The inhibitory effect was, however, much more prominent after IGF1R KD (Figure 1C). To quantify the observed effects we counted the cells and measured DNA synthesis up to 12 days after shRNA induction (Figure 1, D and E). Four days after induction of IGF1R KD, cell growth had ceased. After this time point, cell number started to decline, and by day 12, almost all cells had died. Upon INSR KD, the cellular growth rate was also significantly reduced. However, the decrease in cell number was much slower than upon IGF1R KD. DNA synthesis measured by [methyl-³H]-thymidine incorporation revealed that upon IGF1R KD cell proliferation was inhibited by 90% already at day 4, whereas a reduction of only 35% and 50% was observed upon INSR KD for 4 and 12 days, respectively (Figure 1E). To ensure that the observed effects were specific for IGF1R and INSR KD, we repeated the experiments using an independent set of shRNAs with comparable results (Supplemental Figure 2, C–E). Overall, the characterization of long-term KD confirmed the crucial role of IGF1R and INSR for the proliferation and survival of prostate cancer cells.

IGF1R or INSR KD trigger tumor eradication in vivo

The impact of IGF1R and INSR KD on tumor growth in vivo was investigated in a nude mouse xenograft model. The used approach tested the effect of IGF1R and INSR KD on already established tumors to reflect a clinical situation. Stable inducible shRNA-transduced cells (shLUC, shIGF1R, and shINSR) were injected into the flanks of male nude mice and tumors established in the absence of dox. After the tumors reached a size of about 50-mm³ dox administration via the drinking water was started. Although the volume of shLUC tumors continued to increase, tumor growth had ceased in IGF1R and INSR KD tumors by day 10 and subsequently began to decline thereafter (Figure 2A). Twenty days after treatment start, the volumes of IGF1R and INSR KD tumors had decreased significantly to 50% and 83% of their respective volumes before dox administration. By contrast, control shLUC tumors doubled in volume until the end of the experiment (Figure 2, A and C). Likewise, tumor weight of control tumors was significantly increased in comparison with IGF1R and INSR KD tumors (Figure 2B). The induction of IGF1R or INSR KD not only stopped tumor growth but moreover led to eradication of most of the tumors (Figure 2A, inset). Twenty days after induction of IGF1R and INSR KD, only 1 and 3 of the initial 14 tumors remained, respectively. In most cases, only tumor cell-free matrix remnants were identified upon histopathological investigation of the injection sites (Figure 2D). In agreement with in vitro growth inhibition characteristics, IGF1R KD resulted in faster tumor volume decline and more extensive tumor depletion in comparison with tumors with INSR KD.

Figure 2.

IGF1R and INSR KD eradicated xenograft tumors in nude mice. Seven BALB/c nu/nu mice per group were injected with shIGF1R, shINSR, or shLUC control cells in both flanks. After tumor establishment for 1 week, KD treatment was started by the addition of dox to the drinking water. A, Tumor volume was measured using a Caliper twice a week. Data represent mean ± SEM of 7 mice (14 tumors) (*, P < .05; **, P < .01; ***, P < .001, ANOVA). B, After 20 days, the mice were killed, and tumor weight was measured. C, Representative tumor images of the different treatment groups after preparation. Scale bar, 1 cm. D, Formalin-fixed and paraffin-embedded sections of shLUC, shIGF1R, and shINSR tumors were stained with hematoxylin/eosin (H&E) and scanned for tumor cells. Scale bar, 200 μm.

In summary, IGF1R as well as INSR KD stopped the growth of established prostate xenograft tumors in mice. Most tumors not only diminished but vanished completely upon induction of IGF1R or INSR KD.

IGF1R or INSR KD induces cell death

To further investigate the mechanisms underlying the inhibitory effects of IGF1R or INSR KD, we examined the induction of apoptosis and analyzed different apoptotic events from early to late apoptosis. First, we measured the activity of effector caspases 3/7. Four days after dox addition, shIGF1R cells already showed a strongly enhanced caspase 3/7 activity compared with shLUC control cells (Figure 3A). Increased caspase activity was also detected upon INSR down-regulation but only at the later time point (12 d). At this time point, there were insufficient numbers of viable shIGF1R cells for valid analysis. Thus, IGF1R KD led to fast activation of apoptosis, whereas upon INSR KD, induction of apoptosis was delayed. Consistently, cleaved PARP, a marker for later apoptosis, was detectable only in IGF1R KD cells (Figure 3B). Finally, we analyzed DNA fragmentation, another key feature of apoptosis. A significant amount of fragmentation was observed already after 4 days of IGF1R and INSR KD (Figure 3C). At day 12, the number of cells exhibiting DNA fragmentation was further increased in comparison with shLUC control cells with 72% for IGF1R KD but just 46% for INSR KD.

Figure 3.

Induction of apoptosis upon IGF1R and INSR KD. A, Activity of effector caspases 3/7 was measured in shPC3 cells 4 and 12 days after incubation ± dox (RLU, relative luminescence units). B, Cleaved PARP (cPARP) protein level was analyzed by Western blotting after 4 and 12 days ± dox (representative Western blotting). C, Sub-G₁ phase apoptotic cells were analyzed by flow cytometry after 4 and 12 days ± dox and subsequent staining with propidium iodide. D, Analysis of cell viability after 4 and 12 days ± dox using the water-soluble tetrazolium reagent. Data represent mean ± SD from 4 independent experiments (*, P < .05; **, P < .01; ***, P < .001, ANOVA). Differences between cells without dox treatment were statistically insignificant.

Cell viability measured after receptor KD reflected the induction of apoptosis (Figure 3D). This result was additionally confirmed using an independent set of inducible shRNAs targeting other sites in the IGF1R and INSR mRNAs (Supplemental Figure 2E).

In conclusion these data show that IGF1R and INSR KD inhibited cell growth and induced a strong apoptotic response. Although IGF1R KD resulted rapidly in a strong apoptotic response, INSR KD-induced apoptosis was temporally delayed.

KD of IGF1R or INSR decreases antiapoptotic proteins

The Bcl-2 family is one of the most well-characterized regulators of apoptotic events. To identify potential changes at the protein level upon receptor down-regulation we performed Western blot analyses of the antiapoptotic Bcl-2 members Mcl-1, Bcl-2, and Bcl-xL. Four days after IGF1R KD, protein levels of all 3 members were decreased. After 12 days, this decrease was even greater. In comparison, INSR KD did not significantly change the levels of Bcl-2 and Bcl-xL even after 12 days. However, Mcl-1 protein levels were decreased after 12 days of INSR KD (Figure 4A). Notably, Mcl-1 down-regulation using siRNA (Supplemental Figure 1B) closely recapitulated the effects observed by IGF1R or INSR KD (Figure 4, B–D). For example, cell growth was decreased and DNA fragmentation and caspase 3/7 activity significantly increased in all shPC3 cells (shLUC, shIGF1R, and shINSR) in the absence of dox (ie, in the absence of IGF1R or INSR KD). These findings suggest that down-regulation of Mcl-1 is a critical step in the observed inhibitory effects after IGF1R or INSR KD.

Figure 4.

Mechanisms of tumor cell inhibition in response to IGF1R and INSR KD. A, Alteration of antiapoptotic Bcl-2 family proteins Mcl-1, Bcl-2, and Bcl-xL by Western blot analysis 4 and 12 days after IGF1R and INSR KD. Representative Western blotting images together with mean relative band intensities of 3 independent experiments calculated relative to GAPDH are shown (top). Histograms represent analysis of Bcl-2 protein band intensities after 12 days ± dox. Analysis of DNA synthesis (B), DNA fragmentation (C), and caspase 3/7 activity (D) in shPC3 cells after Mcl-1 KD using siRNAs in the absence of dox. E, Survivin protein levels after IGF1R/INSR KD for 12 days determined by Western blotting. A representative Western blotting image and quantitative analysis of 3 independent experiments are shown. Analysis of DNA synthesis (F) and DNA fragmentation (G) after treatment of shPC3 cells with the survivin inhibitor YM155 (3.5nM [F] and 25nM [G], respectively) for 72 hours in the presence or absence of dox. H, Inhibition of cell growth in response to survivin inhibitor YM155 (3.5nM, 72 h) and Mcl-1 inhibitor UMI-77 (5μM, 72 h). I, Rescue from apoptosis by simultaneous overexpression of survivin and Mcl-1. DNA fragmentation of shIGF1R/INSR cells was measured 72 hours after overexpression of survivin and Mcl-1 in the presence or absence of dox. Histograms represent mean ± SD (*, P < .05; **, P < .01; ***, P < .001, ANOVA).

Besides Bcl-2 family members, inhibitor of apoptosis proteins are potent apoptotic regulators. In particular, survivin has been implicated in the progression and therapy resistance of prostate cancer (21, 22). Upon IGF1R/INSR KD, survivin protein levels were significantly reduced in shPC-3 cells. In contrast to Mcl-1, however, the reduction of survivin was more pronounced upon INSR KD than IGF1R KD (Figure 4E). The role of survivin was further investigated employing the inhibitor YM155, which down-regulates the protein (Supplemental Figure 1C). Treatment of shPC3 cells with YM155 resulted in a robust inhibition of cell growth and induction of apoptosis (Figure 4, F and G, and Supplemental Figure 1D). Notably, survivin inhibition showed a synergistic effect in combination with INSR KD but not IGF1R KD supporting its role as an effector of INSR KD (Figure 4, F and G).

Taken together, the expression levels of antiapoptotic proteins after receptor KD suggested that combined suppression of Mcl-1 and survivin is the central mechanism of tumor cell inhibition. Both proteins have previously been identified as crucial factors for tumor cell survival, and small molecule inhibitors are currently being tested in clinical or preclinical studies in several tumor types (23, 24). In shPC3 cells, the survivin as well as the Mcl-1 inhibitor efficiently reduced cell growth, which was further enhanced significantly upon combined treatment (Figure 4H and Supplemental Figure 1E).

Given the central roles of Mcl-1 and survivin in the execution of receptor KD-induced apoptosis, we tested whether their overexpression would rescue the cells from receptor KD-triggered cell death. Either 1 or both of the 2 antiapoptotic proteins were transiently overexpressed before receptor KD induction in shPC3 (Supplemental Figure 1F). Although overexpression of survivin or Mcl-1 singly could not prevent induction of cell death, concomitant overexpression of both significantly reduced receptor KD-induced DNA fragmentation (Figure 4I).

Taken together, the observed expression pattern of antiapoptotic proteins and the partial rescue from apoptosis by a combination of Mcl-1 and survivin overexpression highlight these 2 proteins as crucial mediators of IGF1R and INSR KD apoptotic effects.

IGF1R KD stimulates autophagy

Autophagy can act as a tumor suppressor by degrading damaged proteins and organelles but is also considered a rescue mechanism supporting survival and promoting tumor growth (25, 26). To investigate whether autophagy is involved in the tumor eradication processes observed upon receptor KD, we analyzed the levels of the autophagic marker LC3, which is converted to membrane-bound LC3-II upon induction of autophagy (27). Increased levels of LC3-II were observed already 4 days after IGF1R shRNA induction, and this marker was further increased after 12 days (Figure 5, A and B). Real-time PCR analysis confirmed mRNA induction of the gene (LC3B) upon IGF1R KD (Supplemental Figure 3A).

Figure 5.

Induction of autophagy upon IGF1R and INSR KD. A and B, Activation of autophagy after induction of receptor KD was detected by Western blot analysis of the autophagy marker LC3-II. Representative Western blotting image (A) and determination of protein band intensities (B). C, IF staining of the autophagy marker LC3-II in cells treated for 48 hours ± dox and in addition treated with autophagy modulators ± CQ (25μM), ± 3-MA (5mM), and ± Mf (10mM). Scale bar, 40 μm. D, IF images of dox treated cells with higher magnification. Scale bar, 20 μm. E, Detection of autophagy marker proteins using Western blotting. F, DNA synthesis of shPC3 cells after 4 days ± dox treated with autophagic modulators for 48 hours. G, Measurement of mitochondrial respiratory O₂ consumption in intact cells (basal state), in cells permeabilized by digitonin and stimulated with metabolites fueling electrons into OXPHOS complexes CI and II (malate, pyruvate, glutamate, and succinate) and maximal ETS capacity in uncoupled cells. H, Mitochondrial mass analyzed by determination of the mitochondrial marker protein cytochrome c oxidase (COX IV) using Western blotting.

In order to analyze autophagy enhancement in greater detail, we employed the autophagic modulators CQ (which prevents fusion of endosomes and lysosomes) and 3-MA (which blocks autophagosome formation) and Mf (which triggers autophagy by AMP-activated protein kinase activation). CQ treatment increased LC3-II levels in shPC3 cells due to autophagosome accumulation (Figure 5, C–E, and Supplemental Figure 3B). Treatment with 3-MA resulted in LC3-II levels comparable with treatment with dox alone, whereas Mf did not induce autophagy in our PC3 cells. p62, an adaptor protein that targets substrates for autophagy (28), increased significantly after IGF1R KD but not after INSR KD (Figure 5E). CQ further enhanced the effect of IGF1R KD on p62 accumulation, whereas 3-MA treatment induced no further change (Figure 5E and Supplemental Figure 3C). Real-time PCR analysis showed no difference on p62 mRNA levels, excluding an induction of gene transcription (Supplemental Figure 3D). Analysis of the autophagy markers Atg7, Atg5–12, and Beclin1 showed high abundance of these proteins but no significant differences between treatments (Figure 5E and Supplemental Figure 3, E–G). None of the autophagy modulators had any significant effect on cell growth (Figure 5F). In conclusion, the observed autophagic effects upon IGF1R KD seemed to accompany but not to be crucial for the initiated processes that finally result in cell death.

The phosphoinositide 3-kinase-Akt-mTOR pathway, a key regulatory pathway of autophagy, and survival was investigated by Western blotting using antibodies detecting the active, phosphorylated forms of Akt-1 and mTOR, and both were found to be constitutively active (Supplemental Figure 4), which is not surprising given the phosphatase and tensin homolog (PTEN)-negative phenotype of PC3 cells (29). KD of IGF1R or INSR had no observable effect on this pathway. Receptor KD also had no significant effect on the activation status of the ERK1/2 MAPK pathway, another frequently activated tumor driver pathway (30), which was shown to be activated in DU145 PCa cells via IGF-1 stimulation for example (31) (data not shown).

IGF1R or INSR KD does not affect mitochondrial energy metabolism

Apoptosis and autophagy are tightly connected to cellular metabolism and energy supply (32). We speculated that IGF1R or INSR KD could potentially result in inhibition of mitochondrial respiration leading to subsequent breakdown of the energy supply and thus in cell death. To address this question, we investigated the effect of receptor KD on mitochondrial respiratory function using high-resolution respirometry to measure respiration and OXPHOS under standardized conditions.

After 4 days of IGF1R or INSR KD, control cells showed a basal oxygen consumption of 69 pmol/(s per 10⁶ cells), IGF1R and INSR KD cells 62 and 83 pmol/(s per 10⁶ cells), respectively (Figure 5G). Substrate-triggered combined respiration of mitochondrial complexes I and II in permeabilized cells was approximately 3-fold higher and maximal ETS respiration approximately 5-fold higher than basal cell respiration, with no significant differences between IGF1R, INSR, or control KD treatments (Figure 5G). Furthermore, determination of the mitochondrial protein cytochrome c oxidase IV by Western blotting revealed no effect on mitochondrial mass upon IGF1R or INSR KD (Figure 5H). Taken together, these data revealed no indication of an involvement of mitochondrial respiratory function in the inhibitory effects of receptor KD.

Both androgen receptor (AR)-negative and AR-positive prostate tumor cells are responsive to IGF1R and INSR KD

The shPC3 cells adopted for long-term inducible receptor KD represent an AR-negative advanced PCa phenotype, which is infrequent in comparison with AR-positive advanced tumors. To demonstrate that the observed downstream effects of IGF1R/INSR KD also apply to AR-positive cells, we confirmed down-regulation of antiapoptotic proteins using an siRNA approach in PC3 cells and the AR-positive cell lines LNCaP and DuCaP (Supplemental Figure 1G). Down-regulation of Mcl-1, Bcl-2, Bcl-xL, and survivin after siRNA receptor KD in PC3 cells showed the same pattern as observed in shPC3 cells and LNCaP cells displayed a very similar pattern. In DuCaP cells, no significant effect on survivin protein was visible at this time point, whereas IGF1R KD induced profound Mcl-1 and weaker Bcl-2 and Bcl-xL down-regulation (Figure 6, A and B). Despite these variations, proliferation of all 3 cell lines was markedly inhibited upon siRNA receptor KD as shown before (11, 12). Likewise, growth of all 3 cell lines was significantly inhibited upon KD of Mcl-1 (Figure 6C and Supplemental Figure 1H) or treatment with the survivin inhibitor YM155 (Figure 6D). In essence, siRNA KD confirmed the strong inhibitory effect of IGF1R or INSR KD and the downstream antiapoptotic effector proteins in AR-negative and AR-positive PCa cell lines.

Figure 6.

Effect of IGF1R and INSR KD in AR-positive LNCaP and DuCaP cells. Down-regulation of antiapoptotic proteins after receptor KD using siRNA targeting was investigated in AR-positive LNCaP and DuCaP cells in comparison with AR-negative PC3 cells. Western blot images (A) and quantitative protein band analysis (B) of apoptotic proteins 72 hours after siRNA transfection. DNA synthesis 72 hours after siRNA Mcl-1 KD (C) and after survivin KD with YM155 (PC3, 3.5nM and 25nM; LNCaP and DuCaP, 50nM and 100nM) (D). Data represent mean ± SD from 3 experiments (*, P < .05; **, P < .01; ***, P < .001, ANOVA and t test [C], respectively).

Discussion

The IGF network regulates several important processes in normal cell physiology. Its central impact on cancer hallmarks such as proliferation, resistance to cell death, and tumor cell invasion makes it a major player in cancer development, therapy resistance, and tumor progression and thus a prime target for the development of new cancer therapies. Previous characterization of the IGF network in AR-positive and AR-negative prostate cancer cell lines revealed sensitive responses to IGF1R and INSR KD in all cell lines by cessation of proliferation, migration, and angiogenesis (11, 12). In the present study, we established cell line models for inducible KD of IGF1R or INSR for investigation of long-term effects of receptor inhibition in vitro and in vivo and characterization of induced changes leading to cell death. The cell line chosen for this study represents advanced, metastatic PCa, which is a focus for the development of new therapeutic approaches. We not only aimed to investigate long-term effects on tumor growth but also to test the therapeutic efficacy of IGF1R and INSR KD on already established tumors in an animal model. With this approach, we mimicked the clinical situation when patients with advanced disease are treated. Furthermore, this setting represents a model to study an AR-independent tumor driver pathway, inhibition of which represents a promising approach in combination with new-generation AR-targeting treatments.

IGF1R as well as INSR KD showed remarkably strong tumor regression effects in this setting. Twenty days upon induction of IGF1R or INSR KD, nearly all xenograft tumors completely vanished. IGF1R, but also the INSR, appears to be crucial for prostate cancer cell survival, because KD directly affected cell survival and tumor growth both in vitro and in vivo. IGF1R KD rapidly led to strong inhibition of cell growth, proliferation, and survival followed by cell death. With INSR KD, these effects were also observed albeit less pronounced and with some delay. Several markers of apoptosis were triggered by both KDs, including activation of effector caspases 3/7 followed by PARP cleavage and DNA fragmentation. Analysis of antiapoptotic proteins identified Mcl-1 and survivin as consistently down-regulated upon KD of IGF1R or INSR. IGF1R KD showed a stronger attenuation of Mcl-1, whereas the reduction of survivin was more pronounced after INSR KD. KD of IGF1R in addition reduced protein levels of Bcl-2 and Bcl-xL. A similar modulation of antiapoptotic proteins after receptor KD was seen in AR-positive PCa cell lines. Bcl-2 and Bcl-xL inhibition in addition to Mcl-1 and survivin is expected to intensify the apoptotic stimulus of IGF1R KD when compared with INSR KD and might be responsible for the more rapid growth inhibition and cell death observed after IGF1R KD.

The combined reduction of Mcl-1 and survivin seem to be the main reason for the observed efficient tumor inhibition by IGF1R or INSR KD. In support of this hypothesis, combined inhibition of both proteins showed a strong inhibitory effect. Although elevation of either Mcl-1 or survivin protein levels by transient overexpression in rescue experiments showed little effect, overexpression of both significantly attenuated induction of apoptosis.

Several members of the Bcl-2 family were reported to be dysregulated in prostate cancer. Overexpression of Bcl-2 has been associated with tumor progression, therapy resistance, and therapy refractory disease (33–35). In agreement with our data, decreased Mcl-1 expression has been shown to induce apoptosis (36, 37). Mcl-1 is reported to be the most abundantly expressed Bcl-2 family member in several cancer cell lines, including those from prostate cancer (38). In prostate cancer, tissue Mcl-1 protein levels are increased, and Mcl-1 was identified as a major prosurvival signal induced by androgen-deprivation therapy (39).

The expression of survivin in prostate cancer is associated with proliferation and tumor aggressiveness (21, 40), and its knockout impairs prostate carcinogenesis in a genetic mouse model (41). In agreement with its reported role as a mediator of growth factor regulated effects (42) and an inhibitor of apoptosis, its KD chemosensitizes prostate tumor cells and inhibits tumor cell growth in vitro and in vivo (43, 44). The small molecule survivin inhibitor YM155 is studied in clinical trials in several tumor entities, including advanced prostate cancer (45).

Apoptosis is closely connected to autophagy and both processes share common pathways (25, 46). In general, autophagy is considered a rescue mechanism to suppress apoptosis and support survival. The autophagy marker LC3-II was elevated in response to IGF1R KD, indicating an autophagic stimulus, but clearly this was insufficient to escape cell death. The concurrent accumulation of p62, an adaptor protein directing proteins into the phagosome, points to blockade in the further execution of the authopagic pathway. The mitochondria play a central role in the initiation and execution of programmed cell death, which is modulated by metabolic stress and compromised energy supply (32). However, mitochondrial respiratory function remained unchanged upon receptor KD, suggesting no such involvement. High-resolution respirometry revealed intact electron transfer of the mitochondrial OXPHOS complexes and mitochondrial masses also remained unchanged after receptor KD.

In conclusion, our data reveal a strong tumor-repressive ability of IGF1R and INSR KD in vitro and particularly in vivo, in an experimental setting mimicking treatment of established tumors. This is contrary to preclinical models and human studies where inhibition of IGF1R did not show sufficient beneficial effects in prostate cancer (47, 48). An explanation for this discrepancy might be the fact that RNA interference (RNAi)-based approaches, as used here, are fundamentally different from pharmacological inhibition (49). For example, treatment with small-molecule protein kinase inhibitors does not prevent protein interaction and receptor complex formation, which is eliminated by RNAi KD. This might reduce the efficacy of chemical inhibitors within a redundant signaling cascade.

The fact that KD of either IGF1R or INSR substantially affected proliferation and survival in vitro as well as in vivo suggests that both receptors are indispensable for cell survival although the network consists of 3 main receptors (IGF1R, INRA, and INSRB), which seem to activate similar pathways. Attempts to identify intracellular downstream signaling differences have yielded conflicting results. Structural analysis of IGF1R and INSR cytoplasmic tails identified different regions responsible for cellular growth, survival, or oncogenic transformation (5, 50). Studies with chimeric receptors, with exchanged regions between IGF1R and INSR, revealed that structural differences provide overlapping as well as distinct intracellular functions (51). Furthermore, the variety of substrates as well as their availability, location, and ratio probably contributed to different outcome effects.

The remarkable antioncogenic effect of IGF1R and INSR KD in the prostate cancer xenograft model, nevertheless, calls for a continuation of efforts to exploit inhibitors of the IGF regulatory network and their application for the therapy of advanced prostate cancer patients, despite the discouraging results of previous clinical trials.