Delayed Activation of Insulin-like Growth Factor-1 Receptor/Src/MAPK/Egr-1 Signaling Regulates Clusterin Expression, a Pro-survival Factor*

Secretory clusterin protein (sCLU) is a general genotoxic stress-induced, pro-survival gene product implicated in aging, obesity, heart disease, and cancer. However, the regulatory signal transduction processes that control sCLU expression remain undefined. Here, we report that induction of sCLU is delayed, peaking 72 h after low doses of ionizing radiation, and is dependent on the up-regulation of insulin-like growth factor-1 as well as phosphorylation-dependent activation of its receptor (IGF-1 and IGF-1R, respectively). Activated IGF-1R then stimulates the downstream Src-Mek-Erk signal transduction cascade to ultimately transactivate the early growth response-1 (Egr-1) transcription factor, required for sCLU expression. Thus, ionizing radiation exposure causes stress-induced activation of IGF-1R-Src-Mek-Erk-Egr-1 signaling that regulates the sCLU pro-survival cascade pathway, important for radiation resistance in cancer therapy.

Ionizing radiation (IR) 1 is commonly used for therapy against many cancers. Elucidation of refractory responses to clinically relevant doses of IR are under intense investigation. Understanding intracellular processes induced by IR that lead to radiation resistance, especially signal transduction pathways that precede gene expression, could reveal interventions that bring about a more favorable clinical outcome. Although it was once thought that the only important cellular responses to IR originated from DNA damage, it is clear that IR creates many different lesions in macromolec-ular targets within the cell that set in motion cascades of responses in both irradiated as well as in neighboring nonirradiated cells. For example, activation of epidermal growth factor and tumor necrosis factor-␣ receptors (i.e. EGFR and tumor necrosis factor receptor, respectively) after IR can stimulate intracellular signaling cascades that lead to gene expression that, in turn, mediates resistance to IR-induced cell death (1). IR may also stimulate the production of ligands (e.g. EGF, TGF-␣, IGF-1, and tumor necrosis factor-␣) that, in turn, activate downstream signaling (2)(3)(4). Both EGF and IGF-1 can activate the Src/MAPK signal transduction pathway (1). Because these signaling cascades can enhance survival of cancer cells, their stimulation can limit radiotherapy efficacy in various human malignancies that overexpress specific receptors. A better understanding of these signal transduction mechanisms is needed to improve radiotherapy.
Expression from the CLU gene encodes two protein forms, a non-glycosylated pro-death cytoplasmic/nuclear form (5) and another highly glycosylated secretory protein (sCLU) that provides cytoprotection after various genotoxic stresses, possibly due to its role as a molecular chaperone (6). sCLU and nuclear CLU proteins originate from separate mRNA species that can be targeted by siRNA approaches (5). sCLU expression is induced by a variety of cytotoxic agents (e.g. topotecan, taxol, and thapsigargin) (7); however, the signal transduction mechanisms regulating its induced expression remain a mystery.
We identified sCLU as an IR-induced protein/transcript (24). sCLU was induced by Ն2 cGy in various breast and colon cancer cells, with peak induction of sCLU noted 24 -72 h after IR, as measured by promoter activity, transcript, and protein levels (7). Although specific transcription factors have been implicated in sCLU regulation after specific genotoxic stresses, including repression by c-Fos (25) and possible activation by c-Myc (19) and AP-1 (26), the signaling pathways involved and transcription factors that directly regulate the expression of this gene after IR have not been elucidated.
Using siRNA, we demonstrated a direct role for sCLU in cell survival after IR. Activation of the IGF-1/IGF-1R/Src/Mek/Erk signaling cascade by clinically relevant doses of IR in MCF-7 breast cancer cells was required for sCLU induction. MAPK signaling was detected within minutes in cells after IR as reported (1); however, a dramatic re-activation of Src/MAPK signaling 24 -72 h after IR was specifically required for CLU promoter activation and elevated endogenous sCLU protein levels. A role for IGF-1/IGF-1R in sCLU induction after IR was elucidated. Because sCLU is a cytoprotective protein, these data define selective radio-resistant tumor cell phenotypes that overexpress and/or activate IGF-1R and offer specific mechanisms for therapeutic intervention.
Cell Culture-Human MCF-7 breast and PC-3 prostate cancer cells were grown in RPMI 1640 medium with 5% fetal bovine serum at 37°C in a humidified incubator with 5% CO 2 , 95% air. MCF-7 1403 cells that contain the human CLU promoter driving luciferase were described (7). Where applicable, MCF-7 cells were transiently transfected using Effectene (Qiagen; Valencia, CA) or Lipofectamine Plus (Invitrogen). All cell lines were mycoplasma-free.
The following experimental conditions were used for growth factor inhibitor treatments. Cells were serum-starved for 24 h and pretreated for 1 h with or without AG1478 or AG1024 in 0.01% Me 2 SO at indicated doses. Cells were mock-or IR-treated (5 Gy) as described (7) or treated with IGF-1 or EGF in serum-free medium. Medium was changed 1 h later to medium containing 1% serum plus inhibitors or ligands described above. After 24 h, medium was changed, and fresh 1% serumcontaining medium was added every 24 h until harvest. For transfections, cells (1 ϫ 10 5 ) were plated onto 35-mm dishes, transfected, and serum-starved. For Western blots, cells (5 ϫ 10 5 ) were plated onto 10-cm dishes and serum-starved. PP1 and U0126 treatments were performed on cells with similar density as noted above. Cells were pretreated for 1 h with PP1, U0126, or Me 2 SO and mock-irradiated or exposed to 5 Gy. In Fig. 5, A and B, medium containing 20 M PP1 was removed 24 h after IR. In all other experiments, continuous exposures of PP1 or U0126 were used, and new PP1 or U0126 in fresh media were added to cells every 24 h until harvest at 72 h or as indicated.
Luciferase Reporter Assays-Luciferase reporter assays using cotransfection of 4250-luc with or without an RSV ␤-gal constitutive reporter were performed using the luciferase assay system (Promega; Madison, WI) as described (7). MCF-7 cells co-transfected with 4250-luc and Src CA, Src KD, dnMek-1, or dnErk1/2 expression constructs were mock-or IR-treated, and CLU promoter-driven luciferase activities were assessed. MCF-7 cells were transfected 24 h before treatment with inhibitors and/or 5 Gy and harvested in 1ϫ reporter lysis buffer at the indicated times. ␤-Galactosidase assays were used to control for transfection efficiency using 4250-luc. All experiments were normalized for protein using Bradford assays (Bio-Rad). Each dose/time point was completed in triplicate, and paired Student's t tests were performed for statistical significance.
Enzyme-linked Immunosorbent Assays-Enzyme-linked immunosorbent assays for IGF-1 were performed as directed (R&D Systems, Minneapolis, MN). MCF-7 cells (1 ϫ 10 5 ) in 35-mm dishes were grown in medium without whole serum. Cells were mock-or IR-treated with 5 Gy 24 h later. At various times (0, 24, 48, and 72 h), medium (1 ml) was removed and stored at Ϫ80°C. Samples (50 l) were compared with an IGF-1 standard curve to determine concentrations. Results are the mean Ϯ S.E. of three experiments, each performed in triplicate.
DNA Pull-down Assays-A biotinylated 1403-bp human CLU promoter was amplified from the 4250-luc plasmid using primers ordered from Integrated DNA Technologies (Coralville, IA): 5Ј-GAT CCA TTC CCG ATT CCT-3Ј and 5Ј-5-biotinylated-AGC CAA GCT TCC TGT GCC-3Ј. Nuclear extracts were harvested from Me 2 SO-, PP1-, or U0126-treated or mock or IR-exposed MCF-7 cells as described (27). Biotinylated CLU promoter (3 g) was incubated with 10 l of streptavidin beads (Oncogene Research Products; Boston, MA) for 1 h at room temperature. Complexes were washed in binding buffer (50 mM Tris-HCL, pH 7.5, 5 mM MgCl 2 , 2.5 mM EDTA, 2.5 mM dithiothreitol, 250 mM NaCl, 0.25 g/l poly(dI-dC)⅐poly(dI-dC), and 20% glycerol), and 100 g of nuclear extract was added at room temperature for 20 min. Binding buffer (1 ml) was added, and complexes were incubated at 4°C overnight. Complexes were washed twice 24 h later in binding buffer, and associated proteins were separated by 12% SDS-PAGE for Western analyses.

Activation of MAPK by IR-IR
can activate both the EGFR and IGF-1R signaling cascades shortly after exposure. Because sCLU is expressed with delayed kinetics in all responding cancer cells (e.g. MCF-7) examined thus far after IR, with induction occurring 48 -72 h after treatment (7, 28), we sought to elucidate whether MAPK signal transduction processes regulated such a delayed-induced gene product. As reported after high doses of IR (29), elevated levels of P-Src (Tyr 416 ), P-Mek (Ser 218/222 ), and P-Erk1/2 (Tyr 204 ) were noted, starting at 0.25 h and lasting for 2 h (Fig. 1, A and B). However, sCLU induction at these early times was not observed. Interestingly, a dramatic re-activation of the Src/Mek/Erk1/2 pathway was noted 24 -72 h after IR that correlated well with sCLU induction ( Fig. 1, A and C). Thus, a delayed functional reactivation of the MAPK cascade correlated well with sCLU protein induction in irradiated MCF-7 cells.
EGFR Is Not an Upstream Activator of Clusterin-As noted above, IR can activate EGFR, resulting in the stimulation of Raf-Mek-Erk, MAPK signaling (4). We, therefore, examined a role for EGFR in CLU induction after IR using MCF-7 cells transiently transfected with a human CLU promoter-reporter construct (4250-luc, a 4250-bp region of the CLU promoter controlling the firefly luciferase reporter). Increasing doses of AG1478, a selective EGFR inhibitor, did not affect IR induction of the CLU promoter in serum-starved MCF-7 cells 72 h post-IR ( Fig. 2A). Furthermore, increasing amounts of exogenous EGF had no affect on CLU promoter activity or sCLU protein synthesis (see Supplemental Fig. 1), suggesting that EGFR did not play a role in sCLU induction. These results were confirmed by Western analyses, wherein sCLU was not induced by EGF or the EGFR inhibitor, AG1478 (Fig. 2B, lanes 4 and 5). Time-course analyses confirmed these results (see Supplemental Fig. 1). Consistently, AG1478 attenuated EGFR activation (Fig. 2C) by IR, as monitored by the autophosphorylation of Tyr 1068 of EGFR, but did not alter IR induction of sCLU (Fig. 2B, lane 3). Therefore, we concluded that EGFR did not play a role in sCLU induction after IR.
Elevated levels of phosphorylated IGF-1R in MCF-7 cells after IR suggested that IGF-1 production may be induced after IR (5 Gy). Media collected from serum-starved MCF-7 cells at various times post-IR showed elevated (ϳ20%) levels of IGF-1, beginning 24 h after IR with sustained levels noted 72 h post-IR (Fig. 3G). Thus, MCF-7 cells have high basal IGF-1 levels, as reported (32), with increased levels of both IGF-1R and IGF-1 noted 24 -72 h post-IR, consistent with enhanced phosphorylation of IGF-1R. These changes are consistent with a putative IR-induced IGF-1/IGF-1R autocrine feedback loop that, in turn, up-regulated sCLU levels.
c-Src Is an Upstream Activator of sCLU Induction after IRc-Src, but not other Src family members, can activate and be activated by IGF-1R (33,34). PP1, a selective c-Src kinase inhibitor, abrogated CLU promoter induction after low doses of IR and partially blocked induction (by 50%) after higher doses of IR (Fig. 4A). As noted, sCLU was induced in MCF-7 cells by low doses of IR (0.1-5 Gy) (7), and PP1 (20 M) blocked this induction (Fig. 4B).
We then determined the lowest dose of PP1 as a continuous 72-h treatment that could inhibit CLU promoter activity after 5 Gy. PP1 (5 M) significantly decreased basal levels of the CLU promoter and sCLU protein and prevented IR-inducible CLU promoter (Fig. 4C) and endogenous sCLU protein levels (Fig. 4D). Phosphorylation of Fak, a known c-Src substrate (Tyr 925 ), decreased with increasing PP1 doses (Fig. 4E). Thus, PP1 effectively blocked IR-inducible c-Src activity at doses that prevented sCLU protein induction.
To confirm a role for c-Src as an upstream regulator of sCLU induction after IR, we transiently overexpressed increasing amounts of constitutively active Src (Y529F) (Src CA) or kinase dead Src (K297R) (Src KD) cDNA constructs in MCF-7 1403 cells. MCF-7 1403 cells stably express 1403 bp of the CLU promoter linked to a luciferase reporter gene. Increasing levels of Src CA stimulated basal and IR-inducible CLU promoter activities (Fig.  4F, lanes 1-6). In contrast, increasing amounts of Src KD repressed IR-induced CLU promoter activity. These data, using PP1 and overexpression of Src CA or Src KD, illustrated a role for c-Src in sCLU induction in MCF-7 cells after IR.
Activation of the MAPK Cascade Is Required for sCLU Induction after IR-c-Src phosphorylates Raf, which in turn activates Mek-1 through phosphorylation (35). A selective Mek-1 kinase inhibitor, U0126 (1 M), completely abrogated IR-inducible CLU promoter activity (Fig. 5A) and sCLU protein levels (Fig. 5B) in MCF-7 cells. Blots are representative of experiments performed three or more times. B and C, relative phospho-protein levels were determined by comparing treated over control levels using the appropriate total protein level as a loading standard. Relative sCLU protein levels were determined by comparing treated to mock-transfected cells, normalizing levels to ␣-tubulin, as in "Experimental Procedures." In B, all data points are presented as the means Ϯ S.E. Earlier times are presented in B, showing early activation of Src, Erk1/2, and Mek-1 but without sCLU induction. Each data point (means Ϯ S.E.) was determined from three or more representative blots.
Log-phase MCF-7 cells were co-transfected with 4250-luc and Src CA, Src KD, dnMek-1 K97A, or dnErk1 (Fig. 5D). In response to 5 Gy, CLU promoter activity was induced ϳ4-fold in MCF-7 cells. Overexpression of Src CA resulted in significant elevation (4-fold) of CLU promoter basal levels compared with control. After IR, a further increase in CLU promoter activity was noted in Src CA-transfected MCF-7 cells. Overexpression of Src KD, dnMek-1, or dnErk1 abrogated IR induction of the CLU promoter. Western analyses confirmed overexpression of the corresponding proteins from the constructs (Fig.  5E). Collectively, these data demonstrated a role for MAPK signaling in sCLU induction in MCF-7 cells after IR.
Egr-1 Is Required for CLU Promoter Activation after IR-Several transcription factors are activated by the IGF-1R/ MAPK-signaling cascade, but only a few are stimulated after IR (36). The CLU promoter contains three potential Egr-1 binding sites. Egr-1 is a known downstream target of MAPK (37) and can be activated by high doses of IR in lymphoid tumor cells (38). Overexpression of Egr-1 increased CLU promoter activity ϳ2-fold over basal levels, and an ϳ3-fold increase was noted in MCF-7 cells after 5 Gy (Fig. 6A). Western blot analyses confirmed expression of Egr-1 in these cells (Fig. 6B). These data suggested that Egr-1 was an important transcription factor that may mediate CLU induction after IR.
DNA pull-down assays were then performed to determine if Egr-1 directly associated with the 1403 CLU promoter and whether IR exposure enhanced Egr-1 DNA binding. Egr-1 DNA binding was minimal in control, mock-irradiated MCF-7 cells. An increase in Egr-1 binding was noted 4 h after IR, with more robust increases observed at 24 h and sustained binding of Egr-1 noted through 72 h post-IR. In contrast, IR-treated MCF-7 cells exposed to either PP1 (5 M) or U0126 (5 M) showed no increase in Egr-1 DNA binding to the CLU promoter DNA after IR. Nuclear extracts (10% input loaded) were separated by SDS-PAGE and probed for PCNA as a loading control (Fig. 6C).
We then examined whether IGF-1R or EGFR inhibition could block IR-inducible Egr-1 binding to the CLU promoter, placing the MAPK pathway in between. As expected, exposure of IRtreated MCF-7 cells with AG1478 (1 M) did not affect Egr-1 DNA binding activity to the CLU promoter (Fig. 6D). In contrast and consistent with our observations of IGF-1R mediating sCLU induction after IR, Egr-1 DNA binding activity was blocked by co-addition of AG1024 (1 M). Egr-1-specific binding to the CLU promoter was demonstrated by competition assays (Fig. 6E).
A role for Egr-1 in sCLU induction was confirmed using Egr-1-specific siRNA (siRNA-Egr-1). Exposure of mock-transfected or MCF-7 1403 cells transfected with scrambled siRNA (scr-siRNA) to IR resulted in an ϳ6-fold induction of the 1403 CLU promoter (Fig. 6F). In contrast, CLU promoter activity was not significantly stimulated in Egr-1 siRNA (20 nM)-transfected MCF-7 1403 cells after 5 Gy. Efficacy of knockdown was confirmed by dramatic decreases in Egr-1 protein levels only

FIG. 2. EGFR is not an upstream activator of sCLU.
A, MCF-7 cells were transiently transfected with 4250-luc and RSV ␤-gal and serum-starved for 24 h. Cells were then treated with 0, 1, and 3 M AG1478 alone (white bars) or with 5 Gy (gray bars) or EGF (black bars) as described under "Experimental Procedures." Cells were harvested at 72 h for relative luciferase activities using RSV ␤-gal for loading. An asterisk denotes a significant statistical difference between irradiated cells and cells treated with AG1478 alone. Experiments were performed independently three times, with means Ϯ S.D. graphed, and p values were determined by paired Student's t tests. RLU, relative luciferase units. B, serumstarved MCF-7 cells were treated with IR, AG1478, EGF, or in combinations as indicated. Whole cell extracts were harvested 72 h, Western blot analyses were performed, and relative levels were determined. Blots are representative of experiments performed three or more times. C, MCF-7 cells were treated as described in Fig. 3B. EGFR was immunoprecipitated (IP) from cellular extracts 48 h after treatment and prepared for Western analyses (WB). Blots are representative of experiments independently performed three or more times. UT, untreated.
siRNA to sCLU in MCF-7 Cells Enhanced IR Lethality-Cancer cells treated with a variety of chemotherapeutic agents induce sCLU, and its expression provides cytoprotection (19, 22).

FIG. 3. Inhibition of IGF-1R abrogates sCLU induction after IR.
A, MCF-7 cells were transiently transfected with 4250-luc and RSV ␤-gal as in Fig. 2A. Cells were incubated 24 h later in serum-free medium for 24 h and then pretreated for 1 h with Me 2 SO or increasing doses of AG1024 as under "Experimental Procedures." Whole cell extracts were harvested 72 h later, and luciferase activities were monitored. A single asterisk denotes a significant statistical difference between irradiated and mock-irradiated samples. Experiments were independently performed three times, with means Ϯ S.D. graphed, and p values were determined by paired Student's t tests. UT, untreated. RLU, relative luciferase units. B, MCF-7 cells were transiently transfected with 4250-luc and serum-starved as in A then treated with increasing doses of IGF-1 (0, 10, 50 ng/ml) alone or with AG1024 as indicated. Cells were harvested at 72 h, and CLU promoter activities were assessed. Single asterisks denote a significant statistical difference between control and IGF-1-treated samples. Double asterisks denote a significant statistical difference between samples treated with AG1024 ϩ IGF-1 (black bars) and samples treated with IGF-1 alone (white bars). Experiments were independently performed three times, with means ϮS.D. graphed, and p values were determined by paired Student's t tests. C, MCF-7 cells were transiently transfected with 4250-luc, serum-starved as in A, then mock-irradiated and exposed to IR or IGF-1 with or without indicated IGFBP3 doses. Cells were harvested 72 h later for CLU promoter activities. The single asterisks denote significant statistical difference between IR or IGF-1 treatments and mock-irradiated control. The double asterisk denotes significant statistical difference between IR-exposed and IR ϩ IGFBP3. The triple asterisk denotes significant statistical difference between IGF-1 alone and IGF-1 ϩ IGFBP3. Experiments were independently performed three times, with means Ϯ S.D. graphed, and p values were determined by paired Student's t tests. D, serum-starved MCF-7 cells were treated as in Fig. 2B, except that AG1024 and IGF-1 were used at the indicated doses. Whole cell extracts were prepared 48 h later, and Western analyses were performed for sCLU levels. Relative levels were determined as treated compared with mock-treated samples as described under "Experimental Procedures." Blots are representative of three or more independent experiments. E, MCF-7 cells were exposed to 5 Gy or mock-irradiated. Cells were harvested 48 h later and separated by 10% SDS-PAGE, and blots were probed for either P-IGF-1R or total IGF-1R. F, serum-starved MCF-7 cells were treated as in D. Cells were harvested for Western analyses 48 h after IR or IGF-1 treatment. Relative levels were determined as treated compared with mock-treated samples as described. G, serum-starved MCF-7 cells were mock-irradiated or exposed to IR, and IGF-1 levels were assessed against an IGF-1 standard curve at various times as indicated. Experiments were performed in triplicate. The asterisk denotes statistical differences in IGF-1 levels between IR-and mock-treated controls (t ϭ 0).
To determine whether sCLU provides a similar cytoprotective role after IR, 20 M siRNA oligomers specific to exon II of the sCLU mRNA were transiently transfected into MCF-7 cells. sCLU-siRNA transfected MCF-7 cells demonstrated a significant increase in clonogenic lethality with increasing IR doses compared with mock-transfected cells (Fig. 7A). In contrast, transfection with scrambled siRNA oligomers (scr-siRNA) did not alter the survival of irradiated MCF-7 cells (Fig. 7A) and only slightly decreased sCLU protein levels (Fig. 7B). sCLU protein levels were decreased 70% in MCF-7 cells using siRNA specific to sCLU (Fig. 7B), whereas only a minor change in nuclear CLU (nCLU) protein levels was observed; nuclear CLU lacks exon II (5). Thus, selective decreases in endogenous sCLU levels significantly increased the lethality of IR-treated MCF-7 cells. DISCUSSION We demonstrated that delayed induction of sCLU, a prosurvival protein, was dependent on the novel re-activation of IGF-1R/Src/Mek/Erk signaling 24 -72 h after IR exposure. This signaling pathway culminated in the transactivation of Egr-1,

FIG. 4. c-Src is an upstream activator of sCLU after IR.
A, MCF-7 1403 cells were pretreated for 1 h with PP1 or vehicle alone (Me 2 SO (DMSO)) and exposed to various IR doses. PP1 was removed 24 h after IR, cells were washed with phosphate-buffered saline, and fresh medium was added. Cells were was harvested 48 h later for CLU promoter activities. Experiments were independently performed three times, with means Ϯ S.D. graphed, and p values were determined by paired Student's t tests. RLU, relative luciferase units. B, MCF-7 cells were treated with PP1 with or without IR as in A. Whole cell lysates were harvested and analyzed by Western analysis. Ku70 was used for loading, and -fold induction was determined as Me 2 SO-irradiated or PP1-irradiated samples compared with Me 2 SO-unirradiated control. C, MCF-7 cells were transiently transfected with 4250-luc and pretreated with various doses of PP1 for 1 h before IR. Medium containing fresh PP1 was added to cells every 24 h, and cells were harvested 72 h for CLU promoter-luciferase assays. Experiments were performed three times, and S.D. was determined by Student's t tests. The asterisk indicates a significant statistical difference between unirradiated PP1-treated versus Me 2 SO-treated cells (white bars) or irradiated PP1-treated versus Me 2 SO-treated cells (black bars). UT, untreated. D, MCF-7 cells were pretreated for 1 h with increasing doses of PP1 and then mock-irradiated or exposed to IR. Medium containing fresh PP1 was added to cells every 24 h. Cells were harvested 72 h for Western analyses. -Fold induction was determined as PP1-treated and irradiated samples compared with Me 2 SO-treated unirradiated control. E, MCF-7 cells were treated as in D, and -fold induction was determined as PP1-treated and irradiated samples compared with Me 2 SO-treated control cells. a known stress-inducible transcription factor (model, Fig. 6H).
Considerable attention has been given to EGFR inhibitors for cancer therapy alone and in combination with IR. EGFR is overexpressed or constitutively activated in many tumors including colorectal, breast, pancreatic, and ovarian cancers (39). EGFR activation may contribute to radio-resistance in various tumors, including glioblastoma multiforme and breast cancer cells by activation of Erk1/2 (40,41). As a result, EGFR-targeted therapies have been proposed, including the use of monoclonal antibodies and small molecule inhibitors that target the kinase domain (42). Interestingly, the selective EGFR inhibitor, AG1478, did not block the pro-survival signal transduction pathway leading to CLU promoter induction or regulate sCLU protein levels after IR (Fig. 2).
In contrast, IGF-1R has been far less studied as a target for refractory radiotherapy. IGF-1R activation resulted in mitogenic growth and cell survival (43), and treatment of cells with IGF-1 provided protection from doxorubicin-and taxol-induced apoptosis (44). AG1024, a selective inhibitor of IGF-1R, blocked sCLU induction after IR. Exposure to IR also activated IGF-1R (Fig. 3) and was required for delayed sCLU expression. These data are consistent with a previous report that AG1024 treatment of MCF-7 cells enhanced cell death after IR (45). The specific cytoprotective role of sCLU after IR (Fig. 7) strongly suggests that activation of IGF-1/IGF-1R signaling is an essential survival pathway that can be targeted for radiotherapy. MCF-7 cells produce and secrete IGF-1 (32), and IGF-1R is often overexpressed in breast cancer (46). Peripheral lymph node stromal cells produce and secrete EGF and IGF-1, which in turn can increase the growth and survival of breast cancer cells (47). Thus, EGF and IGF-1 secretion by lymph nodes may be a factor in the tumorigenesis of neighboring breast tissue, especially for cells that have up-regulated expression of EGFR or IGF-1R by a paracrine mechanism. Consistent with a previous report (43), we showed induction of IGF-1R after IR (Fig.  3D), an increase in IGF-1R phosphorylation (Fig. 3E), and a 20% increase in IGF-1 secretion 24 -72 h post-IR (Fig. 3E). Our data suggest a possible autocrine feedback loop induced by IR, where irradiated cells up-regulate IGF-1 and simultaneously increase IGF-1R synthesis (Fig. 3F). Induction of IGF-1 and its receptor provide a plausible mechanism for the delayed kinetics of endogenous sCLU protein levels in most cells after IR.
Our data suggest that a delayed induction of the Src-Mek-Erk-1/2 pathway, culminating in transactivation of Egr-1, is required for IR activation of the CLU promoter and up-regulation of sCLU protein levels. As previously noted, IR caused an  lanes 1-6), PP1 (lanes 7-12), or U0126 (lanes 13-18) before mock or IR treatment (5 Gy). Nuclear extracts were harvested at various times for Egr-1 DNA binding, and relative levels were determined as irradiated versus control at t ϭ 0 h. PCNA was used as a loading control, and experiments were independently performed at least three times. D, MCF-7 cells were pretreated with Me 2 SO (lanes 1 and 2), AG1024 (lanes 3 and 4), or AG1478 (lanes 5 and 6) for 1 h, then mock-or IR-treated (5 Gy). Nuclear extracts were harvested at 72 h for Egr-1 DNA binding, and relative levels were determined as irradiated versus untreated (UT) samples. PCNA was used as a loading control. E, Egr-1 binding to biotin-labeled 1403 CLU promoter using cells treated as described in Fig. 6C was decreased by increasing amounts (0.0, 0.1, or 5.0 g) of non-biotin labeled 1403 CLU promoter. F, MCF-7 1403 cells were mock-transfected or transfected with siRNA oligomers specific to Egr-1 or scrambled sequence, then mock-or IR-treated. Cells were harvested 72 h later for CLU promoter-luciferase assays. Experiments were independently performed three times, with means Ϯ S.D. graphed, and p values were determined by paired Student's t tests. G, MCF-7 cells were mocktransfected or transfected with scrambled-or Egr-1-specific siRNAs as described under "Experimental Procedures," and 24 h later mock-or IR-treated. Egr-1 steady state levels were determined by Western blotting. Blots are representative of experiments performed three times, and relative Egr-1 levels were determined. H, model depicting delayed IR activation of the IGF-1R and Src/Raf/Mek/Erk cascade, culminating in transactivation of Egr-1 and sCLU gene and protein induction. early activation of MAPK in MCF-7 cells, but that did not result in sCLU up-regulation. Instead, a novel re-activation of this signaling cascade days after exposure correlated with endogenous sCLU level induction. The physiological relevance of biphasic activation of MAPK after IR is unknown, and possible links between these pathways are being explored. Our data strongly suggest that delayed activation of MAPK is critical to sCLU induction. The importance of this delayed IR-stimulated pathway could be related to the fact that IGF-1R production in MCF-7 cells can cause increased resistance to herceptin (a monoclonal antibody to the Her2/neu receptor)-induced cell death (48). Our data strongly suggest a role for IGF-1R signaling in radio-resistance. Thus, tumor cells that survive an initial phase of radiotherapy may develop resistance to EGFR inhibitors as a result of delayed MAPK induction stimulating IGF-1 and IGF-1R synthesis and leading to expression of sCLU, a pro-survival factor (Fig. 7). Small molecule inhibitors of IGF-1R could allow combinatorial therapies to overcome herceptin resistance by manipulating the IGF-1R survival pathway, leading to a down-regulation of sCLU as well as possibly other down-stream factors.
Our data also suggest that therapies specifically focused on down-regulating sCLU levels should augment various cancer therapies. sCLU provides cytoprotection against chemotherapeutic agents in many cancer cell types (19,22), and we show that sCLU provides cytoprotection for IR-exposed MCF-7 cells (Fig. 7). sCLU expression is elevated in many tumors types, including prostate, esophageal, colorectal, renal, and breast cancers (11,(15)(16)(17). IGF-1R and IGF-1 production are also elevated in many of the same tumor types, especially prostate tumors. CLU and insulin-like growth factor binding proteins were suggested as targets for antisense therapy against prostate cancer (22), although a connection between IGF-1R signal-ing and sCLU induction was not examined. Our data strongly suggest that such a connection exists and could be exploited for improved therapy.
It is intriguing to speculate a possible role for sCLU in bystander effects after radiation or chemotherapeutic therapies. An increased production and secretion of sCLU by tumor cells into the lymph or vasculature system may provide a survival effect for neighboring or metastatic cancer cells. Additionally, secretion of IGF-1 by normal stromal or tumor cells may provide a means for sCLU up-regulation.
IGF-1R/Src/MAPK signaling leading to expression of sCLU, a pro-survival protein, may have implications beyond cancer biology and therapy. Increased sCLU expression was noted in Alzheimer patient tissues, after heart attacks, and during replicative senescence. Thus, the pro-survival IGF-1R/Src/MAPK/ Egr-1-signaling pathway may be involved in these processes to regulate sCLU expression needed for debris clearing and survival. Recent development of the rapidly aging, p44 knock-out mice and their reported increase in IGF-1/IGF-1R signaling (49) support this theory, since sCLU overexpression during replicative senescence (50) and after loss of p53 function have been reported (7). Thus, the IGF-1R/Src/MAPK/Egr-1 pathway appears to be an important pro-survival system in many biologically stressful and pathological conditions. At least one end product of this pathway is sCLU.