Simultaneous Inhibition of Epidermal Growth Factor Receptor (EGFR) Signaling and Enhanced Activation of Tumor Necrosis Factor-related Apoptosis-inducing Ligand (TRAIL) Receptor-mediated Apoptosis Induction by an scFv:sTRAIL Fusion Protein with Specificity for Human EGFR*

Epidermal growth factor receptor (EGFR) signaling inhibition by monoclonal antibodies and EGFR-specific tyrosine kinase inhibitors has shown clinical efficacy in cancer by restoring susceptibility of tumor cells to therapeutic apoptosis induction. Tumor necrosis factor-related apoptosis-inducing ligand (TRAIL) is a promising anti-cancer agent with tumor-selective apoptotic activity. Here we present a novel approach that combines EGFR-signaling inhibition with target cell-restricted apoptosis induction using a TRAIL fusion protein with engineered specificity for EGFR. This fusion protein, scFv425:sTRAIL, comprises the EGFR-blocking antibody fragment scFv425 genetically fused to soluble TRAIL (sTRAIL). Treatment with scFv425:sTRAIL resulted in the specific accretion to the cell surface of EGFR-positive cells only. EGFR-specific binding rapidly induced a dephosphorylation of EGFR and down-stream mitogenic signaling, which was accompanied by cFLIPL down-regulation and Bad dephosphorylation. EGFR-specific binding converted soluble scFv425:sTRAIL into a membrane-bound form of TRAIL that cross-linked agonistic TRAIL receptors in a paracrine manner, resulting in potent apoptosis induction in a series of EGFR-positive tumor cell lines. Co-treatment of EGFR-positive tumor cells with the EGFR-tyrosine kinase inhibitor Iressa resulted in a potent synergistic pro-apoptotic effect, caused by the specific down-regulation of c-FLIP. Furthermore, in mixed culture experiments binding Lof scFv425:sTRAIL to EGFR-positive target cells conveyed a potent apoptotic effect toward EGFR-negative bystander tumor cells. The favorable characteristics of scFv425:sTRAIL, alone and in combination with Iressa, as well as its potent anti-tumor bystander activity indicate its potential value for treatment of EGFR-expressing cancers.

The epidermal growth factor receptor (EGFR) 1 is a transmembrane receptor tyrosine kinase comprising an extracellular ligand binding domain, a transmembrane domain, and an intracellular tyrosine kinase domain (1,2). Binding of epidermal growth factor or transforming growth factor-␣ results in EGFR dimerization and subsequent activation of the intrinsic tyrosine kinase activity. Phosphorylated EGFR concomitantly triggers downstream mitogenic signaling via both the p44/42 MAPK and PI3K pathways (3,4). Normal EGFR signaling plays a pivotal role in organ development and repair and in the regulation of cell survival. Aberrant EGFR signaling strongly contributes to the malignant features in cancer including an increased resistance to apoptosis. Aberrant signaling can be the result of EGFR overexpression by EGFR gene amplification, which can lead to very high cell surface expression of up to 10 6 EGFR molecules per tumor cell. Alternatively, various oncogenic mutations of EGFR have been described including EGFRvIII, an EGFR mutant that possesses ligand-independent tyrosine kinase activity (5) and that appears to be selectively expressed in tumor cells since it is not found in normal cells. Recently, also mutations in the EGFR-tyrosine kinase domain have been identified in a subset of lung cancer patients that appear to activate anti-apoptotic pathways (6).
Several targeted strategies have been developed to specifically inhibit aberrant EGFR signaling. Monoclonal antibodies, e.g. mAb C225 (Cetuximab) and mAb 425 (7,8), competitively inhibit the binding of natural ligands to the extracellular ligand binding domain. Small molecule tyrosine kinase inhibitors, e.g. Iressa (also known as ZD1839 or Gefitinib) (9, 10), competitively inhibit with ATP for binding to the EGFR-tyrosine kinase domain. The clinical efficacy of these agents appears to rely on multiple anti-cancer mechanisms, including inhibition of cell cycle progression, inhibition of metastasis, and an increase in the susceptibility of cells to apoptosis.
However, despite promising anti-tumor activity in clinical trials (11)(12)(13)(14), both classes of EGFR-signaling antagonists do not appear to be curative. Therefore, additional EGFR-targeted strategies or combination with other therapeutic approaches * This work was supported by Dutch Cancer Foundation Grant RUG 2002-2668) and Dutch Brain Foundation Grant 10F02.30 (to W. H.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  are under investigation. In this respect strong synergistic tumoricidal effects have been reported for strategies in which EGFR-signaling antagonists are combined with radiation or chemotherapy (12,14,15) and, more recently, with the cytokine TRAIL (16).
TRAIL is normally present as a trimeric type II transmembrane protein (memTRAIL) on various immune effector cells. TRAIL specifically induces apoptosis in cancer cells (17) and virus-infected cells (18) without apparent apoptotic activity toward normal human cells. Homotrimeric memTRAIL initiates apoptosis after cross-linking of the agonistic receptors TRAIL-R1 and TRAIL-R2 (19 -22), leading to activation of the extrinsic apoptotic pathway via the death-inducing signaling complex (23)(24)(25)(26)(27)(28)(29)(30). Assembly of the death-inducing signaling complex sequentially activates initiator caspases (caspase 8 or 10) and effector caspases (e.g. caspase 3 and 7) and ultimately ends in apoptotic cell death. memTRAIL can be proteolytically cleaved into a soluble form (sTRAIL). Several recombinant forms of sTRAIL have been generated that show strong tumoricidal activity in vitro and in xenografted mouse tumor models without toxic side effects (31)(32)(33). Pharmacokinetic studies in cynomolgus monkeys and chimpanzees revealed no TRAILrelated toxicity (34), also indicating a potential role for sTRAIL in human cancer therapy. Nevertheless, several recent reports described apoptotic activity of sTRAIL toward various normal human cells, including primary human hepatocytes (35), keratinocytes (36), prostate epithelial cells (37), and brain tissue (38).
Previously we and others showed that sTRAIL can be genetically fused to a tumor-selective antibody fragment (39,40), resulting in fusion proteins with enhanced and tumor-restricted apoptotic activity. Here we present and analyze the mode of action of a novel and promising strategy that combines EGFR-signaling inhibition with target cell-restricted apoptosis induction using a TRAIL fusion protein with engineered specificity for EGFR.

Monoclonal Antibodies and Inhibitors
TRAIL-neutralizing mAb 2E5 was purchased from Alexis (10P, Breda, The Netherlands). mAb 425 (kindly provided by Merck) is a murine IgG2a with high binding affinity for the extracellular domain of both EGFR and EGFRvIII. mAb 425 blocks EGF binding to EGFR and competes with scFv425 for binding to the same epitope. Total caspase inhibitor Z-VAD-FMK, caspase-8 inhibitor Z-IETD-FMK, and caspase-9 inhibitor Z-LEHD-FMK were purchased from Calbiochem. EGFR-tyrosine kinase inhibitor Iressa was kindly provided by AstraZeneca Inc (Macclesfield, Cheshire, UK). PI3K inhibitor wortmannin was purchased from Sigma-Aldrich. Final working concentrations of inhibitors were diluted in serum-free medium from a stock of 10 mM in Me 2 SO.

Production of scFv425:sTRAIL
Fusion protein scFv425:sTRAIL was constructed and produced essentially as described previously (39). Briefly, in the first multiple cloning site of vector pEE14, the high affinity antibody fragment scFv425 (Vh-(G4S)3-Vl format) (41), kindly provided by Merck, was directionally inserted using the unique SfiI and NotI restriction enzyme sites. In the second MCS a PCR-truncated 593-bp DNA fragment encoding the extracellular domain of human TRAIL (sTRAIL) was cloned in-frame using restriction enzymes XhoI and HindIII, yielding plasmid pEE14-scFv425:sTRAIL. Expression plasmid pEE14-scFv425:sTRAIL was transfected into Chinese hamster ovary K1 cells using FuGENE 6 reagent (Roche Diagnostics) according to the manufacturer's instructions, after which transfectants were selected by the glutamine synthetase system as described (42). Single cell sorting using the MoFlo high speed cell sorter (Cytomation, Fort Collins, CO) established clone 100F1, stably secreting 2.4 g/ml scFv425:sTRAIL into the culture medium.
EGFR-specific Binding of scFv425:sTRAIL EGFR-specific binding of scFv425:sTRAIL was assessed by flow cytometry using the EGFR-positive tumor cell line A431 and the EGFRnegative cell line Jurkat. In short, 1 ϫ 10 6 cells were incubated with scFv425:sTRAIL (300 ng/ml) in the presence or absence of mAb 425 (3 g/ml). Specific binding of scFv425:sTRAIL was detected using phosphatidylethanolamine-conjugated anti-TRAIL mAb B-S23 (Diaclone SAS, Besançon, France) and subsequent fluorescence-activated cell sorting analysis using an EPICS ELITE flow cytometer (Beckman Coulter, Mijdrecht, The Netherlands). Incubations were carried out for 45 min at 0°C and were followed by 2 washes with serum-free medium.

Target Cell-restricted Induction of Apoptosis by scFv425:sTRAIL
Target cell-restricted induction of apoptosis by scFv425:sTRAIL was assessed by analysis of tumor cell viability, loss of mitochondrial membrane potential (⌬), caspase 8 and 3 activation, and PARP cleavage/ DNA fragmentation factor degradation, as described in more detail under "Immunoblot Analysis." Where indicated, treatment with scFv425:sTRAIL was performed in the presence of mAb 425 (3 g/ml) or mAb 2E5 (1 g/ml).

Apoptosis Assessed by Viability Assay
Cells were precultured in a 96-well plate at a density of 3 ϫ 10 4 cells/well. Subsequently, cells were treated for 16 h with the various experimental conditions in a final volume of 200 l. Cell viability of adherent cell lines was determined by crystal violet staining (Sigma) as described previously (40). Cell viability of suspension cell lines was determined using the [3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium assay according to the manufacturer's recommendations (Promega Benelux b.v., Leiden, The Netherlands). Experimental apoptosis induction was quantified as the percentage of apoptosis induction compared with base-line apoptosis in medium control, which was set at 0% apoptosis. Each experimental condition consisted of six independent wells.

Apoptosis Assessed by Loss of Mitochondrial Membrane Potential (⌬)
⌬ was analyzed using the stain DiOC6 (Eugene, The Netherlands) as previously described (39). Briefly, cells were precultured in a 24-well plate at a concentration of 0.5 ϫ 10 6 cells/well. Subsequently, cells were treated for 16 h with the various experimental conditions, after which cells were harvested and incubated for 20 min with DiOC6 (0.1 M) at 37°C, harvested (1000 ϫ g, 5 min), resuspended in phosphate-buffered saline, and assessed for DiOC6 staining using flow cytometry.

Immunoblot Analysis
Cells were precultured at 1.5 ϫ 10 6 cells/well in a 6-well plate, after which cells were incubated with scFv425:sTRAIL in the presence or absence of mAb 425 or mAb 2E5 for the indicated time points. Cell lysates were prepared as described previously (39). Subsequently, 30 g of lysate was separated by SDS-PAGE under reducing conditions and transferred to nitrocellulose by electroblotting.
Apoptosis Signaling-Caspase activation was detected using antibodies directed against caspase-8 and active caspase-3 (Cell Signaling, Beverly, MA). PARP cleavage and DNA fragmentation factor degradation was assessed using anti-PARP mAb F2 and polyclonal anti-DNA fragmentation factor (Santa Cruz Biotechnology Inc., Santa Cruz, CA. Expression of c-FLIP L and Bad phosphorylation was determined using anti-c-FLIP L mAb clone NF6 (Alexis) and polyclonal anti-phospho Bad Ser136 antibody (Cell Signaling).
EGFR Signaling-Expression levels of total and active EGFR were assessed using anti-total EGFR (Cell Signaling) and anti-activated EGFR (Tyr-1173) (Santa Cruz). The MAPK signal transduction pathway was analyzed using polyclonal anti-phospho p44/42 MAPK, and the PI3K signal transduction pathway was analyzed using polyclonal total and anti-phospho-AKT Thr-308 and Ser-437 (Cell Signaling). Equal protein loading was assessed using anti-actin mAb (Roche Applied Science). Specific binding was visualized using appropriate secondary Horseradish Peroxidase-conjugated antibody (DAKO Cytomation, Glostrup, Denmark) and chemiluminescence (Roche Applied Science).

Differential Quantification of Apoptosis in Target and Bystander Cells during Mixed Culture Experiments
Differential cell membrane labeling of target and bystander cells was achieved using the Vybrant multicolor cell-labeling Kit (Molecular Probes). Briefly, Jurkat.EGFRvIII target cells were labeled with the red fluorescent dye DiI. Labeling was performed by incubation of Jurkat.EGFRvIII cells (1 ϫ 10 6 cells/ml in serum free medium) with 5 M DiI (37°C, 5 min) followed by 3 washes with standard medium. DiIlabeled Jurkat.EGFRvIII target and non-labeled Jurkat bystander cells were mixed at the indicated ratio at a final concentration of 0.5 ϫ 10 6 cells/well in a 24-well plate. After treatment differential fluorescent characteristics of target cells and bystander cells were used to separately evaluate induction of apoptosis in both populations by ⌬ as described above.

RESULTS
EGFR-specific Binding of scFv425:sTRAIL-To assess whether scFv425:sTRAIL displayed specific and enhanced binding to EGFR-positive cells, A431 cells were incubated with scFv425:sTRAIL and analyzed for binding by flow cytometry. Strong binding of scFv425:sTRAIL was detected on the cell surface (Fig. 1A, solid line) that could be specifically inhibited by preincubation with parental EGFR-blocking mAb 425 (Fig.  1A, dashed line). In contrast, binding of scFv425:sTRAIL to TRAIL receptors on the cell surface of EGFR-negative Jurkat cells was barely detectable (Fig. 1B). The intensity of scFv425: sTRAIL binding directly correlated to the amount of cell surface-expressed EGFR (data not shown).
Inhibition of EGFR Signaling and Subsequent Sensitization to Apoptosis by scFv425:sTRAIL Treatment-Because scFv425: sTRAIL primarily binds via its EGFR-blocking antibody fragment scFv425, the effect of scFv425:sTRAIL treatment on EGFR signaling was determined. In A431 cells, scFv425:sTRAIL induced a rapid dephosphorylation of EGFR at Tyr-1173 within 10 min, whereas total EGFR levels remained constant (Fig. 3A). The phosphorylation of EGFR observed during normal culture conditions was most likely because of a previously described transforming growth factor-␣-induced autophosphorylation loop (43). Specific inactivation of EGFR signaling was accompanied by a decrease in MAPK pathway activity, which was detected by a dephosphorylation of MAPK at 1 and 3 h of treatment (Fig. 2B).
In addition, the PI3K pathway was markedly inhibited after 1 and 3 h of treatment, as measured by dephosphorylation of Akt at residues Tyr-308 and Ser-473 (Fig. 2B), whereas total Akt levels remained constant.
Resistance to apoptosis by EGFR signaling is partly mediated by its effect on the anti-apoptotic protein cFLIP L and the phosphorylation of Bad via PI3K signaling. Because PI3K signaling was inactivated by scFv425:sTRAIL treatment, cFLIP L expression and Bad phosphorylation were investigated in more detail. At the early time points of 1 and 3 h of treatment a FIG. 3. Inhibition of EGFR signaling and sensitization to apoptosis by scFv425:sTRAIL. A431 cells were challenged with 300 ng/ml scFv425:sTRAIL in the presence of 1 g/ml cycloheximide for 10 min, 1 h, and 3 h. At the elapsed time point 3 h, cells were additionally treated with scFv425:sTRAIL in the presence or absence of mAb 425 or mAb 2E5. A, cell lysates were analyzed for the amount of total and phosphorylated EGFR (pTyr1173) using immunoblot. B, cell lysates were analyzed for the MAPK and PI3K pathway activity by measurement of phosphorylated p44/42 MAPK, total and phosphorylated Akt (pAkt). Actin levels were determined to confirm equal protein loading. C, cell lysates were analyzed for the apoptosis-associated features of caspase-8 activation, the expression level of cFLIP L , and the phosphorylation status of Bad at residue Ser-136. decrease was detected in the expression of the anti-apoptotic caspase 8 homologue cFLIP L (Fig. 3C) that coincided with the activation of caspase 8 (Fig. 3C). Additionally, a marked decrease was observed in phosphorylation of Bad (Fig. 3C), sensitizing the mitochondria to apoptosis. Treatment in the presence of an excess of mAb 425 significantly inhibited EGFR signaling (Fig. 3, A and B), probably due to the well established EGFR inhibitory effect of this antibody (7,44). As expected, treatment in the presence of mAb 2E5, an antibody that specifically prevents binding of TRAIL to TRAIL-R, did not significantly affect EGFR phosphorylation levels (Fig. 3, A and B). Co-incubation with mAb 425 or mAb 2E5 did not restore cFLIP L or Bad phosphorylation to levels of untreated control (Fig. 3C).
Synergistic Induction of Apoptosis by scFv425:sTRAIL and Iressa-Previously, EGFR-signaling inhibition was shown to synergistically enhance TRAIL sensitivity (16). Therefore, the potential synergistic effects of scFv425:sTRAIL with the EGFR tyrosine kinase inhibitor Iressa were assessed on A431 cells and EGFRvIII-transfected Jurkat cells. Treatment of A431 cells with increasing concentrations of Iressa (250 -2000 nM) and a fixed concentration of scFv425:sTRAIL (100 ng/ml) resulted in a dose-dependent synergistic increase in apoptosis (Fig. 4A). Similar results, but with lower concentrations of Iressa (50 -250 nM) and scFv425:sTRAIL (80 ng/ml), were obtained for Jurkat.EGFRvIII cells (Fig. 4B). Dose-response curves of treatment with a fixed concentration of Iressa (250 and 2000 nM, respectively) and increasing concentrations of scFv425:sTRAIL (up to 100 ng/ml) revealed a potent dose-dependent increase in apoptosis in both A431 and Jurkat.EGFR-vIII cells already at 20 ng/ml scFv425:sTRAIL (Fig. 4C). The synergistic pro-apoptotic activity of scFv425:sTRAIL and Iressa was potently inhibited by co-treatment with mAb 425 (Fig. 4C). Parental Jurkat cells subjected to the same experimental conditions were fully resistant to treatment (Fig. 4D).
In control experiments with the solvent Me 2 SO alone or in combination with scFv425:sTRAIL, no significant induction of apoptosis was detected (data not shown).
Synergistic Induction of Apoptosis by scFv425:sTRAIL and Iressa Is Caspase 8-mediated-Next, the mechanism underlying the synergistic pro-apoptotic effect was investigated. Treatment of A431 cells and Jurkat.EGFRvIII cells with scFv425: sTRAIL and Iressa did not significantly alter TRAIL receptor expression (data not shown). Using specific caspase inhibitors, induction of apoptosis was found to be largely caspase 8-dependent since the specific caspase 8 inhibitor Z-IETD-FMK inhibited apoptosis to levels observed for Iressa alone (Fig. 5A). Caspase 9 inhibition by Z-LEHD-FMK only had a minimal effect. Immunoblot analysis further revealed a strong activation of both caspase 8 and 3, resulting in PARP cleavage within 3 h of treatment with scFv425:sTRAIL and Iressa (Fig. 5B). Single agent treatment only marginally activated caspase 8 and caspase 3 (Fig. 5B). Similar results were obtained when A431 cells were treated with scFv425:sTRAIL and Iressa (Fig.  5C). The appearance of apoptotic features was specifically inhibited when treatment was performed in the presence of mAb 425 (Fig. 5, B and C).
Inhibition of EGFR Signaling by Co-treatment with scFv425: sTRAIL and Iressa-Combination treatment of Jurkat.EGFR-vIII cells with scFv425:sTRAIL and Iressa resulted in PI3K inactivation within 2 h in Jurkat.EGFRvIII, as measured by Akt dephosphorylation at Ser-473 (Fig. 6A). No inhibition of MAPK signaling was observed in Jurkat.EGFRvIII cells, which is in line with a previous report showing that EGFRvIII spe- cifically regulates PI3K activity (45). At the concentrations used single agent treatment had no effect on mitogenic signaling in Jurkat.EGFRvIII cells (Fig. 6A). The role of PI3K inhibition in the observed synergistic apoptotic effect on Jurkat.EGFRvIII cells was further analyzed by treatment with scFv425:sTRAIL in the presence of PI3K inhibitor wortmannin. This experiment resulted in levels of apoptosis comparable with those observed for treatment with scFv425:sTRAIL and Iressa (Fig. 6C). For A431 cells, no effect of single agent and co-treatment was detected on PI3K signaling (Fig. 6A). Single agent treatment with Iressa markedly inhibited MAPK signaling, whereas scFv425:sTRAIL treatment alone only had a minimal effect. Combination treatment of A431 cells also inhibited MAPK signaling, but only to an extent comparable with Iressa treatment alone (Fig. 6B).
Treatment with scFv425:sTRAIL and Iressa Induces c-FLIP L Down-regulation-Simultaneous treatment with scFv425: sTRAIL and Iressa markedly reduced the expression of c-FLIP L in both Jurkat.EGFRvIII and A431 cells (Fig. 6D). To a lesser extent treatment with Iressa alone down-regulated c-FLIP L in A431 cells, whereas in Jurkat.EGFRvIII cells no effect of single agent treatment was detected. Treatment in the presence of mAb 425 prevented down-regulation of c-FLIP L in both cell lines. DISCUSSION Here we describe a novel therapeutic approach in which EGFR-signaling inhibition is combined with target cell-restricted apoptosis induction using the new fusion protein scFv425:sTRAIL. Fusion protein scFv425:sTRAIL, comprising the EGFR-blocking antibody fragment scFv425 genetically fused to sTRAIL, clearly demonstrated accretion at the cell surface of EGFR-positive cells, which was specifically abrogated by preincubation with parental EGFR-blocking mAb 425.
Previously, we demonstrated that eukaryotically expressed scFv:sTRAIL fusion proteins are produced as a soluble homogeneous trimer (39). Trimeric scFv425:sTRAIL contains three identical antibody fragment domains and will, therefore, benefit from an associated enhanced avidity effect. Enhanced avidity has been shown to improve the in vivo tumor-targeting efficacy in several antibody-based strategies (46,47). Indeed, relatively high concentrations of the parental mAb 425 were required to competitively inhibit specific binding of scFv425:sTRAIL.
Treatment with scFv425:sTRAIL potently induced apoptosis in EGFR-positive tumor cells that was specifically abrogated by co-incubation with parental mAb 425. Interestingly, the appearance of apoptotic features, such as processing of caspase 8, was preceded by the specific dephosphorylation of EGFR and coincided with dephosphorylation of the PI3K signal transduction pathway and to a lesser extent the MAPK signal transduction pathway.
This rapid inactivation of EGFR signaling clearly points to a role for EGFR inhibition in scFv425:sTRAIL-induced apoptosis. One of the main regulators of TRAIL sensitivity, the anti-apoptotic caspase-8 homologue cFLIP L (48 -50), has previously been shown to be regulated by PI3K signaling (51,52). In A431 cells, inactivation of PI3K signaling was accompanied by a decrease in expression of cFLIP L after 1 and 3 h of treatment. Besides regulating cFLIP L expression, PI3K signaling also influences the phosphorylation status of Bad (53,54). In A431 cells, a marked dephosphorylation of Bad was detected after 1 and 3 h. Therefore, inhibition of PI3K signaling appears to facilitate caspase 8 activation by down-regulating cFLIP L , and sensitizes the mitochondria to induction of apoptosis by dephosphorylation of Bad.
Next to PI3K inhibition, dephosphorylation of the MAPK signal transduction pathway was also detected after 1 and 3 h of treatment with scFv425:sTRAIL. Previously, MAPK activation was shown to protect against TRAIL-induced apoptosis by a mechanism occurring at or above the level of caspase 8 processing that did not involve cFLIP L (55). Conversely, although not formally proven here, MAPK inhibition can sensitize tumor cells toward scFv425:sTRAIL-induced apoptosis at or above the level of caspase 8 processing.
From these data a model for the apoptotic activity of FIG. 6. Inhibition of EGFR signaling by co-treatment with scFv425: sTRAIL and Iressa. Jurkat.EGFRvIII and A431 cells were treated either alone or with a combination of scFv425:sTRAIL and Iressa in the presence or absence of mAb 425. After 2 h of treatment, PI3K pathway and MAPK pathway activity in Jurkat.EGFRvIII (A) and A431 (B) was assessed by immunoblot analysis of total and active phosphorylated Akt (pAkt) and phosphorylated MAPK (pMAPK), respectively. Equal protein loading was confirmed by actin staining. C, Jurkat.EG-FRvIII was treated with scFv425:sTRAIL and either Iressa or the specific PI3K inhibitor wortmannin (WM), after which apoptosis induction was assessed by ⌬. D, cell lysates of A431 and Jurkat.EGFR-vIII treated alone or with a combination of scFv425:sTRAIL and Iressa were analyzed for expression of the anti-apoptotic caspase 8 homologue cFLIP L . scFv425:sTRAIL can be formulated (for schematic representation see Fig. 7). First, binding of scFv425:sTRAIL leads to accretion at the cell surface of EGFR-positive tumor cells only. Subsequently, EGFR-specific binding inhibits EGFR mitogenic signaling via PI3K and MAPK and, thereby, sensitizes tumor cells to apoptosis by down-regulation of c-FLIP L and Bad dephosphorylation. Concomitantly, membrane-bound scFv425: sTRAIL induces apoptosis by reciprocal cross-linking of agonistic TRAIL-receptors on neighboring EGFR-positive tumor cells.
Paracrine activation of TRAIL receptors by scFv425:sTRAIL is not necessarily restricted to EGFR-positive tumor cells but can also be directed toward neighboring tumor cells that have lost EGFR expression. In a recent report, we described a potent anti-tumor bystander effect for an scFv:sTRAIL fusion protein with specificity for the carcinoma-associated cell surface target antigen EGP2 (56). Here, we show that scFv425:sTRAIL potently induced apoptosis in EGFR-negative bystander Jurkat cells in mixed culture experiments with Jurkat.EGFRvIII target cells. This potent anti-tumor bystander effect might help to eliminate EGFR-negative tumor cell mutants that can escape from conventional antibody-mediated strategies.
In a recent report the synergistic effect of the combined treatment with anti-EGFR monoclonal antibody cetuximab and the EGFR-specific tyrosine kinase inhibitor Iressa was described (57). We analyzed combination treatment of scFv425: sTRAIL with Iressa. Potent synergistic induction of apoptosis was observed in both wild type EGFR-positive (A431) and mutant EGFRvIII-positive cells (Jurkat.EGFRvIII). This synergistic pro-apoptotic effect was fully EGFR-restricted and TRAIL-mediated and did not involve modulation of TRAIL receptor expression. Interestingly, inhibition of caspase 8 activity by a specific caspase 8 inhibitor reduced apoptosis induction by scFv425:sTRAIL and Iressa to levels for treatment with Iressa alone. Caspase 9 inhibition only had a minimal effect on apoptosis induction. These data point to an increased processing of caspase 8 as the main cause for the synergistic proapoptotic effect with no or only minimal involvement of the mitochondrial route of apoptosis. When cells were subsequently analyzed for expression of cFLIP L , an important regulator of caspase 8 processing, a marked down-regulation in both Jurkat.EGFRvIII and A431 cells, was observed within 3 h of combination treatment with scFv425:sTRAIL and Iressa.
Down-regulation of cFLIP L coincides with caspase 8 activation and was preceded by inactivation of the PI3K pathway in Jurkat.EGFRvIII cells. In A431 cells combination treatment significantly inhibited MAPK signaling but only to the level observed for treatment with Iressa alone. Based on these results it can be concluded that the synergistic pro-apoptotic effect largely depends on the specific down-regulation of c-FLIP L . For EGFRvIII-positive Jurkat cells, down-regulation of cFLIP L is a consequence of PI3K inhibition. In A431 cells, MAPK dephosphorylation may play a role, but the exact mechanism remains to be further elucidated.
In conclusion, we report for the first time on a recombinant fusion protein that combines the tumoricidal effect of EGFR signal inhibition with target cell-restricted apoptosis induction. The unique characteristics of scFv425:sTRAIL described here indicate its potential therapeutic value alone and in combination with EGFR-tyrosine kinase inhibitor Iressa for the treatment of EGFR-and EGFRvIII-expressing human cancers. FIG. 7. Schematic model of the apoptotic activity of scFv425: sTRAIL. Antibody fragment binding of scFv425:sTRAIL to EGFR inhibits mitogenic signaling by this receptor and its downstream signaling pathways and, thereby, sensitizes tumor cells to apoptosis. Furthermore, antibody fragment binding to EGFR immobilizes soluble scFv425:sTRAIL on the cell surface of EGFR-positive tumor cells and converts soluble scFv425:sTRAIL into a membrane bound form that can efficiently initiate apoptosis by cross-linking of the agonistic TRAIL receptors TRAIL-R1 and TRAIL-R2