Peptabody-EGF: A novel apoptosis inducer targeting ErbB1 receptor overexpressing cancer cells

Authors


Abstract

The epidermal growth factor receptor (EGFR) plays a central role in cell life by controlling processes such as growth or proliferation. This receptor is commonly overexpressed in a number of epithelial malignancies and its upregulation is often associated with an aggressive phenotype of the tumor. Thus, targeting of EGFR represents a very promising challenge in oncology, and antibodies raised against this receptor have been investigated as potential antitumor agents. Various putative mechanisms of action were proposed for such antibodies, including decreased proliferation, induction of apoptosis, stimulation of the immunological response against targeted cancer cells or combinations thereof. We report here the development of an alternative high affinity molecule that is directed against EGFR. Production of this pentameric protein, named peptabody-EGF, includes expression in a bacterial expression system and subsequent refolding and multimerization of peptabody monomers. The protein complex contains 5 human EGF ligand domains, which confer specific binding towards the extracellular portion of EGFR. Receptor binding of the peptabody-EGF had a strong antiproliferative effect on different cancer cell lines overexpressing EGFR. However, cells expressing constitutive levels of the target receptor were barely affected. Peptabody-EGF treated cancer cells exhibited typical characteristics of apoptosis, which was found to be induced within 30 min after the addition of the peptabody-EGF. In vitro experiments demonstrated a significantly higher binding activity for peptabody-EGF than for the therapeutic monoclonal EGFR antibody Mab-425. Furthermore, the antitumor action provoked by the peptabody-EGF was greatly superior than antibody mediated effects when tested on EGFR overexpressing cancer cell lines. These findings suggest a potential application of this high affinity molecule as a novel tool for anti-EGFR therapy. © 2006 Wiley-Liss, Inc.

The epidermal growth factor receptor (EGFR) family, also named ErbB, plays a central role in cell life by controlling growth, differentiation and proliferation.1 This family of tyrosine kinase receptor includes EGFR (ErbB-1/HER-1), ErbB-2 (HER-2/Neu), ErbB-3 (HER-3) and ErbB-4 (HER-4). One of the ErbB receptor studied in great detail is EGFR, a 170 kDa phosphoglycoprotein anchored in the plasma membrane and expressed in all adult tissues except for hematopoietic cells. Activation of the receptor by autophosphorylation occurs upon ligand binding and subsequent receptor dimerization. Among the EGF ligand family, epidermal growth factor (EGF), transforming growth factor α (TGFα), and amphiregulin interact exclusively with EGFR. Betacellulin, heparin binding-epidermal growth factor and epiregulin, called bispecific ligand, are able to bind EGFR or ErbB-4.2

Although EGFR is important in the maintenance of normal cellular function and survival, EGFR expression clearly appears to promote growth and survival of tumor cells. Different studies have demonstrated the role of EGFR in many tumors such as head and neck, breast, colon, lung, prostate, kidney, ovary, brain, pancreas and bladder.3 Dysregulation of ErbB signaling can occur by diverse mechanisms, including gene amplification and ErbB mutations, that increase receptor transcription, translation or protein stability.4

The EGF receptor family is considered an attractive target for cancer treatments because of their overexpression in many cancer cells, their involvement in cell proliferation and survival and their expression on the surface of cells, offering easy access for targeting. Various approaches have been developed to target ErbB family receptors, including specific monoclonal antibodies against the extracellular region, inhibitors of the tyrosine kinase domain (required for receptor activation) and antisense therapy. Among the ErbB family receptors, ErbB-2 has been exploited as target for cancer therapy and Herceptin, a monoclonal antibody raised against ErbB-2, is now being used in clinics with favorable results against metastatic breast cancer with ErbB-2 overexpression.5 Recently, the therapeutic effect of monoclonal antibodies raised against erbB-1/EGFR was assessed for head and neck cancer and several other cancers, and clinical trials are ongoing.6 Such antibodies may inhibit the growth of tumors immunologically through antibody-dependent cellular cytoxicity or complement-dependent cytoxicity.7 Alternatively, antibodies may compete with the binding of a growth factor to its receptor, thereby inhibiting the growth of tumors that express the receptor. Alternative approaches use antibodies raised against tumor antigens, with toxins conjugated to the antibodies to increase the cytotoxic potential of the treatment.8

Few years ago, a promising new type of molecule named Peptabody9 has been developed. This molecule consists of a multimerization domain, which is formed by a part of rat cartilage oligomeric matrix polypeptide (COMP), fused to a camel hinge region or spacer and a target-binding domain consisting of a short peptide ligand placed to the C-terminal part of the hinge. More recently, the same group published a human version of peptabody reacting specifically to ErBb-2 receptor through peptides selected by phage display technology.10

We present the development of an improved version of a peptabody molecule for targeting the EGFR pathway. This multimeric molecule allows the crosslinking of EGF receptors by the interaction of 5 ligands fused to the peptabody with their receptors, leading to apoptosis of cancer cells overexpressing EGFR.

Material and methods

Materials

The following materials were obtained from commercial sources: T4 DNA ligase (Invitrogen, Basel, Switzerland), T4 polynucleotide kinase (Qbiogene, Basel, Switzerland), Ni2+-nitrilotriacetic acid agarose beads (Qiagen, Hombrechtikon, Switzerland), restriction enzymes (Roche, Amersham Pharmacia, Promega), anti-His antibody (Sigma Fluka, Buchs, Switzerland), human EGF (Fluka, Buchs, Switzerland). Oligonucleotide synthesis was carried out at Invitrogen and DNA sequencing at Synergene Biotech GmbH. The human melanoma A-431 and breast cancer MCF-7 cell lines were purchased from ATCC (USA) and human smooth muscle cells (HSMC) were generously provided by Dr Peter Frey (CHUV, Lausanne, Switzerland). The murine monoclonal antibody Mab-425 against the extracellular domain of the human epidermal growth factor receptor was a kind gift of Dr Siegfried Matzku (Merck, Darmstadt, Germany).

Construction of peptabody expression vectors

Peptabody-EGF (p-EGF) and peptabody-Irrelevant (p-IRR) are modifications of previously described peptabodies.10 Expression plasmids were constructed by cloning the different building blocks (Fig. 1a) into the vector pQE-9 (Qiagen).

Figure 1.

(a) Schematic structure of the gene encoding for peptabody-Irrelevant (p-IRR) or peptabody-EGF. (b) Amino acid sequence of p-IRR and p-EGF. Enhancer (Enh) in italic black, histidine tail (HIS6) in underlined grey, human cartillage oligomeric matrix protein (hCOMP) in grey, hinge human antibody fragment (Hinge) in black and human epidermal growth factor (EGF) in black underlined.

Plasmid pAT16210 encoding a His-tagged peptabody containing a 48 amino acid hCOMP (human cartilage oligomeric matrix protein) assembly domain and the 19 amino acid hinge human IgA fragment was used as template to amplify DNA fragments encoding the His6-hCOMP and the hinge region cassettes. Specific PCR primers adding NheI/SpeI cloning sites to the His6-hCOMP fragment and SpeI/BglII cloning sites to the hinge fragment were applied. Different amino acid Enhancer leader (Enh) sequences, truncated forms of peptide SUP-B8,11 were synthesized as DNA duplex containing cohesive NdeI and NheI ends with the following sequences: Enh0 – no added amino acid; Enh1 – YSF; Enh2 – YSFE; Enh3 – YSFED; Enh4 – YSFEDL; Enh5 – SFEDL; Enh6 – FEDL; Enh7 – EDL. The gene encoding the human epidermal growth factor was amplified as BamHI/XhoI fragment (gift of Nils Holler, Institute of Biochemistry, Lausanne, Switzerland). The different DNA building blocks were combined as depicted in Figure 1a to form p-EGF and p-IRR expression plasmids. All constructions were verified by DNA sequence analysis, and the corresponding protein sequences of p-EGF and p-IRR are shown in Figure 1b.

Production and purification of peptabodies

E. coli TG1 strain was used for the production of recombinant proteins. From an overnight preculture, p-EGF and p-IRR clones were grown in 250 ml of 2xTY medium with 100 μg/ml ampicillin at 37°C until O.D. (600 nm) = 0.6. The cultures were cooled to 20°C and protein expression was induced with 250 μM isopropyl β-D-thiogalactoside (IPTG). After 16 h induction at 22°C, bacteria were harvested by centrifugation and the cell pellet was solubilized in 25 ml of denaturing solution (1× PBS, 8 M urea, 10 mM β-mercaptoethanol at pH 7.4) for 2 h at room temperature. Insoluble material was pelleted by centrifugation at 10,000g for 10 min and the supernatant was mixed with 250 μl of Ni2+-NTA agarose beads (Qiagen). After 2 h incubation at room temperature under agitation, the resin was washed 3 times with a washing solution (1× PBS, 5 M urea, 10 mM β-mercaptoethanol, 20 mM Imidazole at pH 7.4). Proteins were eluted in 2.5 ml elution buffer (1× PBS, 5 M urea, 10 mM β-mercaptoethanol, 150 mM Imidazole at pH 7.4).

Refolding of denaturated peptabodies

The refolding protocol was performed at 4°C. A maximum of 1 mg of eluted protein was slowly diluted (1/100) in 250 ml of cold dilution buffer (50 mM Tris, 50 mM NaCl, 10 mM β-mercaptoethanol, 0.1% Triton-X100 at pH 8.2) and placed overnight at 4°C before dialysis against dilution buffer without β-mercaptoethanol. The dialysis process was continued over 3 days against dialysis buffers containing decreasing concentrations of Triton-X100 (0.01, 0.001 and 0.0001%). Finally, the refolded protein was concentrated to a final volume of 250 μl and characterized by SDS-PAGE and Western blot analysis using a specific anti-His primary antibody (1/3,000, Sigma) and a mouse anti-Fab secondary antibody (1/50,000, Sigma), and the ECL peroxidase detection kit (Amersham, Braunschweig, Germany).

Evaluation of peptabody stability

To assess the stability of refolded peptabodies, the molecule was kept at 4°C for up to several weeks or incubated at 37°C for 7 days in conditioned cell culture medium from a confluent A-431 culture. The peptabody was analyzed by SDS-PAGE and its functional integrity was tested by measuring in vitro binding and cytotoxicity activities as described later.

Direct peptabody binding assays on cell cultures

The binding potentials of p-IRR, p-EGF and Mab-425 (a specific anti-EGFR monoclonal antibody) were determined on living cells. The human melanoma A-431 and breast cancer MCF-7 cell lines were maintained in Dulbecco's MEM (DMEM, Gibco) medium supplemented with 5% (v/v) fetal calf serum (FCS, Brunschwig), 1% antibiotics (Penicillin, Streptomycin and Fungizone, Amimed Bioconcept, Allschuiil, Switzerland) at 37°C in a humidified 5% CO2 incubator. HSMC were maintained in Ham's F10 nutrient mixture (HAMF10, Sigma) supplemented with 5% (v/v) fetal bovine serum (FBS, Gibco), 1% antibiotics (Penicillin, Streptomycin and Fungizone, Amined) and 2.5 ng/ml basic fibroblast growth factor (b-FGF, Gibco) at 37°C in a humidified 5% CO2 incubator.

At 70–80% confluence, cells were detached by incubation with 0.5 mM ethylenediaminetetraacetic acid (EDTA) diluted in 1× PBS (EDTA-PBS). 1 × 104 cells in 100 μl were seeded per well of a 96-well plate (Costar) and incubated for 16 h at 37°C in a humidified 5% CO2 incubator.

Cells were washed with PBS–BSA (10 mg/ml) and incubated for 90 min at 4°C with two different concentrations (1 or 10 nM) of p-EGF and p-IRR, or 10 nM of Mab-425. Afterwards, cells were washed 3 times with cold PBS–BSA and bound peptabodies were detected using a primary anti-His antibody (1/3,000) and a secondary mouse anti-Fab antibody (1/50,000). The monoclonal antibody Mab-425 was directly detected with the same secondary mouse anti-Fab antibody. A peroxidase substrate kit (TMB, Pierce, Perbio Lausanne, Switzerland) was used for detection according to manufacture instructions.

Competitive binding assays

The specific anti-EGFR monoclonal antibody Mab-425 was labeled with 125I by the iodogen method according to manufacture instructions (Pierce) and used in a competitive assay against purified p-EGF. Assays were performed in duplicate. Briefly, 1 × 105 cells were incubated for 90 min at 4°C with PBS–BSA (10 mg/ml), containing a constant concentration (50 pM) of radio-iodinated Mab-425 (125I-Mab-425) and increasing concentrations of p-EGF or p-IRR (1.5–200 nM). After 3 washings with cold PBS–BSA, the radioactivity bound to the cells was counted using a gamma reader (COBRAII). Results are expressed as percentage of maximum binding of the radio-iodinated Mab-425.

In vitro cell viability assay using MTT (3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyl tetrazolium bromide) method

A-431 and MCF-7 cells were conditioned for 24 h in a serum-free medium OPTIMEM (Gibco) supplemented with 5% (v/v) fetal bovine serum (FBS, Brunschwig), 1% antibiotics (Penicillin, Streptomycin and Fungizone, Amined) at 37°C in a humidified 5% CO2 incubator. Cells were detached with EDTA-PBS solution and 2,500 cells in 100 μl per well were seeded into 96 wells plates (Costar, Corning, NY). After 16 h fixation, 50 μl cell culture media was added containing human EGF (final concentrations: 5, 25 or 50 nM), p-EGF or p-IRR (final concentrations: 1, 5 or 10 nM) or Mab-425 (final concentrations: 2.5, 12.5, 25 and 50 nM). Lower concentrations of p-EGF (1, 10, 100 pM) were also tested to evaluate a potential stimulating effect at weak dose. MTT cell viability assays were performed in triplicates, after 30 min, at 1, 3, 6 and 24 h of incubation. Briefly, 50 μl of MTT (1 mg/ml) were added onto cells and incubated for 2 h at 37°C in a humidified 5% CO2 incubator. The cell culture medium was replaced with 100 μl DMSO (Sigma). Following the formazan solubilization, absorbance of each well was measured at 490 nm using an ELISA plate reader (Bio-Tek Instruments, EL800, Winooski, VT). Cell viability was expressed in percentage of cells compared to the untreated control.

Similar experiments were performed by incubating HMSC and A-431 cells for 3 days with 2.5 nM of p-EGF or p-IRR. Cell viability was assessed using MTT assay. Pictures of cells were taken with a NIKON Eclipse TE 2000-S microscope connected to an OLYMPUS digital camera C-4040 zoom.

Apoptosis detection by Annexin V assay

A-431 and MCF-7 cancer cells were assayed for induction of apoptosis. Cells (5 × 105 per well in 500 μl DMEM) were incubated for 24 h before the medium was replaced with 500 μl of serum-free medium OPTIMEM containing 10 nM of p-EGF or p-IRR. At different time intervals (0, 30, 60, 180 and 360 min) adherent and floating cells were collected, pooled and analyzed for apoptosis by using an Annexin V-FITC apoptosis detection kit (Apotech, Switzerland) that detects phosphatidylserine on the outer surface of the cell membrane. Briefly, cells were washed with 1× PBS, stained at room temperature for 15 min with 50 μl of Annexin V-FITC (5 μg/ml) and washed once each with 1× PBS and with FACS buffer (1× PBS, 5% FCS, 0.02% NaN3). Cells were then resuspended in 200 μl of FACS buffer, counted and analyzed by FACS.

DNA ladder fragmentation analysis

A-431 cells (106 cells) were seeded in a 100-mm dish, incubated for 24 h and treated for 12 h with EGF (50 nM), p-IRR (10 nM), p-EGF (10 nM) or a positive control Fas-L (1 nM).12 Cells were detached using 1× PBS/0.1 mM EDTA, harvested by centrifugation and lysed in 0.1 ml buffer A [10 mM Tris-HCl (pH 7.5), 400 mM NaCl, 1 mM EDTA, 1% NP40] for 1 min at room temperature. After centrifugation to remove cell debris (1000g for 5 min), supernatants were adjusted to 1% SDS, and treated with RNAse A (5 μg/μl) for 2 h at 37°C and with proteinase K (2.5 μg/μl) for 2 h at 56 °C. The sample was extracted with phenol–chloroform and DNA was precipitated with ammonium acetate. DNA pellets were resuspended in 50 μl of DNA loading buffer and 5 μl of the sample was separated on a 2% agarose gel.

Results

Influence of the N-terminus sequence on the yield of peptabody production

We introduced several modifications (Fig. 1) to previously described versions of peptabodies.9, 10 The EGF ligand was placed at the C-terminal part of the molecule replacing N-terminal binding peptide sequences in previous peptabody versions. This change was necessary to free the C-terminus of EGF, which has to be exposed to retain the binding activity of the growth factor.13 Furthermore, this reorganization allowed the addition of an N-terminal enhancer peptide sequence, increasing protein expression of the fusion protein. Peptide SUP-B8 was previously used in our laboratory to construct a peptabody with the ligand located at the N-terminus (data not published). From the original sequence YSFEDLYRR, the truncated forms were used to analyze their influence on the peptabody synthesis. Huge variations of the production yield were observed between the different enhancers since only Enh2 and Enh4 allowed to produce large amounts of peptabody (data not shown). The presence of an enhancer sequence 4 at the N-terminus leads to a production rate that is 100-fold higher than that without enhancer, while enhancer 2 increases the production yield by 20-fold.

Production of stable peptabody molecules

Both peptabodies, Irrelevant and EGF, were constructed using enhancer 4 (Figs. 1a and 1b) and produced in bacteria at high levels (2–4 mg/l of culture) as inclusion bodies. Denatured monomers showed apparent molecular mass of approx. 10 and 15 kDa for p-IRR and p-EGF, respectively (Fig. 2a). Upon refolding, monomers switched to higher molecular mass (Fig. 2b) of 47 kDa for p-IRR and 76 kDa for p-EGF, in agreement with calculated molecular mass of putative pentameric structures. Refolded proteins appeared as single band, indicating a high efficacy of the refolding protocol even for a complex protein like EGF with several intramolecular disulfide bridges. Peptabodies were found to be particularly stable. Prolonged incubation at 4 or 37°C (see Material and Methods) in either dialysis buffer or conditioned culture medium did not affect the mature pentameric structure or the functional characteristics of the molecule.

Figure 2.

12% SDS-PAGE analysis of p-EGF and p-IRR under (a) reducing and (b) nonreducing conditions.

The N-terminal addition of an enhancer sequence, with YSFEDL amino acids, resulted in a 10-fold increase of peptabody-EGF production. This observation was also made with several other peptabody constructions for which low expression levels were observed when expressed without enhancer sequence (data not shown).

Peptabody-EGF binds efficiently to cancer cells expressing EGF receptor

The binding activity of peptabodies was evaluated on cultured cell lines overexpressing different levels of EGF receptor: epidermoid cancer cells A-431 with a very high level of expression, breast cancer cells MCF-7 with a level of expression slightly superior to normal cells and normal bladder smooth muscle cells expressing constitutive levels of EGFR. Peptabody binding was compared with the binding of the monoclonal antibody anti-EGFR Mab-425. Peptabody-EGF showed dose-dependent binding on all 3 cell lines (Fig. 3a) and the binding correlated with their respective EGF receptor expression levels. The peptabody control molecule p-IRR lacking the EGF-ligand domain but possessing the enhancer sequence did not exhibit specific binding on tested cell lines, which demonstrated the strictly EGF dependent type of binding of the peptabody molecule. However, binding on A-431 cells expressing very high levels of EGFR was only about twice as high for 10 nM of p-EGF compared to 10 nM of Mab-425, a big difference in binding activity was observed using MCF-7 cells, where p-EGF binding was found to be about 8 times higher than for the antibody. This could be explained by the double detection used for peptabody, which consists of anti His Antibody followed by mouse anti-Fab secondary antibody while Mab431 was revealed by only one antibody. However, this high binding activity should also be due to an expected cooperative binding type of the 5 EGF ligands present on the peptabody molecule, which contrasts the binding of the 2 arms of the antibody. When tested on cells expressing a low level of EGFR as normal smooth muscle bladder cells, p-EGF binding dropped to levels that were close to binding observed with the monoclonal antibody.

Figure 3.

(a) ELISA illustrating the direct binding of p-IRR, Mab-425 and p-EGF on A-431, MCF-7 and HSMC cells. (b) Competitive binding assay between p-EGF and the specific monoclonal antibody anti-EGFR (Mab-425) on A-431 cancer cells.

Peptabody-EGF binding occurs through interaction with EGF receptor

Peptabody-EGF demonstrated a strong binding affinity for cells overexpressing EGF receptor. To affirm that p-EGF binds specifically to the EGF receptor, we examined the ability of p-EGF to outcompete a specific anti-EGFR monoclonal antibody. Displacement of 125I-labeled Mab-425 bound on A-431 cells was determined with increasing concentrations of p-EGF and results clearly showed that p-EGF competes for the same binding site as Mab-425. About 10 nM of p-EGF were needed to displace 50% (IC50) of 125I-labeled Mab-425 (Fig. 3b) bound to A-431 cells. In addition, this binding occurs specifically through the EGF part of p-EGF since p-IRR did not compete with 125I-labeled Mab-425. Finally, p-EGF exhibited a higher activity of 125I-labeled Mab-425 displacement than the nonlabeled version of the monoclonal antibody itself (not shown).

Multimerization of EGF is required to kill cancer cells

The cytocidal effect on A-431 and MCF-7 cancer lines was assessed for p-EGF, monomeric EGF and Mab-425 (Fig. 4). The epidermoid and breast cancer cell lines A-431 and MCF-7 were highly sensitive against p-EGF action. The effect was dependent on the concentration and time of drug exposure. At a concentration as low as 5 nM, close to 100% of cancer cells were killed within 3–24 h. As expected, the reaction was quicker in A-431 cells with a higher level EGF receptor expression than in MCF-7. Addition of p-IRR did not affect cell growth (not shown).

Figure 4.

Cytoxicity in various tumor cells induced by exposure to different concentrations of peptabody-EGF. Comparison with monomeric EGF and monoclonal antibody raised against EGFR (Mab-425). Results are represented in percentage of inhibition compared to untreated control cells.

Comparison of action by p-EGF with monomeric EGF (1 nM of peptabody-EGF was considered equivalent to 5 nM of monomeric EGF) indicated that EGF multimerization is required for efficient killing of cells. Monomeric EGF inhibited growth of A-431 cells by no more than 40% and further increasing the concentration from 5 to 50 nM EGF did not significantly change growth inhibition. In addition, no cell death was observed in culture plate, proving an inhibition action rather than killing action. In contrast to A-431, MCF-7 cells were not sensitive against up to 50 nM EGF.

Interestingly, no effect was observed at very low dose of p-EGF (down to pM), demonstrating the cutting action of p-EGF in comparison to monomeric EGF.

Incubation of cells in presence of Mab-425 showed no significant effect of the antibody at up to 6-fold higher concentrations than p-EGF.

Peptabody-EGF shows in vitro specificity for EGFR overexpressing cells

To evaluate the in vitro effect of peptabody-EGF on cells expressing a constitutive level of the EGF receptor, human bladder smooth muscle cells were exposed over a prolonged period of time to p-EGF. The cells were incubated in the presence of p-EGF or p-IRR for 3 days, before MTT viability assays and visualization of in vitro cultures were performed. However hardly any effect could be detected on the bladder smooth muscle cells after 3 days of treatment, the action of 2.5 nM p-EGF on cells overexpressing EGF receptors (A-431) was very strong. Already 24 h after addition of p-EGF, close to 100% of cells were dead (Fig. 5). P-IRR did not cause any growth effect.

Figure 5.

Visualization of in vitro culture of epidermoid cancer cells A-431 or normal bladder muscle cells HMSC in serum-free DMEM medium after 3 days of treatment with p-EGF or p-IRR.

Peptabody-EGF induces rapid apoptosis of cancer cells

To assess a potential apoptotic action of p-EGF, Annexin V was employed to detect exposed phosphatidylserine. The apoptotic response to drug exposure was measured by flow cytometry. Representative flow scan presentations of the light scatter characteristics of Annexin V stained cells before and after p-EGF exposure are presented (Fig. 6). Significant increases in fluorescence intensity, indicative of apoptosis, were observed after treatment of A-431 with 10 nM p-EGF within 30 min of incubation. No effect was found in treating EGFR overexpressing cells with the control molecule p-IRR (not shown). Induction of apoptosis on MCF-7 cells expressing lower levels of the receptor than A341 was slower and a significant increase in fluorescence intensity was detected after 1 h.

Figure 6.

Apoptosis detection by Annexin V staining performed by FACS analysis. A-431 and MCF-7 cells are incubated for 30 min, 1 h, 3 h and 4 h with p-IRR (dotted line), EGF monomeric form (thin line), p-EGF pentameric form (thick line) and control (full area).

Beside Annexin V staining, DNA fragmentation is another characteristic of apoptotic cells. A-431 cells were treated with 10 nM p-EGF for 12 h and DNA fragments were separated by agarose gel electrophoresis (Fig. 7). Formation of a characteristic DNA ladder pattern was specifically induced by p-EGF and the positive control FasL, and no DNA degradation was detectable after p-IRR or EGF exposure. However we clearly demonstrated that p-EGF induces apoptosis of cancer cells overexpressing EGFR. The apoptotic pathways involved in this process remain to be elucidated.

Figure 7.

Electrophoretic separation of fragmented DNA from A-431 cancer cells after treatment with p-IRR, monomeric EGF, p-EGF, and Fas-L.

Discussion

Previously published peptabodies were rather simple molecule structures with a short polypeptide serving as ligand.9, 10, 14 Fusing a complex protein molecule such as the human epidermal growth factor to the peptabody core represented a big challenge mostly because of various intra- and intermolecular disulfide bridges that had to be formed correctly. Each hEGF domain consisting of 53 amino acids harbors 3 internal disulfide bridges and the core structure of the peptabody formed by the 5 hCOMP domains is stabilized by 5 intermolecular disulfide bridges. It was reported previously that human EGF forms inclusion bodies when produced in bacteria and required denaturation/renaturation to recover its biological activity.15 Concerning the peptabody core molecule, Clement et al.14 recently described the production of several versions, which appeared on SDS-PAGE as 5 major bands corresponding to monomeric and multimeric forms, demonstrating a nonperfect oxidation. In contrast to previously described peptabody-peptides, our peptabody-EGF molcule cannot be produced as soluble form and a refolding protocol had to be established to obtain stable, active proteins. Upon refolding, peptabodies appeared as a single band on SDS-PAGE, corresponding to the mature pentameric structure of the molecule. The refolded protein showed high stability and no change of its characteristics on SDS-PAGE analysis or its in vitro activity could be detected even after prolonged incubation at 4 or 37°C in culture media. This refolding protocol showed suitability for several other peptabodies (data not shown) and might prove suitable for the development of new versions of peptabody for other applications. This work demonstrates the feasibility to fuse a protein in complex tertiary structure onto the peptabody backbone and to produce large amount in prokaryotic system. Previous publication from Holler et al.16 reported the production of several 3 D structures (Fas, TRAIL-R1, TRAIL-R2, TRAIL-R3, TNF-R1, CD40) fused to COMP. However, these productions were made in eukaryotic cell line CHO, which provided a low level of expression in comparison to our system.

The insertion of an enhancer sequence and the transfer of the EGF ligand domain to the C-terminus of the molecule lead to a drastic increase of the production level of the peptabody when compared to original constructs lacking these features.

Previously, it was observed that the modification of the N-terminal sequence influences bacterial expression of peptabody molecules and a 12 amino acid segment, a peptide specific for the surface immunoglobulin receptor of the human Burkitt's lymphoma cell line SUP-B8,11 was found to increase significantly the expression levels of a peptabody construct (Jean Pierre Mach, Personal communication). We subcloned fragments of this sequence and tested their potential in boosting production levels of p-EGF in E. coli (data not shown). The enhancer sequence conferring the biggest impact on expression consisted of an YSFEDL sequence. Addition of this sequence to other peptabody constructs leads to a similar increase in production levels. A similar influence of the composition of the N-terminal ligand peptide fragment sequence on peptabody expression levels was also recently reported by Clement et al.14 However, further investigation will be made to determine if the increase of the synthesis occurs at the transcription or at the translation level.

The effect of exogenous EGF on tumor cells is a controversial subject and it has been reported to both increase and decrease proliferation and apoptosis.17 Maruyama described that local injections of 2 mg EGF significantly suppressed in vivo tumor growth of human breast cancer MX-1 and primary breast tumor UM-1.18 EGF was also shown to inhibit growth of MKN-28 human gastric carcinoma transplanted into nude mice,19 but other reports were unable to confirm any obvious effect on growth of human gastric cancer cell neither in vitro nor in vivo.20 In contrast, it was demonstrated that EGF increases the metastatic behavior of pancreatic cancer cell lines21 or stimulates the growth of thyroid cancer xenografts in nude mice.22 However the effect of human EGF on tumor growth depends heavily on the choice of cell lines and in vivo models; it undoubtedly plays an important role in cancer and it is widely accepted that the EGF might be used as scaffold to develop new therapies.

We speculated that the peptabody-EGF with its 5 EGF ligand domains should be an efficient tool to target the EGF receptor that is overexpressed or constitutively activated in a variety of human malignancies (see recent reviews23, 24, 25). In vitro experiments on several cell lines confirmed that the peptabody-EGF binds strongly cancer cells via the EGF receptor. To evaluate the interest of Peptabody EGF for a potential clinical outcome, it was compared with the murine monoclonal antibody Mab-425, which is under development as anticancer drug targeting EGFR under the name of EMD55900 (EMD Pharmaceuticals/ Merck). Mab-425 is an antibody directed against the ligand-binding domain of human EGFR and competes for receptor binding with EGF and other ligands. In vitro, Mab-425 is known to inhibit EGFR tyrosine kinase activity and proliferation of EGFR-overexpressing A-431 cell lines, but is not able to induce apoptosis as shown for p-EGF. In preclinical studies and in clinical trials, it has also been shown that Mab-425 requires synergy with others therapies such as chemotherapy (with doxorubicin/cisplatin) or radiotherapy.26

Interestingly, peptabody-EGF showed binding selectivity between a cell line with a constitutive level of EGFR expression such as HSMC and cells displaying a moderate overexpression such as the breast cancer cell line MCF-7. This selectivity is further supported by the cytotoxic effect of p-EGF killing, specifically A-431 or MCF-7 cells, but not the normal HSMC bladder smooth muscle cells. In addition, low concentrations of p-EGF did not provoke the stimulation of cell growth, demonstrating the safety of the molecule for using it in clinical development. Further experiments to modulate the threshold of selectivity by changing the length of hinge region and thereby adjusting the binding of the peptabody to cells expressing different levels of receptor densities on their surface are currently ongoing.

The rapid induction of cell death seen with p-EGF suggested activation of apoptosis as the most likely mode of action. Indeed, cells treated with peptabody-EGF showed typical characteristics of apoptosis, as confirmed by analysis of Annexin V staining and fragmentation of chromosomal DNA. This finding is in agreement with previous reports that demonstrated an apoptotic action of high doses of EGF when applied on cells overexpressing the EGF receptor, including the epidermoid cancer cell line A-431.27, 28, 29 Interestingly, the rate of apoptosis induction by peptabody-EGF seems to be dependent on EGF receptor density, since A-431 entered in apoptosis more rapidly than MCF-7. P-EGF showed a tremendous apoptotic potential and most of the cells entered in apoptosis within 1 h of peptabody exposure.

Although the use of antibodies and antibody fragments as antitumor agents has shown promise, there are disadvantages that must be overcome. For example, monoclonal antibodies are produced in hybridomas. However, the use of hybridoma cell lines in clinical production is less efficient when compared to other expression systems such as bacteria. Moreover, the cost of production of therapeutic proteins in mammalian cells remains an obstacle for their development, whereas production in bacteria is extremely cost-efficient. Compared with antibodies, the peptabody has the advantage that it can be produced at high level in bacteria. Furthermore, because of the very low effective concentration of the molecule, small quantities of peptabody should be sufficient to assure a potential therapeutic effect.

Blocking of the epidermal growth factor receptor and its ligands is a promising approach to treat human tumors, offering a noncytotoxic alternative to conventional chemotherapeutic cancer treatment. However, different antibodies developed during the last decade have a limited potential to induce apoptosis and combination with other treatment options is currently evaluated to increase antitumor effects of antibody drugs. In contrast, p-EGF is able to kill cancer cells without combining it with other treatments, which represents a major improvement over the antibody technology.

We reported the development of a new antitumor tool capable of inducing specific apoptosis of cancer cells overexpressing EGF receptor. The peptabody-EGF technology offers new perspectives in the targeting of EGF receptor pathway and might propose an alternative to existing monoclonal antibodies in cancer treatment.

Ancillary