Abstract
The ErbB2 receptor tyrosine kinase is overexpressed in ∼30% of breast tumor cases and its overexpression correlates with an unfavorable prognosis. A major contributor for this course of the disease is the insensitivity of these tumors toward chemotherapy. Monoclonal antibodies, inhibiting the ligand-induced activation of the receptor and tyrosine kinase inhibitors acting on the intrinsic enzymatic activity of the intracellular domain, have been developed as targeted drugs. Both have been shown to be beneficial for breast cancer patients. We targeted a third aspect of receptor function: its association with intracellular signaling components. For this purpose, we selected peptide aptamers, which specifically interact with defined domains of the intracellular part of the receptor. The peptide aptamers were selected from a random peptide library using a yeast two-hybrid system with the intracellular tyrosine kinase domain of ErbB2 as a bait construct. The peptide aptamer AII-7 interacts with high specificity with the ErbB2 receptor in vitro and in vivo. The aptamers colocalized with the intracellular domain of ErbB2 within cells. We investigated the functional consequences of the aptamer interaction with the ErbB2 receptor within tumor cells. The aptamer sequences were either expressed intracellularly or introduced into the cells as recombinant aptamer proteins. The phosphorylation of p42/44 mitogen-activated protein kinase was nearly unaffected and the activation of signal transducers and activators of transcription-3 was only modestly reduced. In contrast, they strongly inhibited the induction of AKT kinase in MCF7 breast cancer cells treated with heregulin, whereas AKT activation downstream of insulin-like growth factor I or epidermal growth factor receptor was not or only slightly affected. High AKT activity is responsible for the enhanced resistance of ErbB2-overexpressing cancer cells toward chemotherapeutic agents. Peptide aptamer interference with AKT activation resulted in the restoration of regular sensitivity of breast cancer cells toward Taxol. (Mol Cancer Res 2006;4(12):983–98)
- breast cancer
- peptide aptamers
- protein transduction
- interference with signaling components
- chemotherapy
Introduction
Cancer research has provided detailed insights into the processes and components involved in tumor initiation and progression. This knowledge serves as a basis for the development of highly specific drugs with improved efficacy and minimal side effects. They are routinely used for the treatment of cancer patients and constitute a major achievement of modern molecular medicine. The ErbB family of receptor tyrosine kinases plays a major role in the formation of human tumors and has been used as a target for the development of advanced cancer drugs. The epidermal growth factor (EGF) receptor (EGFR) and ErbB2 have both been known as oncoproteins for over 20 years (1). ErbB2 is found to be overexpressed in invasive breast, ovarian, stomach, bladder, salivary, and lung cancers (2-8).
After the discovery of the ErbB2 oncoprotein, research focused on the elucidation of the signaling pathways activated by this receptor and their role in cancer formation. The ErbB family of receptor tyrosine kinases are activated on binding of growth factors that belong to the EGF family (9). The binding of EGF-like ligands to ErbB receptor monomers induces a conformational change in the extracellular domain of the receptor. The conformational change of EGFR exposes a dimerization domain, which enables two monomers to interact with each other and form an activated receptor dimer (10). The ErbB2 receptor exhibits structural features that promote its oncogenic potential. In contrast to EGFR, this receptor is always present in the plasma membrane in an open conformation and the propensity to form dimers does not require the binding of a ligand (11); it heterodimerizes with other family members and can be activated (e.g., through heregulin and dimer formation with ErbB3). ErbB3 plays a crucial role in breast tumor cell division: it couples active ErbB2 to the phosphatidylinositol 3-kinase-protein kinase B/AKT pathway. ErbB2/ErbB3 dimers function as an oncogenic unit driving breast tumor cell proliferation (12).
Activated receptor dimers cause the subsequent activation of intracellular signaling pathways. This is achieved through the intrinsic tyrosine kinase activities present in the intracellular COOH-terminal tail of ErbB receptors. In a receptor dimer, the tyrosine kinase domain of one dimerization partner is able to phosphorylate specific tyrosine residues in the COOH-terminal part of the other monomer, a mechanism termed cross-phosphorylation (13). The phosphotyrosines then serve as binding sites for cytoplasmic signaling molecules and represent the vantage point of signaling pathways inducing proliferation, antiapoptotic responses, angiogenesis, and metastasis formation (reviewed in ref. 14).
Two pathways induced by ErbB2-containing heterodimers are the mitogen-activated protein kinase (MAPK) and AKT pathways, which are involved in cellular proliferation. AKT activation also provides a strong antiapoptotic response and protects cells against inducers of apoptosis (reviewed in ref. 15). AKT causes the direct inhibition of the apoptotic inducer Bad through its phosphorylation (16). AKT also phosphorylates and thereby inactivates the FOXO-1 transcription factor, which controls the expression of cell cycle inhibitors and proapoptotic proteins (17). AKT activation also causes the induction of nuclear factor κB–mediated antiapoptotic responses and the repression of tumor suppressors (e.g., phosphatase and tensin homologue; ref. 18).
ErbB2 activation is therefore associated with an antiapoptotic phenotype mediated by multiple pathways. This is reflected in the inefficiency of chemotherapy in patients with ErbB2-overexpressing tumor cells and the insensitivity of cultured tumor cells in vitro toward such drugs (19). The strong antiapoptotic signaling in combination with the high metastatic potential of ErbB2-overexpressing cells is a major factor responsible for the unfavorable prognosis of patients. Drugs targeting the ErbB2 receptor not only interfere with cell proliferation but also result in an improvement of the therapeutic efficacy of chemotherapy (20, 21).
Two strategies have been exploited to block the function of ErbB receptors. Drugs can act either through interaction with the extracellular domain and prevent the activation or dimerization, or they can inhibit the tyrosine kinase activity in the intracellular part of the receptor. Antibodies against the extracellular ErbB2 domain were derived many years ago (22), but it took until 1998 for the first ErbB2 antibody trastuzumab (Herceptin) to be approved as a drug for the treatment of metastatic breast cancer. It binds specifically to extracellular domain IV of the ErbB2 receptor (11) and acts by initiating the endocytosis of ErbB2, leading to its degradation in the lysosome (23). Its activity is further enhanced by the induction of an antitumoral immune response (24). In the meantime, it has become clear that this monoclonal antibody shows considerable therapeutic benefit. It leads to reductions in the rates of recurrence of tumors and in the number of deaths of patients (20, 21, 25). To target the tyrosine kinase domain, chemical compounds have been developed that bind the ATP binding pocket (e.g., PKI-166) and act as competitive binding inhibitors. For this reason, tyrosine kinase inhibitors have to overcome the high concentrations of ATP in cancer cells (26).
Both strategies, based on Herceptin and tyrosine kinase inhibitor, showed encouraging results with considerable patient benefit. Still, only a fraction of the patients respond and the responses are largely limited in duration. The combination of drugs inhibiting ErbB2 function with conventional chemotherapeutic agents seems to be promising. ErbB2-targeted approaches might further be enhanced by the interference with ErbB2 properties related to its signal transduction functions. Such a strategy has been outlined when investigators tried to adapt antibody binding to intracellular conditions. A single-chain derivative of a monoclonal antibody was used to sequester the receptor from the cell surface (27). These single-chain antibodies are composed of heavy and light chain variable domains connected by a flexible peptide linker. When expressed in ErbB2-transformed cells, the single-chain antibodies bound to the receptor, prevented its transit through the endoplasmic reticulum, and resulted in the reversion of the transformed phenotype (27).
The use of intracellular antibodies has been shown to be limited by the restrictions on folding imposed by the reducing milieu of the cytoplasm. For this reason, we have employed a strategy based on peptide aptamers, rather than on intracellular antibodies, to interfere with intracellular ErbB2 functions. In these constructs, a variable peptide region is displayed by a scaffold protein (e.g., thioredoxin; ref. 28). This protein has a defined structure under intracellular conditions and peptide sequences can be presented in the active loop structure of the protein. Peptide aptamers represent random sequences of 12 to 40 amino acids in length (29). They can be selected for the interaction with a target protein in a yeast two-hybrid system (30). Peptide aptamers have been shown to specifically bind and functionally interfere with intracellular proteins, such as transcription factors E2F and signal transducers and activators of transcription 3 (Stat3), viral oncoprotein E6, and growth factor receptors like EGFR and others (31-35).
Delivery of proteins that inhibit intracellular targets requires the passage through the cellular membrane. Recent advances have shown that recombinant proteins can be provided with a protein transduction domain (PTD), which greatly enhances their exogenous uptake (36) and improves their therapeutic potential. The most frequently used PTD is derived from the HIV-TAT protein, a positively charged sequence (37). The positive charge allows the interaction of the fusion protein with negatively charged heparin sulfate glycosaminoglycans on the surface of plasma membranes (38) and leads to the uptake of the conjugated proteins by macropinocytosis (39, 40), an actin-mediated form of endocytosis.
In the present study, we isolated peptide aptamers that specifically interact with the intracellular domain of ErbB2 and are able to inhibit selected functions of this receptor. Peptide aptamers were either expressed after gene transfer into tumor cells or delivered as recombinant proteins fused with a PTD. The aptamers were able to inhibit the activation of AKT on heregulin induction of tumor cells and thereby sensitize chemoresistant breast cancer cells for the treatment with paclitaxel. These peptide aptamers might become useful in the combination therapy of ErbB2-dependent cancers.
Results
Isolation of Peptide Aptamers Interacting with the Intracellular Domain of the ErbB2 Receptor from Random Peptide Libraries
The extracellular inhibition of ligand binding and dimerization through monoclonal antibodies and the intracellular inhibition of the tyrosine kinase activity are strategies that have been used to block the function of the ErbB2 receptor. Our aim was to find peptides that are able to interfere with subtle intracellular signaling aspects. We used the yeast two-hybrid system to identify specific binding sequences. In this screening system, the aptamer-target protein interaction can be analyzed under intracellular conditions. We derived bait constructs composed of sequences of the ErbB2 kinase domain fused to the Gal4-DBD. The expression of the ErbB2 bait constructs in the yeast strain KF1 (41) was shown by Western blotting experiments (Fig. 1 ).
Expression of Gal4-DBD ErbB2 fusion proteins in yeast. Left, fragments of the intracellular domain of ErbB2 were fused to the Gal4 DNA binding domain and served as bait constructs in the two-hybrid selection procedure in yeast cells. Schematic representation of the ErbB2 receptor with extracellular domains (ED) I-IV, the α-helical transmembrane domain (TM), the intracellular tyrosine kinase domain (TK), and the COOH-terminal part (Ct). The four fragments of the intracellular domain of ErbB2 integrated into the bait constructs and used in the yeast two-hybrid screen are indicated. The Gal4-DBD was fused to the entire cytoplasmic domain (CT) comprising amino acids 678 to 1,254; the tyrosine kinase domain KDg, comprising amino acids 678 to 987; and fragments of the tyrosine kinase domain KDI, comprising amino acids 678 to 852, or KDII, comprising amino acids 853 to 987. Right, the Gal4-DBD ErbB2 fusion constructs were expressed in the KF1 yeast strain and detected with an antibody specific for the Gal4 domain by Western blot analysis. The expression of cytoplasmic domain (80 kDa; lane 1), KDg (45 kDa; lane 2), and KDII (31 kDa; lane 4) was detected. No expression of KDI (lane 3) was found. The bait expressing yeast cells were used for yeast two-hybrid screening assays of the fusion constructs encoding the peptide libraries.
The prey vectors encode a 12-mer peptide library inserted into the thioredoxin scaffold, which is fused with the Gal4-AD. These prey vectors were transformed into the bait expressing cells. The yeast strain KF1 used for the yeast two-hybrid system has three selectable auxotrophic markers (Ade2, His3, and Ura3) under the control of Gal4 upstream activating sequences integrated into the genome. Due to differences in the promoter regions of the three genes, His3 expression is already induced when the bait and prey proteins weakly interact, whereas Ade2 and Ura3 expressions require stronger interactions. The use of the KF1 strain therefore gives an indication about the strength of the ErbB2-aptamer interaction.
In the yeast two-hybrid experiments, a total of 3.6 × 108 yeast transformants were screened. Thirteen different peptide aptamers were initially isolated that were able to interact with the kinase domain of ErbB2 and induce growth of the yeast cells in selective medium lacking histidine. The three aptamers (AI-1, AII-7, and Ag-11) that induced growth under stringent selection conditions are shown in Table 1 . These aptamers bind to different regions of the ErbB2 intracellular region. AI-1 was selected using fragment I of the kinase domain as a bait (see Fig. 1). Ag-11 selection was carried out with the entire kinase domain as a bait. These two baits include the active site (ATP-binding pocket) of the kinase domain. AII-7 interacts with fragment II of the kinase domain, which does not include the active site.
Sequences of Peptide Aptamers Interacting with the Kinase Domain of ErbB2
The sequences of the isolated aptamers were determined and compared with known protein-protein interaction motifs. These homology searches using the Scansite 2.0 (42) software revealed tyrosine residues within AI-1 and Ag-11, which could potentially be phosphorylated. The Scansite software searches databases of peptides that have been isolated in the past with phage display analysis and interact with a certain target protein (e.g., EGFR). The potential phosphorylation site within the aptamer sequence AI-1 is predicted to interact with an SH2 domain of phospholipase Cγ. The potential phosphorylation site in aptamer Ag-11 is predicted to interact with the SH2 domain of Src. It is possible that the aptamers could serve as substrates for the ErbB2 tyrosine kinase and interact as phosphopeptides with the identified SH2 domains. Aptamer AII-7, which interacts with sequences outside the active site in the kinase domain, was chosen for further analysis. We hypothesized that a potential inhibitory function might be based on a mechanism distinct from interference with the tyrosine kinase activity.
Confirmation of the Specificity of Peptide Aptamer Interactions with the Intracellular Domain of the ErbB2 Receptor In vivo and In vitro
The peptide aptamers obtained in the yeast two-hybrid screening procedure were further analyzed and the specificity of their interaction properties was studied in in vitro and in vivo experiments. For this purpose, they were tested in vivo by reintroduction into yeast cells expressing bait constructs encoding fragments of different receptors. The specificity of the in vitro interactions was investigated in glutathione S-transferase (GST) pull-down experiments. These experiments were carried out to examine the cross-reactivity of the peptide aptamers with other receptor tyrosine kinases.
KF1 yeast cells were cotransformed with constructs encoding the selected aptamer sequences and the complete tyrosine kinase domains (KDg) of insulin-like growth factor receptor (IGF-IR) or EGFR. IGF-IR belongs to the type II receptor tyrosine kinases. The kinase domain shows only a moderate homology with the ErbB2 receptor kinase. EGFR is a member of the type I ErbB receptor tyrosine kinase family and its kinase domain shares 82% homology with the kinase domain of ErbB2. Growth of the yeast cells on medium lacking histidine indicates that aptamers Ag-11 and AII-7 interact with the tyrosine kinase domain of the ErbB2 receptor, but not with IGF-IR (Fig. 2 ). Only a very weak interaction of aptamer AII-7 with the kinase domain of EGFR was observed (Fig. 2).
Interaction specificity analysis of peptide aptamers with intracellular growth factor receptor domains in vivo; yeast two-hybrid experiments. The KF1 cells were cotransformed with bait constructs encoding the complete kinase domain of ErbB2 (ErbB2-KDg), IGF-IR (IGF-IR-KDg), or EGFR (EGFR-KDg) and with aptamer encoding prey constructs [Ag-11, AII-7, and thioredoxin (Trx)] to determine the interaction specificity of individual aptamers. The receptor domain encoding constructs comprised a Gal4-DBD and the aptamers or empty thioredoxin fused to the Gal4-AD. The transformed yeast cells were grown overnight in liquid cultures and plated the next day on media lacking leucin and tryptophan (−LT) as control for transformation of both vectors and on media lacking leucin, tryptophan, and histidine (−LTH). For plating, the A600 of the culture was adjusted to 0.7 from which 1:10 serial dilutions were made. The KF1 yeast cells are only able to synthesize histidine on intracellular interaction of the receptor domain and peptide aptamer, which in turn allows growth under selective conditions (−LTH). The construct encoding the thioredoxin scaffold protein without a peptide sequence served as a control and does not interact with ErbB2-KDg (lane 1). Ag-11 interacts with the ErbB2-KDg domain (lane 2) but not with the EGFR-KDg (lane 3). AII-7 interacts with ErbB2-KDII (lane 4) but not with the kinase domain of IGF-IR (lane 5) and only slightly with the EGFR-KDg (lane 4).
To further confirm the specificities of interaction of the peptide aptamers, we conducted GST pull-down experiments. For these experiments, the aptamer proteins were overexpressed as GST fusion proteins in bacteria and purified by affinity chromatography. The purified proteins were incubated with cell extracts enriched in ErbB2 or EGFR. For this purpose, membrane fractions were prepared from SKBR3 cells, known to express high levels of ErbB2, or from Renca-EGFR cells, known to express high levels of EGFR (Fig. 3 ). Two independent bacterial lysates of GST-aptamer fusion proteins (lysates 1 and 2) were used in these experiments. The thioredoxin scaffold protein without an aptamer sequence served as a negative control. A peptide aptamer, which was selected for binding to IGF-IR1 and which interacts with ErbB2, served as a positive control. As can be seen in Fig. 3A, the aptamer-GST fusion proteins AI-1 and AII-7 were able to efficiently pull down the ErbB2 receptor, but were not able to interact with the EGFR extracted from Renca-EGFR cells (Fig. 3B). Other aptamers (AI-2, AI-3, AII-5, AII-6, and AII-8) tested in this pull-down experiment showed a weak interaction with EGFR and were not used for further analyses.
Interaction specificity analysis of peptide aptamers with intracellular growth factor receptor domains in vitro; GST fusion protein pull-down experiments. Aptamers (AI-1 and AII-7) and the scaffold protein thioredoxin were fused to GST and the fusion proteins were expressed in bacteria. Bacterial lysates were prepared. Bottom, expression of the aptamers in the lysates was shown by SDS-PAGE. These lysates were incubated with SKBR3 cell extracts, which contain high levels of ErbB2 receptor (A), or with Renca EGFR cell extracts, which contain high levels of EGFR (B). The GST fusion proteins were pulled down by glutathione beads and the retained complexes were analyzed. A. Two independently prepared aptamer-containing bacterial lysates (Lys1 and Lys2) were used. An aptamer known to interact with the ErbB2 receptor was used as a positive control (C+). Proteins retained by the glutathione beads were recovered and analyzed by gel electrophoresis. The ErbB2 receptor, pulled down in a complex with the GST-aptamer fusion proteins C+, AI-1, and AII-7, was visualized by Western blot analysis with an ErbB2-specific antibody. B. To control the specificity of isolated aptamers, the same GST pull-down experiment was done with EGFR-containing cell extracts from Renca EGFR cells. Retained proteins were recovered and analyzed with an EGFR-specific antibody. Although the kinase domains of both receptors share 82% sequence homology, the aptamers tested (AI-1, AI-2, AI-3, AII-6, AII-7, and AII-8) show no, or only very slight, interactions with EGFR.
These data show that the specific interaction of the aptamers AI-1, AII-7, and Ag-11 with ErbB2 could be verified in vivo by cotransformation with different bait constructs and in vitro by GST pull-down experiments with extracts enriched in ErbB2 or EGFR. These three aptamers were selected for further functional analyses.
Recombinant Purified Peptide Aptamers Can Be Introduced into Cells via a PTD and Interact Intracellularly with ErbB2
To evaluate the function of peptide aptamers in cell culture or animal models, it is necessary to introduce them into cells from the outside. To achieve uptake into target cells, the peptide aptamers were fused with a PTD, consisting of a homopolymer of nine arginine residues. The fusion proteins were bacterially expressed and purified. PTD fusion proteins are taken up by macropinocytosis and subsequently released from early endosomes into the cytoplasm (39, 43).
The PTD-aptamer proteins were furnished with a Flag epitope to facilitate their detection (Fig. 4A ), expressed in bacteria and purified from bacterial lysates by fast protein liquid chromatography under denaturing conditions (Fig. 4B). The denatured proteins were refolded in an arginine-containing buffer and dialyzed, which is necessary for these recombinant proteins to maintain their specific binding characteristics. The purity and concentration of the proteins was investigated by gel electrophoresis and Coomassie staining (Fig. 4C). The correct refolding and the binding properties were investigated by coimmunoprecipitation experiments. Cell extracts were prepared from SKBR3 breast carcinoma cells, which overexpress the ErbB2 receptor. The extracts were incubated with 0.5 μmol/L recombinantly expressed, purified peptide aptamers (Fig. 4D). Subsequently, the receptor was bound to a specific ErbB2 antibody and protein complexes were isolated with Protein-A magnetic beads. The presence of the peptide aptamers in the immunopurified ErbB2 complex was detected by Western blot with a Flag-specific antibody. Peptide aptamer AII-7 was prominently detected, indicating its strong interaction with ErbB2. The empty thioredoxin protein and aptamer Ag-11 only showed weak interactions (Fig. 4D).
Expression and purification of recombinant aptamer-PTD fusion proteins and their interaction with the ErbB2 receptor in vitro and in vivo on protein transduction into SKBR3 cells. A. Schematic representation of the domains of bacterial expression vector, pFlag-2: Tac bacterial promoter, Flag tag (for antibody recognition), Trx-apta (thioredoxin scaffold protein with inserted peptide aptamer sequence), PTD (consisting of nine arginine residues), and six histidine residues used for protein purification. B. Aptamer fusion proteins AI-1 and AII-7 and the thioredoxin scaffold protein were expressed in bacteria and purified under denaturing conditions (8 mol/L urea in PBS) by fast protein liquid chromatography. Individual fractions were assayed by SDS-PAGE and the gels were stained with Coomassie blue. C. The proteins present in fractions 2 and 3 (thioredoxin and AI-1) or fractions 3 and 4 (AII-7) were pooled, renatured, and protein aliquots (10 and 20 μL) were analyzed by SDS-PAGE to determine the purity and the concentration of the aptamer-PTD fusion proteins after dialysis. Aliquots of 1, 2, 3, and 5 μg of bovine serum albumin (BSA) were used as a standard. D. The purified aptamer fusion proteins interact with the ErbB2 receptor in vitro. To show that the recombinant aptamer fusion proteins are able to interact with the ErbB2 receptor on renaturation, 0.5 μmol/L of purified aptamers were incubated with SKBR3 cell extracts containing high levels of ErbB2-receptors. Aptamer-ErbB2 receptor complexes were immunoprecipitated with an ErbB2-specific antibody coupled to Protein-A magnetic beads. On elution of the bound protein and separation of the proteins by SDS-PAGE, aptamers were detected by Western blotting with a Flag-specific antibody. The presence of the receptor in the extract was analyzed by an immunoprecipitation reaction (IP) with an ErbB2-specific antibody. Ten percent of the purified aptamers were used as input control (IN). As a negative control, the purified aptamers were incubated with the magnetic beads without the SKBR3 cell extracts (Beads). E. The purified aptamer fusion proteins interact with the ErbB2 receptor after protein transduction into SKBR3 cells. Protein transduction of aptamer-PTD fusion proteins was accomplished by incubation of SKBR3 cells with 2 μmol/L of purified aptamers or thioredoxin proteins for 3 h. The cells were lysed and ErbB2 was complexed with an ErbB2-specific antibody and bound to magnetic Protein-A beads. The retained proteins were analyzed by gel electrophoresis and Western blot analysis with Flag-specific (top) and ErbB2-specific (bottom) antibodies. Ten percent of the total cell extract were used as input to visualize the uptake of the transduced aptamer proteins. IP, proteins isolated by the antibody interaction. SKBR3 cell extracts without prior incubation of recombinant peptide aptamers were used as a negative control (C−).
To investigate the ability of the recombinant aptamer protein to enter the cells and recognize its target protein intracellularly, we transduced SKBR3 cells with the PTD-aptamer protein and did a coimmunoprecipitation experiment with cytosolic extracts. SKBR3 cells were incubated with 2 μmol/L purified aptamers and 100 μmol/L chloroquine for 4 h to allow the peptide aptamers to enter the cytoplasm. The chloroquine treatment enhances the escape of the aptamers from the endosomes. After cell lysis, immunoprecipitation experiments were done with ErbB2-specific antibodies and the ErbB2 and PTD-aptamer proteins were visualized in the isolated protein complexes. The presence of these proteins in the blot analysis confirmed the uptake of aptamers into the cells and the intracellular interaction of the transduced AII-7 aptamer protein with ErbB2 (Fig. 4E). The intracellular binding of aptamer Ag-11 could not be detected. No interaction was observed with the empty thioredoxin protein, which served as a negative control.
We confirmed the specific intracellular interaction of the transduced peptide aptamers with the ErbB2 receptor by direct visualization in immunofluorescence microscopy experiments. For this experiment, we used NIH#3.7 cells, which stably express the ErbB2 receptor (44), and transduced them with thioredoxin or AII-7 for 4 h. The cells were treated with 100 μmol/L chloroquine to enhance the release from the endocytotic vesicles. They were subsequently washed with 0.2 mol/L acidic acid to remove proteins associated with the cell surface. ErbB2 and peptide aptamers were detected by confocal laser scanning microscopy using specific antibodies and secondary antibodies tagged with a fluorescent dye. The NIH#3.7 cells could be efficiently transduced and the thioredoxin protein, as well as peptide aptamer AII-7, was detected. The empty thioredoxin protein was mainly present in the cytoplasm of the cells (Fig. 5 ). In contrast, a significant fraction of peptide aptamer AII-7 was found at the cell membrane, where it colocalized with the ErbB2 receptor. Colocalization was further verified by comparing the fluorescent intensity values of the signals detected with a microscope (Fig. 5). These results confirmed the data obtained in the coimmunoprecipitation experiments and strongly suggest that the aptamers are able to enter the cell by protein transduction and interact with ErbB2 intracellularly.
Colocalization analysis of transduced recombinant aptamer-PTD fusion proteins and the ErbB2 receptor by confocal laser scanning microscopy. NIH#3.7 cells are NIH 3T3 cells stably transfected with an ErbB2 expression construct. They were transduced with 2 μmol/L of purified aptamers or thioredoxin proteins. After an acid wash, which removes the untransduced peptide aptamers from the cell surface, the cells were subjected to immunofluorescence staining. The presence of the ErbB2 receptor (red) and aptamer AII-7 or thioredoxin (green) was visualized by confocal laser scanning microscopy. Colocalization (yellow fluorescence) of ErbB2 and peptide aptamers was determined through an overlay of the green and red fluorescence. Fluorescence intensities were measured in a line of interest (LOI) and are shown in an intensity/LOI diagram (right).
Intracellular Interaction of Peptide Aptamer AII-7 with ErbB2 Inhibits AKT Activation after ErbB2/ErbB3 Induction with Heregulin
The experiments shown above indicate the specific intracellular interaction between the peptide aptamers and the ErbB2 receptor but do not allow predictions about the functional consequences. We analyzed the effects of the aptamers on the biological functions of the ErbB2 receptor on ligand activation. The ErbB2 receptor assumes a special role among the EGFR family. It acts as a preferred heterodimerization partner with the other family members (45) but it is not able to bind an EGF-like ligand with high affinity by itself (46). The ErbB3 receptor can be activated by heregulin binding and forms a signaling unit with ErbB2 in tumor cells (12). The ErbB3 receptor, however, does not have an active tyrosine kinase domain and its signaling activity is dependent on the ErbB2 dimerization partner. This dimer can be induced by heregulin binding to ErbB3 and ErbB2/ErbB3 activation causes a strong induction of the phosphatidylinositol 3-kinase-AKT pathway. This pathway in turn results in an antiapoptotic phenotype (reviewed in ref. 15).
MCF7 cells are breast carcinoma cells that express moderate levels of ErbB2 and ErbB3. When the cells are induced for a short period of time (10 min) with heregulin, ErbB2/ErbB3 dimers are formed and phosphorylation of the AKT kinase can be observed. This cell system was used to analyze the functional consequences of peptide aptamer binding to the ErbB2 receptor. Initial experiments were done in which the peptide aptamers were intracellularly expressed on transfection of a gene construct. Although the transfection efficiency was only ∼30% (data not shown), a clear reduction in AKT phosphorylation was detectable when MCF7 cells transfected with aptamer AII-7 were compared with cells transfected with the control thioredoxin construct (Fig. 6 ). Aptamer AI-1 showed no reduction in AKT phosphorylation and aptamer Ag-11 showed only a weak reduction of AKT phosphorylation. No AKT phosphorylation was detected in cells not treated with heregulin (Fig. 6). These results are in accordance with the relative binding affinities for the different aptamers observed above.
Peptide aptamers suppress heregulin-induced AKT phosphorylation intracellularly on aptamer introduction into cells by gene transfection. MCF7 cells were transiently transfected with pRC/CMV-VP22 aptamer vectors encoding thioredoxin, Ag-11, AI-1, or AII-7 and a pEGFP vector. The transfection efficiency was determined by counting the fraction of EGFP-positive cells (∼30%). Transfected MCF7 cells were kept in medium without serum and then induced with heregulin-1β (HRG) for 15 min. Cell lysates were prepared. pAKT, AKT, and β-tubulin were visualized after gel electrophoresis and Western blotting with specific antibodies.
To obtain more quantitative results and to achieve peptide aptamer expression in a large majority of the analyzed cells, the peptide aptamer sequences were cloned into a lentiviral expression vector (Fig. 7A ). Lentiviruses are able to infect many cell lines with a very high efficiency (Fig. 7B). Lentiviral infection also provides for a lasting expression of the aptamer constructs in the MCF-7 target cells. Viral transduction efficiencies were measured through the coexpressed green fluorescent protein (GFP) marker and 90% to 93% infected cells could be observed. The expression of the peptide aptamer constructs in the virally infected cells was also confirmed by Western blot analysis (Fig. 7C). Although the expression level of the fusion proteins with aptamer sequences was found to be lower than the level observed with the empty thioredoxin control protein, their expression persisted over long periods of time.
Efficient gene transfer and aptamer expression in 293T and MCF7 cells on lentiviral gene transduction. A. The aptamer fusion proteins were integrated into lentiviral expression vectors, pSiEW, with 5′and 3′ long terminal repeat (LTR), spleen focus-forming virus (SFFV) promoter, Flag-tagged aptamer-thioredoxin expression cassette, internal ribosome entry site (IRES), EGFP coding sequence, and woodchuck posttranscriptional regulatory element (WPRE). B. Lentiviral particles encoding thioredoxin, Ag-11, and AII-7 were obtained from transfected helper cells and used for the infection of 293T, MCF7, MCF7-neo, MCF7-her2, and SKBR3 cells. Infection efficiency was measured by EGFP expression of infected cells in a fluorescence activated cell sorter. C. 293T cells infected with the lentiviral expression vectors for thioredoxin, Ag-11, and AII-7 were lysed and the virally encoded proteins were visualized after gel electrophoresis and Western blot analysis.
We also investigated the effects of aptamer expression on AKT phosphorylation. The presence of the peptide aptamers in these cells caused an even more pronounced inhibition in the induction of AKT phosphorylation by heregulin (Fig. 8A ) than observed after transfection. This is consistent with the higher percentage of cells expressing aptamer AII-7 when compared with the transfected cells analyzed in Fig. 6. AKT phosphorylation in MCF7 cells expressing aptamer AII-7 was completely prevented. The MCF-7 control cells expressing only the thioredoxin scaffold protein were not affected in their AKT induction. We compared the effects of aptamer AII-7 on AKT activation with the activity of PKI-166 (Novartis AG, Basel, Switzerland), a tyrosine kinase inhibitor that inhibits the enzymatic activities of EGFR and ErbB2. At 1 μmol/L concentration, PKI-166, similarly to AII-7, was able to prevent AKT phosphorylation on heregulin treatment of the cells.
Peptide aptamers expressed in MCF7 cells on lentiviral infection or introduced into MCF7 cells by PTD-mediated uptake of recombinant proteins suppress heregulin-induced AKT phosphorylation. A. Lentivirally transduced MCF7 cells expressing thioredoxin, Ag-11, and AII-7 were grown in the absence of serum and induced with heregulin (HRG) for 15 min. Cell lysates were prepared and pAKT, AKT, and β-tubulin were visualized after gel electrophoresis and Western blotting with specific antibodies. Uninfected, noninduced (−HRG), and induced (+HRG) cells were used as controls. Cells were also treated with 1 μmol/L of the EGFR/ErbB2 tyrosine kinase inhibitor PKI-166 (PKI). B. MCF7 cells were incubated with 1.5 or 2 μmol/L of the purified aptamer-PTD fusion proteins thioredoxin, Ag-11, and AII-7 for 4 h and then induced with heregulin for 15 min. Nontransduced, noninduced (−HRG), and induced (+HRG) cells were used as controls. After cell lysis, the levels of pAKT, AKT, and transduced aptamers were visualized in Western blot analysis with specific antibodies. C. MCF7 cells were transduced with aptamer AII-7 (1.5 μmol/L) and induced with heregulin. Cell lysates were prepared and Western blot analysis was carried out. Specific antibodies were used to visualize AKT, p-AKT, MAPK, p-MAPK, Stat3, and p-Stat3 to detect the levels of phospho-Stat3 and phospho-p42/p44-MAPK. A Flag-specific antibody was used to detect the amount of transduced AII-7 present in the lysate.
If the peptide aptamer proteins should become useful for therapeutic applications, it will be necessary to achieve functionally relevant intracellular concentrations on protein transduction. The binding experiments described above (Figs. 4 and 5) showed that the recombinant aptamers are able to cross the plasma membrane and interact with the ErbB2 receptor within the cell. To determine if the transduced proteins can also exert inhibitory effects, MCF7 cells were transduced with 1.5 and 2 μmol/L of peptide aptamers. The cells were induced with heregulin and the extent of phosphorylated AKT was analyzed in Western blot experiments (Fig. 8B). Aptamer AII-7 was able to inhibit AKT phosphorylation in a dose-dependent manner. A slight reduction in AKT phosphorylation was also achieved with 2 μmol/L of aptamer Ag-11. The empty thioredoxin protein, used as a control, had no effect. As expected, the addition of 1 μmol/L PKI-166 also prevented AKT phosphorylation.
We carried out experiments to investigate the potential effects of the ErbB2-specific peptide aptamers on other signaling pathways induced in MCF-7 on treatment with heregulin. The kinase domain is able to phosphorylate multiple tyrosine residues in the COOH-terminal tail of the receptor. These residues then form docking sites for adaptor proteins such as growth factor receptor binding protein 2, which leads to the activation of the Ras/Raf/MAPK pathway on growth factor stimulation. The transcription factor Stat3 is found associated with the receptor and, in this case, tyrosine phosphorylation leads to the direct activation of Stat3, which then dimerizes and translocates to the nucleus. We analyzed the effects of the peptide aptamer AII-7 on the activation of these signaling molecules. MCF-7 cells were incubated with AII-7 (1.5 μmol/L) for 4 h and induced with heregulin for 15 min. Cell extracts were prepared and the amounts of p42/44-MAPK, its activated form phospho-p42/44-MAPK, Stat3, and its activated form phospho-Stat3 were compared in Western blots with specific antibodies. The phosphorylation of p42 (extracellular signal–regulated kinase 1) was only slightly affected when extracts from heregulin-induced cells treated with AII-7 and control cells were compared, whereas phosphorylation of p44 (extracellular signal–regulated kinase 2) was clearly unaffected. The induction of Stat3 phosphorylation was slightly reduced. These results indicate that aptamer AII-7 inhibits receptor function in a pathway-specific manner. It interferes only with a subset of signaling molecules and most likely does not lead to a complete loss of tyrosine kinase activity of the ErbB2 kinase domain. Our experiments show that this effect is dependent on the endogenous or exogenous provision of the peptide aptamer. Transfected cells, virally transduced cells, or cells into which the peptide aptamers had been introduced as recombinant proteins showed a similar behavior on ligand activation of the ErbB2 receptor.
EGFR- and IGF-IR–Mediated Induction of AKT Is Not Affected by Peptide Aptamer AII-7
The results presented above indicate that aptamer AII-7 inhibits AKT signaling induced by the activation of the ErbB2 receptor. We carried out additional experiments to confirm the specificity of AII-7 action with respect to the inhibition of signaling downstream of the ErbB2 receptor. Figures 2 and 3 show the preferential binding of aptamer AII-7 to the ErbB2 receptor in vitro and in vivo. In vitro GST pull-down experiments showed that the aptamer strongly binds the ErbB2 receptor, but not the EGFR. In yeast cells, the AII-7 aptamer binds preferentially to the ErbB2 receptor kinase domain and only very weakly to EGFR. In both experiments, additional sequences, Gal4-AD or GST, are present in the aptamer fusion proteins, which might have influenced the observed specificity. To further confirm the functional specificity, we investigated if the recombinant purified peptide aptamer AII-7, not containing Gal4-AD or GST sequences, affects the activation of AKT induced through IGF-IR or EGFR.
In this experiment, we induced MCF-7 cells with IGF-I (Fig. 9A ). AKT phosphorylation was low when the cells were kept in low serum overnight and could be induced after addition of IGF-I to the medium (Fig. 9A, lanes 1-3). Increasing amounts of aptamer AII-7 were added to the medium of the cells (lanes 4-6), but no effects on the induction of AKT phosphorylation were observed. This confirms the interaction studies shown in Fig. 2. The aptamer AII-7 does not interfere with AKT activation downstream of IGF-IR.
Transduction of purified peptide aptamer AII-7 has no influence on the phosphorylation of AKT activated via EGFR or IGF-IR. MCF-7 cells (A), A431 cells (B), or Renca-EGFR (C) were grown to 80% confluency and were then starved overnight. Peptide aptamers fused with a PTD were fast protein liquid chromatography purified using a histidine tag. The same batch of protein was used in Fig. 8C. The peptides were added in the medium of the cells in concentrations as indicated in each figure. After 4 h, EGF (10 ng/mL) or IGF-I (100 ng/mL) was added for 15 min as indicated. Cells were washed twice with ice-cold PBS and radioimmunoprecipitation assay extracts were prepared. Proteins were blotted and membranes were incubated with AKT, MAPK (p42/44) antibodies, or antibodies recognizing the phosphorylated form of these proteins (pAKT and pMAPK). The Flag antibody was used to detect the amount of transduced peptide aptamers in the cell lysates. A β-tubulin antibody was used to verify equal loading of the samples.
To study the effect of AII-7 on EGFR signaling, A431 cells and Renca-EGFR cells were used. A431 cells express high amounts of EGFR and low amounts of ErbB2. When A431 cells were kept in low serum overnight and the activation status of AKT was monitored (Fig. 9B, lane 1) and induced with EGF (lane 2), we found that AKT phosphorylation was not down-regulated. A431 cells are able to self-activate this signaling pathway, probably by an autocrine mechanism (Fig. 9B). In contrast, MAPK phosphorylation was clearly reduced under these conditions and could be reinduced by the provision of EGF. MAPK phosphorylation was not affected by the addition of increasing aptamer concentrations (lanes 3-5).
Because AKT phosphorylation could not be down-regulated in these cells, we also investigated the effects of AII-7 on the induction of EGFR in Renca-EGFR cells. These murine renal caricinoma cells have been stably transfected with the human EGFR gene and express high amounts of the human EGFR, whereas ErbB2 expression is very low. Starvation reduced the phosphorylation of AKT to very low levels (Fig. 9C, lane 1) and the addition of EGF to the medium resulted in AKT phosphorylation and activation (lane 2). The addition of increasing amounts of aptamer AII-7 only slightly decreased pAKT. These results confirm that IGF-I- or EGF-dependent activation is not, or only very slightly, affected by aptamer AII-7; this is in contrast to heregulin-mediated AKT activation via the ErbB2/ErbB3 heterodimer.
Aptamer AII-7 Sensitizes Chemoresistant MCF-7 (her2) Cells toward Paclitaxel Treatment
AKT activity is a well-recognized survival signal in cancer cells. Because aptamer AII-7 is able to suppress AKT activation, we tested if the sensitivity of ErbB2-expressing tumor cells toward chemotherapeutic drugs could be modulated by this peptide aptamer. It has recently been shown that the overexpression of ErbB2 in MCF-7 cells changes its sensitivity toward chemotherapeutic drugs. Stable transfection of the ErbB2 gene and its concomitant overexpression increases the resistance of MCF7 cells toward paclitaxel, doxorubicin, and 5-fluorouracil (47). These agents induce apoptosis in proliferating cells. The drug resistance in ErbB2-overexpressing cells (MCF7-her2 cells) is dependent on AKT signaling. The antiapoptotic effect of AKT dampens the induction of apoptosis. We used MCF7-her2 cells to investigate if the prevention of AKT phosphorylation by aptamer AII-7 changes their sensitivity toward paclitaxel treatment. MCF7-her2 and MCF7-neo control cells were infected with lentiviral vectors encoding the empty thioredoxin, aptamer Ag-11, and aptamer AII-7. These cells were then subjected to paclitaxel treatment and cytotoxicity was assayed. The cells were treated with increasing concentrations of paclitaxel for 5 h. The percentage of living cells was determined with a 2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt (XTT) assay 3 days later (Fig. 10 ).
Aptamer-mediated inhibition of AKT enhances the sensitivity of MCF7-her2 against Taxol treatment. MCF7-neo and MCF7-her2 cells were compared in their sensitivity toward Taxol in the absence and presence of peptide aptamer proteins. A. MCF7-neo cells were infected with lentiviruses encoding thioredoxin or the aptamer proteins Ag-11 or AII-7 (5 × 103 cells per 24-well plate). The cells were treated with paclitaxel, 0.01, 0.04, 0.4, and 4 μmol/L, for 5 h. Control cells were not treated with paclitaxel or with paclitaxel and PKI-166 (1 μmol/L). Viable cell numbers were determined 3 days after paclitaxel treatment by XTT proliferation analysis. B. MCF7-her2 cells were infected with the lentiviral vectors and treated as described in (A). C. MCF7-neo cells (5 × 103 per 24-well plate) were transduced with the aptamer-PTD fusion proteins thioredoxin, Ag-11, or AII-7 (2 μmol/L) before and during the 5-h paclitaxel treatment. Control cells were treated with PKI-166 (1 μmol/L) and paclitaxel. Viable cells numbers were determined 3 days after paclitaxel treatment by XTT proliferation analysis. D. MCF7-her2 cells (5 × 103 per 24-well plate) were transduced with the indicated aptamer-PTD fusion proteins and treated as described in (C).
In the experiments shown in Fig. 10, the untreated cells were considered 100% viable. MCF7-neo cells transduced with aptamers or thioredoxin showed the same response to paclitaxel treatment as untransduced MCF7-neo cells (Fig. 10A). Aptamer transduction has no unspecific effects on MCF7-neo cells and does not affect the sensitivity toward paclitaxel. MCF7-her2 cells were clearly more resistant to paclitaxel treatment (Fig. 10B) when compared with MCF7-neo control cells (Fig. 10A). These results confirm the previously published results (47). Thioredoxin had only a very minor effect on the viability of MCF7-her2 cells in the presence of increasing concentrations of paclitaxel. However, when Ag-11 and AII-7 were virally transduced, a clear increase in sensitivity was observed. The percent of living cells was reduced to 50% and 40%, respectively. Aptamer Ag-11 sensitizes the cells to a lesser extent than AII-7, consistent with its binding properties reported above. PKI-166 was used for comparison in the paclitaxel treatment experiments. Cytotoxicity was only modestly affected in MCF7-her2 cells. Without paclitaxel treatment, the proliferation of MCF7-her2 cells transduced with aptamers or thioredoxin was nearly identical to that of untransduced MCF7-her2 cells. The aptamers did not change the proliferative properties of the cells.
Recombinant aptamer proteins were used to determine if the induced sensitivity toward paclitaxel could also be conferred by protein transduction. For this purpose, MCF7-her2 and MCF7-neo cells were incubated with 2 μmol/L of purified peptide aptamers 2 h before and during the paclitaxel treatment (Fig. 10C and D). PKI-166 (1 μmol/L) was used as a control. Consistent with the results obtained in the experiments in which virally infected MCF7-her2 cells were used, we observed that transduced aptamer AII-7 sensitizes these cells to paclitaxel. About 37% of the MCF7-her2 cells are viable on treatment with AII-7 and 0.4 μmol/L paclitaxel, compared with 70% viable MCF7-her2 cells treated with paclitaxel alone. It did not affect proliferation and paclitaxel sensitivity of MCF7-neo cells. Aptamer Ag-11 again shows moderate effects on MCF7-her2 cells (55% viable MCF7-her2 cells treated with 0.4 μmol/L paclitaxel) and the percentage of viable cells after incubation with empty thioredoxin (65% of viable MCF7-her2 cells treated with 0.4 μmol/L paclitaxel) was similar to that of MCF7-her2 cells incubated without recombinant proteins. The incubation of PKI-166 also sensitized the cells toward paclitaxel treatment (50% viable MCF7-her2 cells treated with 0.4 μmol/L paclitaxel) and had a similar effects as aptamer Ag-11.
The reduction of AKT phosphorylation by aptamer AII-7 sensitizes the relatively more resistant MCF7-her2 cells to paclitaxel treatment. This observation was made after infection of the cells with a viral expression vector encoding aptamer AII-7 and on protein transduction of recombinant AII-7 protein. The weaker sensitizing effect was exerted by aptamer Ag-11. The effects caused by the aptamers are most likely mediated via relatively subtle influences on the function of the ErbB2 receptor; no consequences on the growth of MCF7-her2 cells were observed. PKI-166 was able to sensitize the cells similarly to Ag-11.
Discussion
Breast cancer patients have strongly benefited from the development of targeted drugs. The overexpression of the ErbB2 receptor was realized more than 20 years ago and has subsequently been systematically exploited to improve therapy. The introduction of Herceptin showed the usefulness of the ErbB2 receptor as a drug target and the application of recombinant proteins as drugs. Although Herceptin has important effects on disease-free and overall survival, there is still room for improvement. Therapeutic improvements can be achieved through other ErbB2-directed drugs with distinct mechanisms of action or the combination of Herceptin with chemotherapeutic agents. Targeting individual signaling aspects of ErbB2 also seems to be a promising strategy. Examples for the combination of targeting strategies of the closely related EGFR have been described. The monoclonal antibody Erbitux (cetuximab) and the tyrosine kinase inhibitors Iressa (gefitinib) and Tarceva (erlotinib) showed synergistic effects in various cancer cells lines (48).
Up to now, only the ATP binding pocket in the tyrosine kinase domain of the ErbB2 receptor has been targeted to inhibit its intracellular function. This prevents the phosphorylation of multiple tyrosine residues in the COOH-terminal tail of the receptor and therefore affects multiple signaling pathways at once. This, however, could be one of the reasons for the side effects observed on usage of such inhibitors. In the present study, we show that not only the enzymatic activity but also the individual signaling aspects of the ErbB2 receptor can be targeted to influence tumor cell properties. We employed a peptide aptamer that interacts with the intracellular domain of the ErbB2 receptor but does not involve the ATP binding site directly. The peptide aptamer AII-7 recognizes a region between amino acids 853 and 987, which lies outside the ATP binding site (amino acids 726-734). Aptamer Ag-11 was identified using amino acids 678 to 986 as bait. The exact binding positions for these aptamers have not been localized yet.
The isolated aptamers, initially selected in the yeast two-hybrid system, were found to retain their specificity for ErbB2 in other interaction assays (e.g., GST pull-down experiments and coimmunoprecipitation analysis). No cross-reactivity with the highly related EGFR or IGF-IR tyrosine kinase domain was detected in these assays. Specificity was confirmed by showing that AKT or MAPK signaling was unaffected if activated through these receptors. Although no binding constants have been established, these properties indicate high affinities. Aptamer AII-7 interacts stronger with the ErbB2 receptor when compared with aptamer Ag-11. To further develop these aptamers, we might have to further increase the binding affinity, which could be achieved through mutational analysis. For this purpose, the evaluation of the aptamer structure in conjunction with the binding site on the receptor will be important.
In our coimmunoprecipitation experiments, we evaluated interactions between the ErbB2 receptor and transduced peptide aptamers or by mixing cell lysates with purified peptide aptamers. The interaction with the receptor showed that we are able to refold recombinantly expressed peptide aptamers correctly. Transduction procedures and subsequent analysis of the localization of the peptide aptamers by confocal microscopy confirmed this conclusion. These experiments motivated us to pursue the functional evaluation of the aptamers and to investigate their intracellular effects. The efficiency of protein transduction is most likely a limiting factor if the application of peptide aptamers as protein therapeutics is being considered. The efficacy of protein transduction can be enhanced by chloroquine, which disrupts intracellular vesicles and thereby releases the PTD-fused proteins (43). Chloroquine was used in our protein transduction experiments; however, due to its toxicity, use might be restricted to cell culture experiments. The use of a HA2 peptide sequence has recently been proposed to circumvent this problem. This peptide was combined with a PTD and enhanced the efficiency of delivery (39, 49). HA2 peptides are able to form pores in vesicular structures at low pH and release enclosed proteins into the cytoplasm of the cell. The principle has been shown in cell culture (50), but its effective use still has to be confirmed in animal models.
In addition to the transduction of purified proteins, virally based gene transfer methods were used to evaluate the aptamer function. Both methods showed that aptamer AII-7 reduced the induction of AKT phosphorylation after ErbB2 activation in MCF7 cells. Although aptamer AII-7 is not binding to the active site, the aptamer interaction with the kinase domain interferes with the activation of a kinase substrate. The phosphatidylinositol 3-kinase-AKT pathway activates an antiapoptotic response in the cell (reviewed in refs. 15, 51) and its activation seems to be responsible for a drug resistance phenotype, which can be reversed by phosphatidylinositol 3-kinase inhibitors (52, 53). AKT activation can either be mediated by constitutive active mutations of phosphatidylinositol 3-kinase or AKT or by components of upstream signaling like the ErbB2/ErbB3 heterodimer. Therefore, the inhibition of this pathway, directly at the receptor level, could improve the efficacy of chemotherapy and reduce side effects.
MCF7-her2 cells show increased resistance toward chemotherapeutic agents such as paclitaxel, doxorubicin, 5-fluorouracil, etoposide, and camptothecin compared with the parental MCF7 cells (47). The correlation of ErbB2 expression and chemoresistance has also been observed in other breast and ovarian cancer cell lines (54, 55). Chemoresistance has also been correlated with the overexpression of other receptor tyrosine kinases (e.g., platelet-derived growth factor receptor) and could be prevented through inhibition of the tyrosine kinase activity (56). This enhanced drug resistance is dependent on the activation of AKT and can be blocked through dominant negative AKT variants or phosphatidylinositol 3-kinase inhibitors (47). Inhibition of this pathway sensitizes these cells for apoptosis induced by the chemotherapeutic agents. The aptamer-mediated AKT inhibition has been used to influence the efficacy of chemotherapy of ErbB2-overexpressing cancer cells and the regular chemosensitivity of these cells could be restored (Fig. 11 ). The presence of the aptamers did not cause other measurable effects on cell proliferation. This was confirmed by XTT assays and was also shown by the fact that the Stat3 and MAPK pathways are only moderately affected after aptamer treatment. Proliferation of cancer cells is important if chemotherapeutic agents are being used because they are only effective on dividing cells. This idea is supported by the finding that the inhibitor PKI-166 and aptamer AII-7 showed comparable inhibitory effects on the induction of AKT phosphorylation, but, interestingly, the sensitizing effect of the aptamer in the cytotoxicity assays was clearly more pronounced. PKI-166 inhibits the kinase domain more effectively, leading to the down-regulation of multiple pathways, including those involved in proliferation. These results suggest that the combination of the specific aptamer-mediated inhibition of AKT signaling and chemotherapy might become useful.
Model for the inhibition of ErbB2 signaling by peptide aptamers. ErbB2 interacting peptide aptamers isolated by yeast two-hybrid screening analysis were introduced into cancer cells by infection with a lentiviral vector (A) or by direct transduction of a recombinantly produced aptamer protein (B). Within the cell, peptide aptamer AII-7 was able to interact with the ErbB2 receptor, inhibiting selected signaling functions without binding to the ATP-binding site. Aptamer AII-7 binding to ErbB2 leads to a decrease in AKT phosphorylation on heregulin induction of the receptor dimer and increases the sensitivity of MCF7-her2 cells toward paclitaxel. Our results show that a treatment targeting the ErbB2 receptor with aptamers in combination with chemotherapy might become useful.
Although the results are encouraging and identify peptide aptamers as potential therapeutics, many steps have to be optimized before they can be applied to complement or compete with established drugs. Production, administration, stability, and other in vivo properties have to be investigated and optimized. The aptamer protein, embedded in its scaffold, has a mass of ∼20 kDa. This is comparable to camel antibodies called nanobodies, which show excellent tumor-penetrating properties (57). These properties are not altered if an albumin-binding peptide is added to the compounds. However, the use of the albumin-binding peptide strongly reduces the clearance of the nanobodies and results in a prolonged blood circulation. The pharmacokinetic variables of the protein might also be affected by the addition of a PTD, a question that has not yet been addressed. The use of PTD-conjugated proteins in vivo has shown promising leads. After i.v. injection, the proteins can be found in every tissue including the brain (58). These results illustrate that it might be possible to deliver therapeutic proteins to a variety of organs and cell types within an organism.
If delivery of protein therapeutics can be targeted to specific cell types via surface receptors, side effects of such drugs might be further minimized (59). Additional targeting domains in such recombinant proteins might improve their properties (60, 61). The same is true for albumin-binding sequences and HA2 peptides. Our results show that the combination of peptide aptamers with PTDs allows the interference with specific signaling events and that it is thus possible to target crucial protein-protein interactions in cancer cells. Further in vivo experiments will have to complement our results obtained with cell culture models.
Materials and Methods
Cell Lines and Cell Culture
SKBR3 breast cancer cells and A431 epidermoid carcinoma cells were maintained in DMEM containing 10% heat-inactivated FCS and 4 mmol/L glutamine. SKBr3-SF-thioredoxin (Trx), SKBr3-SF-Ag-11, and SKBr3-SF-AII-7 cells were generated by infection of lentiviral particles containing pSiEW-Flag-Trx or pSiEW-Flag-aptamer vectors. Cells were maintained in DMEM containing 10% heat-inactivated FCS and 4 mmol/L glutamine. Renca EGFR cells were maintained in RPMI containing 10% heat-inactivated FCS, 4 mmol/L glutamine, and 0.5 mg/mL G418. NIH#3.7 cells were maintained in DMEM containing 10% heat-inactivated FCS, 4 mmol/L glutamine, and 0.5 mg/mL G418. MCF7 cells were maintained in DMEM containing 10% heat-inactivated FCS and 4 mmol/L glutamine. MCF7-her2 and MCF7-neo cells were maintained in DMEM containing 10% heat-inactivated FCS, 4 mmol/L glutamine, and 0.5 mg/mL G418. MCF7-SF-Trx, MCF7-SF-Ag-11, MCF7-SF-AII-7, MCF7-her2-SF-Trx, MCF7-her2-SF-Ag-11, MCF7-her2-SF-AII-7, MCF7-neo-SF-Trx, MCF7-neo-SF-Ag-11, and MCF7-neo-SF-AII-7 cells were generated by infection of lentiviral particles containing pSiEW-Flag-Trx or pSiEW-Flag-aptamer vectors. Cells were maintained in DMEM containing 10% heat-inactivated FCS, 4 mmol/L glutamine, and 0.5 mg/mL G418.
Peptide Aptamer Screening
The screening was done in the yeast strain KF1 (MATα Trp-901 Leu2-3112 His3-200 gal4Δ gal80Δ LYS2::GAL1-HIS3 GAL2ADE2 MET2::GAL7-lacZ SPAL10-URA3; ref. 31). As baits, parts of the ErbB2 intracellular domain (for cytoplasmic domain, amino acids 678-1,254; for KDg, amino acids 678-987; for KDI, amino acids 678-852; and for KDII, amino acids 853-987) were fused to the GAL4-DNA binding domain into the pPC-97 vector. Screenings were done with a randomized 12-mer and 20-mer peptide aptamer library and selection procedures were done as previously described (31).
GST Pull-Down Assay
For GST pull-down experiments, aptamers were inserted in the pGex 4T3 vector (Amersham Biosciences, Freiburg, Germany) and expressed in Escherichia coli Bl21 (DE3) lysS (Stratagene, Amsterdam, the Netherlands) cells as a GST fusion protein. Aptamer expression was induced with isopropyl-1-thio-β-d-galactopyranoside (1 mmol/L) for 2 h at 30°C. Proteins were purified and used in the pull-down experiment as previously described (32).
Bacterial Expression and Purification of PTD Peptide Aptamers
For protein transduction experiments, a PTD of nine arginine residues was fused to the COOH terminus of the peptide aptamers, which were inserted into the pET30a+ vector (Novagen, Schwalbach, Germany). Aptamer expression was induced with isopropyl-1-thio-β-d-galactopyranoside (1 mmol/L) for 4 h at 30°C. Proteins were purified under denaturing conditions as previously published (62).
In vitro and In vivo Coimmunoprecipitation Experiments
For detection of the peptide aptamer interaction with full-length ErbB2 in in vitro coimmunoprecipitation experiments, SKBR3 cell lysates were incubated with recombinantly expressed peptide aptamers (0.5 μmol/L) for 1 h at 4°C. Afterwards, ErbB2 antibody (Santa Cruz Biotechnology, Heidelberg, Germany) was added and incubated for 1 h at 4°C. Immunocomplexes were collected with Protein-A DynaBeads (Invitrogen, Karlsruhe, Germany) and washed with PBS. Bound proteins were eluated by boiling in sample buffer and subjected to Western blot analyses. For detection of protein transduction through in vivo coimmunoprecipitation experiments, SKBr3 cells were incubated with recombinantly expressed peptide aptamers (2 μmol/L) for 4 h at 37°C. Cells were solubilized with radioimmunoprecipitation assay (RIPA) lysis buffer [50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 1% NP40, 0.5% sodium desoxcholate, 1 mmol/L EDTA, protease inhibitors] and incubated on ice for 12 min. Lysates were clarified by centrifugation at 16,000 × g for 10 min and incubated with anti-ErbB2 antibody (Santa Cruz Biotechnology) for 1 h. Immunocomplexes were collected as described above and subjected to Western blot analyses.
Western Blot Analyses
Cells were solubilized in Triton extraction buffer [50 mmol/L Tris (pH 7.5), 5 mmol/L EGTA, 150 mmol/L NaCl, 1% Triton X-100, protease inhibitors] or RIPA lysis buffer [50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 1% NP40, 0.5% sodium desoxcholate, 1 mmol/L EDTA, protease inhibitors] and incubated on ice for 10 min. Lysates were clarified by centrifugation at 16,000 × g for 10 min, subjected to SDS-PAGE, and blotted on polyvinylidene difluoride membranes. After blocking in 5% TBS with Tween 20 [50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 0.05% Tween 20], membranes were probed with specific antibodies and proteins were visualized with peroxidase-coupled secondary antibodies using the enhanced chemiluminescence system (Amersham Biosciences). Antibodies were used to detect ErbB2, EGFR, and IGF-IR (Santa Cruz Biotechnology), peptide aptamers (anti Flag) and β-tubulin (Sigma-Aldrich, Hannover, Germany), and AKT, pAKT, Stat3, pStat3, p42/44-MAPK, and phospho-p42/p44-MAPK (Cell Signaling, Frankfurt, Germany).
Immunofluorescence Imaging
Cells were grown on coverslips and protein transduced (1 μmol/L) for 4 h at 37°C. Afterwards, untransduced proteins were removed from the cell surface through acid wash (0.2 mmol/L acidic acid for 5 min at 4°C), fixed with 95% methanol, and permeabilized with 0.1% Tween 20 in PBS. Cells were washed, blocked (0.5% cold fish gelatin, 0.1% ovalbumin in PBS), and incubated with primary antibodies at 4°C overnight. After intensive washing, incubation with fluorescence-labeled antibodies (Molecular Probes, Karlsruhe, Germany) was done for 1 h at room temperature in the dark. The stained proteins were visualized with a confocal laser scanning microscope (Leica, Bensheim, Germany).
Transfection of Eukaryotic Cells
Transfections were done with LipofectAMINE 2000 (Life Technologies, Karlsruhe, Germany) according to the manufacturer's protocol. For expression of peptide aptamers in eukaryotic cells, the thioredoxin cassette was subcloned into the eukaryotic expression vector pRc/CMV (Invitrogen) and fused 3′ to the herpes simplex virus VP22 gene (pRC/CMV-VP22-Trx).
Encoding of Peptide Aptamers in Lentiviral Particles and Lentiviral Transduction in Eukaryotic Cells
For stable expression of peptide aptamers in eukaryotic cells, the Flag-Trx cassette was subcloned into the lentiviral expression vector pSiEW. Peptide aptamer expression is driven by a spleen focus-forming virus promoter, which also drives the expression of enhanced GFP (EGFP) through an internal ribosome entry site. Lentiviral particles were produced by triple transfection of pSiEW-Flag-Trx, R.8.1, and pM2 in 293T cells. Viral particles were collected in 5-mL culture supernatants 2 days after triple transfection. Virus titers were estimated by titration of 1:1,000, 1:100, and 1:10 culture supernatants of 293T cells. The percentage of infected cells was measured 3 days after infection by fluorescence-activated cell sorting analyses.
AKT Phosphorylation Assay
For AKT phosphorylation analyses, after lentiviral transduction, 1 × 106 MCF7 cells expressing thioredoxin or aptamers were plated and starved overnight. Specific ErbB2/ErbB3 heterodimer activation was induced by 5 ng/mL heregulin-1β for 15 min and stopped with ice-cold PBS. MCF7 cells were lysed with radioimmunoprecipitation assay lysis buffer [50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 1% NP40, 0.5% sodium desoxcholate, 1 mmol/L EDTA, protease inhibitors] and incubated on ice for 10 min. Lysates were clarified by centrifugation at 16,000 × g for 10 min and subjected to Western blot analysis. For protein transduction experiments, 1 × 106 MCF7 cells were incubated with 1.5 and 2 μmol/L of purified peptide aptamers, as well as with 100 μmol/L of chloroquine, 4 h before heregulin-1β induction.
Taxol Treatment and Cytotoxicity Assays
Cytotoxicity assays were done by plating 5 × 103 MCF7-her2 or MCF7-neo cells in 24-well plates. After lentiviral transduction, MCF7-her2 or MCF7-neo cells expressing thioredoxin or aptamers were incubated with increasing paclitaxel concentrations (0.01, 0.04, 0.4, and 4 μmol/L) for 5 h in charcoal-stripped DMEM. For protein transduction, cells were incubated with 2 μmol/L purified PTD-aptamers 4 h before and during paclitaxel treatment. Cell viability was measured 3 days after paclitaxel treatment with the XTT-based proliferation kit II (Roche Molecular Biochemicals, Mannheim, Germany).
Proliferation and Viability Assays
Relative viable cell numbers were quantified using the XTT-based proliferation kit II according to the manufacturer's protocol, which assesses cell viability via bioreduction of a tetrazolium compound by measuring absorbance at 490 nm in a 96-well plate reader (Roche Molecular Biochemicals).
Acknowledgments
We thank K. Butz and F. Hoppe-Seyler for providing the yeast KF1 strain; Z. Fan for providing the MCF7-neo/her2 cells; Novartis for providing the tyrosine kinase inhibitor PKI-166; M. Grez, S. Stein, and H. Kunkel for providing the lentiviral expression vectors and for helpful support in the handling of viral particles; W. Wels, S. Hyland, B. Daelken, and C. Hartmann for support in bacterial expression technologies and ErbB2 functional assays; K. Nagel-Wolfrum, B. Brill, C. Shemanko, N. Palomino-Castro, C. Bähr, B. Sanchez De Juan, R. Pick, J. Tiefenbach, C. Koenigs, and M. Humbert for helpful discussions; and N. Delis for technical assistance.
Footnotes
↵1 Baehr and Groner, unpublished results.
Grant support: Wilhelm-Sander-Stiftung, Förderantrag Nr. 2000.021.1; Deutsche Krebshilfe, Dr. Mildred Scheel Stiftung für Krebsforschung, Projekt Nr. 10-1626-Gr 2; and Deutsche Forschungsgemeinschaft Gr 536/4-1 (B. Groner).
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 U.S.C. Section 1734 solely to indicate this fact.
- Accepted October 5, 2006.
- Received February 15, 2006.
- Revision received September 12, 2006.
- American Association for Cancer Research