Aptamer to ErbB-2/HER2 enhances degradation of the target and inhibits tumorigenic growth

Abstract

Aptamers, oligonucleotides able to avidly bind cellular targets, are emerging as promising therapeutic agents, analogous to monoclonal antibodies. We selected from a DNA library an aptamer specifically recognizing human epidermal growth factor receptor 2 (ErbB-2/HER2), a receptor tyrosine kinase, which is overexpressed in a variety of human cancers, including breast and gastric tumors. Treatment of human gastric cancer cells with a trimeric version (42 nucleotides) of the selected aptamer (14 nucleotides) resulted in reduced cell growth in vitro, but a monomeric version was ineffective. Likewise, when treated with the trimeric aptamer, animals bearing tumor xenografts of human gastric origin reflected reduced rates of tumor growth. The antitumor effect of the aptamer was nearly twofold stronger than that of a monoclonal anti–ErbB-2/HER2 antibody. Consistent with aptamer-induced intracellular degradation of ErbB-2/HER2, incubation of gastric cancer cells with the trimeric aptamer promoted translocation of ErbB-2/HER2 from the cell surface to cytoplasmic puncta. This translocation was associated with a lysosomal hydrolase-dependent clearance of the ErbB-2/HER2 protein from cell extracts. We conclude that targeting ErbB-2/HER2 with DNA aptamers might retard the tumorigenic growth of gastric cancer by means of accelerating lysosomal degradation of the oncoprotein. This work exemplifies the potential pharmacological utility of aptamers directed at cell surface proteins, and it highlights an endocytosis-mediated mechanism of tumor inhibition.

The epidermal growth factor related protein (ErbB) family of receptor tyrosine kinases plays an important role in epitheliogenesis and, accordingly, serves as a major therapeutic target in several cancers. The family comprises four transmembrane receptors and 11 ligands that induce homodimerization or heterodimerization upon binding to the respective receptor (1). ErbB-1 (also called the epidermal growth factor receptor; EGFR) and ErbB-4 share some ligands, whereas no similar ligand is so far known for ErbB-2. Overexpression and mutations of ErbB family members lead to a multitude of malignancies. To date, synthetic tyrosine kinase inhibitors (e.g., Erlotinib and Gefitinib), as well as monoclonal antibodies (mAbs; e.g., Cetuximab and Trastuzumab), have been developed to inhibit pathological signaling or recruit the immune system to cancer cells (2). Aptamers might represent an alternative therapeutic modality. These molecules are small, single-stranded DNA or RNA molecules (3). RNA aptamers were described for the first time in 1990 by two laboratories (4, 5). Since then, aptamers against a multitude of different organic and inorganic, small and macromolecular, targets were developed. In addition, high-affinity aptamer binding ranging from picomolar to low nanomolar concentrations have been documented (6). Aptamers are selected in an evolutionary process called systematic evolution of ligands by exponential enrichment (SELEX). A DNA or RNA library containing single-stranded random sequences, flanked by two primer-binding regions, is allowed to bind to a specific target. In several selection rounds, binders are amplified and nonspecific binders are removed in a partitioning step. Selected sequences can be modified after selection to improve their stability in different chemical environments (e.g., serum). Different therapeutic aptamers, which are antiviral, anticoagulation, antiinflammatory, or antiangiogenic, are already in clinical trials (7). The first clinically approved aptamer is an RNA molecule, called Macugen, which effectively inhibits macular degeneration. In addition to pharmacological applications, aptamers can be exploited for transducing a binding event into a signal. As a consequence, aptamers have been adapted to a variety of bioanalytical methods (8).

Several anticancer aptamers have been developed, including an aptamer against Nucleolin, which led to phase 2 clinical trials (9). So far, several anti-ErbB–specific aptamers have been developed (10–14). They generally show high affinity and specificity to their targets and, in the case of ErbB-1– and ErbB-3–specific aptamers, they also inhibit the proliferation of cultured cancer cells (10, 14). Notably, an aptamer against ErbB-1/EGFR was able to inhibit tumor growth in a mouse model (12), and ErbB-2–specific aptamers were used to deliver siRNAs targeting B-cell lymphoma 2 (Bcl-2) (15). In our present work, we demonstrate specific binding of several aptamers to ErbB-2. The SELEX selection focused on specific binders, which were randomly mutated (using PCR) and represented shorter versions than the random sequences derived from the DNA library. To improve efficacy of specific ErbB-2 binders and potentially enable cross-linking of ErbB-2 molecule on the surface of cancer cells, we applied trimeric versions (42 nucleotides) of the original sequence (14 nucleotides). Multimerization enhanced binding and accelerated degradation of ErbB-2 in lysosomes, which led to inhibition of growth of a cultured human tumor cell line, and a reduction of tumor mass in an animal model.

Results

Selected Monomeric Aptamers Display Specificity to ErbB-2.

ErbB-2–specific aptamers were selected in five SELEX selection rounds from a single-stranded DNA library (see scheme in Fig. S1). Each aptamer of the library comprises two constant primer-binding regions and a 50-nt–long random insert. A fluorescein-labeled PCR primer enabled ligand quantification following each selection round. Antibody-immobilized, native ErbB-2 protein from human N87 gastric cancer cells, which overexpress the oncogenic protein, served as a selection target. To test target specificity of selected aptamers, we applied a series of assays, including titration and competition tests (Fig. S2 A and B), as well as a double-immunoblot analysis. For this latter assay, N87 cells were lysed, extracted proteins were resolved (using gel electrophoresis), and blotted to a nitrocellulose membrane. Three selected, biotinylated aptamers, along with a control biotinylated primer from the SELEX rounds, were incubated with the membrane. To exclude nonspecific interactions, we introduced a second blotting step, which electrophoretically transferred aptamer–ErbB-2 complexes to a second membrane. This second membrane was incubated with streptavidin-HRP before developing a chemiluminescence signal (Fig. 1A). The results presented in Fig. 1A clearly show that all selected aptamers (B212, 2-2, and 2-1) were able to specifically bind with ErbB-2. In contrast, the primer control (PR) failed to bind with ErbB-2.

Fig. 1.

Selected aptamers show specificity to ErbB-2. (A) Extracts of N87 cells were resolved by using gel electrophoresis, blotted onto a membrane, and after a blocking step, the blot was incubated with the indicated biotinylated aptamers, along with the biotinylated primer (PR; each at 100 nM). Bound oligonucleotides were reblotted to a second membrane and detected with HRP-labeled streptavidin. (B) Biotinylated aptamers 2-2 and the control primer (PR) were immobilized on streptavidin beads and used to pull down ErbB-2 from extracts of N87 cells (500 μg/mL). For control, we used extracts of A431 cells. Pull downs were resolved by using gel electrophoresis, blotting to a nitrocellulose membrane and ErbB-2 detection with a rabbit polyclonal antiserum and an HRP-labeled secondary antibody.

In a second attempt to examine specificity, a single aptamer, denoted 2-2 (14 nucleotides), was tested in a pull-down experiment. The biotinylated aptamer, along with the biotinylated control primer, PR, were incubated with whole extracts derived from either ErbB-2–overexpressing N87 cells, or from ErbB-1–overexpressing A431 cells. After three washing steps, streptavidin magnetic beads were allowed to bind with ErbB-2–aptamer complexes. Pulled-down complexes were resolved by gel electrophoresis and blotted onto a nitrocellulose membrane. The membrane was blocked, and the aptamer-tagged ErbB-2 was detected by using a kinase domain-specific rabbit polyclonal antiserum, along with a secondary HRP-labeled antibody. This experiment demonstrated high ErbB-2 specificity of aptamer 2-2, relative to the control primer (Fig. 1B, Left). In addition, we showed that aptamer 2-2 cannot bind to the overexpresed ErbB-1 of A431 cells (Fig. 1B, Right). In summary, the aptamer we selected (2-2) showed high specificity to human ErbB-2. No other protein, including the closest family member of ErbB-2, namely ErbB-1, could similarly bind with the selected aptamer.

Trimeric Version of the Selected Aptamer Is Endowed with Improved Binding to ErbB-2.

To enhance the avidity of the selected 2-2 aptamer (14 nucleotides), and also enable cross-linking of ErbB-2 molecules expressed on tumor cells, we designed a trimeric version (42 nucleotides). Superior binding of the trimeric aptamer was first verified by a dot blot assay. Monomeric and trimeric biotinylated aptamers were allowed to interact with spotted whole extracts of ErbB-2–overexpressing N87 cells or extracts derived from ErbB-1–overexpressing A431 cells. Next, bound aptamers were exposed to streptavidin-HRP and chemiluminescence was measured. As shown in Fig. 2A, strong binding of monomeric aptamer 2-2 was detected only at the higher amount (20 μg) but not at the 10-fold lower amount of spotted cell lysate. However, the trimeric version of the aptamer clearly displayed specific binding with extracts derived from N87 cells at both high and low amounts of cell lysates. The relatively low level of ErbB-2 expressed by A431 cells likely underlies the minor signal we observed with these cells. Flow cytometry analyses similarly supported enhanced binding by the trimeric aptamer. Biotin-labeled trimeric aptamers 2-2(t) and PR(t) were incubated with N87 cells, before treatment with phycoerythrin streptavidin and fluorescence-based cell analysis. As indicated by Fig. 2B, the trimeric aptamer 2-2(t) was able to efficiently bind with ErbB-2–overexpressing cells, whereas the controls demonstrated no binding.

Fig. 2.

A trimeric aptamer displays improved ErbB-2 binding. (A) Whole extracts were prepared from ErbB-2–overexpressing N87 cells and ErbB-1–overexpressing A431 cells. Samples (20 and 2 μg for N87 and 2 μg for A431) were spotted onto a nitrocellulose membrane. Biotinylated trimeric [2-2(t)] or monomeric aptamers, each at 100 nM, were applied in duplicates. HRP-labeled streptavidin was used for detection. (B) N87 cells were treated with trypsin, washed, and incubated at 4 °C for 30 min with aptamers 2-2(t) or PR(t) in saline containing 0.1% albumin. Thereafter, cells were washed and phycoerythrin-conjugated streptavidin was added at 4 °C for 30 min. Fluorescence was measured by using FACSORT cytometer. (C) The indicated biotinylated aptamers (1 μM) were conjugated to magnetic beads and incubated with extracts (500 μg/mL) from N87 or A431 cells. Extensively washed precipitates were separated by using gel electrophoresis and then blotted onto a nitrocellulose membrane. The membrane was blocked and probed with an ErbB-2–specific polyclonal antiserum, followed by an HRP-labeled secondary antibody.

Next, we used a pull-down assay to directly compare the trimeric and monomeric versions of the selected aptamer. Biotinylated aptamers 2-2(t) and PR(t) were allowed to bind extracts of either N87 or A431 cells. Streptavidin magnetic beads were applied and the aptamer–receptor complexes were first pulled down and then resolved by using gel electrophoresis and immunoblotting. Unlike the control primer, aptamer 2-2(t) could efficiently precipitate large amounts of ErbB-2 from extracts of N87 cells, and much smaller receptor amounts were precipitated from A431 cells (Fig. 2C). Thus, in comparison with the monomeric aptamer, the trimeric version of aptamer 2-2 displayed improved binding, while preserving high specificity to ErbB-2. Conceivably, the extended length of aptamer 2-2(t) does not interfere with binding, implying that key structural elements needed for target binding remain unchanged after multimerization.

Immunofluorescence and Biochemical Analyses Reveal Aptamer-Induced Internalization and Lysosomal Degradation of ErbB-2.

To follow the fate of aptamer-bound ErbB-2 molecules, we incubated N87 cells for 3 d with the trimeric aptamer 2-2(t) (100 nM) and then used immunofluorescence to localize ErbB-2. Unlike untreated cells, which displayed the characteristic localization of ErbB-2 at the cell surface, aptamer-treated cells displayed no membranal signal (Fig. 3). Instead, multiple puncta of ErbB-2, which avoided the nucleus, were displayed by aptamer-treated cells. This observation proposed that aptamer-mediated cross-linking of ErbB-2 molecules targets the oncogenic receptor to the well-studied pathway of endocytosis, culminating in delivery to lysosomes for degradation (reviewed in ref. 16). To examine inducible degradation, N87 cells were incubated with the trimeric aptamer 2-2(t), or with the trimeric control primer PR(t), extracted at 24-h intervals, and resolved by immunoblotting. This analysis clearly showed that ErbB-2 underwent extensive degradation following treatment with the trimeric 2-2 aptamer, whereas the control primer exerted no impact on protein stability (Fig. 4A). As shown before, ErbB-2 undergoes degradation in lysosomes, once perturbed by growth factors or monoclonal antibodies (17, 18). Therefore, we tested the effect of a specific inhibitor of lysosomal degradation, namely chloroquine, on ErbB-2 degradation. As shown in Fig. 4B, the drug almost completely inhibited the effect of aptamer 2-2(t) on ErbB-2 degradation. Importantly, we could show that the monomeric aptamer 2-2 exerted no significant effect on ErbB-2 degradation (Fig. 4C). Hence, on the basis of the presented lines of evidence, we concluded that the trimeric aptamer 2-2(t) effectively targets ErbB-2 to intracellular degradation, likely through a mechanism that involves receptor cross-linking.

Fig. 3.

The trimeric aptamer promotes translocation of surface ErbB-2 into intracellular puncta. N87 cells growing on fibronectin were treated with the trimeric aptamer 2-2 (100 nM) for 3 d. Immunofluorescence localization of ErbB-2 used a rabbit antibody, as well as a secondary, fluorescently labeled antibody. DAPI was used to visualize nuclei. (Scale bars: 5 μm.)

Fig. 4.

The trimeric aptamer 2-2 enhances lysosomal degradation of ErbB-2. (A) N87 cells (10⁵) were plated on 24-well plates and incubated for 1–3 d with the trimeric aptamer 2-2(t) (10 μM), or with the trimeric primer PR(t). Thereafter, cells were extracted and subjected to electrophoresis and immunoblotting with an ErbB-2–specific polyclonal antiserum. Antibodies to ERK were used to test equal loading, and HRP-labeled secondary antibodies were used for detection. (B) Cells were treated with aptamers as in A, except that chloroquine (Chl.; 0.01 μM) was coincubated with the aptamer. (C) N87 cells were plated as in A and incubated for 3 d with the monomeric or trimeric aptamers 2-2 and 2-2(t).

Trimeric Aptamer Inhibits Growth of Gastric Cancer Cells both in Vitro and in Vivo.

While at the cell surface, ErbB-2 delivers strong oncogenic signals, but this phenomenon does not occur when ErbB-2 is localized in intracellular compartments (19). Hence, aptamer-induced internalization is expected to reduce the ability of an overexpressed ErbB-2 to deliver mitogenic and oncogenic signals in tumor cells. To test this prediction, we treated N87 cells (10³ per well) for 7 d with either the trimeric aptamer 2-2(t) or with the trimeric primer PR(t). As demonstrated in Fig. 5, only aptamer 2-2(t) could efficiently inhibit cell proliferation, whereas control untreated cells, as well as cells treated with the monomeric 2-2 aptamer, and cells treated with either the monomeric or trimeric versions of the control primer, displayed no significant effect on proliferation.

Fig. 5.

The trimeric aptamer inhibits growth of gastric cancer cells both in vitro and in animals. (A) N87 cells (10³ cells per well) were plated on 96-well plates, and 24 h later, monomeric (2-2 and PR) or trimeric aptamers [2-2(t) and PR(t)] at 10 μM were added. Subsequently, cells were incubated at 37 °C for 7 d, and the medium was refreshed every other day. Cell proliferation was determined by using a commercial kit. (B) CD-1 nude mice were inoculated with 5 × 10⁶ N87 cells. Once tumors became palpable, mice were treated i.p., once per week (for eight times) with the ErbB-2–specific mAb431 (160 μg per week), the trimeric, ErbB-2–specific aptamer 2–2(t) (40 μg per week) or with PR(t) (40 μg per week). Untreated tumor-bearing mice were similarly monitored; their rate of tumor growth was statistically indistinguishable from the rate displayed by the PR(t)-treated group. Shown are the results of one of three experiments. Data represent mean ± SEM of seven mice per group.

As a prelude for testing the effects of the trimeric aptamer in animals bearing N87 tumors, we determined the stability of aptamer 2-2(t) in mouse serum. By using PCR and measuring fluorescence of aptamer-containing serum samples, we concluded that 2-2(t) is stable for at least 48 h, before emergence of the first signs of degradation (Fig. S3). On the basis of these results, we implanted N87 cells in CD-1 immunocompromised mice and started treatments once tumors became palpable. Mice were left untreated or they were treated weekly for eight times with the control PR(t) or aptamer 2-2(t). A third group was treated with an ErbB-2–specific monoclonal antibody, mAb431, because this antibody can partially inhibit tumorigenic growth of N87 cells in mice (20). In line with this approach, three independent experiments indicated that mAb431 significantly reduced tumor growth in comparison with untreated animals and also relative to mice treated with the PR(t) control oligonucleotide (P < 0.01; Fig. 5B). Importantly, the trimeric 2-2 aptamer consistently displayed, in all three experiments, greater long-term antitumor effects than the antibody we tested. Moreover, the inhibitory effects exerted by mAb431 and aptamer 2-2(t) were initially quite similar, as long as we kept injecting the respective agents, but they separated upon cessation of treatment (approximately day 70). This long-term difference might be attributed to either pharmacokinetics or mechanisms of action.

In summary, a consistent picture emerged from the studies performed in vitro and in animals. Accordingly, the trimeric aptamer (42 nucleotides) we selected on the basis of specificity to ErbB-2 can effectively suppress growth of N87 human gastric cells, and this effect exceeds the inhibition observed in a xenograft model treated with a mAb to ErbB-2. Predictably, combining the trimeric aptamer with chemotherapy, antibodies, or other drugs, along with optimizing schedule of delivery, might enhance the antitumor activity of the aptamer, thereby offering a unique strategy to treat ErbB-2–overexpressing human cancer.

Discussion

The application of aptamers in personalized cancer therapy might offer unique combinatorial approaches with immunotherapies and chemotherapeutics. So far, the most promising aptamer in the cancer field is AS1411, a nucleolin-specific truncated version of an aptamer called GRO29A, which is in advanced clinical trials (9). Targeting the ErbB family of receptor tyrosine kinases with monoclonal antibodies (e.g., Cetuximab, Trastuzumab), as well as with low molecular weight kinase inhibitors (e.g., Erlotinib, Gefitinib), has established this family of receptors as an effective target for novel drugs. Thus, the therapeutic use of anti-ErbB aptamers in oncology could represent a promising alternative to the currently used monoclonal antibodies and kinase inhibitors. Aptamers can be conveniently selected and synthesized, with no reliance on recombinant systems. In addition, their relatively low production cost and low batch-to-batch variability make aptamers an attractive endeavor for clinical application.

This study developed an aptamer with specificity to ErbB-2/human epidermal growth factor receptor 2 (HER2), a master regulator of a signaling network essential for progression of several different types of carcinoma, including one subtype of gastric cancer. After selecting several aptamers we decided, based on specificity tests, to concentrate on one, aptamer 2-2, which was subsequently multimerized to form a 42-nucleotide-long aptamer, 2-2(t). Aptamer multimerization served two purposes: first, enhancing the avidity of binding to ErbB-2, and second, cross-linking the receptor on the surface of living cells. Accordingly, however, we demonstrated that the trimeric aptamer 2-2(t) is endowed with improved binding to ErbB-2, in comparison with the monomeric version. However, our previous studies of oligoclonal mixtures of mAbs against surface receptors (20, 21) predicted that a multimeric aptamer would enable improved internalization of ErbB-2 due to the formation of multimolecular complexes at the cell surface. In the case of a combination of antibodies to ErbB-1/EGFR, it was shown that antibody mixtures not only effectively internalize the receptor, but they also divert it from a recycling route to a pathway leading to intracellular degradation (22). By using immunofluorescence and Western blotting, we confirmed that aptamer 2-2(t) can induce translocation of ErbB-2 from the plasma membrane to cytoplasmic vesicles (Fig. 3). Furthermore, the ability of aptamer 2-2(t), but not the monomeric aptamer 2-2, to accelerate degradation of the oncogenic target was validated by using immunoblot analyses (Fig. 4). The inability of monomeric aptamer 2-2 to induce internalization and degradation of ErbB-2 might be considered as a further indication that cross-linking of ErbB-2 molecules by the trimeric aptamer 2-2(t) underlies functionality. In light of the observation that the degradation of ErbB-2 was arrested by an inhibitor of lysosomal hydrolases (i.e., chloroquine), this set of results leads us to posit that the biological activity of aptamer 2-2(t) relates to the ability to cross-link and sort ErbB-2 to lysosomal degradation.

The trimeric aptamer, in comparison with the monomer, is also characterized by superior antiproliferative capacity, when tested in vitro on gastric cancer cells overexpressing ErbB-2 (Fig. 5A). The lack of antiproliferative activity of the respective monomer might be related to its lower binding or inability to cross-link ErbB-2 molecules on the cell surface. Importantly, the inhibitory effect of the trimeric aptamer extended to a xenograft model that used immunocompromised mice. Notably, aptamer 2-2(t) displayed superior antitumor activity in comparison with a monoclonal anti–ErbB-2 antibody, namely the 431 antibody we generated and tested in animals (20). Antibody 431 engages the epitope targeted by Trastuzumab, the humanized antibody approved for treatment of breast and gastric cancer. Interestingly, mAb431 inhibited tumor growth more efficiently during the first weeks of treatment, whereas the effect of aptamer 2–2(t) was more sustained, an observation we relate to differential stability or clearance of the drugs. Another difference entails the ability of mAb431 to recruit, at least to some extent, the immune system. It is generally accepted that antibody-dependent cellular cytotoxicity plays a major role in immunotherapy, such as in protocols that make use of Trastuzumab (23). However, the growth-inhibitory activities of aptamer 2-2(t) we observed in vitro and in animals, as well as the chemical nature of a DNA aptamer, do not favor immunological mechanisms of action.

In summary, this study offers a unique and generally applicable strategy to target cell surface receptors in the context of cancer therapy. Accordingly, aptamers are selected on the basis of effective binding to the surface receptor or transporter of interest. Once target specificity is validated, the aptamer might be converted into a linear oligomer, capable of polyvalent binding to the target. In the next step, the ability of the extended aptamer to translocate the target molecule from the cell surface to intracellular compartments (e.g., endosomes or lysosomes) is verified, and animal models are used to demonstrate antitumor effects in vivo. Notably, the increase in effectiveness upon multimerization of a monomeric aptamer may not be limited to trimers, and it might be even more successful with higher oligomers. Extension of the outlined strategy to oncogenic proteins other than ErbB-2, as well as translation to clinical practice, are expected to pave the way to a new therapeutic modality in oncology. Combination treatments with chemotherapeutics, biologics (e.g., antibodies and decoy receptors), as well as therapeutic radiation, might augment clinical efficacy, thereby represent a future alternative to conventional therapies. In addition, because of their target specificity and versatile modifications, aptamers might be harnessed as carriers of toxic loads. This has been demonstrated by applying RNA aptamers and drug-encapsulated nanoparticles on prostate tumor xenografts (24), which exemplifies yet another potential application of the anti–ErbB-2 aptamer we identified in this study.

Materials and Methods

Materials.

Cell lines were purchased from the American Type Culture Collection and grown in RPMI medium 1640 (Biological Industries) supplemented with 10% (vol/vol) FCS. Antibodies were purchased from Santa Cruz Biotechnology. Oligonucleotides were purchased from Hylabs. Phycoerythrin streptavidin was from BD Biosciences.

SELEX Screens.

The outline of the screen relates to the so-called “One–Pot” experiment, where the whole process takes place in a single PCR tube (25). The surface of the tube was coated at 4 °C with an ErbB-2–specific, rabbit polyclonal antiserum. Following washing and blocking (2 h) with 3% (wt/vol) milk powder, we added cleared extracts of N87 cells (500 μg/mL; 50 μL). Two hours later, we incubated the tube (37 °C; 60 min) with a single-stranded DNA library (1 μM; from Invitrogen) of aptamers, containing two constant primer binding regions and a 50-nucleotide random sequence (5′-ATACCAGCTTATTCAATT-N40-AGATAGTAAGTGCAATCT-3′) (26). Unbound DNA strands were removed, and PCR reagents (50 μL) were added. A fluorescein-labeled primer (1 μM; 5′-FL-ATACCAGCTTATTCAATT-3′) was used for the leading strand and a biotinylated primer (5′-Biotin- AGATTGCACTTACTATCT-3′) for the lagging strand. After PCR, biotinylated, double-stranded products were incubated (1 h at 4 °C) with streptavidin-coated beads. Thereafter, the beads were washed twice, resuspended in saline, and heat denatured at 95 °C for 5 min, to release fluorescein-labeled leading strands. The supernatant was collected, and the biotinylated strands were removed. Fluorescence of labeled leading strands was measured, and selected aptamers were used for a new SELEX round. After four iterative selection rounds and one counter selection round (to remove background binders), the selected aptamer sequences were subcloned into a pGEM-T vector and transformed into bacteria. After sequencing, only aptamers with shorter sequences (due to random PCR deletions) than the original random sequence in the library were selected for further characterization and functional tests. All aptamer sequences were synthesized without primers. The sequence of aptamer 2-2 is GCAGCGGTGTGGGG.

Synthesis of DNA Aptamers.

The monomeric and trimeric aptamers were synthesized by applying a DNA synthesizer. The synthesizer is equipped with a computer-controlled delivery system of reagents, which initially attaches the first base to a solid support (glass or polystyrene beads). Subsequently, additional nucleotides are introduced step by step. Our aptamers (monomers and trimers) were synthesized as single, linear DNA molecules without spacers or linkers. Wherever indicated, single biotin residues were conjugated to the 5′ nucleotide of monomeric or trimeric aptamers.

ELISA and Dot Blot Assays.

A heterogeneous, noncompetitive, direct ELISA binding assay was used to test direct binding of aptamer 2-2(t) to native ErbB-2 from N87 cell extracts. An ErbB-2–specific, polyclonal rabbit antibody (to the kinase domain) was used to coat a solid plate surface (Nunc) and incubated overnight at 4 °C. After three washing steps, the plate was blocked with 3% (wt/vol) milk powder, and extracts of N87 cells were applied (at 100 μg/mL). After three additional washing steps, several concentrations of a biotinylated aptamer 2-2(t) (up to 100 nM) were allowed to bind for 2 h. Once again, the plate was washed, and streptavidin-HRP conjugate was applied as a secondary detection reagent. The plate was washed, and tetra-methyl-benzidine substrate (Enco) was applied to the plate. Color development was stopped after 10 min, with 1M H₂SO₄, and light absorbance (450 nm) was measured by using an ELISA reader (Lumitron). The competition assay was similarly performed with 10 nM biotinylated, trimeric aptamer 2-2(t) and different concentrations of the unlabeled aptamer 2-2(t). For dot blots, 0.2-mL extracts of either N87 (0.1 and 0.01 mg of protein per milliliter) or A431 (0.1 mg of protein per milliliter) cells were spotted onto a nitrocellulose membrane. The membrane was blocked and biotinylated, monomeric, and trimeric aptamers (100 nM) were applied together with 1% milk powder to avoid nonspecific binding. After three washing steps, HRP-labeled streptavidin was used. The ECL detection kit (Amersham Pharmacia Biotech) was used to detect signals by chemiluminescence.

Regular and Double Immunoblotting.

N87 cell extracts were resolved by using gel electrophoresis and blotted onto a first nitrocellulose membrane. The membrane was blocked with 3% (wt/vol) milk powder and subsequently incubated for 2 h with specific aptamers (100 nM). After three washing steps, the membrane was reblotted to a second membrane, which was blocked, incubated with HRP-labeled streptavidin, washed three times, and chemiluminescence of specific bands was determined.

Pull-Down Assays.

Biotinylated, monomeric, and trimeric aptamers (1 μM), or the primer from the SELEX rounds, were incubated overnight at 4 °C with extracts from N87 or A431 cells (500 μg/mL). Streptavidin magnetic beads (10 μL; Novagen) were applied to allow pull down, and the beads were washed three times in saline. Samples were resolved by using gel electrophoresis and blotted to a nitrocellulose membrane (Rhenium). The membrane was blocked with milk [3% (wt/vol), in saline], and an ErbB-2–specific, rabbit polyclonal antiserum specific to the kinase domain (Santa Cruz Biotechnology) was used to detect ErbB-2. After three washing steps, an HRP-labeled, secondary antibody (Jackson ImmunoResearch Laboratories) was used for colorimetric staining.

Cell Proliferation Assays.

Cells (10,000 cells per milliliter) were grown for 24 h in 96-well plates in DMEM/F12 (1:1; 100 μL per well). Next, the medium was removed, and aptamers (10 μM) were applied in triplicates. The medium contained 1% serum, and it was refreshed every 2 d with aptamer-containing medium. Cell proliferation was determined 7 d later, by using an 2,3-bis-(2-methoxy-4-nitro-5-sulfophenyl)-2H-tetrazolium-5-carboxanilide (XTT) proliferation assay kit (Biological Industries).

FACS Analysis.

N87 cancer cells were treated with trypsin, washed, and incubated at 4 °C for 30 min with the biotinylated ErbB-2–specific aptamer 2-2(t) or a trimeric primer as control (each at 1 μM) in saline (containing 0.1% albumin). Thereafter, cells were washed and phycoerythrin streptavidin was added for 30 min at 4 °C. After washing, bound aptamer molecules were detected and cells were analyzed by using a fluorescent-activated cell sorter.

In Vivo Experiments.

Female CD-1 nude mice were inoculated intradermally with N87 cells (5 × 10⁶) human gastric cancer cells. Treatment with the indicated agents started 7 d after inoculation, when tumors became palpable and measurable. Groups of seven mice were injected i.p. eight times (once per week) with either aptamers (40 μg per mouse) or an antibody (160 μg per mouse).

Immunofluorescence.

N87 cells were plated on fibronectin-coated coverslips, 24 h before treatment with a fluorescein-tagged aptamer. Following 3 d of treatment, cells were washed, permeabilized (0.03% Triton X-100), and fixed [3% (vol/vol) paraformaldehyde]. A polyclonal anti–ErbB-2 antibody (Santa Cruz Biotechnology) and a fluorescent secondary antibody (Invitrogen) were used to visualize ErbB-2. Nuclei were visualized by using DAPI counterstaining. Microscopy used the DeltaVision System (Applied Precision) and a 100×/1.4 objective.

Statistical Analysis.

The two-way ANOVA multiple comparison test was used to analyze differences between groups.