Functional expression of a single-chain antibody to ErbB-2 in plants and cell-free systems

Abstract

Background: Aberrant signaling by ErbB-2 (HER 2, Neu), a member of the human Epidermal Growth Factor (EGF) receptor family, is associated with an aggressive clinical behaviour of carcinomas, particularly breast tumors. Antibodies targeting the ErbB-2 pathway are a preferred therapeutic option for patients with advanced breast cancer, but a worldwide deficit in the manufacturing capacities of mammalian cell bioreactors is foreseen.

Methods: Herein, we describe a multi-platform approach for the production of recombinant Single chain Fragments of antibody variable regions (ScFvs) to ErbB-2 that involves their functional expression in (a) bacteria, (b) transient as well as stable transgenic tobacco plants, and (c) a newly developed cell-free transcription-translation system.

Results: An ScFv (ScFv800E6) was selected by cloning immunoglobulin sequences from murine hybridomas, and was expressed and fully functional in all the expression platforms, thereby representing the first ScFv to ErbB-2 produced in hosts other than bacteria and yeast. ScFv800E6 was optimized with respect to redox synthesis conditions. Different tags were introduced flanking the ScFv800E6 backbone, with and without spacer arms, including a novel Strep II tag that outperforms conventional streptavidin-based detection systems. ScFv800E6 was resistant to standard chemical radiolabeling procedures (i.e. Chloramine T), displayed a binding ability extremely similar to that of the parental monovalent Fab' fragment, as well as a flow cytometry performance and an equilibrium binding affinity (Ka approximately 2 x 10(8) M(-1)) only slightly lower than those of the parental bivalent antibody, suggesting that its binding site is conserved as compared to that of the parental antibody molecule. ScFv800E6 was found to be compatible with routine reagents for immunohistochemical staining.

Conclusion: ScFv800E6 is a useful reagent for in vitro biochemical and immunodiagnostic applications in oncology, and a candidate for future in vivo studies.

Figure 1

Cloning strategy and map of ScFv 800E6 constructs. Diagram illustrating the construction of ScFvs for expression in plants (stable and transient expression: panels A and B, respectively), and in cell-free transcription-translation systems (C–G). A linker peptide sequence (Gly₄Ser)₃, multiple cloning sites (MCS), the T7 promoter, the IPTG-inducible pLac promoter, a 3' transcription terminator, a proteolytic cleavage site (Xa), a ribosome binding site (RBS), the Strep II and His-tags, the kanamycin resistance NPT II gene, the 35S-CaMV promoter, the NOS terminator region for transcript stabilization, the RNA-dependent RNA-polymerase binding site (RdRb), the viral movement proteins (M1, M2 and M3), and the viral coat protein (CP) are present in different plasmids. A synopsis of the different ScFvs and tag positions is provided in panel H.

Figure 2

Flow cytometry and immunoprecipitation with ScFv 800E6 produced in E. coli and transgenic plants. Lysates (50 μl) from bacteria expressing either ScFv800E6 or ScFvαCTV, and the parental mAb 800E6 were incubated with 32D-ErbB-2 transfectants and ErbB-2-negative 32D cells (panels A and B, respectively), and revealed by an FITC-labeled antibody to whole murine Ig. Panel C: two-fold dilutions of bacterial lysates containing ScFv800E6 were tested as above on ErbB-2⁺ SK-BR-3 cells. ScFvαCTV and mAb 800E6 are reported for comparison. Panel D: metabolically radiolabeled SK-BR-3 cells were lysed and immunoprecipitated with the indicated antibodies and ScFvs. Asterisks mark a group of closely migrating background bands of unknown origin present in several lanes. These are unlikely to represent MHC class I heavy chains that are poorly, if at all, expressed by SK-BR-3 cells [30]. Panel E: Lysates (50 μl) from tobacco plantulas stably (stbl) or transiently (trns) expressing ScFv800E6, and from transiently mock-infected plants (as a representative negative control) were incubated with SK-BR-3 cells, and ScFv binding revealed by flow cytometry as above.

Figure 3

Biochemical analysis of tagged ScFv 800E6 produced in a cell-free transcription-translation system. Panel A: His(6x)-ad-N-ScFv800E6 and urokinase (UK) were transcribed-translated either in a conventional mix (disulfide -), or in a mix promoting disulfide bonding (disulfide +), as indicated, with and without Brij 35, or the chaperone DnaK. Transcription-translation was also carried out in parallel in the absence of template DNA. Supernatants (5 μl) of transcription-translation mixtures were run on a SDS-PAGE slab, side by side with two-fold dilutions of MW standards containing known amounts of a 31 kD protein, and the gel was stained by Coomassie blue. Panel B: Smaller volumes (1 μl) of the same supernatants were run as above, and electroblotted. The filter was stained with anti His-tag and anti Ig antibodies. Open and closed arrowheads mark UK and ScFv800E6 polypeptides.

Figure 4

Comparative flow cytometry analysis of different ScFvs to ErbB-2 and epitope blocking. Panel A: mAb 100A4 (0.1 mg/ml) and mAb 800E6 (at the different, indicated concentrations) were incubated for 30 min with SK-BR-3 cells. His(6x)-ad-N-ScFv800E6 from transcription-translation mixtures (at a final concentration of 1 μg/ml) was then added and tested for its ability to bind SK-BR-3 cells using a rabbit antibody to the His-tag followed by an FITC-labeled antibody to rabbit Ig. Binding of the His-tagged ScFv in the absence of competing antibody (no mAb), and background staining in the absence of ScFv (but in the presence of both anti His-tag and FITC-labeled antibodies; no ScFv) are also shown. Mean fluorescence intensities (m.f.i.). Four selected experimental points of the experiment in panel A (including maximal inhibition by mAb 800E6 at 0.1 mg/ml) are shown in panel B. Panel C: five-fold dilutions of the indicated ScFv and UK preparations were tested by flow cytometry for their ability to bind SK-BR-3 cells, and revealed by FITC-labeled anti Ig antibodies. Panel D: Strep-N-ScFv800E6, Strep-C-ScFv800E6, and a mock transcription-translation mixture (-) were tested in flow cytometry for their ability to bind SK-BR-3 cells using either PE-Strep-Tactin or PE-streptavidin (thick and thin lines, respectively).

Figure 5

Flow cytometry and Scatchard plot analysis with ScFv800E6 produced in transcription-translation systems and tobacco leaves. Panel A: Equimolar amounts of His(6x)-ad-N-ScFv800E6, Fab' 800E6 and parental mAb 800E6 were compared by flow cytometry for their ability to bind SK-BR-3, using a conventional FITC-labeled secondary antibody to whole murine Ig. Panel B: the binding of ¹²⁵I-labeled mAb 800E6 and His(6x)-ad-N-ScFv800E6 to SK-BR-3 cells at equilibrium was expressed in the form of bound (B) ligand versus the ratio of bound over free (B/F) ligands. The slope of the best-fit curve of individual determinations is directly proportional to the observed binding affinity (Ka: association binding constant). The intercept on the abscissa is an extrapolation to infinity of the number of ErbB-2 epitopes per cell multiplied by the valence of the ligand.

Figure 6

Immunohistochemistry of breast carcinoma lesions with different preparations of ScFv800E6. Semi-serial cryostatic sections of a primary ductal breast carcinoma lesion were stained with (A) 50 μl of His(6x)-N-ScFv800E6 transcription-translation mixes; (B) 50 μl of extracts from plants transiently expressing ScFv800E6; (C) 50 μl of extracts from wild-type plants; and (D) 50 μl of mAb 800E6 (50 μg/ml) in PBS/5% FCS. ScFv/mAb binding was revealed as described in Materials and Methods.