Hanna Björkelund

Gefitinib Induces Epidermal Growth Factor Receptor Dimers Which Alters the Interaction Characteristics with 125I-EGF

The tyrosine kinase inhibitor gefitinib inhibits growth in some tumor types by targeting the epidermal growth factor receptor (EGFR). Previous studies show that the affinity of the EGF-EGFR interaction varies between hosting cell line, and that gefitinib increases the affinity for some cell lines. In this paper, we investigate possible mechanisms behind these observations. Real-time interaction analysis in LigandTracer® Grey revealed that the HER2 dimerization preventing antibody pertuzumab clearly modified the binding of 125I-EGF to EGFR on HER2 overexpressing SKOV3 cells in the presence of gefitinib. Pertuzumab did not affect the binding on A431 cells, which express low levels of HER2. Cross-linking measurements showed that gefitinib increased the amount of EGFR dimers 3.0–3.8 times in A431 cells in the absence of EGF. In EGF stimulated SKOV3 cells the amount of EGFR dimers increased 1.8–2.2 times by gefitinib, but this effect was cancelled by pertuzumab. Gefitinib treatment did not alter the number of EGFR or HER2 expressed in tumor cell lines A431, U343, SKOV3 and SKBR3. Real-time binding traces were further analyzed in a novel tool, Interaction Map, which deciphered the different components of the measured interaction and supports EGF binding to multiple binding sites. EGFR and HER2 expression affect the levels of EGFR monomers, homodimers and heterodimers and EGF binds to the various monomeric/dimeric forms of EGFR with unique binding properties. Taken together, we conclude that dimerization explains the varying affinity of EGF – EGFR in different cells, and we propose that gefitinib induces EGFR dimmers, which alters the interaction characteristics with 125I-EGF.

Comparing the Epidermal Growth Factor Interaction with Four Different Cell Lines: Intriguing Effects Imply Strong Dependency of Cellular Context

The interaction of the epidermal growth factor (EGF) with its receptor (EGFR) is known to be complex, and the common over-expression of EGF receptor family members in a multitude of tumors makes it important to decipher this interaction and the following signaling pathways. We have investigated the affinity and kinetics of 125I-EGF binding to EGFR in four human tumor cell lines, each using four culturing conditions, in real time by use of LigandTracer®. Highly repeatable and precise measurements show that the overall apparent affinity of the 125I-EGF – EGFR interaction is greatly dependent on cell line at normal culturing conditions, ranging from KD≈200 pM on SKBR3 cells to KD≈8 nM on A431 cells. The 125I-EGF – EGFR binding curves (irrespective of cell line) have strong signs of multiple simultaneous interactions. Furthermore, for the cell lines A431 and SKOV3, gefitinib treatment increases the 125I-EGF - EGFR affinity, in particular when the cells are starved. The 125I-EGF - EGFR interaction on cell line U343 is sensitive to starvation while as on SKBR3 it is insensitive to gefitinib and starvation. The intriguing pattern of the binding characteristics proves that the cellular context is important when deciphering how EGF interacts with EGFR. From a general perspective, care is advisable when generalizing ligand-receptor interaction results across multiple cell-lines.

Deciphering complex protein interaction kinetics using Interaction Map

Cellular receptor systems are expected to present complex ligand interaction patterns that cannot be evaluated assuming a simple one ligand:one receptor interaction model. We have previously evaluated heterogeneous interactions using an alternative method to regression analysis, called Interaction Map (IM). IM decomposes a time-resolved binding curve into its separate components. By replacing the reductionistic, scalar kinetic association rate constant k(a) and dissociation rate constant k(d) with a two-dimensional distribution of k(a) and k(d), it is possible to display heterogeneous data as a map where each peak corresponds to one of the components that contribute to the cumulative binding curve. Here we challenge the Interaction Map approach by artificially generating heterogeneous data from two known interactions, on either LigandTracer or Surface Plasmon Resonance devices. We prove the ability of IM to accurately decompose these man-made heterogeneous binding curves composed of two different interactions. We conclude that the Interaction Map approach is well suited for the analysis of complex binding data and forecast that it has a potential to resolve previously uninterpretable data, in particular those generated in cell-based assays.

Selection and in vitro characterization of human CD44v6-binding antibody fragments

The cluster of differentiation (CD) 44v6 antigen has been suggested to be involved in tumor formation, invasion, and metastasis formation, and has been observed in a majority of primary and metastatic squamous cell carcinomas of the head and neck. Probes specifically binding to this region may be utilized as tools for the challenging tasks of early detection and targeted treatments of small residual disease. In this project, an epitope‐guided phage display selection of human fragment antigen‐binding (Fab) fragments with affinity to the v6 sequence was performed. A selected set of Fab fragments was shown to specifically recognize increasingly complex forms of the target sequence, both in the form of a short synthetic or recombinant peptide and in the context of a purified extracellular domain of CD44. The binding was independent of known v6‐sequence variation and posttranslational modifications that are common in the CD44 protein family. Furthermore, real‐time interaction measurements on antibody fragments labeled with 125I showed specific and high‐affinity binding to the antigen present on cultured head and neck squamous cell carcinoma cells. There was no cross‐reactivity toward cells that lack the target protein. As hypothesized, characterization of the interaction between Fab fragments and the targets using the mathematical tool Interaction Map revealed more heterogeneous interactions on cells than with pure proteins analyzed by surface plasmon resonance. One main candidate Fab fragment with optimal affinity for all forms of the target sequence was identified. The flexible recombinant source of the Fab fragments might aid the development of tailored molecules adapted for therapeutic or diagnostic applications in the future.

The non-toxic and biodegradable adjuvant Montanide ISA 720/CpG can replace Freund's in a cancer vaccine targeting ED-B-a prerequisite for clinical development

We have recently shown that immunization against the extra domain-B (ED-B) of fibronectin, using Freunds adjuvant, reduces tumor growth in mice by 70%. In the present study we compare the immune response generated against ED-B using the non-toxic and biodegradable adjuvant Montanide ISA 720/CpG with the response elicited by Freunds adjuvant. Montanide ISA 720/CpG induced anti-ED-B antibodies with higher avidity and less variable levels between individuals than Freunds. Moreover, the duration of the immune response was longer and the generation of anti-ED-B antibodies in naïve mice was faster, when Montanide ISA 720/CpG was used. We conclude that it is possible to replace the mineral oil based adjuvant Freunds with an adjuvant acceptable for human use, which is a prerequisite for transfer of the ED-B vaccine to the clinic.

Circumventing the requirement of binding saturation for receptor quantification using interaction kinetic extrapolation

Quantification of the number of receptors per cell (NRPC) is important when assessing whether a tumor surface biomarker is suitable for medical imaging. One common method for NPRC quantification is to use a binding saturation assay, which is time consuming and requires large amounts of reagents. The aim of this study was to evaluate an alternative method based on kinetic extrapolation (KEX) and compare it with the classical manual saturation technique with regard to accuracy as well as time and reagent consumption. Epidermal growth factor receptor (EGFR) and HER2 receptor surface expression were quantified on five tumor cell lines using three 125I-labeled and 131I-labeled ligands (cetuximab and EGF for EGFR, trastuzumab for HER2 receptor) for both techniques. The KEX method involved interaction measurements in the LigandTracer, followed by KEX through computerized real-time interaction analysis to correct for nonsaturation on cells. Variability and NRPC estimates of the EGFR and HER2 receptor levels using the KEX method were comparable with the results from the classical saturation technique. However, the ligand consumption for the KEX method was 26–46% of the classical saturation technique. Furthermore, the KEX method reduced the workload radically. From the observations described in this study, we believe that the KEX method enables fast, credible, and easy NRPC quantification with a reduction in reagent consumption.

Avoiding false negative results in specificity analysis of protein–protein interactions

The competition measurement using simultaneous incubation of labeled and unlabeled Ligand is a common method to assess the specificity of a biomolecular interaction. In this paper we show that invalid assumptions about the interactions may lead to improper experimental setups which in turn can result in inaccurate conclusions about the specificity. To improve understanding of competition measurements, simulations in MATLAB as well as real‐time interaction analysis using LigandTracer have been performed. We show that use of a concentration of unlabeled Ligand of at least 10 × KD is necessary for assay accuracy. Increasing the incubation time to assure equilibrium, adding a pre‐incubation phase, and a general understanding of the reversibility of an interaction may also improve the reliability of the measurement and the conclusions drawn about specificity. These findings may lower the risk of false negative results as well as reducing the amount of reagent needed. Copyright

Conjugation Effects on Antibody-Drug Conjugates : Evaluation of Interaction Kinetics in Real Time on Living Cells

Antibody-drug conjugates (ADC) have shown promising effects in cancer therapy by combining the target specificity of an antibody with the toxicity of a chemotherapeutic drug. As the number of therapeutic antibodies is significantly larger than those used as ADCs, there is unused potential for more effective therapies. However, the conjugation of an additional molecule to an antibody may affect the interaction with its target, altering association rate, dissociation rate, or both. Any changes of the binding kinetics can have subsequent effects on the efficacy of the ADCs, thus the kinetics are important to monitor during ADC development and production. This paper describes a method for the analysis of conjugation effects on antibody binding to its antigen, using the instrument LigandTracer and a fluorescent monovalent anti-IgG binder denoted FIBA, which did not affect the interaction. All measurements were done in real time using living cells which naturally expressed the antigens. With this method the binding profiles of different conjugations of the therapeutic anti-EGFR antibody cetuximab and the anti-CD44v6 antibody fragment AbD15171 were evaluated and compared. Even comparatively small modifications of cetuximab altered the interaction with the epidermal growth factor receptor (EGFR). In contrast, no impact on the AbD15171-CD44v6 interaction was observed upon conjugation. This illustrates the importance to study the binding profile for each ADC combination, as it is difficult to draw any general conclusion about conjugation effects. The modification of interaction kinetics through conjugation opens up new possibilities when optimizing an antibody or an ADC, since the conjugations can be used to create a binding profile more apt for a specific clinical need.

Evaluating real-time immunohistochemistry on multiple tissue samples, multiple targets and multiple antibody labeling methods

BackgroundImmunohistochemistry (IHC) is a well-established method for the analysis of protein expression in tissue specimens and constitutes one of the most common methods performed in pathology laboratories worldwide. However, IHC is a multi-layered method based on subjective estimations and differences in staining and interpretation has been observed between facilities, suggesting that the analysis of proteins on tissue would benefit from protocol optimization and standardization. Here we describe how the emerging and operator independent tool of real-time immunohistochemistry (RT-IHC) reveals a time resolved description of antibody interacting with target protein in formalin fixed paraffin embedded tissue. The aim was to understand the technical aspects of RT-IHC, regarding generalization of the concept and to what extent it can be considered a quantitative method.ResultsThree different antibodies labeled with fluorescent or radioactive labels were applied on nine different tissue samples from either human or mouse, and the results for all RT-IHC analyses distinctly show that the method is generally applicable. The collected binding curves showed that the majority of the antibody-antigen interactions did not reach equilibrium within 3 hours, suggesting that standardized protocols for immunohistochemistry are sometimes inadequately optimized. The impact of tissue size and thickness as well as the position of the section on the glass petri dish was assessed in order for practical details to be further elucidated for this emerging technique. Size and location was found to affect signal magnitude to a larger extent than thickness, but the signal from all measurements were still sufficient to trace the curvature. The curvature, representing the kinetics of the interaction, was independent of thickness, size and position and may be a promising parameter for the evaluation of e.g. biopsy sections of different sizes.ConclusionsIt was found that RT-IHC can be used for the evaluation of a number of different antibodies and tissue types, rendering it a general method. We believe that by following interactions over time during the development of conventional IHC assays, it becomes possible to better understand the different processes applied in conventional IHC, leading to optimized assay protocols with improved sensitivity.

Real-time immunohistochemistry analysis of embedded tissue

We present a novel analysis of membrane-protein expression in tissue sections based on semi-automatic real-time measurement using LigandTracer(®) technology. A commercial antiHER2 antibody developed for immunohistochemistry used in this setup was revealed to have sub-optimal interaction with tissue when analyzed as recommended for immunohistochemistry. We therefore think that real-time measurement of tissue, offering direct and quantitative membrane-protein interaction analysis, can lead to improved reproducibility and eliminate the subjective operator dependences that classical immunohistochemsitry suffers from.