(T0930-06-31) Non-Tissue Based Screening Methods for Selection of Immunohistochemistry Antibody Candidates

Purpose: Screening antibody reagents for use in immunohistochemistry (IHC) assays is a challenging and expensive process. This screening relies on tissue samples which are expensive and depending on the tissue type can be hard to procure. Furthermore, the use of orthogonal methods, such as western blot, may not accurately reflect antibody binding to the target protein in IHC assays which includes steps such as formalin fixation. Therefore, it becomes important to investigate other orthogonal non-tissue-based screening methods for IHC antibodies that are both informative and economical. In this study, we utilized biolayer interferometry (BLI) for the biophysical assessment of antibodies in an IHC-like process. We further combined this with surface plasmon resonance (SPR) for linear epitope mapping. Lastly the antibodies were screened on gastric cancer tissue samples using IHC for comparison to the BLI orthogonal method. The study aims to establish a high-throughput and cost-effective non-tissue screening process that enables informed selection of IHC antibodies based on their potential functional performance.

Methods: In this study, three commercial monoclonal antibodies (mAbs) targeting a cancer related type-II transmembrane serine protease were evaluated for potential use on gastric cancer tissue. The initial screening process involved conjugating the antigen of interest at a concentration of 20 μg/ml onto amine reactive tips using the Octet HTX instrument. Subsequently, these antigen-conjugated tips were incubated overnight in either 1X-HBS-EP buffer or formalin solution. The following day, formalin-treated tips were subjected to steps commonly applied in the IHC process such as baking, treatment with Bond Dewax, ethyl alcohol, Bond Wash and Bond Epitope Retrieval solutions. Both IHC-treated and untreated antigen conjugated tips were included in a typical binding assay on the Octet HTX platform. For linear epitope mapping, a peptide library was generated spanning the full sequence length of the antigen of interest. On the Biacore 8K+ SPR instrument, antibodies were conjugated on a CM5 chip and equilibrated overnight. This was followed by sequential injections of peptides diluted in 1X-HBS-EP buffer to test for binding to antibody reagents. For IHC, consecutive sections of 4 µm thickness were used for comparison. The slides were loaded into the Leica BondRx and processed using the standard online dewaxing and rehydration protocol. Antigen retrieval was performed using Bond Epitope Retrieval Solution 2 (ER2, EDTA buffer, pH 9.0) at 95°C for 20 minutes. The primary antibody was used at a concentration of 5 µg/mL for Abcam mAb 1 and Cell Signaling Technology mAb, and 8.3 µg/mL for Abcam mAb 2. The Cell Signaling antibody diluent was used for primary antibody dilution and incubated for 15 minutes. Detection was carried out using the Bond Polymer Refine Detection Kit.

Results: In the traditional BLI assay format, Abcam mAb 1 and mAb 2 demonstrated a binding signal below the 0.1 nm threshold indicating low/no binding. However, with application of IHC-like conditions to the BLI assay, positive binding signals were observed for both Abcam antibodies. This suggests that exposure to unique IHC assay conditions like formalin fixation and epitope retrieval contribute to unmasking of epitopes inaccessible in the native form of the target protein. A positive binding signal was also observed for the Cell Signaling Technology (CST) antibody under both traditional and IHC-like BLI assay. Further investigation of antigenic determinants by SPR revealed that Abcam mAb 1 and CST mAb bound the same peptide sequence, whereas Abcam mAb 2 bound two peptide sequences with a 10 amino acid overlap. Both epitope sites were found to be located on the protein surface, however the N-term of Abcam mAb 2 epitope was located on the interior of the target protein. This could potentially affect accessibility to epitope for antibody binding and result in poor performance in the IHC assay. IHC comparison of the three antibodies showed the best sensitivity and specificity with the CST mAb.

Conclusion: The characterization of antibodies used in IHC assays is important for developing robust assays. Our report establishes an economically viable and high-throughput orthogonal approach for screening and epitope mapping of antibodies. This work also has the potential to substantially enhance our understanding of critical quality attributes that define a good performing IHC antibody. By understanding the relationship between epitope characteristics (e.g. protein surface vs. interior) and IHC performance, this work may help inform antibody campaign immunogen design and reagent selection, ultimately leading to better antibodies specifically for use in IHC assays.

Acknowledgements: The authors declare the following competing financial interest(s): All authors were employed at Bristol-Myers Squibb (BMS) at the time of conceptualization, data generation, analysis, and abstract writing/submission. This work was supported by the Translational Sciences and Diagnostics division of Bristol Myers Squibb.

Figure 1A. The IHC-like process on Octet HTX system identified a binding partner for target protein I) Flow chart illustrating the IHC process treated and untreated antigen binding assay. II) Binding signal of selected antibodies to IHC treated and untreated antigen (n=2).

Figure 1B. IHC-based comparison of selected antibodies in gastric cancer tissue.