Deep mining of a complex antibody pool binding Chronic Lymphocytic Leukemia cells

Abstract

Monoclonal antibodies represent an attractive approach for treatment of cancer. Phage display is a very powerful technique to display large antibody libraries, which can be used to isolate therapeutic antibodies. Panning of phage display libraries against cells generates complex antibody pools, and direct screening of such pools typically identifies high-frequent antibodies against highly expressed cell surface proteins. To be able to identify also low-frequent clones binding lower expressed proteins, different deep mining techniques are needed. In this study, we evaluated three deep mining techniques, with the aim to find new antibody/target combinations from a complex antibody pool raised against primary CLL cells. These techniques involved selection on recombinant proteins (Fishing) but also selection on cells blocked with antibodies in IgG format, previously isolated from the pool (Blocking) and pre-selection against proteins identified as targets for these IgGs (Protein Depletion). Fishing selections generated antibodies against three new targets, not previously discovered using direct screening. We also found that Blocking and Protein Depletion generated antibodies against new targets (not yet deconvoluted) on CLL cells. Taken together, these results show that antibodies with new specificities can be isolated from an antibody pool by going beyond direct screening using deep mining techniques to reach deeper down in the pool. These techniques can hopefully be used to find new antibody/target combinations with clinical relevance in many drug discovery programs.

Popular Abstract

Deep mining of complex antibody pools

Antibodies are promising drugs for treatment of cancer as they can selectively target tumor cells and elicit a variety of responses without causing intolerable side-effects in patients. Large antibody libraries containing billions of different antibodies can be constructed and used to isolate therapeutic antibodies using different techniques, such as phage display.

Antibodies (more specifically pools of antibodies) are isolated from phage display libraries with specific proteins, or entire cell populations, as bait. When using cells as bait the isolated antibody pools become extremely complex, containing thousands of different antibodies binding a variety of cell surface antigens. In the pool, each unique antibody is present in different numbers (from a few to millions), which adds to the complexity of the pool. The most frequent antibodies are most easily found when screening randomly selected antibodies from the pool. To be able to find and characterize also less frequent antibodies, that are ‘hidden’ deep down in the antibody pool, different techniques are needed, so-called deep mining techniques.

A complex antibody pool has previously been generated at BioInvent from a phage display library, using cancer cells from patients with chronic lymphocytic leukemia (CLL) as bait. The most frequent antibodies from the pool have been characterized and found to bind a handful of different targets on CLL cells. An antibody drug against one of these targets has been developed and is currently tested on CLL patients in a clinical trial. Deep sequencing of the CLL antibody pool has shown that it contains many less abundant clones that remain to be isolated. In this study the CLL antibody pool was used to evaluate three deep mining techniques (Fishing, Blocking and Protein Depletion) with the aim of finding new antibody/target combinations for CLL.

In Fishing, antibodies were pulled out from the complex pool by the use of specific proteins as bait. These proteins were selected by analyzing the expression of genes in CLL cells compared to healthy cells. Cell surface proteins that seemed up-regulated on CLL cells, and not previously identified as targets for antibodies in the pool, were chosen as bait. In Blocking, whole CLL cells were first blocked with previously identified antibodies and then these blocked cells were used as bait to pull out new antibodies from the complex pool. In Protein Depletion, antibodies binding to previously identified targets on CLL cells were first subtracted (depleted) from the CLL pool using these target proteins as bait and then new antibodies from the remaining pool were pulled out with CLL cells as bait.

Antibodies against three of nine proteins used for Fishing were identified, and also Blocking and Protein Depletion generated a number of new antibodies with currently unknown specificity that bind to new targets, previously not discovered in the antibody pool. Importantly, we have shown that all of these methods are useful deep mining techniques. Hopefully some of the new antibody/target combinations identified in this study will be of therapeutic interest and in the long run they could possibly show clinical success and help cure CLL and other B-cell malignancies.

Master’s Degree Project in Molecular Biology 60 credits 2017
Department of Biology, Lund University

Advisors: Anne Ljungars & Mikael Mattsson
BioInvent International AB

• author: Svensson, Carolin
• supervisor: Anne Ljungars; Mikael Mattsson
• organization: Degree Projects in Molecular Biology
• course: MOBT01 20162
• year: 2017
• type: H2 - Master's Degree (Two Years)
• subject: Biology and Life Sciences
• language: English
• id: 8925111
• date added to LUP: 2017-09-08 15:08:47
• date last changed: 2017-09-08 15:08:47