Engineering Cell-Based Selections for Translatable Ligand Discovery

Loading...
Thumbnail Image

Persistent link to this item

Statistics
View Statistics

Journal Title

Journal ISSN

Volume Title

Title

Engineering Cell-Based Selections for Translatable Ligand Discovery

Published Date

2021-12

Publisher

Type

Thesis or Dissertation

Abstract

Engineered protein ligands with specific, high-affinity binding to a biomarker that are differentially expressed in a disease state have been applied in a variety of therapeutic and diagnostic applications. Yeast surface display libraries coupled with high-throughput selection strategies have shown effectiveness in discovering and maturing ligands towards a variety of target molecules. These high-throughput selection strategies often require soluble protein as a target molecule. This requires that cell surface biomarkers with transmembrane domains, constituting a large class of interesting targets, be produced as recombinant extracellular domains due to the hydrophobic nature of the transmembrane domain. However, a variety of factors including poor stability, improper folding, incorrect post-translational modifications, the addition of chemical purification tags, and the lack of plasma membrane may result in additional non-natural epitopes or the masking of native epitopes. Thus, ligand discovery campaigns performed using recombinantly-produced extracellular domains may result in ligands that bind to the recombinant target but fail to recapitulate that binding towards full-length target on target-expressing cells or tissues. The use of either whole cells or detergent-solubilized cell lysate expressing full-length target has been successfully applied as an alternative to recombinant target in discovering ligands that translate binding to target-expressing cells and tissues in the context of cancer and blood-brain barrier targets. However, these selections lack the throughput to effectively screen full-sized yeast surface display libraries and are limited in their ability to select ligands from naïve libraries with limited affinity if overexpressing cell lines are not available. Finally, the heterogeneous nature of the mammalian cell surface often results in non target-specific ligands dominating the campaign, making the isolation of target-specific ligands difficult. All these factors limit the wide-spread use of cell-based selections. The work presented below aims to tackle each of these issues, as well as to elucidate the factors that affect successful cell-based selections and isolate panels of ligands with specific, high-affinity binding to biomarkers overexpressed in cancer. Naïve affibody and fibronectin libraries were sorted against cluster of differentiation 276 (CD276 or B7-H3) and cluster of differentiation 90 (CD90 or Thy1) by five selection strategies using recombinant extracellular domains and target-expressing cells. Cellular selection strategies provided a higher frequency of ligands that translate to binding on target-expressing cell monolayers, albeit with a relatively high degree of non target-specific binding. Sequential depletion on target-negative cell monolayers was insufficient to deplete these non target-specific binders, but pre-blocking yeast populations with disadhered target-negative cells provided significant depletion. Directed evolution through helix walking of a preliminary affibody molecule with modest but specific binding to CD276 (AC2, Kd = 310 ± 100 nM) resulted in a panel of CD276-specific ligands, including a sub-nanomolar binder (AC12, Kd = 0.9 ± 0.6 nM). Next, the use of mammalian cell-magnetic bead conjugates was investigated for use as effective cell-based pulldown agents to provide a new method of cell-based selection. This method displayed an order of magnitude higher throughput than traditional adherent cell panning, putting it on par with recombinant target magnetic-activated cell sorting (MACS), and was effective in enriching ligands under the same conditions as adherent cell panning in an EGFR model system, but failed to provide sufficient enrichment in a CD276 model system. Additionally, the use of an extended 641-amino acid linker was investigated to provide more consistent yeast-mammalian cell engagement and enhanced avidity. This extended linker provided enhanced enrichment in a >600-nM affinity ligand, 106 EGFR per cell system where the original 80-amino acid linker failed to provide effective enrichment (23 ± 7 vs. 0.8 ± 0.2, p = 0.004). This enrichment benefit was generalizable to a CD276 model system and mathematical modelling of the linkers as random chain polymers confirmed that this enhanced enrichment was likely due to the ability of an increased number of ligands to access the extracellular environment. Lastly, a method of high-throughput clonal specificity screening was developed using deep sequencing to observe clonal frequency in populations differentially panned on target-expressing and target-negative populations in the context of insulin-like growth factor receptor (IGF1R) and insulin receptor isoforms A (InsRA) and B (InsRB). Adherent cell panning yielded affibodies that were preferentially enriched on IGF1R-expressing cells relative to IGF1R-negative cells and affibodies and fibronectins that were preferentially enriched on InsR-expressing cells relative to parental HEK293T cells, but with limited isoform specificity. Deep sequencing of the IGF1R populations revealed several affibody sequences with specificity towards IGF1R-expressing cells. In total, the results contained in this thesis elucidate the factors that dictate successful cell-based panning and provide new methods to increase the throughput, enrichment, and specificity of cell-based panning to motivate wider adoption, as well as panels of compelling molecules with high-affinity, specific binding to cancer-relevant biomarkers for therapeutic and diagnostic applications.

Description

University of Minnesota Ph.D. dissertation. 2021. Major: Chemical Engineering. Advisor: Benjamin Hackel. 1 computer file (PDF); 177 pages.

Related to

Replaces

License

Collections

Series/Report Number

Funding information

Isbn identifier

Doi identifier

Previously Published Citation

Other identifiers

Suggested citation

Lown, Patrick. (2021). Engineering Cell-Based Selections for Translatable Ligand Discovery. Retrieved from the University Digital Conservancy, https://hdl.handle.net/11299/226388.

Content distributed via the University Digital Conservancy may be subject to additional license and use restrictions applied by the depositor. By using these files, users agree to the Terms of Use. Materials in the UDC may contain content that is disturbing and/or harmful. For more information, please see our statement on harmful content in digital repositories.