(427d) Advancing the Preferential Enrichment of Matrix Metalloproteinase-9 Inhibitors and Their Characterization Using Yeast Display

Matrix metalloproteinases (MMPs) are a major group of extracellular enzymes that are found in the tumor microenvironment and are known contributors to tumor progression, regulating signaling pathways that control cell growth, inflammation and even angiogenesis. MMP inhibitory antibodies, discovered from phage display libraries, have shown some promising tumor-suppressing effects in in vivo studies. Discovery of such antibodies, however, remains challenging and is the result of exhaustive screening efforts that combine animal immunization and phage display techniques. The goal of this work is to explore the conditions that lead to the discovery of MMP inhibitors through less resource-intensive processes and in high throughput. The inhibitory binder DX2802, the non-inhibitory binder M0076, and the non-binding control FAPB2.3.6 were cloned into the yeast display vector pCTCON2-Aga2p and displayed on the yeast surface as single chain variable fragments (scFv). Each displayed construct was evaluated for MMP-9 binding via flow-cytometry techniques to create a basis for the assessment of model sorts. Four model libraries with different ratios of these constructs were created to simulate distinct library compositions. To preferentially enrich for inhibitory binders over the other two constructs we varied incubation time and order of incubation of magnetic bead-based sorting, as well as gating strategies during fluorescence activated cell sorting (FACS). Incubating cells for 30 minutes with biotinylated antigen in solution before immobilizing on magnetic beads yielded up to 60-fold increase in the enrichment ratio for inhibitory binders compared to standard bead-based sorting conditions. Under these same altered conditions, we observed up to two-fold decrease in the enrichment of non-inhibitory binders, implying a preferential enrichment for inhibitors. Similarly, FACS results showed that sorting through a “high gate” yields up to 40-fold increase in the enrichment ratio of inhibitory binders compared to a lower gate, and at least a 2-fold decrease in the enrichment ratio of non-inhibitory binders. Finally, through carefully considered sets of assumptions and protocols we obtained values for inhibitory constants (IC₅₀) directly on the yeast surface, a method that, to our knowledge, has not been previously explored. Our results show that the IC₅₀ obtained in yeast display format (24.59±5.42 nM) falls within the confidence interval of their soluble counterparts (20.74±4.66 nM), indicating that it is possible to obtain this parameter on the yeast surface while eliminating the arduous process of expressing and purifying proteins in solution. In summary, we have identified a more effective strategy that leads to the preferential enrichment of inhibitory binders, and successfully characterized inhibitor IC₅₀ values directly on the yeast surface. This model system establishes key strategies for pursuing more efficient discovery of complex target inhibitors in high throughput and evaluating inhibitory properties directly in surface display format.