(56f) Engineering Multi-Epitopic Antibodies to Interrogate and Manipulate Receptor Trafficking

• Targeting multiple epitopes on a single transmembrane protein modifies its trafficking dynamics
• Antibody-induced down-regulation depends on orientation, epitope, and affinity

Introduction: Dysfunction of normal immune system responses contributes to cancer development and progression. A combination of the tumor microenvironment and surface receptor expression on cancer cells inhibits effector T cell activity while simultaneously boosting T_Reg activity. Therefore, immune modulation is a promising avenue for cancer treatment, and is generally accomplished either by therapies which stimulate T effector cell activation or by blockade of immune checkpoint receptors, which are responsible for suppressing T cell activity. Immune checkpoint blockade in particular has demonstrated improved patient outcomes and effects complete cures in some patients across several types of cancer. Selective blockade of natural inhibitory checkpoint receptors, like PD-L1, increase rates of T-cell activation and provoke anti-tumor responses. However, the percentage of patients who exhibit durable responses to immune checkpoint blockade is limited, highlighting the need for further development in this area. Here we demonstrate the development of multi-epitopic internalizing antibodies targeting PD-L1 which simultaneously block ligand-mediated activation and downregulate receptor expression from the surface, potentially improving the anti-tumor immune response.

Methods: Antibodies targeting PD-L1 were isolated from existing yeast surface display libraries using soluble extra-cellular receptor domains. Single domain shark antibodies (IgNAR) were isolated from B cells of sharks immunized with PD-L1. Binding competition analysis with existing checkpoint blockade antibodies (Atezolizumab) was carried out using the Octet system. Single and multi-specific antibodies were cloned into the gWiz vector and produced in HEK-293 cells, and purified using protein G and seize exclusion chromatography.

Downregulation assays were carried out as follows. Cells were seeded in 96 well plates and allowed to adhere overnight in a CO₂ incubator. The next day, media was aspirated and replaced with antibody treatments diluted in media and returned to the incubator. Up to 24 hours later, the cells were removed and relative surface receptor levels identified by flow cytometry. In some cases, monensin was added to some wells to inhibit receptor recycling.

For pulse-chase assays, cells were plated as in downregulation assays. Antibodies directly conjugated with Alexa-Fluor 488 were added to the media and left for 2-3 hours to allow time for internalization to occur. The cells were washed and treated with quenching antibody for 30 minutes at 4C, then chased in the presence of quenching antibody at 37C for varying lengths of time. Non-quenched fluorescence was subsequently detected by flow cytometry.

For fluorescent imaging, cells were allowed to adhere to a slide and then treated with fluorescently labeled antibodies for 6 hours. The cells were then washed and fixed with PFA, permeabilized with Tween, and stained for endosomal and/or lysosomal markers. Cells were imaged using a Zeiss LSM 710 confocal microscope.

Results: Surface expression of PD-L1 in high expressing cells is significantly reduced after treatment with most bi or tri-specific antibodies compared to both untreated control, and treatment with the component mono-specific antibodies. The extent of downregulation varies with similarity of epitope between binding domains, and orientation of the binding domains within the antibody construct. While structural data is not available, steric or allosteric competition between domains and/or accessibility of intra-molecular binding based on orientation and epitope are likely to reduce receptor clustering and internalization. In contrast, low expressing cells show little or no significant difference in PD-L1 expression after treatment. We hypothesize that low receptor surface density may prevent efficient receptor clustering and subsequent internalization by the bi-specific antibody. Microscopy experiments to probe this explanation are currently ongoing. On high expressing cells, confocal microscopy shows a relatively diffuse distribution of PD-L1 when treated with Atezolizumab, but treatment with bispecific antibodies results in more and/or larger clusters both intra-cellularly and on the surface. Kinetic downregulation experiments show that most downregulation occurs within the first 2 hours, with minimal change after that up to 24 hours. Recycling and pulse-chase assays suggest that rates of receptor recycling after internalization vary with the construct, which may be due to saturation of degradation pathways in constructs with higher internalization rates and more efficient clustering compared to those with lower internalization rates.

Here we have identified downregulating antibodies which offer potential synergy with immune checkpoint blockade for the treatment of cancer. This approach might also be extended to other cell surface targets to develop a new class of antibody therapeutics which act to downregulate cell surface receptors, as opposed to existing antibody modalities which focus on receptor blockade, receptor pathway activation or targeted drug delivery. Additionally, surface clustering of receptors without internalization may be a mechanism for increasing avidity of specific cell-cell interactions. Finally, these internalizing antibodies are a novel platform for the study of receptor trafficking mechanisms, which remain poorly understood.