This presentation will focus on the convergence of a new high throughput antibody discovery platform capable of screening hundreds of millions of antibodies with machine learning to accelerate the full discovery process. This work is resulting in the identification of high affinity, developable modalities fit for therapeutic use in accelerated time frames while generating significant amounts of data, further refining our algorithms and models.

The design of potent multispecific immune cell engagers, based on the Ichnos BEAT platform, relies on the identification of diverse and developable Fabs incorporating a common light chain (cLC). We have developed a high-fidelity mammalian display system for the enrichment of antibodies bespoke for the desired binder profiles. This enables NGS analysis of multi-dimensional FACS gated populations and the creation of rich datasets for antibody optimisation by AI/ML methods.

We present a new strategy that combines atomistic design calculations and machine learning to design repertoires of multipoint mutants highly enriched in stable, foldable, and functional variants. Applying high-throughput screening to designed libraries yields variants with large changes in activity profiles including orders of magnitude improvement in antibody affinity, catalytic activity, or specificity.

GPR65 is a proton-sensing G-protein coupled receptor associated with multiple immune-mediated inflammatory diseases. To probe its biology, a phage display antibody discovery campaign was performed using soluble chimeric ApoE3 scaffolds to present the extracellular loops of GPR65. Loop-specific antibodies were identified and their ability to bind the wild-type receptor generated confidence in the use of chimeric soluble proteins to act as efficient surrogates for membrane protein extracellular loop antigens.

Deep screening leverages the Illumina HiSeq platform for massively parallel sequencing, display, and rapid affinity screening at the level of >10e8 individual antibody-antigen interactions. Deep screening enabled the discovery of mid- to high-picomolar single-chain Fv (scFv) antibody leads directly from unselected, synthetic scFv repertoires augmented by machine learning. Deep screening promises to accelerate antibody discovery for a wide range of targets.

Snakebite envenoming leads to over 100,000 fatalities annually, with additional victims suffering from long-term complications. Today, the only existing specific therapy against envenoming is polyclonal, plasma-derived antivenoms from immunized animals. While life-saving, these medicines come with the risk of causing adverse reactions. To address this, recombinant monoclonal antibodies and nanobodies that target medically relevant toxins in the snake venoms are being explored as therapeutic alternatives. In this presentation, our strategies for the development of broadly-neutralising recombinant antibodies and nanobodies against snake venom toxins will be presented. In particular, the presentation will include a deep dive into the use of in vitro display technologies, in silico design of protein binders, and methods for producing defined recombinant antivenoms based on oligoclonal mixtures of antibodies.

Targeting pHLA class I and II at therapeutic resolution has been largely restricted to approaches building on the native TCR ligand as biologics or cell therapy. We here show how the special pIX phage display system has been optimised and combined with in silico guidance and deep sequencing as a unique platform allowing TCR-Like antibodies to enter this stage beyond the current state of the art.

We use chicken immunisation in combination with yeast surface display and a competitive FACS screening for the isolation of single chain Fv fragments that functionally block a therapeutic antibody. N-terminal fusion with an MMP-9 cleavable linker results in variants with more than 1000-fold attenuated affinity, where proteolytic demasking enables regain of antibody function in the tumour microenvironment. Examples are presented for masking a FcγRIIB blocking and a T cell receptor targeting antibody, as well as an internalising antibody that locates a cargo to the cell cytoplasm.

Affinity proteins are crucial for life, for building structures, performing reactions, and for signaling purposes. By using combinatorial protein engineering and protein library technologies, small protein scaffolds can be engineered and thereby equipped with various functions. To develop protein-based systems for protein diagnostic and therapeutic purposes we are utilising small, well characterised domains. To suit the intended use, the biodistribution, half-life as well as the cell internalisation can be manipulated. Here, the development, evaluation, and use of these affinity domains will be discussed.