Glycomimetic Langerin Ligands for Langerhans Cell Targeting

Abstract

Immune evasion represents an important hallmark of cancer progression. In this context, the induction of tumor-specific cytotoxic T cell immunity to overcome immunological tolerance has become a focal point of cancer immunotherapy. Langerhans cells (LCs) constitute a dendritic cell (DC) subset residing in the epidermis of the human skin. They have been recognized for their capacity to endocytose and cross-present exogenous antigens via MHC-I to efficiently prime naïve CD8+ T cells. Langerin, an endocytic C-type lectin receptor (CLR) involved in the Ca2+-dependent recognition of both pathogen- and self-associated glycans, displays an expression profile highly restricted to LCs. Hence, the targeted delivery of tumor-associated antigens (TAAs) to Langerin represents an intriguing approach to develop novel vaccination strategies. Over the last decades, liposomes have emerged as versatile delivery platforms that can be targeted to LCs via the conjugation to Langerin ligands. As glycan interactions with CLRs are typically weak and highly promiscuous, liposomal targeting required the discovery of potent and specific glycomimetic ligands. However, the onerous synthesis of carbohydrate analogs as well as the hydrophilicity and high solvent exposure of carbohydrate binding sites render glycomimetic ligand design challenging. The structure-based in silico analysis of 21 X-ray structures presented in this dissertation corroborated the classification of CLRs as undruggable or challenging targets. Druggable secondary binding pockets adjacent to the carbohydrate binding site were exclusively identified for CLRs of limited therapeutic relevance. Several strategies were employed to address the challenges outlined above and to discover potent carbohydrate analogs for Langerin. The structure-based in silico screening of substituents in C2 of the mannose (Man) scaffold served to design an initial focused glycomimetic library. The structure-activity relationship (SAR) of this scaffold was further elucidated via determination of affinities for an existing library of Man analogs derivatized in C1 and C6. Alternatively, a structure-based design strategy guided the exploration of substituents in C2 of glucsoamine-2-sulfate (GlcNS). These investigations ultimately led to the discovery of potent Man (KI = 0.25±0.07 mM) and GlcNS (KI = 0.24±0.03 mM) analogs displaying a 40- to 42-fold affinity increase over naturally occurring carbohydrate ligands. The multivalent organization of carbohydrates or their synthetic analogs to match the geometry of the Langerin trimer represents an attractive strategy to optimize their specificity and potency. The GlcNS analog was conjugated to nucleic acid scaffolds to design divalent glycomimetics (IC50 = 23±2 µM) resulting in an additional distance-dependent 12-fold avidity increase. In an alternative approach, trivalent Man-bearing glycoclusters (IC50 = 0.20±0.08 mM) displayed an 80-fold avidity increase over naturally occurring carbohydrate ligands. The development of a sensitive 19F R2-filtered nuclear magnetic resonance (NMR) assay enabled the determination of KI and IC50 values for Langerin. Importantly, the optimization of the assay setup with respect to throughput and material consumption proved instrumental for the discovery of potent carbohydrate analogs and multivalent glycomimetics. Moreover, the implementation of an explorative 19F R2-filtered NMR fragment screening led to the identification of the first non-carbohydrate inhibitor reported for Langerin. The assay was successfully transferred to the CLR DC-SIGN to evaluate the specificity of designed carbohydrate analogs. While 19F R2-filtered NMR experiments served as the primary screening and characterization assay, affinities were validated via saturation transfer difference and 15N heteronuclear single quantum coherence NMR. Furthermore, an integrated strategy combining these NMR experiments with molecular docking studies was implemented to analyze the Ca2+-dependent binding mode of the designed Man and GlcNS analogs. These investigations enabled the development of suitable conjugation strategies for liposomal formulations. The GlcNS analog displayed remarkable specificity against DC-SIGN in 19F R2-filtered NMR experiments and was thus utilized for the preparation of targeted liposomes. Flow cytometry studies were employed to optimize liposomal formulations and to validate the binding of these liposomes to Langerin+ model cells in vitro. Finally, ex vivo experiments demonstrated their capacity to specifically target LCs in the human skin. The liposomes were efficiently endocytosed and thus represent a promising TAA delivery platform. In conclusion, the integration of carbohydrate chemistry, structure-based in silico methods and NMR experiments enabled the discovery of carbohydrate analogs and multivalent glycomimetics as potent and specific ligands of the endocytic CLR Langerin. These ligands were demonstrated to specifically target liposomes to LCs in the human skin and to promote endocytosis. Consequently, the findings presented in this dissertation constitute an important advancement for the research field of DC immunology and the development of novel cancer immunotherapies.