Skip to main content

Multimodal glycoconjugates: a molecular Lego approach for antitumoral immunotherapy

Periodic Reporting for period 3 - LEGO (Multimodal glycoconjugates: a molecular Lego approach for antitumoral immunotherapy)

Reporting period: 2018-09-01 to 2020-02-29

Cancer remains a major cause of mortality worldwide. Despite significant progress in treatment modalities, current therapeutic regimens are still deficient due to intolerable side effects, while stray cancer cells often escape destruction. Immunotherapy-based approaches have been proposed recently and certainly represent the most promising alternatives in this area although very limited approaches are available and major issues remain to be addressed. We have demonstrated in this interdisciplinary program that fully synthetic structures with unprecedented combinations and complexity can immunological properties against cancers. We have developed a “molecular LEGO” approach to construct synthetic molecules capable of redirecting endogenous antibodies present in the human bloodstream against tumors without preliminary immunization. By using supramolecular chemistry, molecular engineering, biochemistry, immunochemistry and glycoscience approaches, we have designed compounds with: 1/ accurate molecular definition; 2/ unprecedented molecular composition; 3/ efficient and selective antitumoral properties.

In the absence of immunization, different immunoglobulin subtypes are present in the bloodstream of all individuals. The idea of redirecting these natural antibodies (Abs) against human pathogens emerged in the 90’s. This approach relies on the utilization of simple bimodular molecules composed of two distinct recognition domains, one for these antibodies and one for moieties expressed on the pathogen surface. The presence of these two modules allows the simultaneous binding of Abs and pathogens to trigger immune-mediated clearance. While pioneering studies were highly promising, they also clearly pointed out critical conceptual issues related to the imperfect molecular composition of these structures and the absolute necessity of structural optimization to reach clinical trials. The LEGO project has provided innovative and concrete solutions to these problems. We have designed numbers of unprecedented synthetic molecular constructions, namely Antibody Recruiting Glycodendrimers (ARGs) embedded with key parameters that were not considered before and identify one optimal ARG that: 1/ recruit natural Abs through clusters of oligosaccharides (antibody binding module) with high affinity by means of multivalent interactions; 2/ target cancer related receptors selectively expressed at the surface of tumors with peptide tumor binding module (TBMs); 3/ promotes the formation of a ternary complex between natural Abs and cancer cells, which is required for the stimulation of the cytotoxic immune response in vivo; 4/ promotes up to 70% of selective cytotoxicity towards cancer cells.
The first task of the project is dedicated to the development of novel antibody binding modules (ABMs). For this purpose, we first synthesized functionalized carbohydrates (L-Rha, NeuGC, isoGb3 and cheaper monosaccharides as models), peptides (cyclopeptides, polylysine dendrimers) and other building blocks (cyclotriphosphazene-based cores) that were combined in different manner using orthogonal ligations strategies (oxime ligation and/or Cu(I) catalyzed alkyne-azide cycloaddition) to afford a variety of glyco-clusters and dendrimers displaying identical or different sugars (Chem. Eur. J., 2017, 23, 16283). In this task and before using human sera which are more delicate to manipulate than purified proteins, we systematically use model lectins/monosaccharide partners to validate our methodology (ChemPlusChem, 2017, 82, 390; Bioconjugate Chem., 2018, 29, 83). In parallel, several Rha-based constructs have shown promising binding properties by ELISA with endogenous antibodies. However, such assays require large quantities of both ligand and sera and only give limited information. For this reason, we are developing innovative microarray-based assays to screen immobilized multivalent structures. Before being equipped with the microspotter, our first experiments have validated the feasibility of this approach (Org. Biomol. Chem., 2017, 15, 5135). On the basis of this preliminary study, we have constructed more complex arrays displaying various glycodendrimers by using microspotter. We have been able to discriminate and identify nanomolar ligands for Helix pomatia agglutinin. A similar approach is currently under investigation to identify optimal multivalent ligands for endogenous IgG and IgM in serum from healthy or ill patients. For the moment, Rha and Gal-based ligands seems to achieve promising effect.
The objective of the second task is to identify new selective ligands and to combine them covalently to ensure simultaneous interactions with cancer cells (TBM, tumoral binding modules). In a first part, we synthesized and characterized functionalized peptide building blocks for the preparation of randomized combinatorial libraries of ‘neoantibodies’. Cyclopeptide sequences have been selected from previous studies as mimetics of the discontinuous regions of Rituximab. If all the peptide sequences are ready, this task was hierarchized as a lower priority and will be finalized later in the project. Instead, we have focused our attention to the development of phage display experiments, a new methodology in my group, to discover peptide sequences specific to protein and tumor cell lines. The selection process, including DNA sequencing, was highly time consuming but we have now identified and synthesized several peptides that have been multimerized onto cyclopeptide scaffold by using thiol-chloroacetyl coupling. Binding experiments between these constructs and targeted proteins will start in the next days by BLI and/or ITC. Another innovative aspect concerns the utilization of carbohydrate binding domains of MGL (specific for the Tn antigen) in a multivalent fashion onto different peptide scaffolds as TBM (see 1.2).
In parallel to the preparation of innovative TBMs, we have decided to synthesize the first ARM prototypes to demonstrate the proof of concept of our immunotherapy approach. These compounds are composed of ABMs identified in Task 1 (i.e. Rha-based conjugates) and the well-known cRGD (specific for the αvβ3 integrins) or biotin as TBMs. Recognition potency with diverse cancer cell lines is currently investigated by ELISA, flow cytometry analysis and cytotoxic experiments. This study is crucial and will guide the design of compounds and biological experiments for the final part of the project (task 3).
The first unconventional methodology of my project is the development of glycodendrimer-arrays to prepare and screen multivalent ligands varying in their shape, geometry and sugar density for endogenous antibodies. Standard ELISA assays require large quantities of both ligand and sera and only give limited information. For this reason, we are developing innovative microarray-based assays to screen immobilized multivalent structures. This approach is advantageous because it requires lower quantities of both ligand and protein and is easy to set-up. In addition, glycodendrimer-arrays allow the controlled presentation of sugars in a spatially defined arrangement whereas density variations of monovalent structures only differ in the average spacing between glycans. Thus, we are able to identify multivalent ligands with optimal affinity and selectivity for diverse carbohydrate binding proteins (eg lectins and Abs) where rational design is impossible. Moreover we can have access to a deepened analysis of the interaction parameters and binding mechanisms. Another highly interesting aspect which is being investigated is the utilization of an indirect approach to fully assemble complex glycodendrimers on slide by using simple synthetic building blocks and an innovative ‘stacking’ strategy. This will enable to strongly accelerate and facilitate the screening process by providing readily addressable arrays on which sugars will be linked chemoselectively at defined plots, thus allowing binding studies with any kind of carbohydrate binding protein. Secondly, we use the recent Bio-Layer Interferometry (BLI) methodology, a new equipment available in our department, to study interactions between proteins (lectins, antibodies, biomarkers) and multivalent carbohydrate and/or peptides and to confirm binding constants measured on microarrays. As far as we know, there are very few examples on the utilization of BLI for studying multivalency. Our first experiments with lectins seems to indicate the reliability and robustness of this methodology and confirm binding data obtained by ITC and microarrays (two manuscripts in preparation). The multimerization of carbohydrate-binding domains of lectins onto synthetic scaffold is also a highly innovative approach to target cancer cells with both efficiency and selectivity. Carbohydrate binding domains functionalized with reactive functions have been expressed in the group of our collaborators (F. Fieschi, IBS, Grenoble, international expert in C-type lectins expressed by dendritic cells such as DC-SIGN and MGL,) and we expect to obtain the first neolectins by using sortase-mediated ligation in the next 3 months. Finally, the utilization of heterodimer-based molecular TBMs for dual targeting of cancer cells to improved tumor targeting efficacy compared to the single receptor targeting peptide monomers. This approach can also be considered as unconventional because the knowledge in this field is very limited but highly promising. With these methodologies, we expect to develop ARMs with optimized affinity and selectivity for both antibodies and tumors to trigger potent immunothereputical effects against cancers.