Discovery of novel glycomimetic ligands for Siglec-8

This thesis principally focuses on the development of glycomimetic ligands with improved affinity to Siglec-8. Siglec-8 is a member of the Siglec family, I-type lectins that contain a sialic acid-binding domain and that play different roles in cell-cell interactions and cell signaling. Siglec-8 is a CD33-related protein expressed exclusively on mast cells and eosinophils, and weakly on basophils. Cross-linking of Siglec-8 with antibodies leads to apoptosis of eosinophils and inhibition of degranulation of mast cells. Apoptosis has also been induced when eosinophils were treated with a glycopolymer (6’-sulfo-sLex-polyacrylamide polymer). Therefore, Siglec-8 represents a very interesting new pharmacological target for the treatment of eosinophil- or mast cell-associated diseases, such as asthma, chronic rhinosinusitis, chronic urticaria, hypereosinophilic syndromes, mast cell and eosinophil malignancies, and eosinophilic gastrointestinal disorders. Recently, the NMR solution structure of the Siglec-8 lectin domain was solved in complex with its preferred ligand, the tetrasaccharide 6′-sulfo sialyl Lewisx (Neu5Acα2-3[6S]Galβ1-4[Fucα1-3]GlcNAc). The main interactions involve a salt bridge between Arg109 and the carboxylate of sialic acid and a second salt bridge between Arg56 and Gln59 and the sulfate at the 6-position of the galactose moiety. Interestingly, the sulfate plays a key role in both binding and specificity to Siglec-8. In addition, hydrogen bonds exist between hydroxyl groups 7, 8 and 9 of sialic acid and Tyr7, Ser118 and Gln122. In contrast, the fucose and glucosamine subunits show only minor interactions. This led to the suggestion that the disaccharide substructure 6-sulfo-Sia-Gal might represent the minimal binding epitope for Siglec-8.
We first synthesized and tested this disaccharide, which showed a 2-fold lower affinity compared to the parent tetrasaccharide, but considering the simplified structure and the much easier synthetic approach, 6-sulfo-Sia-Gal can be considered as the minimal binding epitope for Siglec-8 and represents a very good starting point for further optimization processes. Bioisosters and a deoxygenation strategy were employed to improve its affinity to Siglec-8. By replacing the galactose unit with a sulfo-substituted hydroxymethyl-cyclohexane, we obtained a new lead compound with 2-fold higher affinity compared to the minimal binding epitope, mainly caused by the reduced desolvation costs due to the decreased polarity. Finally, the introduction of a sulfonamide substituent in position 9 of the sialic acid led to the identification of a potent ligand with a further 17-fold improvement in binding affinity. Homology modeling was used to get more structural information about Siglec-8 and about the binding mode of our ligands (Chapter 2).
In a previous unpublished work, a molecule where the sialic acid of the minimal binding epitope was replaced by a lactic acid derivative showed interesting affinity to Siglec-8. Since it would be the first molecule to bind a Siglec without containing sialic acid, we decided to resynthesize and test it with different assays, together with a small library of derivatives. However, we could not confirm the activity of the reference compound and one of our derivatives showed questionable results in the applied assays. We think that the compound probably interacts with Siglec-8, but with different site(s) of the protein. Further experiments, such as 1H-15N HSQC NMR analysis, might be useful in the future to further investigate this hypothesis (Chapter 2).
In Chapter 3, we present two different virtual screening approaches. In the first case, by looking at the docking pose of our lead compound, we could observe an empty pocket close to position 5 of the sialic acid. We therefore decided to explore this cavity by virtually combining different fragments with the core structure of the lead compound, docking the obtained molecules into the Siglec-8 NMR solution structure and evaluating their poses. The most promising ones, where fragments were attached via amine or triazole linkers, were synthesized and tested. However, none of them showed any affinity to the target protein, probably because the absence of an amide could be detrimental for the binding. In the second approach, we performed a virtual screening of commercially available molecules to identify new possible hits with no carbohydrate moieties. Also in this case, none of them showed affinity to our target protein. However, the introduction of substituents via amide linker by rational design, modifying the ligand in the binding pocket, was more successful. In particular, the introduction of a methoxypropionamide in position 5 showed an almost 2-fold enhanced affinity compared to the parent compound, revealing that modifications in this position indeed are beneficial to improve the binding potency.
In Chapter 4, we tested the stability of our most important Siglec-8 ligands towards neuraminidase 2 (NEU2), a human enzyme able to hydrolyze sialic acids connected via α2-3 linkages. An LCMS method was developed to follow the possible hydrolysis, by checking the consumption of the substrate and the formation of the product. Our observations indicated that glycomimetic ligands show higher stability compared to the more natural minimal binding epitope. However, the positive control used is quite labile, therefore to definitively confirm our results the use of an additional control compound should be considered. The assay is anyway validated and transferable to other compounds and enzymes.
Finally, as a small side project, we synthesized and tested some ligands towards Galectin-8, which has gained attention as a potential new pharmacological target for the treatment of various diseases, including cancer, inflammation, and disorders associated with bone mass reduction. The structure of our lactic acid derivatives described in Chapter 2 is very similar to a recently published Galectin-8 ligand, the 3-O-[1-carboxyethyl]-β-D-galactopyranoside. Therefore, we decided to test the corresponding non-sulfated derivatives and to optimize them for binding to Galectin-8. Affinity data measured by fluorescence polarization showed that the most potent compound reached a KD of 12 μM. Furthermore, reasonable selectivity versus other galectins was achieved, making the highlighted compound a promising lead for the development of new selective and potent ligands for Galectin-8 as molecular probes to examine the protein’s role in cell-based and in vivo studies (Chapter 5).