Abstract

All cells, whether prokaryotic or eukaryotic, have a rich and diversified external glycosylation layer, forming the immediate dominant face in relation to their environment. They result from complex enzymatic processes linking sugars to each other and to proteins or lipids. Variations of the "glycome" can appear in certain pathologies. Cancers are the most frequent pathologies with abnormalities in these glycosylations. These alterations are almost systematic on the surface of cancer cells. Among them, the Thomsen-new antigen (Tn), an N-acetylgalactosamine (GalNAc) on a serine or threonine, is strongly expressed in 90% of mammary carcinomas as well as in cancers of the bladder, cervix, ovary, colon, stomach and prostate. The ubiquitous presence of the Tn antigen in many cancers, combined with its absence in healthy cells, makes it a target of choice for targeted therapy or synthetic anti-tumor vaccines. No antibody targeting the Tn antigen is currently available because of the difficulty in developing an antibody with such specificity. Thus, we were interested in an alternative targeting strategy, based on the use of a molecule capable of recognizing the Tn antigen. C-Type lectins are a family of proteins capable of specifically and reversibly binding to certain carbohydrates in the presence of calcium. Macrophage galactose lectin (MGL) is a C-type lectin with a high affinity for GalNac and its derivatives such as the Tn antigen. This work consisted, initially, in the use of a soluble recombinant form of MGL to validate the potential of this tool for the targeting of cancer cells. The different experiments, in vitro and in vivo, involving MGL, demonstrated the latter's ability to specifically target human tumors via the Tn antigen. The extracellular portion of MGL is therefore a very good vector candidate for the diagnosis and imaging of human tumors and potentially for drug delivery. In a second step, various strategies for the development of a bifunctional tool exploiting this lectin were explored. The goal was to create a peptide platform that could be functionalized on one hand with several lectin domains, in order to control recognition affinity, and on the other hand with functional groups that could be variable according to the application (diagnostic, therapeutic, ...). The different coupling strategies employed allowed us to attach several lectin CRDs to a peptide support, while preserving the three-dimensional and functional state of the proteins. The characterizations carried out show a significant increase in affinity directly related to the number of lectins added to the platform. This work paves the way to new customizable sugar-targeting systems.