Hyaline Fibromatosis Syndrome (HFS) is a rare inherited disease that is characterized by an accumulation of an unidentified hyaline material, largely affecting connective tissues. Patients afflicted with HFS present a wide range of clinical symptoms such as nodules, papules, hyperplasia of the gingiva, joint contractions, thickness and hyperpigmentation of the skin, and osteopenia. Depending of the severity and the delay of appearance of these symptoms, the life expectancy varies between 2 years and > 30 years. Death earlier in life is due to recurrent diarrhea and chest infections of unknown origin. In 2003, the gene responsible for HFS was identified to be Capillary Morphoqenesis Gene 2 (CMG2), a gene upregulated during capillary morphogenesis in three-dimensional collagen matrices. This gene encodes a type I transmembrane glycoprotein composed of a von Willebrand (vWA) domain followed by an immunoglobulin (Ig)-like domain in its extracellular part, a single α-helix spanning transmembrane domain, and a 148 amino acids cytoplasmic tail predicted as unstructured. Even though the physiological function of CMG2 remains to be elucidated, it has been shown that this protein induces endothelial proliferation, and dictates cell polarization during embryogenesis in a zebrafish model. Besides its physiological function, CMG2 was found to be an anthrax toxin receptor, thus its alternative name ANTXR2, and has been extensively studied in this context. Genetic analysis has led to the identification of 34 CMG2 mutations in HFS patients. The molecular mechanisms underlying the disease have however not yet been identified. During the five years of my thesis, we were interested in understanding the cellular consequences of these mutations and in getting a better understanding of the physiological function of CMG2. I first showed that most of the HFS mutations in the vWA domain lead to alterations in folding of the domain, either in terms of kinetics or of structure, leading to recognition of the protein by the ER quality control (ERQC) and subsequent degradation by the ERAD pathway. A direct consequence of this retention is the loss of function at the cell surface or other cellular sites. In a second study, I studied the effect of novel HFS mutations that our collaborators and we identified in the CMG2 Ig-like domain. Since the structure of this domain was uncharacterized, we first generated a model of the domain and found that it has an Ig-like fold, despite the very low sequence homology, and harbors two disulfide bonds essential for folding and stabilization of the domain. HFS mutations localized to this domain led to aberrant intermolecular disulfide binding, ER retention and enhanced ERAD when compared to the vWA domain HFS mutations as shown in patients fibroblasts. The consequential loss of CMG2 could be partly alleviated by treating cells with proteasome inhibitor, such as Bortezomib, illustrating that ERAD components are potential therapeutic targets for HFS. As most of the HFS mutations in the ectodomain of CMG2 trigger this ER retention during the CMG2 biosynthesis, we then addressed this process by studying one of the most described post-translational modifications occurring in the ER for protein folding, namely the N-glycosylation. We found that the two potential CMG2 N-glycosylation sites, N250 and N260, are both occupied. Neither of the two sites is essential for ER exit. Glycosylation of CMG2 on N260 is however important since the expression levels of the N260A mutant were significant lower that for the WT protein, despite equivalent or higher synthesis, suggesting that glycosylation at this site is important for efficient folding. Finally, in order to better understand the CMG2 function, we investigated what the physiological ligands could be, assuming it is a receptor as suggested by its role in anthrax infection. We identified type VI collagen as a potential ligand in-vitro that activates CMG2 in-vivo, by phosphorylation of CMG2 cytoplasmic tail. This CMG2 activation upon type VI collagen binding could be a signal for ligand-triggered endocytosis, as is has been described with the protective antigen (PA) of anthrax toxin. The results of this work allow us to gain insights in the molecular mechanisms underlying HFS. Our studies indicate that the identification of the genotype-phenotype correlation of this disease is important for proper diagnosis and clinical approaches, and also that a therapeutic strategy based on proteasome inhibitors, such as Bortezomib, could be envisioned to treat some HFS symptoms in order to improve the life conditions of the patients.