Mapping of epitopes in discoidin domain receptor 1 critical for collagen binding.

The binding and activation of the discoidin domain receptor 1 by collagen has led to the conclusion that proteins from the extracellular matrix can directly induce receptor tyrosine kinase-mediated signaling cascades. A region in the extracellular domain of DDR1 homologous to the Dictyostelium discoideum protein discoidin-I is also present in the secreted human protein RS1. Mutations in RS1 cause retinoschisis, a genetic disorder characterized by ablation of the retina. By introducing point mutations into the discoidin domain of DDR1 at positions homologous to the retinoschisis mutations, ligand binding epitopes in the discoidin domain of DDR1 were mapped. Surprisingly, some residues only affected receptor phosphorylation, whereas others influenced both collagen-binding and receptor activation. Furthermore, two truncated DDR1 variants, lacking either the discoidin domain or the stalk region between the discoidin and transmembrane domain, were generated. We showed that (i) the discoidin domain was necessary and sufficient for collagen binding, (ii) only the region between discoidin and transmembrane domain was glycosylated, and (iii) the entire extracellular domain was essential for transmembrane signaling. Using these results, we were able to predict key sites in the collagen-binding epitope of DDR1 and to suggest a potential mechanism of signaling.

A soluble recombinant multimeric anti-Rh(D) single-chain Fv/CR1 molecule restores the immune complex binding ability of CR1-deficient erythrocytes.

CR1 (CD35, the C3b/C4b receptor) is a widely distributed membrane glycoprotein with a unique cluster conformation on the surface of erythrocytes (E). CR1 on E is responsible for the transport of immune complexes (IC) to liver and spleen. As a cofactor of the C3b cleavage by factor I, CR1 is also a potent inhibitor of C activation and inflammation. In some diseases (systemic lupus erythematosus, hemolytic anemia, AIDS, etc.) an acquired low level of CR1 on E has been observed, leading to an impaired clearance of IC. The aim of this study was to design a heterofunctional molecule that will bind to E and restore a normal or a supranormal CR1 density on E that could mimic the unique distribution pattern of CR1 on normal E. For that purpose a new multimerizing system based on the properties of the C-terminal part of the alpha-chain of the C4 binding protein (C4bp) was used. We first produced a multimeric soluble CR1 that proved to be a better inhibitor of in vitro C activation than the monomeric form of CR1, then a heteromultimeric molecule made of CR1 and single-chain Fv anti-Rh(D) valences able to attach E and providing E with as much as a 10-fold increase in CR1 density with the same CR1 distribution pattern as native E. CR1/single-chain Fv anti-Rh(D)-treated E were able in vitro to attach as many opsonized IC as native E. These data open the way for future use of multimeric and heteromultimeric forms of soluble recombinant CR1 as therapy of IC diseases.

No quantitative relationship between CR1 and Lutheran expression on erythrocytes: In(Lu) gene product is not a common regulator of CR1 expression on erythrocytes.

The density of CR1, the C3b/C4b receptor (CD35), on erythrocytes (E) (CR1/E) is genetically determined. However, the broad distribution of CR1/E within a given genotype suggests that other genetic elements might contribute to the regulation of CR1/E. In some pathological conditions, including systemic lupus erythematosus (SLE), AIDS and hemolytic anemia, CR1 deficiency parallels the severity of the disease. When compared to healthy individuals, an accelerated decrease in CR1/E in these patients has been demonstrated, but other mechanisms interfering with CR1 density regulation during erythropoiesis might also contribute. In exceptional circumstances, CR1/E can be dramatically decreased in healthy individuals by the effect of a regulatory gene, In(Lu), that switches off various surface molecules on E, the structure genes of which are located on four different chromosomes, suggesting a transcription regulatory role for In(Lu) gene products. The hypothesis that products of this gene could physiologically regulate the surface density of all these molecules has been tested by determining Lub density on E (Lub/E) using quantitative flow cytometry. Lub antigenic sites were then compared to CR1/E among healthy individuals of the different CR1 density phenotypes, SLE patients with and without CR1 deficiency, and an exceptional SLE patient totally lacking CR1/E and reticulocytes. No quantitative relationship was found between CR1 and Lub expression in either normal or pathological conditions. These data establish that In(Lu) products are not involved in normal or pathological CR1 density regulation.

In-vivo delivery of therapeutic proteins by genetically-modified cells: comparison of organoids and human serum albumin alginate-coated beads.

We have designed a self-assembling multimeric soluble CD4 molecule by inserting the C-terminal fragment of the alpha chain of human C4-binding protein (C4bp alpha) at the C-terminal end of human soluble CD4 genes. This CD4-C4bp alpha fusion protein (sMulti-CD4) and two other reference molecules, a fusion protein of human serum albumin (HSA) and the first two domains of CD4 (HSA-CD4) and monomeric soluble CD4 (sMono-CD4), were delivered in vivo by genetically modified 293 cells. These cells were implanted in mice as organoids and also encapsulated in HSA alginate-coated beads. sMulti-CD4 showed an apparent molecular weight of about 300-350 kDa, in accordance with a possible heptamer formula. sMulti-CD4 produced either in cell culture or in vivo in mice appeared to be a better invitro inhibitor of HIV infection than sMono-CD4. Plasma levels of sMulti-CD4, HSA-CD4, and sMono-CD4 reached approximately 2,300, 2,700, and 170 ng/mL, respectively, 13 weeks after in-vivo organoid implantation, which had formed tumours at that time. This suggests that the plasma half-life of sMulti-CD4 is much longer than that of sMono-CD4. The 293 xenogeneic cells encapsulated in HSA alginate-coated beads remained alive and kept secreting sMono-CD4 or HSA-CD4 continuously at significant levels for 18 weeks in nude mice, without tumour formation. When implanted in immunocompetent Balb/c mice, they were rejected two to three weeks after implantation. In contrast, encapsulated BL4 hybridoma cells remained alive and kept secreting BL4 anti-CD4 mAb for at least four weeks in Balb/c mice. These results suggest the clinical potential of the C4bp-multimerizing system, which could improve both the biological activity and the poor in-vivo pharmacokinetic performance of a monomeric functional protein like soluble CD4. These data also show that a systemic delivery of therapeutic proteins, including immunoglobulins, can be obtained by the in-vivo implantation of engineered allogeneic cells encapsulated in HSA alginate-coated beads.

A recombinant human scFv anti-Rh(D) antibody with multiple valences using a C-terminal fragment of C4-binding protein.

Monomeric recombinant molecules prove generally unsatisfactory for in vivo use. Most biological systems are indeed multivalent either structurally, associating different chains, or functionally, when cross-linked by their ligands. Mimicking natural molecules for immune intervention implies the need for multimerizing systems to create multivalent molecules capable of interfering with physiological processing. A multivalent anti-Rh(D) recombinant protein has been designed by reconstructing the antibody binding site of a human monoclonal anti-Rh(D) antibody as a single chain Fv mini antibody, then multimerizing it by inserting at its C-terminal end the C-terminal part of the C4 binding protein (C4bp) alpha chain, which is responsible for the octamer multimerization of that molecule. This soluble multivalent recombinant molecule was functional, bound red blood cells (RBCs), agglutinated them, and did not activate complement. This demonstration model opens the way for future in vivo use of multivalent molecules associating antibody valences and other functional molecules for cell targeting, imaging, or removal of cells such as Rh(D)-positive RBCs for preventing Rh alloimmunization.

Catabolism of the human erythrocyte C3b/C4b receptor (CR1, CD35): vesiculation and/or proteolysis?

Human erythrocytes (E) react by exocytosis of membrane vesicles to various stresses including the fixation of the membrane attack complex of Complement. E from normal individuals loose a notable proportion of their initial number of surface CR1 molecules during the ageing process. An acquired decrease of CR1 on E also occurs in pathological conditions such as Systemic Lupus Erythematosus or AIDS. The present study investigated whether calcium ionophore A23187 (Ca-ion) induced vesicle formation of human E in vitro is responsible for a preferential loss of CR1 as well as whether CR1 molecules at the surface of Ca-ion treated E or vesicles are: (i) functional, (ii) native or protease degraded, or (iii) more clustered than CR1 on native E. A study of E from 137 normal individuals showed that a one-hour Ca-ion induced vesicle formation preferentially removed one third of E surface CR1. Kinetic experiments suggested that all surface CR1 could be removed from E upon longer incubation times. CR1 molecules on vesicles were still able to inhibit Complement activation, and were found in larger clusters than on native E. These data suggest that a significant part of surface CR1 molecules may be removed from E by vesicle formation during the life of E in normal individuals. This phenomenon could be exacerbated in pathological conditions.