Identification And Characterization Of A Human Platelet Glycoprotein Vi Peptidomimetic Permitting Molecular Imaging Of Fibrosis

Abstract Glycoprotein VI (GPVI), the main receptor for platelet activation by collagen, has been shown to play an important role in thrombosis, vascular remodelling and atherothrombosis. GPVI which belongs to the immunoglobulin receptor family, binds to fibrillar type I and type III collagens of vascular as well as non-vascular origin. 9O12.2, a high affinity monoclonal antibody directed to the GPVI extracellular domain, blocks GPVI binding to collagen and possess antithrombotic properties (Ohlmann et al J.Thromb. Haemost. 2008,6:1013). We have hypothesized that the 9O12.2 epitope overlaps, at least in part, with the collagen-binding site on GPVI and (ii) that molecules mimicking the 9O12.2 epitope can be expected to be antithrombotic by competing with platelet GPVI for binding to collagen and/or to act as tracers for collagen in vivo. A bacterial random 12 mer cyclic peptide library was screened against the 9O12.2 IgG. Twenty clones were selected. Sequencing the inserts revealed 9 peptidic motifs with 7 identical residues. One sequence was selected to synthesize a biotin-coupled constrained peptide. (designated collagelin). Surface plasmon resonance (SPR) analysis showed that 9O12.2 IgG bound to immobilized collagelin (KD 10−6M) and that binding was inhibited in the presence of soluble recombinant (sr)GPVI or after disulfide bridge reduction as expected for a molecule mimicking the 9O12.2 epitope known to be conformational (Lecut et al. J.Biol.Chem.2004, 279:52293). Using SPR and solid phase assays, we observed that collagelin bound to immobilized fibrillar collagen (KD10−7M) and that binding was inhibited by 9O12.2 IgG and by rsGPVI, indicating that collagelin mimics at least in part the collagen-binding site of GPVI. Collagelin did not inhibit collagen-induced platelet aggregation in vitro. However, histochemical analysis demonstrated that it bound to collagen on sections of rat aortas and of rat tail tendon. We then hypothesized that collagelin could be retained in vivo at sites of collagen accumulation, thus allowing isotopic imaging of fibrosis. Collagelin and a control peptide (same size and cyclic,) were labeled either indirectly using 99mTc-streptavidin or directly with 99mTc and iv injected into rats presenting fibrotic scars of myocardial infarction. Radiolabeled collagelin uptake in fibrosis areas was demonstrated in vivo by planar and tomographic scintigraphy. Mean heart-to-lung ratios were of 2.76±0.36 and 2.08±0.17 for 99m Tc-streptavidin-coupled collagelin and 99m Tc-collagelin respectively. Ex vivo, autoradiography on frozen heart sections showed a clear uptake of labeled-collagelin in the infarct collagen-rich scars with mean scar to remote myocardium activity ratios of 2.52±0.2 and 2.92±.053 for 99mTc-streptavidin-coupled collagelin and 99mTc-collagelin respectively as compared with 1.82±0.32 and 1.61±0.23 for the control peptides (p<0.006 and 0.01). In conclusion, we have produced a peptide which partly mimics the collagen binding site of GPVI, specifically binds to collagen and appears to be a specific tool for direct targeting of collagen in vitro and in vivo. Collagelin or derived molecules thus potentially have a large field of applications, as a tracer of fibrotic lesions, in non-invasive vascular as vascular pathologies.