Dominique André Rothenfluh
Osteoarthritis (OA) is one of the most common causes of pain and disability. Currently, non-surgical treatment is limited to analgesic and anti-inflammatory drugs, which are not capable of either modifying the disease process or regenerate damaged parts in articular cartilage. From a clinical point of view, small focal lesions in articular cartilage are frequent incidental findings mainly in the hip and knee but also in other joints during open and arthroscopic joint preserving surgical procedures. There is therefore an obvious demand for (bio)-pharmaceutical treatment options to complement surgical procedures in order to regenerate damaged parts in articular cartilage and generally attenuate disease progression. However, pharmaceutical approaches are generally associated with low bioavailabilities of the therapeutic molecules in the cartilage matrix, the primary site of the disease process in OA. The low bioavailability is due to the fact that the joint assumes a remote location being connected to the systemic circulation via the synovial fluid, which is an ultrafiltrate of blood secreted into the joint space. Articular cartilage, i.e. the lining of the joint surfaces providing nearly frictionless motion, is avascular and therefore not connected to the systemic circulation at all and thus depends on nutrients from the synovial fluid being transported to the cells, i.e. chondrocytes, by convective transport, which is inherently impaired due to its dense extracellular matrix. In addition, small molecules which may end up in the joint space via the synovial fluid systemically or after local administration by intra-articular injection are rapidly cleared out of the joint through lymphatic uptake. In order for (bio)-pharmaceutical treatment options to be efficient and effective, the therapeutic needs to reach the site of the disease process, here articular cartilage, and stay there long enough to exert its desired function. At first, we sought to engineer a nanoparticle-based drug delivery system which is capable of targeting the cartilage matrix and is subsequently immobilized there to release its cargo over time. Towards this end, we used a technique called peptide-on-phage display to carry out rounds of affinity selection against bovine cartilage in order to discover potential 6-mer peptides as bioaffinity ligands to a cartilage extracellular matrix component. Phage display relies on the genotype-phenotype linkage of inserts in the gene coding for minor coat protein of the phage virion, which is then displayed as a fusion protein on its surface. The sequence displayed by recovered binding phages can therefore be determined by sequencing of the phage DNA. After five rounds of affinity selection, i.e. recovering the binding phages, amplifying and re-exposing them to bovine articular cartilage, the peptide sequence WYRGRL could be discovered, which appeared in 94 out of 96 sequenced phage clones. WYRGRL was shown to bind specifically to articular cartilage and not to the synovial membrane. In addition, adding synovial fluid does not impair its capability of binding to articular cartilage. In a competetive binding assay of phages displaying WYRGRL and its synthesized free peptide an IC50 of binding to articular cartilage of 140 nM could be calculated. The free peptide was subsequently immobilized in a thiopropyl sepharose column for affinity purification in order to identify the binding partner of WYRGRL in the cartilage extracellular matrix. Affinity purification yielded one single protein band in SDS-PAGE which was identified to be collagen II alpha1 by in-gel trypsin digestion and analysis of protein fragments by mass-spectrometry. The peptide WYRGRL therefore provides means of immobilizing macromolecules in the collagen II network of articular cartilage. As a proof of principle, the peptide was conjugated to the surface of poly(propylene sulfide) nanoparticles, which were in average between 31 to 38 nm in size. It was covalently bound to the emulsifier for nanoparticle synthesis which remains displayed on the surface after polymerization. To visualize the nanoparticles in vivo by confocal laser scanning microscopy, they were labelled with a fluorophore for visualization. The nanoparticles were injected into the knee joints of mice, which were sacrificed after 24, 48 and 96 h. Compared to control nanoparticles displaying the scrambled peptide sequence, a 44.8-fold higher accumulation of nanoparticles in the extracellular matrix after 24 h could be observed, which was 71.7-fold higher after 48 h and 27.6-fold higher after 96 hours. Conjugation of WYRGRL to nanoparticles therefore efficiently immobilized them in the extracellular matrix of articular cartilage, an effect that was not present in nanoparticles with a scrambled peptide sequence attached to their surface. Nanoparticles could also be observed in chondrocytes in vivo, although to a lesser extent than in the extracellular matrix. In the light of potential adverse effects on cells, internalization of nanoparticles in immature murine articular chondrocytes (IMAC) and their effect on cell viability as well as phenotype specific gene-expression was studied. In accordance with the observations in vivo, conjugated and non-conjugated nanoparticles were efficiently internalized by cultured IMACs ranging from 74.8% to 77.5% after 24 h. Nanoparticle internalization was shown to be associated with elevetad rates of cell death and a decrease in the phenotypic gene expression, especially collagen II and aggrecan, but also collagen I, whereas cathepsin B as a de-differentiation marker and lysosomal enzyme was not downregulated. After 96 h, gene expression levels seemed to recover and approach baseline values. At the concentrations used, internalized nanoparticles seemed to be toxic to IMACs and interfere with phenotypic gene expression and potentially alter the cell phenotype, which needs to be investigated further for potential clinical applications. As it was shown that the peptide WYRGRL effectively immobilized poly(propylene sulfide) nanoparticles in the extracellular matrix of articular cartilage, we sought to extend this concept to protein-based therapeutics. Of interest here is FGF-18 which had been shown to be an important trophic factor on chondrocytes, inducing synthesis of matrix molecules. It had also been shown that FGF-18 was capable of repairing focal cartilage lesions in a model of rapidly progressing OA in rats, which required bi-weekly intra-articular injections, however. A fusion protein of human FGF-18 and the collagen II-binding peptide WYRGRL was expressed in E. coli and purified by heparin-affinity chromatography. It was demonstrated in vitro by surface plasmon resonance, that the fusion of FGF-18 and WYRGRL exhibits slower release after binding to immobilized collagen II than wild-type FGF-18. In a model of surgically induced knee joint instability in mice, injection of the fusion protein of FGF-18 at the time of surgery and after 2 weeks prevented the onset of OA in these animals over 5 weeks and demonstrated significantly improved histological scores compared to the wild-type protein and control animals. Enhanced retention of FGF-18 in the cartilage matrix resulted therefore to shift from bi-weekly to less than bi-monthly intra-articular administration of the therapeutic. In conclusion, the present thesis introduces an enhanced retention technology for articular cartilage which has been demonstrated to immobilize nanoparticles and protein-based therapeutics in the cartilage matrix by binding to the collagen II network. The potential reduction of intra-articular injections as a result of immobilizing either the carrier or the therapeutic is of significant clinical relevance and may prove to be pivotal for future (bio)-phamaceutical approaches for cartilage regeneration.
EPFL2009
