Regenerative medicine aims to repair body tissues when natural mechanisms fail. Growth factors (GFs) are essential proteins that promote tissue regeneration upon injury, and as such, are very promising therapeutic drugs in regenerative medicine. Due to their strong therapeutic potential, many GFs have been tested in clinical trials, but most of them were not effective enough to be translated into clinical therapies. On the other hand, the few that are used in the clinics have often been associated with serious life-threatening side-effects. Both the lack of efficacy of GFs and their safety-related issues are the consequences of poor pharmacodynamic properties; indeed, GFs are quickly cleared from the treated area upon bolus delivery. Physiologically, GFs are not only secreted by cells into wounds, but they are also sequestered there by the extracellular matrix (ECM). The ECM then control the spatiotemporal release of GFs, which makes them highly effective at very low dose and allows proper tissue morphogenesis, in contrast to the clinically delivered GFs. Therefore, in this thesis, we developed a new approach to improve the pharmacodynamic properties of GFs by engineering them for super-affinity to the ECM, thus making them more effective and safe for clinical application. We first discovered a domain in the placental growth factor-2, called PlGF-2ECM, that naturally display an exceptional affinity to various components of the endogenous ECM. By fusing this domain to GFs that bear clinical translation limitations, we proved that PlGF-2ECM-engineered GFs greatly enhanced tissue healing and were more effective than wild-type ones at low doses. In total, we applied this protein engineering concept to five different GFs and one protease inhibitor, and evaluated them in multiple disease models that require tissue regeneration, notably in skin chronic wound, critical-size bone defects, limb ischemia, osteoarthritic cartilage and inner ear degeneration. In all these models, we consistently found that PlGF-2ECM-engineered GFs were highly and locally retained in the ECM compared to wild-type GFs. This increased retention translated into significant enhancement of skin and bone regeneration, and trends toward improved recovery after limb ischemia and slower cartilage degeneration in a model of knee osteoarthritis. In addition, we adapted our technology to the inner ear, and implemented novel GF-delivery nanosystems by sequestering PlGF-2ECM-engineered GFs onto ECM-mimetic nanoparticles. In contrast to most of GF-delivery systems that rely on biomaterials, our technology is particularly versatile since super-affinity proteins were successfully delivered within natural and synthetic biomaterials, or even directly into the endogenous ECM, as a carrier-free system. Although biomaterials might be beneficial from a tissue regeneration perspective, carrier-free delivery is more translatable into the clinics. Therefore, we also performed a pre-formulation study of the PlGF-2ECM-engineered GFs for topical delivery in diabetic chronic wounds, pushing our technology toward a specific clinical application. We believe that this simple and broadly applicable concept to engineering super-affinity therapeutic proteins will have a strong impact in several regenerative medicine applications.