(A-7) Applying Directed Evolution and Rational Design to Generate Affinity-Based Delivery Systems for Angiogenic Growth Factors

Introduction:: Complex musculoskeletal injuries can involve damage to multiple tissue types. Angiogenesis, the formation of new vascular tissue, is integral to the process of wound recovery, as it promotes the delivery of oxygenated blood and nutrients to the injury site. The angiogenic cell signaling cascade has been well-characterized, requiring sequential presentation of growth factors. Within this signaling cascade, three therapeutic growth factors are known to be critical: vascular endothelial growth factor (VEGF), fibroblast growth factor-2 (FGF-2), and platelet derived growth factor (PDGF). In the case of a critically-sized wound, the presentation of these growth factors at physiological concentrations can be altered or inhibited, resulting in a poor healing outcome.
Current strategies to reconstitute the angiogenic signaling cascade have applied various methods for delivering growth factors at the wound site. Among these approaches, hydrogel-mediated protein delivery provides an attractive approach for accelerating regenerative cell signaling cascades. However, hydrogels can suffer from uncontrolled burst release of protein payloads, limiting their therapeutic potential as delivery vehicles. To address this, small protein binders have shown efficacy for controlling rates of protein release from hydrogel scaffolds. As such, the goal of this work was to develop mini-protein binders called affibodies for conjugation to hydrogels as affinity-based delivery systems to achieve sequential, rate-controlled delivery of VEGF, FGF-2, and PDGF.


Materials and Methods:: A yeast surface display library of affibody variants was subjected to magnetic- and fluorescence-activated cell sorting to identify affibodies for VEGF, FGF-2, and PDGF with variable affinities characterized by flow cytometry. Computational protein modeling tools Alphafold2, ZDOCK, and PyMol were used to predict affibody folding and binding. Site-directed mutagenesis and Rosetta computational design protocols were applied to generate variable affinity affibodies for PDGF and VEGF in silico. Scoring metrics for designed affibody-growth factor interactions were analyzed using GoogleColab packages, and bound structures were visually inspected. Affibody sequences from directed evolution and rational design were appended with an N-terminal adjacent hexa-histidine tag and cysteine and cloned into BL21 DE3 E. coli for protein expression. Soluble affibodies were purified using metal affinity chromatography and size exclusion chromatography. Rate constants (i.e., k_off, k_on, K_D) for affibody-growth factor interactions were determined by biolayer interferometry using 6-point concentration gradients of either soluble growth factor or affibody to validate dissociation constants (K_D) obtained from flow cytometry. Affibodies were conjugated into 5% w/v polyethylene glycol-maleimide (PEG-mal, 4-arm, 20KDa) hydrogels using thiolene click chemistry and crosslinked with dithiothreitol (DTT). Hydrogels were loaded with growth factors at a 500:1 affibody:growth-factor stoichiometric ratio. Protein release was performed at 37 ℃ over 21 days and quantified using enzyme linked immunosorbent assays.


Results, Conclusions, and Discussions:: Results: Seven unique affibodies with different affinities for VEGF, FGF-2, and PDGF were initially discovered using yeast surface display and were characterized by flow cytometry (VEGF: K_{D, high}=58.3 nM, K_{D, medium}=307 nM, K_{D, low}=6470 nM; FGF-2: K_{D, high}=3.08 nM, K_{D, medium}=121 nM, K_{D, low}=4550 nM; PDGF: K_{D, medium}=855 nM) (Fig.1A). ZDOCK and Rosetta computational modeling suggest that PDGF affibodies bind to PDGF at the PDGF receptor binding epitope and VEGF affibodies bind to VEGF at the VEGFR epitope (Fig.1B). This suggests that VEGF and PDGF affibodies may function as inhibitors for VEGF and PDGF signaling pathways. Three unique low affinity affibodies for VEGF were rationally designed from a starting high affinity affibody by site-directed mutagenesis and characterized by flow cytometry (K_{D, D28A}=1835 nM, K_{D, D32A}=4186 nM, K_{D, D36A}=109 nM) (Fig.1C). Two unique high affinity affibodies for PDGF were generated using Rosetta rational design and characterized using biolayer interferometry (K_{D, 0010}=6.5 nM, K_{D, 0057}=1.7 nM) (Fig.1D). Conjugation of affibodies to PEG-mal hydrogels decreased angiogenic growth factor release as compared to hydrogels without affibodies. PDGF high affinity affibody-conjugated hydrogels decreased growth factor release compared to medium, low, and no affibody hydrogels (Fig.1E).
Conclusions: We have demonstrated that VEGF, FGF-2, and PDGF specific affibodies discovered by yeast surface display and rational design demonstrate a range of affinities values capable of tuning growth factor release rates from hydrogel. We have demonstrated the efficacy of site-directed mutagenesis and Rosetta rational design approaches in engineering affibodies for target proteins across a micromolar to nanomolar affinity range. ZDOCK and Rosetta docking protocols may indicate the therapeutic potential of VEGF and PDGF affibodies as antagonists through binding at the receptor epitopes. This work demonstrates the efficacy of both directed evolution and rational design approaches in engineering affinity-based binders as controlled release systems for angiogenic growth factors.
Discussions: Our future work will probe the potential of affibodies to locally modulate the bioactivity of angiogenic growth factors, demonstrate controlled release of multiple growth factors from a single hydrogel, and investigate amplified vascular network formation using microvascular fragments as an in vitro model of angiogenesis.


Acknowledgements (Optional): :

References (Optional): :