검색 상세

Deep-learning-guided design of a VEGFR2 agonistic peptide that promotes angiogenesis and wound repair

초록/요약

Vascular endothelial growth factor (VEGF) is a central regulator of angiogenesis and maintains vascular homeostasis through its interaction with receptor tyrosine kinases, in particular vascular endothelial growth factor receptor 2 (VEGFR2). Therapeutic modulation of VEGF–VEGFR2 signaling offers significant potential in clinical contexts where impaired vascularization underlies disease, such as chronic wounds, ischemic pathologies, and tissue degeneration. However, direct use of VEGF or gene- based strategies faces important limitations that include short half-life, off-target effects, delivery challenges, and limited control over the intensity and duration of angiogenic responses. These constraints have encouraged the search for alternative approaches that retain the proangiogenic benefits of VEGF while reducing safety and delivery problems. In this thesis, a deep-learning-guided strategy is used for the design and evaluation of VEGF Mimetic Peptide 3 (VMP3), a synthetic peptide that acts as a VEGFR2 agonist. Long short-term memory (LSTM) networks guided the extraction of sequence motifs within known VEGF agonist peptides and supported rational design of new candidates. Molecular dynamics simulations (MD) of the VMP3–VEGFR2 complex indicated structural stability and favorable interaction energetics, which supported receptor engagement at the atomic level. In vitro functional assays demonstrated that VMP3 promotes endothelial cell proliferation, migration, and invasion in a pattern that resembles native VEGF. Ex vivo aortic ring assays confirmed robust microvessel sprouting, while in vivo Matrigel plug assays in C57BL/6J mice showed enhanced neovascularization by gross, histological, and immunohistochemical analysis. The proangiogenic effect of VMP3 was further examined in a streptozotocin-induced diabetic mouse wound model. VMP3 treatment accelerated wound closure and improved tissue regeneration relative to disease and vehicle controls. Histological and immunohistochemical evaluation of wound tissue (H&E, Masson’s trichrome, Picrosirius red, and CD31 staining) indicated enhanced re-epithelialization, collagen deposition, and neovascularization across VMP3-treated groups. These findings support the concept that VMP3 functions as a VEGF-mimetic peptide that activates VEGFR2 signaling and promotes angiogenesis and matrix remodeling in a diabetic wound environment. Overall, the thesis combines deep learning, molecular modeling, and multi-level experimental validation to identify VMP3 as a promising VEGFR2 agonistic peptide. The results suggest that such computationally designed peptides may serve as translational candidates for angiogenic therapy in indications that require controlled vascular regeneration, including diabetic wounds and tissue repair settings.

more

목차

Chapter 1 Introduction 1
1.1 Angiogenesis and its biological significance 1
1.2 Pathological angiogenesis and therapeutic opportunities 2
1.3 Limitations of current VEGF-based therapies 2
1.4 Peptide-based VEGFR2 agonists and computational design 3
1.5 Rationale of the present study 4
1.6 Aim and objectives: 4
Chapter 2 Literature Review 6
2.1 Overview of Angiogenesis 6
2.2 VEGF and Its Receptor System 7
2.2.1 The VEGF family 7
2.2.2 Molecular basis of VEGF–VEGFR2 interaction 8
2.3 Angiogenesis in Disease 8
2.3.1 Insufficient angiogenesis 8
2.3.2 Excessive angiogenesis 9
2.3.3 Limitations of VEGF-Based Proangiogenic Therapies 9
2.3.4 Short half-life and instability 9
2.3.5 Lack of spatial and temporal control 9
2.3.6 Off target effects and immunogenicity 9
2.3.7 Delivery challenges 9
2.4 Peptide-Based VEGFR2 Agonists 10
2.4.1 Rationale for peptide therapeutics 10
2.4.2 QK and related peptides 10
2.5 Computational Design of Angiogenic Peptides 11
2.5.1 Machine learning for peptide design 11
2.5.2 Integration of deep learning with molecular dynamics 12
2.6 Ex vivo and In vivo Models for Angiogenesis Research 12
2.6.1 Endothelial cell-based assays. 12
2.6.2 Aortic ring assay 12
2.6.3 Matrigel plug assay 13
2.6.4 Diabetic wound healing models 13
2.7 Gaps Addressed by the Present Study 13
Chapter 3 Materials and Methods 15
3.1 Computational Design of VEGF Mimetic Peptides 15
3.2 LSTM-based deep learning framework 15
3.3 Physicochemical filtering and structural optimization 16
3.4 Molecular Docking and Molecular Dynamics Simulations 17
3.5 Initial receptor–peptide modeling 17
3.6 MD system construction 17
3.7 Energy minimization and equilibration 17
3.8 Production MD simulations 17
3.9 Experimental Materials, Reagents, and Cell Culture 19
3.9.1 In vitro functional assays 19
3.9.2 Assessment of cell viability by MTT assay 19
3.9.3 Evaluation of cell proliferation 20
3.9.4 Scratch wound healing assay 21
3.9.5 Transwell chemotactic migration assay 22
3.9.6 VEGFR2 Activation and Downstream Signaling 22
3.9.7 Western blot analysis of VEGFR2 and signaling intermediates 23
3.9.8 PathScan phospho-VEGFR2 (Tyr1175) ELISA 25
3.9.9 Rationale for dual method signaling analysis 26
Chapter 4 Results 27
4.1 Computational Design and Selection of VMP3 27
4.1.1 LSTM-based sequence generation 27
4.2 Molecular Docking and Molecular Dynamics Simulations 28
4.2.1 Docking of VMP3 to VEGFR2 28
4.2.2 Structural stability during 100 ns MD simulations 28
4.2.3 Extended 300 ns triplicate MD simulations 30
4.3 MM/PBSA binding energy calculations 30
4.4 Thermodynamic Insights into the Interaction Between VMP3 and VEGFR2 30
4.5 In Vitro Evaluation of VMP3 in Endothelial Cells 34
4.5.1 Selection of VMP3 as the Lead VEGF-Mimetic Peptide. 34
4.5.2 Biochemical and Biophysical Characterization of VMP3–VEGFR2 Interactions 37
4.6 Ex vivo Angiogenesis 41
4.6.1 Assessment of Ex vivo Angiogenesis Using the Mouse Aortic Ring Assay 41
4.7 Assessment of In vivo Angiogenesis Using the Mouse Matrigel Plug Assay 44
4.8 In vivo Angiogenesis 47
4.8.1 Pro-angiogenic and Pro-repair Actions of VMP3 in Diabetic Wounds 47
Chapter 5 Discussion 52
5.1 The significance of a deep-learning-guided approach in peptide engineering 52
5.2 Structural stability and receptor engagement of VMP3 53
5.3 Mechanistic insights into VEGFR2 activation by VMP3 53
5.4 Functional validation of VMP3 in endothelial biology 54
5.5 Ex vivo and in vivo angiogenic activity: relevance to physiological environments 55
5.6 Therapeutic potential of VMP3 in diabetic wound healing 55
5.7 Comparison of VMP3 with existing VEGF mimetics 56
5.8 Broader implications and future directions 57
Chapter 6 Conclusion and Future Work 58
6.1 Structural validation and optimization 59
6.2 Pharmacokinetics, biodistribution, and stability 59
6.3 Immunogenicity and safety assessment 60
6.4 Mechanistic profiling beyond VEGFR2 60
6.5 Evaluation in advanced in vivo models 60
6.6 Formulation and delivery strategies 61
6.7 Final Remarks 61
References. 62
Supplementary Material (Appendices) 68

more
반출 Meta View 목록