Therapeutic angiogenesis: controlled delivery of angiogenic factors

Abstract

Therapeutic angiogenesis aims at treating ischemic diseases by generating new blood vessels from existing vasculature. It relies on delivery of exogenous factors to stimulate neovasculature formation. Current strategies using genes, proteins and cells have demonstrated efficacy in animal models. However, clinical translation of any of the three approaches has proved to be challenging for various reasons. Administration of angiogenic factors is generally considered safe, according to accumulated trials, and offers off-the-shelf availability. However, many hurdles must be overcome before therapeutic angiogenesis can become a true human therapy. This article will highlight protein-based therapeutic angiogenesis, concisely review recent progress and examine critical challenges. We will discuss growth factors that have been widely utilized in promoting angiogenesis and compare their targets and functions. Lastly, since bolus injection of free proteins usually result in poor outcomes, we will focus on controlled release of proteins.

Figure 1. Alginate and hyaluronic acid hydrogels

(A) (i) CD31 staining of tissue sections from ischemic hindlimbs 6 weeks post-injection of VEGF-alginate gels. (ii) Comparison of blood-vessel density: no injection (−−); bolus VEGF (−+); alginate gel without VEGF (+−); or alginate gel with VEGF (++). The statistical significance indicates enhanced vessel formation in the VEGF-alginate gel group. *p <0.02. (B) ELISA measures HGF release from alginate microbeads with or without sulfated alginate. Cumulative release percentage is calculated by dividing the amount of the released HGF at a given time point by the total amount of HGF. (C) Design of a hyaluronic acid (HA) hydrogel and its use in vascularization. Top: Acrylated HA (AHA) hydrogel. (i) AHA polymers are first modified with and stromal cell derived-a RGD-containing cell-adhesion peptide followed by mixing with a GF containing VEGF₁₆₅, FGF2, TNF-α factor-1. (ii) An AHA hydrogel is obtained by crosslinking AHA polymers with MMP-sensitive linkers. Bottom: (i) ECFCs are encapsulated in the AHA hydrogel containing multiple biological signals. (ii) ECFCs in the AHA hydrogel form vacuoles 3–6 h postencapsulation. Continuing to grow in size, vacuolated ECFCs coalesce to form large lumens at day 1. (iii) From day 2, the ECFC tube structure forms via sprouting and branching, and it is accompanied by degradation of the AHA hydrogel. (iv) By day 3, a stable vascular network is observed in the AHA hydrogel. ECFC: Endothelial colony-forming cells; GF: Growth factor mixture; MMP: Matrix metalloproteinase. (A) Reprinted with permission from [83] © Wiley (2007). (B) Reprinted with permission from [84] © Elsevier (2010). (C) Reprinted with permission from [91] © American Society of Hematology (2011).

Figure 2. Collagen, fibrin and polyethylene glycol hydrogels

(A) Injection of VEGF or the CBD-modified VEGF to the infarcted heart. (i) VEGF or CBD-VEGF was injected in the border zone indicated by dots. (ii) 3 h later, Western blot assessed the exogenous level of VEGF. (iii) Quantitative comparison revealed that CBD-VEGF was better retained in the border zones. **p <0.01. (B) Plasmin catalyzes degradation of fibrin gels as detected by reduction of absorbance at 405 nm. (i) Aprotinin conjugation inhibits fibrin gel degradation. Degradation of fibrin gels (1 mg/ml) is performed at plasmin (100 nM) containing 5 nM (● soluble or ○ aprotinin-conjugated) or 10 nM (◆ soluble or ◇ aprotinin-conjugated) of fibrinogen. Control: fibrin (▲) and 30 nM of aprotinin-conjugated fibrin (△) not subjected to plasmin; plasmin contained no aprotinin (■). (ii) Concentration effect in plasmin inhibition. Fibrin gels are degraded by plasmin containing 10 nM (◆ soluble or ◇ aprotinin-conjugated), 30 nM (● soluble or ○ aprotinin-conjugated) or 50 nM (■ soluble or □ aprotinin-conjugated) of fibrinogen. (C) Design of a PEG-based bioartificial matrix. The MMP-degradable sequence, adehesive ligands and growth factors are modified with PEG-acrylate groups. Photopolymerization crosslinks PEGylated precursors and generates in a hydrogel with multiple signals. CBD: Collagen-binding domain; MMP: Matrix metalloproteinase; PEG: Poly(ethylene glycol). (A) Reprinted with permission from [95] © American Heart Association (2009). (B) Reprinted with permission from [110] © American Chemical Society (2007). (C) Reprinted with permission from [119] © National Academy of Sciences, USA (2010).

Figure 2. Collagen, fibrin and polyethylene glycol hydrogels

(A) Injection of VEGF or the CBD-modified VEGF to the infarcted heart. (i) VEGF or CBD-VEGF was injected in the border zone indicated by dots. (ii) 3 h later, Western blot assessed the exogenous level of VEGF. (iii) Quantitative comparison revealed that CBD-VEGF was better retained in the border zones. **p <0.01. (B) Plasmin catalyzes degradation of fibrin gels as detected by reduction of absorbance at 405 nm. (i) Aprotinin conjugation inhibits fibrin gel degradation. Degradation of fibrin gels (1 mg/ml) is performed at plasmin (100 nM) containing 5 nM (● soluble or ○ aprotinin-conjugated) or 10 nM (◆ soluble or ◇ aprotinin-conjugated) of fibrinogen. Control: fibrin (▲) and 30 nM of aprotinin-conjugated fibrin (△) not subjected to plasmin; plasmin contained no aprotinin (■). (ii) Concentration effect in plasmin inhibition. Fibrin gels are degraded by plasmin containing 10 nM (◆ soluble or ◇ aprotinin-conjugated), 30 nM (● soluble or ○ aprotinin-conjugated) or 50 nM (■ soluble or □ aprotinin-conjugated) of fibrinogen. (C) Design of a PEG-based bioartificial matrix. The MMP-degradable sequence, adehesive ligands and growth factors are modified with PEG-acrylate groups. Photopolymerization crosslinks PEGylated precursors and generates in a hydrogel with multiple signals. CBD: Collagen-binding domain; MMP: Matrix metalloproteinase; PEG: Poly(ethylene glycol). (A) Reprinted with permission from [95] © American Heart Association (2009). (B) Reprinted with permission from [110] © American Chemical Society (2007). (C) Reprinted with permission from [119] © National Academy of Sciences, USA (2010).

Figure 3. Peptide nanofibers

(A) Injectable peptide NF delivers PDGF-BB in a controlled fashion. (i) The binding capacity of peptide NFs to PBS, BSA, PDGF-BB, VEGF, FGF2 or Ang-1 was determined by ELISA assays. (ii) PDGF-BB was injected alone or with peptide NFs in the border zones post-MI. The result suggests that PDGF-BB without peptide NFs was undetectable after 3 days, whereas PDGF-BB delivered by peptide NFs was still detectable after 14 days. **p <0.001. (B) Approach of immobilizing IGF-1 on peptide NFs. (i) IGF-1 and self-assembling peptides are biotinlyated first and then tethered together via tetravalent streptavidin. (ii) Western blot reveals that biotinylated IGF-1 only binds the BPs (b-IGF-1+BP) and not the non-biotinylated peptides (b-IGF-1+P). In addition, the affinity of b-IGF-1 to biotinylated peptides does not change before or after peptides assembly (b-IGF-1+BP premix vs b-IGF-1+BP outside). (C) The release of rhodamine-FGF2 (gray curve) from the PA-heparin gel is significantly slower than that from the PA-Na₂HPO₄ gel (black curve). (D) Laser Doppler imaging compares blood perfusion at ischemic hindlimbs among all treatment groups, VEGF PA, VEGF peptide, mutated VEGF PA and saline control. (i) Perfusion ratios (ischemic hindlimb/non-ischemic hindlimb) reveal that the VEGF PA group has significantly higher perfusion than any other groups from day 14 to 28. (ii) Laser Doppler images show the perfusion of the same animals from day 0 to 28. BP: Biotinylated self-assembling peptide; BSA: Bovine serum albumin; MI: Myocardial infarction; NF: Nanofiber; NS: No significance; PA: Peptide amphiphile; PBS: Phosphate-buffered saline. (A) Reprinted with permission from [128] © American Society for Clinical Investigation (2006). (B) Reprinted with permission from [129] © National Academy of Sciences, USA (2006). (C) Reprinted with permission from [132] © American Chemical Society. .(D) Reprinted with permission from [134] © National Academy of Sciences, USA (2011).

Figure 3. Peptide nanofibers

(A) Injectable peptide NF delivers PDGF-BB in a controlled fashion. (i) The binding capacity of peptide NFs to PBS, BSA, PDGF-BB, VEGF, FGF2 or Ang-1 was determined by ELISA assays. (ii) PDGF-BB was injected alone or with peptide NFs in the border zones post-MI. The result suggests that PDGF-BB without peptide NFs was undetectable after 3 days, whereas PDGF-BB delivered by peptide NFs was still detectable after 14 days. **p <0.001. (B) Approach of immobilizing IGF-1 on peptide NFs. (i) IGF-1 and self-assembling peptides are biotinlyated first and then tethered together via tetravalent streptavidin. (ii) Western blot reveals that biotinylated IGF-1 only binds the BPs (b-IGF-1+BP) and not the non-biotinylated peptides (b-IGF-1+P). In addition, the affinity of b-IGF-1 to biotinylated peptides does not change before or after peptides assembly (b-IGF-1+BP premix vs b-IGF-1+BP outside). (C) The release of rhodamine-FGF2 (gray curve) from the PA-heparin gel is significantly slower than that from the PA-Na₂HPO₄ gel (black curve). (D) Laser Doppler imaging compares blood perfusion at ischemic hindlimbs among all treatment groups, VEGF PA, VEGF peptide, mutated VEGF PA and saline control. (i) Perfusion ratios (ischemic hindlimb/non-ischemic hindlimb) reveal that the VEGF PA group has significantly higher perfusion than any other groups from day 14 to 28. (ii) Laser Doppler images show the perfusion of the same animals from day 0 to 28. BP: Biotinylated self-assembling peptide; BSA: Bovine serum albumin; MI: Myocardial infarction; NF: Nanofiber; NS: No significance; PA: Peptide amphiphile; PBS: Phosphate-buffered saline. (A) Reprinted with permission from [128] © American Society for Clinical Investigation (2006). (B) Reprinted with permission from [129] © National Academy of Sciences, USA (2006). (C) Reprinted with permission from [132] © American Chemical Society. .(D) Reprinted with permission from [134] © National Academy of Sciences, USA (2011).

Figure 4. Representative microparticles

(A) Poly(lactic acid-co-glycolic acid) (PLGA) scaffolds enable different release kinetics of growth factors. VEGF is incorporated into a porous PLGA scaffold by mixing with PLGA followed by processing into a scaffold. This approach results in VEGF absorption on the scaffold. On the other hand, PDGF is incorporated by pre-encapsulation of PDGF into PLGA microspheres followed by processing into a scaffold. Compared with VEGF, PDGF has a significantly slower release kinetic. These two approaches can be combined to deliver dual factors with distinct kinetics. (B) Implantation of PLGA microparticles that release VEGF, HGF, and Ang-1 is combined with intravenous injection of ECFCs to enhance neovascularization in mice with hindlimb ischemia. (i) 14 days postadministration, relative blood flow (ischemic/nonischemic hindlimb) assesses vascular function of each group. (ii) Top: necrosis in toes has disappeared in the group receiving growth factor-carrying microparticles and ECFCs. Bottom: fluorescent imaging by indocyanine green injection compares blood perfusion in the toes. Red and yellow indicate high and low perfusion, respectively. (C) Two methods determine effect of VEGF-containing liposomes in angiogenesis post-MI. Panels A–C: 3,3′-diheptyloxacarbocyanine iodide (green) stains perfused vessels. Panels D–F: CD31 staining (brown) labels endothelial cells. Both results reveal that MI decreases the quantity of blood vessels and VEGF-containing liposomes are able to induce neovasculature. ECFC: Endothelial colony-forming cells; MI: Myocardial infarction; PBS: Phosphate-buffered saline. (A) Reprinted with permission from [148] © Nature Publishing Group (2001). (B) Reprinted with permission from [151] © American Heart Association (2010). (C) Reprinted with permission from [167] © Federation of American Societies for Experimental Biology (2009).

Figure 4. Representative microparticles

(A) Poly(lactic acid-co-glycolic acid) (PLGA) scaffolds enable different release kinetics of growth factors. VEGF is incorporated into a porous PLGA scaffold by mixing with PLGA followed by processing into a scaffold. This approach results in VEGF absorption on the scaffold. On the other hand, PDGF is incorporated by pre-encapsulation of PDGF into PLGA microspheres followed by processing into a scaffold. Compared with VEGF, PDGF has a significantly slower release kinetic. These two approaches can be combined to deliver dual factors with distinct kinetics. (B) Implantation of PLGA microparticles that release VEGF, HGF, and Ang-1 is combined with intravenous injection of ECFCs to enhance neovascularization in mice with hindlimb ischemia. (i) 14 days postadministration, relative blood flow (ischemic/nonischemic hindlimb) assesses vascular function of each group. (ii) Top: necrosis in toes has disappeared in the group receiving growth factor-carrying microparticles and ECFCs. Bottom: fluorescent imaging by indocyanine green injection compares blood perfusion in the toes. Red and yellow indicate high and low perfusion, respectively. (C) Two methods determine effect of VEGF-containing liposomes in angiogenesis post-MI. Panels A–C: 3,3′-diheptyloxacarbocyanine iodide (green) stains perfused vessels. Panels D–F: CD31 staining (brown) labels endothelial cells. Both results reveal that MI decreases the quantity of blood vessels and VEGF-containing liposomes are able to induce neovasculature. ECFC: Endothelial colony-forming cells; MI: Myocardial infarction; PBS: Phosphate-buffered saline. (A) Reprinted with permission from [148] © Nature Publishing Group (2001). (B) Reprinted with permission from [151] © American Heart Association (2010). (C) Reprinted with permission from [167] © Federation of American Societies for Experimental Biology (2009).

Figure 5. Coacervate: design and application

(A) (i) The crystal structure of the FGF–heparin–FGFR complex. The proteins are shown as coils and heparin as a stick model. The heparin-binding domains of FGFR and FGF are highlighted in pink and yellow, respectively. Both analyses showed that the heparin-binding regions contain a high density of positively charged amino acid residues such as arginine. (ii) A possible model of the matrix formed by ionic interactions between an arginine-based synthetic polycation and a heparin–growth factor complex. (B) Comparison of the number of blood vessels in a given size range between free and coacervate FGF2 groups as previously described; the value represents the cumulative number of all the slides examined. The coacervate induced more blood vessel formation than free FGF2. Furthermore, the coacervate group contained more large vessels (>1000 μm², likely associated with arterioles and venules). PEAD: Poly(ethylene argininylaspartate diglyceride). Reprinted with permission from [172] © National Academy of Sciences, USA (2011)