Supramolecular nanostructures that mimic VEGF as a strategy for ischemic tissue repair

Supramolecular nanostructures that mimic VEGF as a strategy for ischemic tissue repair Matthew J. Webbera,1, Jörn Tongersb,c,1, Christina J. Newcombd, Katja-Theres Marquardtb, Johann Bauersachsc, Douglas W. Losordob,1,2, and Samuel I. Stuppd,e,f,g,1,2 a Department of Biomedical Engineering, Northwestern University, Evanston, IL 60208; bFeinberg Cardiovascular Research Institute, Feinberg School of Medicine, Northwestern University, Chicago, IL 60611; cCardiology and Angiology, Hannover Medical School, 30625 Hannover, Germany; dDepartment of Materials Science and Engineering, Northwestern University, Evanston, IL 60208; eDepartment of Chemistry, Northwestern University, Evanston, IL 60208; f Department of Medicine, Northwestern University, Chicago, IL 60611; and gInstitute for Bionanotechnology in Medicine, Chicago, IL 60611

I

schemic tissue disease remains one of the foremost causes of morbidity and mortality worldwide (1). There is tremendous need for new therapeutic approaches that regenerate ischemic tissue. One target is to enhance microvasculature perfusion in ischemic tissue by delivering proangiogenic signals, termed therapeutic angiogenesis. The mechanisms of angiogenesis have been extensively studied and its regulation involves complex cascades of signaling molecules and growth factors (2). VEGF is among the most potent, yet rate limiting, of these angiogenic factors (3). Thus, efforts toward therapeutic angiogenesis have focused on VEGF to enhance microvasculature in ischemic tissue. However, clinical trials to date have not convincingly demonstrated efficacy (4, 5). One potential obstacle for the success of these therapies is inadequate retention of protein in the target zone because protein retention in tissue is on the order of minutes to hours, depending on delivery route (6–8). Maintaining a therapeutic level of protein within ischemic tissue could require serial treatments over time, making these therapies more invasive and necessitating cost-prohibitive quantities of protein (4). Recent work has demonstrated the therapeutic potential of a platform of self-assembling filamentous nanofibers formed from customizable peptide amphiphile (PA) molecules (9–12). PAs consist of a hydrophobic alkyl segment covalently linked to a peptide sequence that contains at least two domains: an amino acid sequence that drives self-assembly of the molecules into nanofibers through the formation of β-sheets and a customizable bioactive domain designed to interact with specific proteins, receptors, biopolymers, or other cellular targets. Hydrophobic collapse drives the alkyl tails into the core of the nanofiber, resulting in the presentation of the bioactive domain on the fiber surface. www.pnas.org/cgi/doi/10.1073/pnas.1016546108

These nanofibers have dimensions similar to filamentous structures in the extracellular matrix and can form gel networks at low concentrations in aqueous media, allowing for threedimensional entrapment of cells presuspended in aqueous PA solutions (13). Their high aspect ratio and the high surface area of displayed signals at controlled density likely facilitate their enhanced biological signaling, while their extensive internal hydration also offers the necessary dynamics to promote interaction with receptors and ligands (14–16). Furthermore, these PA-based therapies can be delivered noninvasively as easily injectable liquids that become solid nanostructures in situ and are biocompatible with desirable rates of degradation and tissue clearance over a period of weeks (17–19). Advancements in the design of biomaterials have enabled some of the issues of bolus VEGF protein therapy to be overcome through the use of materials to control the spatial and temporal delivery of VEGF (20). These scaffolds have been designed to control the release kinetics through scaffold design (21–23), the use of heparin to specifically bind VEGF (24, 25), or through the covalent attachment of VEGF for on-demand cell-mediated VEGF delivery (26, 27), among many other strategies. The general concept aims to use materials to control the kinetics of protein bioavailability. However, each strategy is reliant on recombinant proteins and, although some are injectable, some require invasive surgical implantation within the site where angiogenesis is to be modulated. In addition, a number of synthetic strategies have emerged to modulate angiogenesis through small molecules or peptides and have been developed to antagonize receptors for cancer therapies or to promote angiogenesis by signaling through these receptors (28, 29). Currently, only one synthetic oligopeptide has been demonstrated to mimic VEGF through the activation of its receptors (30, 31). This peptide was designed based on the native α-helical receptor-binding domain of VEGF and was shown to mimic native VEGF through activation of VEGF receptors. This synthetic approach, although potentially addressing issues related to the production of recombinant proteins, does not address the poor tissue retention for protein-based therapies. Here, the customizable presentation of bioactive signals on our PA platform was leveraged to present this VEGF-mimetic epitope in a polyvalent fashion on the surface of high aspect ratio nanofibers. This strategy could provide a feasible synthetic Author contributions: M.J.W., J.T., J.B., D.W.L., and S.I.S. designed research; M.J.W., J.T., C.J.N., and K.-T.M. performed research; M.J.W. and J.T. analyzed data; and M.J.W., J.T., D.W.L., and S.I.S. wrote the paper. The authors declare no conflict of interest. This article is a PNAS Direct Submission. 1

M.J.W. and J.T. contributed equally to this work.

2

To whom correspondence may be addressed: E-mail: [email protected] or [email protected].

This article contains supporting information online at www.pnas.org/lookup/suppl/ doi:10.1073/pnas.1016546108/-/DCSupplemental.

PNAS Early Edition ∣

1 of 6

ENGINEERING

There is great demand for the development of novel therapies for ischemic cardiovascular disease, a leading cause of morbidity and mortality worldwide. We report here on the development of a completely synthetic cell-free therapy based on peptide amphiphile nanostructures designed to mimic the activity of VEGF, one of the most potent angiogenic signaling proteins. Following selfassembly of peptide amphiphiles, nanoscale filaments form that display on their surfaces a VEGF-mimetic peptide at high density. The VEGF-mimetic filaments were found to induce phosphorylation of VEGF receptors and promote proangiogenic behavior in endothelial cells, indicated by an enhancement in proliferation, survival, and migration in vitro. In a chicken embryo assay, these nanostructures elicited an angiogenic response in the host vasculature. When evaluated in a mouse hind-limb ischemia model, the nanofibers increased tissue perfusion, functional recovery, limb salvage, and treadmill endurance compared to controls, which included the VEGF-mimetic peptide alone. Immunohistological evidence also demonstrated an increase in the density of microcirculation in the ischemic hind limb, suggesting the mechanism of efficacy of this promising potential therapy is linked to the enhanced microcirculatory angiogenesis that results from treatment with these polyvalent VEGF-mimetic nanofibers.

APPLIED BIOLOGICAL SCIENCES

Edited by David A. Tirrell, California Institute of Technology, Pasadena, CA, and approved July 7, 2011 (received for review November 3, 2010)

alternative to protein therapies that would address issues of tissue retention and bioavailability. Moreover, polyvalent epitope display could enhance the dimerization-dependent signaling of the VEGF receptors. Thus, we have evaluated here PA presentation of this VEGF-mimetic sequence and its possible use as therapy for ischemic tissue regeneration. Results Design and Characterization of a VEGF-Mimetic PA. The VEGF PA (Fig. 1A) was designed to display on the surface of nanostructures a peptide sequence that mimics VEGF, KLTWQELYQLKYKGINH2 (30). At the N terminus of this peptide, we covalently attached a K 3 G sequence to facilitate solubility and a V2 A2 β-domain followed by a C16 alkyl chain to promote self-assembly into cylindrical nanostructures through intermolecular hydrogen bonding and hydrophobic collapse (Fig. 1G). Cryogenic transmission electron microscopy (TEM) of this PA reveals the formation of self-assembled high aspect ratio cylindrical nanostructures (Fig. 1B). At a concentration used for therapeutic studies and in the presence of divalent counterions, VEGF PA nanostructures form entangled nanofiber gels (Fig. 1C). The reported bioactive secondary structure of the VEGF-mimetic sequence is α-helical (30, 31). CD of the VEGF PA revealed a signal characteristic of α-helix (Fig. 1D). This α-helix had greater conformational stability when incorporated in PA molecules (Fig. 1E) compared to the peptide alone (Fig. 1F), evident by less change in the 220-nm α-helical CD signature upon heating. Conjugation to an alkyl tail is known to stabilize α-helical peptides (32) and supramolecular effects could also explain the increased thermal stability of the α-helical epitope when presented on the PA. Such stabilization of the bioactive conformation of the peptide by PA conjugation could enhance the potency of the epitope. In addition to the VEGF PA and epitope-only peptide control, a PA was synthesized with systematic replacement of four specific residues

Fig. 1. The chemical structure of the VEGF-mimetic peptide amphiphile (A), designed to assemble into cylindrical nanostructures (G). The VEGF PA forms nanofibers, visualized by cryogenic TEM (B), and entangled nanofiber gel networks, imaged by SEM (C). Circular dichroism for the VEGF PA demonstrating α-helical secondary structure (D), and melting analysis performed about the 220-nm α-helical signature for VEGF PA (E) and the peptide epitope control (F). 2 of 6 ∣

www.pnas.org/cgi/doi/10.1073/pnas.1016546108

known to be near the peptide–receptor-binding interface with structurally distinct amino acids. Structures of all molecules used in this study are shown (Fig. S1). VEGF PA Specifically Activates VEGF Receptors in Vitro. VEGF signal transduction is initiated by phosphorylation of several tyrosines on the intracellular receptor domain (33). In order to determine VEGF-mimetic signaling of the VEGF PA, human umbilical vein endothelial cells (HUVECs) were stimulated with VEGF PA and then quantified for the amount of phosphorylated VEGF receptor 1 (VEGFR1) or phosphorylated VEGF receptor 2 (VEGFR2), the two primary VEGF receptors implicated in its angiogenic signaling. Stimulation with the VEGF PA resulted in an amount of phosphorylated VEGFR1 (Fig. 2A) that was 1.63 times greater than an untreated control group (P < 0.001), showing a significant enhancement in receptor phosphorylation. Interestingly, the level of phosphorylation for the VEGF PA was also significantly greater than for the bioactive VEGF peptide (P < 0.001) and mutant PA (P < 0.001). The bioactive peptide also induced significantly greater phosphorylation than untreated controls. For VEGFR2 phosphorylation (Fig. 2B), the VEGF PA (1.58 times increase) again demonstrated phosphorylation levels significantly greater than an untreated control (P < 0.001). The value for the VEGF peptide was again less than that for the VEGF PA. The mutant PA again demonstrated no effect on VEGFR2 phosphorylation, establishing this as an ideal material control for the VEGF PA. Both the VEGF PA and the VEGF peptide signal similarly to the VEGF assay control for both

Fig. 2. Results from an ELISA assay for the receptor phosphorylation of VEGFR1 (A) and VEGFR2 (B) as well as a time course of phosphorylation for both VEGFR1 and VEGFR2 (C). Significance is shown relative to VEGF PA treatment. VEGF protein is shown as an assay control for verification of phosphorylation and was not included in statistical analysis as a comparative group.

Webber et al.

VEGF PA Induces Angiogenesis in Vivo. An established in vivo angiogenesis model, the chicken chorioallantoic membrane (CAM) assay, was used to evaluate the angiogenic activity of the VEGF PA. When VEGF PA was coated onto a glass coverslip and applied to the CAM (Fig. 3), we saw a 229% increase in the blood vessel density over the following 3 d. For comparison, this increase was significantly greater (P < 0.001) than CAM treated with the VEGF peptide (139%), mutant PA (149%), or saline (132%). This result suggests a strong angiogenic response from treatment with the VEGF PA. This effect is visualized in the density of blood vessels at the point of CAM stimulation, in addition to indications of vascular remodeling and leakage and the spokelike pattern radiating from the center of the coverslip where the material was applied. Representative images from the various controls do not display a similar effect. This assay confirms the angiogenic properties of our VEGF PA using an in vivo model and reinforces proangiogenic findings in vitro. VEGF PA Enhances Repair of Ischemic Hind-Limb Tissue. The murine hind-limb ischemia model was used to evaluate the potential of the VEGF-mimetic PA nanostructures as a therapy for ischemic disease. VEGF PA or control treatments were administered by an intramuscular injection 3 d after the induction of critical ischemia by ligation and excision of the right femoral artery and all superficial and deep branches. To assess functional recovery after critical hind-limb ischemia, animals were assessed for limb salvage and limb motor function via established semiquantitative scoring methods. In terms of tissue necrosis and amputation of ischemic limb, we saw significant improvement (P < 0.05) in Webber et al.

tissue salvage (i.e., less necrosis) in animals treated with VEGF PA (Fig. 4A) at both day 21 and day 28 compared to animals treated with VEGF peptide, mutant PA, and saline. When scoring animals based on active limb motor function (Fig. 4B), we again saw a significant effect for treatment with the VEGF PA at day 21 (P < 0.05) and day 28 (P < 0.01) compared to treatment with the VEGF peptide, mutant PA, and saline. This assessment of limb use suggests that treatment with the VEGF PA leads to a more functional phenotype, which was further supported in results subjecting animals to a walking endurance test using a Rota Rod treadmill at day 28 (Fig. 4C). Animals treated with VEGF PA walked significantly longer prior to failure (150.5 s) than animals treated with the VEGF peptide (115.6 s, P < 0.05), mutant PA (106.7 s, P < 0.05), and saline (90.4 s, P < 0.001). Laser Doppler perfusion imaging (LDPI) was performed to assess tissue perfusion in the ischemic hind limb (Fig. 4D). Treatment with VEGF PA significantly enhanced the recovery of tissue perfusion following treatment. At 14 d after induction of ischemia, animals treated with VEGF PA had a perfusion ratio (0.76) significantly greater than that for animals treated with the VEGF peptide (0.54, P < 0.01), mutant PA (0.42, P < 0.01), or a control injection of saline (0.53, P < 0.05). At day 28, animals treated with VEGF PA continued to have a significantly higher perfusion ratio (0.72) than animals treated with the VEGF peptide (0.52, P < 0.05), mutant PA (0.48, P < 0.05), and saline (0.52, P < 0.05). Histological tracking of fluorescently tagged PA and peptide in muscle tissue of the ischemic hind limb revealed that the PA is retained significantly longer than the peptide control (Fig S3). Harvesting ischemic limb tissue at 2, 7, and 14 d after administration of treatment revealed large quantities of VEGF PA remained in the ischemic hind limb. Even small quantities could be seen at 28 d after treatment. A thorough histological search for fluorescently tagged VEGF peptide did not indicate any material remained in the tissue at any of these follow-up points. It is known PNAS Early Edition ∣

3 of 6

APPLIED BIOLOGICAL SCIENCES

Fig. 3. Quantified results from the CAM assay beginning on embryonic day 10 (t ¼ day 0) and extending for 4 d along with representative images from day 3 for treatments of VEGF PA, VEGF peptide, mutant PA, and an untreated control. Significance shown is for VEGF PA treatment compared to all other treatment groups. The scale bar shown corresponds to 2 mm in all micrographs.

ENGINEERING

receptors, confirming reports on the discovery of the epitope (30). Examining the effect of VEGF PA stimulation over time (Fig. 2C), we observed an initial rise in the levels of both phosphorylated VEGFR1 and VEGFR2, followed by a decrease beginning at 10 min of stimulation to levels below the baseline of an untreated control. The response and time frame of signaling by VEGF PA is consistent with the known temporal response for VEGF receptor activation followed by subsequent cleavage of phosphates and ubiquitination of the receptor (34). VEGF signaling is known to enhance, among other cellular functions, the proliferation, survival, and migration of endothelial cells (3, 33). Therefore we also evaluated these phenotypic outcomes in HUVECs upon exposure to VEGF PA (Fig. S2). Prolonged stimulation with VEGF PA significantly increased cell number 1.37 times compared to an untreated control (P < 0.001). In addition, HUVECs that were serum-starved exhibited significantly improved survival in the presence of the VEGF PA (10.2% apoptotic) compared to an untreated group (31.7% apoptotic, P < 0.001). Neither the VEGF peptide nor the mutant PA group significantly enhanced cell number or survival. In addition, treatment with VEGF PA resulted in a significant increase in migration into a denuded scratch (68.7% closure) compared to an untreated control (25.2% closure, P < 0.05), whereas treatment with the VEGF peptide did not significantly enhance migration relative to the untreated control group. The robust effect on HUVEC function in vitro, combined with VEGF-specific receptor phosphorylation, supports the VEGFmimetic activity of the VEGF PA nanofibers. The VEGF peptide, meanwhile, did not show similar effects on endothelial cell function in vitro in spite of a demonstrated increase in receptor phoshorylation. As the phosphorylation studies were performed for only a single phosphorylation cycle, it is possible that the peptide does not retain potency over longer times in culture. Alternatively, it could be that the incremental increase in phosphorylation by VEGF PA compared to the VEGF peptide seen in one phosphorylation cycle results in a cumulative effect on cell activity over the time course of many phosphorylation cycles afforded by longer functional in vitro experiments.

that proteins, such as VEGF, have short retention times in tissue (6–8). The enhanced retention observed for VEGF PA could underlie its therapeutic benefit compared to the peptide epitope control. In order to determine whether the beneficial effects of VEGF PA treatment on recovery of blood flow, motor function, and tissue salvage were associated with an effect on the microcirculation of the ischemic limb muscle, we quantified the number of CD31þ capillaries in the ischemic limb harvested at day 28 (Fig. 5). There was a significant increase in the number of CD31þ capillaries in animals treated with VEGF PA (1;582∕mm2 ) compared to treatment with VEGF peptide (949∕mm2 , P < 0.001), mutant PA (954∕mm2 , P < 0.001), and saline (893∕mm2 , P < 0.001). This proangiogenic effect of VEGF PA on the microcirculation is consistent with its angiogenic activity in the CAM assay, and reinforces its therapeutic efficacy for ischemic tissue repair. Of note, there was no effect on the number of CD31þ ∕smooth muscle actinþ arterioles, also known as muscularized or mature capillaries. VEGF protein treatment can result in the development of immature vasculature, and this is one known limitation surrounding VEGF therapy (35). We do not know at this time if a lack of mature capillaries will be a limitation in the translation of VEGF PA. However, the versatility of supramolecular PA systems allows for multiplexing of bioactive signals, and future efforts could explore the development of PA mimics of other growth factors (i.e., FGF-2) that, in combination with VEGF signaling, are known to promote more mature vasculature. Overall, the improvements in tissue perfusion, limb salvage, motor function, and capillarization point to the therapeutic utility of VEGF PA for ischemic tissue repair. Thus, VEGF PA nanos-

Fig. 4. Results from in vivo hind-limb ischemia study examining the tissue salvage score (A) and the motor function score (B) of the various treatment groups over time, as well as the endpoint analysis at day 28 of failure time for a Rota Rod motor functional performance test (C). Laser Doppler perfusion imaging (D) for mouse hind-limb ischemia studies quantified as the perfusion ratio of treated to untreated limb along with representative LDPI images from the same animal at day 0 and day 28 for treatments of VEGF PA, VEGF peptide, mutant PA, and an untreated control. Significance is shown for the VEGF PA relative to other treatments (A, B, and D) and for other treatments compared to the VEGF PA treatment (C). 4 of 6 ∣

www.pnas.org/cgi/doi/10.1073/pnas.1016546108

Fig. 5. Results from quantification of immunohistological staining for CD31þ capillary forms in the ischemic tissue at day 28 of the hind-limb ischemia study, as well as representative immunohistological images staining for CD31 (green), smooth muscle α-actin (red), and nuclei (blue) for treatments of VEGF PA, VEGF peptide, mutant PA, and an untreated control. Significance is shown relative to VEGF PA treatment, and the scale bar shown corresponds to 100 μm in all micrographs.

tructures are identified here as a promising synthetic therapeutic strategy for ischemic cardiovascular disease. Although the primary objective of this work was to establish that supramolecular epitope display could enhance the potency of a VEGF-mimetic epitope, it remained of interest to see how this PA therapy would compare to a recombinant protein approach. Thus, following up on the promising therapeutic potential demonstrated, a second in vivo study comparing the VEGF PA to a bioactive treatment of VEGF165 protein in the ischemic hind limb was performed (Fig. S4). Because there exists no literature consensus on an effective dose of VEGF protein for intramuscular delivery in a mouse ischemic hind limb, and to be sure that an effective protein dose was used, a relatively high dose of 20 μg per animal was selected. For comparison, intramuscular delivery of 3 μg was shown to be ineffective in a mouse hind-limb model when delivered as a bolus without a biomaterial to control its release (21), whereas a 50-μg intramuscular bolus was effective in a similar model in much larger rats (ca. 300 g) (36). For these studies, the control group receiving a saline injection and the group receiving VEGF PA were repeated to account for variability in the model or in instrumentation for functional assessment. As shown, both the VEGF PA and VEGF165 performed similarly on the basis of LDPI perfusion ratio, with both showing a significant increase (P < 0.05) compared to the control (Fig. S4A). Scoring for limb necrosis indicated that only the VEGF PA group significantly (P < 0.05) enhanced tissue salvage compared to the control, with a trend for improvement in the VEGF protein group that was not significant from the control (Fig. S4B). Scoring for motor function in the hind limb indicated that both the VEGF PA group (P < 0.01) and the VEGF protein group (P < 0.05) significantly improved limb motor function compared to the control group (Fig. S4C). The measure that was most affected by VEGF PA treatment was histological capillary density in the ischemic hind-limb muscle. Treatment with VEGF PA resulted in significantly (P < 0.001) more capillaries in the hind limb than for treatment with either VEGF protein or the control (Fig. S4D). VEGF protein also demonstrated a significant (P < 0.001) Webber et al.

Webber et al.

PNAS Early Edition ∣

5 of 6

APPLIED BIOLOGICAL SCIENCES

Discussion Here we have demonstrated the use of bioactive and biodegradeable nanostructures as a strategy for therapeutic angiogenesis. The display on the surface of these nanofibers of a peptide mimic of VEGF showed enhanced signaling and bioactivity by activation of specific VEGF receptors and consequent functional outcomes for endothelial cells in vitro. The proangiogenic activity of this system was further substantiated in vivo using the CAM assay. Evaluation of the therapeutic potential of these VEGF PA nanostructures in a murine hind-limb ischemia model revealed improved tissue perfusion, limb motor function, limb salvage, and capillarization of the ischemic limb. The demonstrated efficacy suggests further consideration of these systems as an alternative therapy to protein-based strategies currently being evaluated for ischemic cardiovascular diseases. The material we have evaluated here is similar to that demonstrated previously with a different class of self-assembling peptides, where a VEGF-mimetic epitope and a cell adhesion epitope (RGDS) were evaluated for their ability to promote proliferation, migration, and tubulogenesis of cultured HUVECs (37). In this previous study, the VEGF epitope was not found to be in the required α-helical conformation by circular dichroism and its overall in vitro bioactivity was not markedly different from an RGDS fibronectin epitope. This result suggests to us that perhaps the peptide is not acting in a truly VEGF-mimetic way when presented on these β-sheet ribbon assemblies, and could be instead acting as an extracellular matrix as opposed to a protein mimic. The studies we have described in this work, however, establish that the epitope is in its appropriate conformation when presented on our cylindrical nanofibers and also that this epitope specifically acts in a mimetic fashion by activating VEGF receptors. Presentation on highly hydrated cylindrical supramolecular assemblies could afford more dynamics for efficient and potent receptor-mediated signaling that may not be possible on flat ribbon-like assemblies. In addition to functional in vitro evaluations, we have gone well beyond this previous work in our use of two different in vivo models to establish angiogenic bioactivity and therapeutic efficacy. There are several potential explanations underlying the enhanced epitope potency that we have demonstrated by our VEGF PA nanostructures compared to the soluble mimetic peptide control. As our CD melting curves established, the bioactive secondary structure of the epitope, an α-helix, is stabilized when incorporated within supramolecular aggregates of PA molecules. Stabilization of the bioactive conformation would likely translate into enhanced epitope bioactivity. The enhanced retention observed here could also explain in part the increased efficacy seen in vivo, though presumably retention would not be an issue for the short in vitro studies. A feature of the PA that could enhance its bioactivity compared to the peptide control is the polyvalency of epitopes on the nanostructure surface. It is well established that polyvalency, the presentation of multiple bioactive binding sites, significantly enhances binding strength for biological interactions, a phenomenon known as avidity (38). This phenomenon has a natural basis, vital to the function of binding proteins such as IgM with 10 binding sites (39) or the biological adhesion of many viruses (40), and even the extracellular matrix is polyvalent (41). Polyvalency is used frequently in the design of synthetic bioactive molecules to enhance their binding strength and has been used in the creation

of bioactive peptides, organic molecules, carbohydrates, nucleotides, antibiotics, and phage mimics, among others (42–46). Polyvalency has even been explored for biological signaling initiated by receptor dimerization, as is the case for VEGF receptor signaling. Examples include synthetic multivalent mimics of erythropoietin and thrombopoietin, both of which require receptor dimerization that is enhanced by multivalent signals (47, 48). Native VEGF, as a homodimeric cysteine knot protein, is multivalent (specifically divalent) with two bioactive domains that interact with receptor dimers (3), a feature that is not recreated in the soluble VEGF peptide mimic. The polyvalent and dynamic PA nanofiber could facilitate the necessary dimerization of receptors in a highly efficient manner. For the VEGF PA nanofibers investigated here, dynamic features of the nanostructures could help match the lengths necessary to promote receptor dimerization. PA nanostructures are highly hydrated, giving epitopes flexibility in both order and spacing within the assembled nanostructure. Given what is known about polyvalency in biological signaling, it is reasonable that avidity afforded by a dynamic PA assembly plays a role in the enhancement in bioactivity observed for the VEGF PA nanostructures relative to the soluble single peptide mimic. Regardless of the mechanistic details, because two bioactive signals must be present for VEGF receptor dimerization and activation, something which is the case for native VEGF protein with two cysteine-tethered bioactive domains, it is logical that PA molecules programmed for aggregation would be more apt to colocalize epitopes than would a soluble peptide, even if an enhanced avidity is not postulated. As mentioned previously, some issues raised with the clinical application of VEGF or other recombinant proteins are specific target tissue retention, limited production resources, and cost. Bioactive PAs provide a potential means to overcome these obstacles, and we were especially encouraged that the VEGF PA compared favorably to a high dose of recombinant protein. Though PAs are biodegradable by design and thus will be eventually broken down into natural products, they have been shown here to remain in the ischemic tissue for over 2 wk after injection. This result is a substantial improvement when compared to the retention time reported for VEGF protein on the order of a few hours. Another possible consideration is cost, which has been speculated to be prohibitively expensive to the clinical implementation of efficacious protein-based therapies (4). The prolonged retention and bioavailability of the PA in the target tissue could address both of these issues, circumventing the need for serial protein deliveries and the additional pain and suffering, along with cost, entailed therein. In many ways, the therapeutic mechanism observed here for the PA could be similar to observations for biomaterials that facilitate the slow release of VEGF protein. Nanofiber geometry and dynamics make it unlikely that a high percentage of the total epitopes presented on the PA are signaling at any given time, which makes the exact bioactive dose of PA difficult to determine. However, increased retention and bioavailability allows for more continuous and lasting signaling of the epitopes presented on the PA and, as nanofibers break apart in vivo, smaller aggregates could diffuse and signal in the tissue surrounding the injection site. This effect is similar to a scaffold-based controlled release approach, except here the effect is achieved using a defined single component synthetic system that can be delivered by a minimally invasive injection, which is not the case for several of the polymeric growth factor delivery scaffolds reported to date. Overall, the results we have demonstrated for bioactivity and therapeutic efficacy of this proangiogenic PA designed to signal through VEGF receptors point to the translational potential of this strategy. We have demonstrated that a polyvalent self-assembling nanofiber displaying a known VEGF mimicking sequence is efficacious in a hind-limb ischemia model of cardiovascular disease. The observed functional recovery is likely linked to

ENGINEERING

increase in capillary density compared to the control. This dramatic effect in capillarization could result from the prolonged retention and activity of the VEGF PA in the muscle tissue compared to the VEGF protein. Overall, the robust therapeutic effect seen previously for VEGF PA was conserved in these studies and the PA performed as well or better than the recombinant protein in every measured outcome.

the proangiogenic, VEGF-mimetic behavior of the VEGF PA nanostructures established both in vitro and in vivo. Presentation of the mimetic epitope on the polyvalent nanofiber leads to more efficient and effective VEGF signaling compared to the bioactive peptide alone. Further, the PA compares favorably to a high-dose injection of recombinant protein. We conclude that these bioactive nanostructures are a promising synthetic therapeutic strategy to regenerate microcirculation and restore perfusion to ischemic tissue in cardiovascular diseases and could provide an alternative to VEGF protein-based strategies.

through a microscope. The murine hind-limb ischemia model was performed in wild-type mice. The right femoral artery was ligated and excised, and the material was administered 3 d after induction of ischemia. Animals were scored on a semiquantitative scale for limb salvage (scale 1–6), motor function (scale 1–5), and were imaged by LDPI. Tissue perfusion is the ratio of ischemic to nonischemic limb perfusion from LDPI. Endurance testing was performed on a Rota Rod. At the end of the study, tissue was stained with antibodies to CD31 and smooth muscle actin and DAPI.

Detailed experimental methods can be found in the SI Text. Briefly, PAs and peptides for this study were synthesized by solid phase methods and purified by reversed phase HPLC. Nanofibers were imaged by cryogenic TEM, gels were imaged using SEM, and CD was preformed using standard methods. Primary HUVECs were purchased and used at passage 4. The receptor phosphorylation studies were performed using commercially available kits. Proliferation was assessed by a standard DNA-based assay, survival was assessed using flow cytometry, and migration was assessed by a scratch assay. The CAM assay was performed on fertilized chicken eggs, evaluating blood vessel density surrounding a coated coverslip in digital images captured

ACKNOWLEDGMENTS. The authors thank Dr. Yuri Velichko for assistance with molecular graphics. We gratefully acknowledge funding support from National Institutes of Health (NIH), specifically Award 1R01-EB003806-04 (to S.I.S.) and Awards HL-53354, HL-57516, HL-77428, HL-63414, HL-80137, and P01HL-66957 (to D.W.L.). This work was also supported by an Institute for Bionanotechnology in Medicine—Baxter Research Incubator Grant (J.T. and D.W.L.). Support for M.J.W. was provided by the Northwestern Regenerative Medicine Training Program NIH Award 5T90-DA022881 and support for J.T. by the German Heart Foundation, Solvay Pharmaceuticals, and the American Heart Association Midwest Affiliate. Peptide synthesis and purification was performed at the core facility of the Northwestern Institute for BioNanotechnology in Medicine. Instrumentation was used at the Northwestern Electron Probe Instrumentation Center (SEM), Biological Imaging Facility (TEM), and Keck Biophysics Facility (CD).

1. Lloyd-Jones D, et al. (2009) Heart disease and stroke statistics—2009 update: A report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 119:480–486. 2. Folkman J (2007) Angiogenesis: An organizing principle for drug discovery? Nat Rev Drug Discov 6:273–286. 3. Ferrara N, Gerber HP, LeCouter J (2003) The biology of VEGF and its receptors. Nat Med 9:669–676. 4. Gupta R, Tongers J, Losordo DW (2009) Human studies of angiogenic gene therapy. Circ Res 105:724–736. 5. Yla-Herttuala S, Rissanen TT, Vajanto I, Hartikainen J (2007) Vascular endothelial growth factors: Biology and current status of clinical applications in cardiovascular medicine. J Am Coll Cardiol 49:1015–1026. 6. Laham RJ, et al. (1999) Intracoronary and intravenous administration of basic fibroblast growth factor: Myocardial and tissue distribution. Drug Metab Dispos 27:821–826. 7. Lu E, et al. (2003) Targeted in vivo labeling of receptors for vascular endothelial growth factor: Approach to identification of ischemic tissue. Circulation 108:97–103. 8. Post MJ, Laham R, Sellke FW, Simons M (2001) Therapeutic angiogenesis in cardiology using protein formulations. Cardiovasc Res 49:522–531. 9. Hartgerink JD, Beniash E, Stupp SI (2001) Self-assembly and mineralization of peptideamphiphile nanofibers. Science 294:1684–1688. 10. Hartgerink JD, Beniash E, Stupp SI (2002) Peptide-amphiphile nanofibers: A versatile scaffold for the preparation of self-assembling materials. Proc Natl Acad Sci USA 99:5133–5138. 11. Webber MJ, Kessler JA, Stupp SI (2010) Emerging peptide nanomedicine to regenerate tisues and organs. J Intern Med 267:71–88. 12. Cui H, Webber MJ, Stupp SI (2010) Self-assembly of peptide amphiphiles: From molecules to nanostructures to biomaterials. Biopolymers 94:1–18. 13. Webber MJ, et al. (2010) Development of bioactive peptide amphiphiles for therapeutic cell delivery. Acta Biomater 6:3–11. 14. Silva GA, et al. (2004) Selective differentiation of neural progenitor cells by high-epitope density nanofibers. Science 303:1352–1355. 15. Muraoka T, Koh CY, Cui H, Stupp SI (2009) Light-triggered bioactivity in three dimensions. Angew Chem Int Ed Engl 48:5946–5949. 16. Tovar JD, Claussen RC, Stupp SI (2005) Probing the interior of peptide amphiphile supramolecular aggregates. J Am Chem Soc 127:7337–7345. 17. Ghanaati S, et al. (2009) Dynamic in vivo biocompatibility of angiogenic peptide amphiphile nanofibers. Biomaterials 30:6202–6212. 18. Tysseling-Mattiace VM, et al. (2008) Self-assembling nanofibers inhibit glial scar formation and promote axon elongation after spinal cord injury. J Neurosci 28:3814–3823. 19. Webber MJ, et al. (2010) Capturing the stem cell paracrine effect using heparinpresenting nanofibres to treat cardiovascular diseases. J Tissue Eng Regen Med 4:600–610. 20. Chen RR, et al. (2007) Integrated approach to designing growth factor delivery systems. FASEB J 21:3896–3903. 21. Silva EA, Mooney DJ (2007) Spatiotemporal control of vascular endothelial growth factor delivery from injectable hydrogels enhances angiogenesis. J Thromb Haemost 5:590–598. 22. Silva EA, Mooney DJ (2010) Effects of VEGF temporal and spatial presentation on angiogenesis. Biomaterials 31:1235–1241. 23. Garbern JC, Hoffman AS, Stayton PS (2010) Injectable pH- and temperature-responsive poly(N-isopropylacrylamide-co-propylacrylic acid) copolymers for delivery of angiogenic growth factors. Biomacromolecules 11:1833–1839. 24. Tae G, Scatena M, Stayton PS, Hoffman AS (2006) PEG-cross-linked heparin is an affinity hydrogel for sustained release of vascular endothelial growth factor. J Biomater Sci Polym Ed 17:187–197.

25. Rajangam K, et al. (2006) Heparin binding nanostructures to promote growth of blood vessels. Nano Lett 6:2086–2090. 26. Zisch AH, et al. (2003) Cell-demanded release of VEGF from synthetic, biointeractive cell ingrowth matrices for vascularized tissue growth. FASEB J 17:2260–2262. 27. Zisch AH, Schenk U, Schense JC, Sakiyama-Elbert SE, Hubbell JA (2001) Covalently conjugated VEGF—fibrin matrices for endothelialization. J Control Release 72:101–113. 28. D’Andrea LD, Del Gatto A, Pedone C, Benedetti E (2006) Peptide-based molecules in angiogenesis. Chem Biol Drug Des 67:115–126. 29. Ferrara N, Kerbel RS (2005) Angiogenesis as a therapeutic target. Nature 438:967–974. 30. D’Andrea LD, et al. (2005) Targeting angiogenesis: Structural characterization and biological properties of a de novo engineered VEGF mimicking peptide. Proc Natl Acad Sci USA 102:14215–14220. 31. Diana D, et al. (2008) Structural determinants of the unusual helix stability of a de novo engineered vascular endothelial growth factor (VEGF) mimicking peptide. Chemistry 14:4164–4166. 32. Malkar NB, Lauer-Fields JL, Juska D, Fields GB (2003) Characterization of peptideamphiphiles possessing cellular activation sequences. Biomacromolecules 4:518–528. 33. Matsumoto T, Claesson-Welsh L (2001) VEGF receptor signal transduction. Sci STKE 2001:re21. 34. Duval M, Bedard-Goulet S, Delisle C, Gratton JP (2003) Vascular endothelial growth factor-dependent down-regulation of Flk-1/KDR involves Cbl-mediated ubiquitination. Consequences on nitric oxide production from endothelial cells. J Biol Chem 278:20091–20097. 35. Yancopoulos GD, et al. (2000) Vascular-specific growth factors and blood vessel formation. Nature 407:242–248. 36. Kofidis T, et al. (2002) Restoration of blood flow and evaluation of corresponding angiogenic events by scanning electron microscopy after a single dose of VEGF in a model of peripheral vascular disease. Angiogenesis 5:87–92. 37. Wang X, Horii A, Zhang S (2008) Designer functionalized self-assembling peptide nanofiber scaffolds for growth, migration, and tubulogenesis of human umbilical vein endothelial cells. Soft Matter 4:2388–2395. 38. Vance D, Martin J, Patke S, Kane RS (2009) The design of polyvalent scaffolds for targeted delivery. Adv Drug Deliv Rev 61:931–939. 39. Czajkowsky DM, Shao Z (2009) The human IgM pentamer is a mushroom-shaped molecule with a flexural bias. Proc Natl Acad Sci USA 106:14960–14965. 40. Douglas T, Young M (2006) Viruses: Making friends with old foes. Science 312:873–875. 41. Hynes RO (2009) The extracellular matrix: Not just pretty fibrils. Science 326:1216–1219. 42. Galeazzi S, et al. (2009) Multivalent supramolecular dendrimer-based drugs. Biomacromolecules 11:182–186. 43. Mammen M, Choi SK, Whitesides GM (1998) Polyvalent interactions in biological systems: Implications for design and use of multivalent ligands and inhibitors. Angew Chem Int Ed Engl 37:2755–2794. 44. Rao J, Lahiri J, Isaacs L, Weis RM, Whitesides GM (1998) A trivalent system from vancomycin. D-ala-D-Ala with higher affinity than avidin.biotin. Science 280:708–711. 45. Helms BA, et al. (2009) High-affinity peptide-based collagen targeting using synthetic phage mimics: From phage display to dendrimer display. J Am Chem Soc 131:11683–11685. 46. Vance D, Shah M, Joshi A, Kane RS (2008) Polyvalency: A promising strategy for drug design. Biotechnol Bioeng 101:429–434. 47. Cwirla SE, et al. (1997) Peptide agonist of the thrombopoietin receptor as potent as the natural cytokine. Science 276:1696–1699. 48. Johnson DL, et al. (1997) Amino-terminal dimerization of an erythropoietin mimetic peptide results in increased erythropoietic activity. Chem Biol 4:939–950.

Materials and Methods

6 of 6 ∣

www.pnas.org/cgi/doi/10.1073/pnas.1016546108

Webber et al.

Supporting Information Webber et al. 10.1073/pnas.1016546108 SI Materials and Methods Synthesis and Purification of Peptide Amphiphiles (PAs). We synthe-

sized two PAs and one peptide for this study having the following amino acid sequences, with PAs covalently linked to a 16-carbon alkyl segment: C16 -V2 A2 K 3 GKLTWQELYQLKYKGI-NH2 (VEGF PA), Ac-KLTWQELYQLKYKGI-NH2 (VEGF peptide), and C16 -V2 A2 K3 GKLTAQELVFLKVKGI-NH2 (mutant PA). The structure of VEGF PA is shown in Fig. 1A. All peptides were synthesized by standard solid phase Fmoc chemistry. Fmocprotected amino acids, 4-methylbenzhydrylamine rink amide resin, and O-(benzotriazol-1-yl)-N,N,N’,N’-tetramethyluronium hexafluorophosphate were purchased from NovaBiochem and all reagents were purchased from Mallinckrodt. The resulting product was purified using standard reversed-phase HPLC. The purity and accurate mass for each PA was verified using liquid chromatography/MS on an electrospray ionization quadripole time-of-flight mass spectrometer (Agilent). Additionally, peptide content analysis was performed on the purified product (Commonwealth Biotechnologies) to ensure concentration accuracy and consistency for all experiments. Structural Characterization. Cryogenic transmission electron microscopy (TEM) was performed on a JEOL 1230 microscope with an accelerating voltage of 100 kV. A Vitrobot Mark IV equipped with controlled humidity and temperature was used for plungefreezing samples. A small volume (7 μL) of 1 wt∕vol % VEGF PA dissolved in pure water was deposited on a copper TEM grid with Quantifoil support film (Electron Microscopy Sciences) and held in place with tweezers mounted to the Vitrobot. The specimen was blotted in 90–95% humidity and plunged into a liquid ethane reservoir cooled by liquid nitrogen. The vitrified samples were transferred into liquid nitrogen and inserted into a Gatan 626 cryoholder through a cryotransfer stage. Samples were imaged using a Hamamatsu ORCA CCD camera. SEM was performed using a Hitachi S4800 scanning electron microscope with a 5-kV accelerating voltage. To prepare samples for imaging, VEGF PA was dissolved at 1.5 wt∕vol % in water and mixed with 10 mM Na2 HPO4 to induce hydrogelation at 0.75 wt∕vol % final PA concentration. This concentration is consistent with that used in the in vivo portion of the study. The sample was fixed in 2% glutaraledyde and 3% sucrose in PBS for 30 min at 4 °C followed by sequential dehydration in ethanol. It was then dried at the critical point and coated with 7 nm OsO4 prior to imaging. Circular dichroism was performed on a Jasco J-815 CD spectrophotometer complete with Peltier sample holder for precise temperature control. Samples were analyzed at 0.15 mM in water, correcting for absolute peptide content. Measurements were collected over a wavelength range of 260–180 nm with a step size of 0.5 nm and five total accumulations for each scan. For thermal denaturation studies, samples were heated from 25 to 95 °C at a step of 5 °C and held for 10 min prior to each reading. Cell Culture. Human umbilical vein endothelial cells (HUVECs) and complete endothelial cell growth media (EGM) were purchased (Genlantis), passaged two times after receipt, and cryopreserved in media with 5% DMSO. Cells were thawed as needed and grown to confluence in 75-mm2 flasks (VWR Falcon) prior to plating for experiments. VEGF Receptor (VEGFR) Phosphorylation Assays. Phosphorylation of both VEGFR1 and VEGFR2 was assayed using commercially Webber et al. www.pnas.org/cgi/doi/10.1073/pnas.1016546108

available assay kit (R&D Systems) following the recommended protocol. First, confluent HUVECs plated in 24-well plates were starved for 5 ho in serum-free defined media (SFDM, Genlantis) specifically designed to maintain cells in a growth-factor-free setting for growth factor signaling and metabolic assays. Following starvation, the various treatments were dissolved in SFDM to a concentration of 1 μM. Native VEGF165 (100 ng∕mL, Peprotech) diluted in SFDM was used pursuant to assay recommendations in order to serve as an internal assay control for VEGF phosphorylation patterns and a group was treated with plain SFDM as a baseline control. Cells were exposed to treatment for 2 min. To evaluate the time course of phosphorylation, the same protocol was followed and cells were exposed to VEGF PA for 0, 1, 3, 5, 10, 15, 20, 25, 30, 40, 50, and 60 min. Functional in Vitro Assays. To evaluate proliferation, HUVECs were plated at 5,000 per well in a 96-well plate. Four hours after plating, EGM was exchanged for fresh EGM supplemented with 1 μM of VEGF PA, VEGF peptide, or mutant PA (n ¼ 8∕group). Additionally, unsupplemented growth media was used as a control. After 48 h, cell number was quantified using CyQUANT-NF (Invitrogen) and a standard fluorescent microplate reader. Cell number is expressed relative to the group treated with unsupplemented growth media. To evaluate cell survival, endothelial basal media (EBM) without growth factors (Lonza) was used to induce a serum starvation. Cells were plated in 12-well plates and grown to confluence with standard growth media. Cells were then washed twice with PBS and then treated with EBM containing 1 μM of VEGF PA, VEGF peptide, or mutant PA along with an untreated control receiving only EBM. Cells were grown in these conditions for 24 h. Survival was quantified using Annexin V: phycoerythrin (PE) staining with 7-aminoactinomycin vital staining (BD Biosciences) following provided assay instructions and analyzed on a DakoCytomation CyAn. Survival was assessed by determining the fraction of cells that were apoptotic (Annexin V-PE positive). To evaluate cell migration, EBM without growth factors (Lonza) was supplemented with 0.5% FBS in order to prevent apoptosis but not promote significant migration or proliferation. Cells were grown to confluence in a 12-well plate and a 1-mL pipette tip was used to create a denuded scratch. The surface was washed twice with PBS to remove detached cells. The average scratch width at the initial time point was 926.3
103.0 μm. Markings were placed on the underside of the plate to ensure the same region of the scratch was recorded in each image. Following scratch creation, the cells were treated with 1 μM VEGF PA, VEGF peptide, mutant PA in EBM with 0.5% FBS. The total pixel area of the scratch at the initial time point was recorded using ImageJ analysis software, and the percent migration was determined from the reduction in denuded area at 18 h. Chorioallantoic Membrane Angiogenesis Assay. In order to evaluate the angiogenic potential of the VEGF PA, we used well-established assays for in vivo angiogenesis, the chicken chorioallantoic membrane (CAM) assay. This assay utilizes the extraembryonic allantois, a tissue derived from the mesoderm that develops into a densely vascularized membrane. A common deviation from the traditional assay is to remove the shell and conduct the assay on a shell-less embryo, termed the shell-less CAM assay, as we have performed here (1, 2). Fertilized white leghorn chicken eggs (Phil’s Fresh Eggs) were received and cultured in a temperature controlled, humidified egg incubator. On embryonic day 3, eggs 1 of 4

were cracked within a sterile tissue culture hood into round 100-mm Petri dishes. Fertilized embryos were then transferred to a water-jacketed CO2 incubator, set to 37.5 °C, 1% CO2 , and 100% relative humidity. On embryonic day 10, the material treatment was dissolved at 2 mM in PBS, evaporated onto the surface of a 5-mm round glass coverslip, and placed facing down on top of the CAM (n ¼ 16∕group). Digital images were captured through the eyepiece of a Nikon stereomicroscope and vessel density was quantified by the number of intersections of vessel structures with the edge of the coverslip and expressed relative to the initial time point. Images were captured daily, beginning at embryonic day 10 and culminating on embryonic day 13, the standard range over which the CAM assay is performed. Additionally, images were assessed qualitatively for morphological differences in the CAM vasculature including spoking, branching, and leakage. Mouse Hind-Limb Ischemia Model. In order to assess the therapeutic potency of our developed VEGF PA for ischemic tissue repair, we chose the hind-limb ischemia (HLI) model, an established model for critical tissue ischemia. For the HLI procedure, 8-wk-old male FVB wild-type mice (Charles River) were used. By means of a dissecting microscope, the femoral nerve was carefully separated from the vessel bundle. The right femoral artery was ligated and excised, including all superficial and deep branches (3). Critical limb ischemia was immediately verified by laser Doppler imaging (LDPI, MoorLDI-SIM, Moor Instruments) to ensure the ratio (ischemic/nonischemic limb) was ≤0.20. At postoperative day 3, outliers with low ischemia were triaged based on LDPI (ratio ischemic/nonischemic 0.30) as were outliers with extreme ischemia (necrotic demarcation of entire limb), determined by macroscopic evaluation. After triage, all remaining mice were treated by a single transcutaneous intramuscular injection (25 μL) of 2 mM VEGF PA, VEGF peptide, mutant PA, or saline (PBS) as control. For follow-up, animals underwent reevaluations with LDPI at postoperative day 7, 14, and 28 before animals were killed (CO2 asphyxia) for tissue harvest at day 28. At each time point, tissue perfusion was measured via LDPI, measuring blood flow in both the ischemic and nonischemic limb and reporting results as the ratio of these two measurements. All LDPI measurements were taken on a 37 °C heating pad to control body temperature. In addition, motor function and tissue damage was semiquantitatively assessed on postoperative day 7, 14, 21, and 28 by established scoring systems. Tissue damage in the ischemic limb (limb salvage score) was graded as full recovery (grade 6), minor necrosis or nail loss (grade 5), partial toe amputation (grade 4), total toe amputation (grade 3), partial/total foot amputation (grade 2), or partial/total limb amputation (grade 1) (modified from ref. 4). Limb motor function was graded as unrestricted (grade 5), no active use of toe(s) or spreading (grade 4), restricted foot use (grade 3), no use of foot (grade 2), or no use of limb at all (grade 1. Auerbach R, Kubai L, Knighton D, Folkman J (1974) A simple procedure for the long-term cultivation of chicken embryos. Dev Biol 41:391–394. 2. Auerbach R, Lewis R, Shinners B, Kubai L, Akhtar N (2003) Angiogenesis assays: A critical overview. Clinical Chemistry 49:32–40. 3. Couffinhal T, et al. (1998) Mouse model of angiogenesis. Am J Pathol 152:1667–1679. 4. Rossig L, et al. (2005) Histone deacetylase activity is essential for the expression of HoxA9 and for endothelial commitment of progenitor cells. J Exp Med 201:1825–1835. 5. Heil M, Schaper W (2004) Influence of mechanical, cellular, and molecular factors on collateral artery growth (arteriogenesis). Circ Res 95:449–458.

Webber et al. www.pnas.org/cgi/doi/10.1073/pnas.1016546108

1) (modified from ref. 5). Finally, walking capacity was measured via a Rota Rod apparatus. The rotational velocity was steadily increased and time at which the mouse failed to keep up with the treadmill was recorded. The mean of two assessments was used as for failure time for each animal. This in vivo experiment was carried out for 28 d, because the experience we have with this model suggests this to be an appropriate time course to assess whether a therapeutic target results in functional difference and this is a frequently used time course for these studies (6, 7, 8). These studies were approved by the Northwestern University Animal Care and Use Committee. Follow-up studies comparing the PA to VEGF165 (Peprotech) used the same PA dose (2 mM) and compared this to injection of 20 μg VEGF165 delivered in 25 μL saline. The procedure was performed identically to that done previously and the same functional assessments were conducted at the same follow-up times. Ischemic groups were also prepared for the injection and tracking of fluorescently labeled VEGF PA and peptide. To obtain fluorescently labeled PA nanofibers, a PA conjugated with FITC was mixed with the VEGF PA at 2 mol %, as described previously (9). For fluorescently labeled VEGF peptide, the peptide was functionalized at the N terminus with FITC and all peptides were labeled. For the fluorescent tracking study, muscle tissue was harvested at 2, 7, 14, and 28 d after injection, fixed in methanol, and processed for histology with DAPI staining, following which an extensive histological search was performed on multiple sections throughout the tissue. Immunohistology for Capillary Quantification. At day 28, muscle tissue from the ischemic limb was harvested, fixed in methanol, paraffin-embedded, and cross-sectioned (6 μm) for histological immunostaining. Briefly, sections were blocked with 10% donkey serum (30 min, room temperature). Primary antibodies were diluted in PBS containing BSA, and applied to tissue slices for 2 h at 37 °C. Sections were stained for CD31, an endothelial-specific marker, using rat anti-CD31 antibodies (BD Pharmingen) and smooth-muscle α-actin (αSMA), a vascular smooth-muscle marker, using rabbit-anti-αSMA (Sigma-Aldrich). For immunofluorescent detection, primary antibodies were resolved with AlexaFluor-conjugated secondary antibodies (Invitrogen Corporation dilution) and nuclei were counterstained with DAPI (Research Organics). Slides were imaged using fluorescent microscopy (Zeiss), and CD31þ capillary forms and CD31þ ∕sMAþ mature microvessels/arterioles were quantified in three separate highpower fields (20×) from three independent sections in each animal (nine images per animal). Statistics and Data Analysis. All error bars indicate the standard error of the mean. Differences between groups were determined using ANOVA with a Bonferroni multiple comparisons post hoc test using GraphPad InStat v3.0b. 6. Cao L, Arany PR, Wang YS, Mooney DJ (2009) Promoting angiogenesis via manipulation of VEGF responsiveness with notch signaling. Biomaterials 30:4085–4093. 7. Pola R, et al. (2003) Postnatal recapitulation of embryonic hedgehog pathway in response to skeletal muscle ischemia. Circulation 108:479–485. 8. Yamaguchi J, et al. (2003) Stromal cell-derived factor-1 effects on ex vivo expanded endothelial progenitor cell recruitment for ischemic neovascularization. Circulation 107:1322–1328. 9. Ghanaati S, et al. (2009) Dynamic in vivo biocompatibility of angiogenic peptide amphiphile nanofibers. Biomaterials 30:6202–6212.

2 of 4

Fig. S1. Stuctures of VEGF PA, mutant PA, and peptide control used in this study, with arrows noting the residues that were modified in the mutant PA.

Fig. S2. Results from in vitro assays on HUVECs, assessing the effects of various treatments on cell number (A), survival during serum starvation (B), and migration into a denuded scratch (C). Significance is shown relative to VEGF PA treatment.

Fig. S3. Histological imaging of ischemic muscle at 2, 7, 14, and 28 d following injection of either fluorescent VEGF PA nanofibers (Upper) or fluorescently tagged VEGF peptide (Lower), where green indicates fluorescence from the injected compound. Tissue was counterstained with DAPI (blue).

Webber et al. www.pnas.org/cgi/doi/10.1073/pnas.1016546108

3 of 4

Fig. S4. Results from in vivo hind-limb ischemia study examining the day 28 LDPI perfusion ratio (A), tissue salvage score (B), motor function score (C), and histological quantification of CD31þ capillaries for treatment with VEGF PA compared to a 20-μg bolus injection of VEGF165 protein. Significance is shown relative to the untreated control.

Webber et al. www.pnas.org/cgi/doi/10.1073/pnas.1016546108

4 of 4