Tissue-Factor Targeted Peptide Amphiphile Nanofibers as an Injectable Therapy To Control Hemorrhage Courtney E. Morgan,†,‡ Amanda W. Dombrowski,†,⊥ Charles M. Rubert Pérez,†,⊥ Edward S.M. Bahnson,†,‡ Nick D. Tsihlis,†,‡ Wulin Jiang,†,‡ Qun Jiang,†,‡ Janet M. Vercammen,†,‡ Vivek S. Prakash,†,‡ Timothy A. Pritts,¶,# Samuel I. Stupp,†,⊥,▲,∥,§ and Melina R. Kibbe*,†,‡ †
Simpson Querrey Institute for BioNanotechnology, ‡Department of Surgery, Feinberg School of Medicine, and §Department of Medicine, Feinberg School of Medicine, Northwestern University, Chicago, Illinois 60611, United States ⊥ Department of Materials Science & Engineering, ▲Biomedical Engineering, and ∥Department of Chemistry, Northwestern University, Evanston, Illinois 60208, United States ¶ Department of Surgery and #Institute for Military Medicine, University of Cincinnati, Cincinnati, Ohio 45220, United States S Supporting Information *
ABSTRACT: Noncompressible torso hemorrhage is a leading cause of mortality in civilian and battlefield trauma. We sought to develop an i.v.-injectable, tissue factor (TF)-targeted nanotherapy to stop hemorrhage. Tissue factor was chosen as a target because it is only exposed to the intravascular space upon vessel disruption. Peptide amphiphile (PA) monomers that self-assemble into nanofibers were chosen as the delivery vehicle. Three TF-binding sequences were identified (EGR, RLM, and RTL), covalently incorporated into the PA backbone, and shown to self-assemble into nanofibers by cryotransmission electron microscopy. Both the RLM and RTL peptides bound recombinant TF in vitro. All three TF-targeted nanofibers bound to the site of punch biopsy-induced liver hemorrhage in vivo, but only RTL nanofibers reduced blood loss versus sham (53% reduction, p < 0.05). Increasing the targeting ligand density of RTL nanofibers yielded qualitatively better binding to the site of injury and greater reductions in blood loss in vivo (p < 0.05). In fact, 100% RTL nanofiber reduced overall blood loss by 60% versus sham (p < 0.05). Evaluation of the biocompatibility of the RTL nanofiber revealed that it did not induce RBC hemolysis, did not induce neutrophil or macrophage inflammation at the site of liver injury, and 70% remained intact in plasma after 30 min. In summary, these studies demonstrate successful binding of peptides to TF in vitro and successful homing of a TF-targeted PA nanofiber to the site of hemorrhage with an associated decrease in blood loss in vivo. Thus, this therapeutic may potentially treat noncompressible hemorrhage. KEYWORDS: peptide amphiphile, hemorrhage, animal model, hemostasis, self-assembly, nanotherapy, tissue factor are not readily available in the field due to limited shelf life and stringent storage requirements. While other therapies, such as recombinant Factor VIIa (i.e., Novoseven), have been shown to reduce the need for blood products in trauma patients,6 outcomes with respect to adverse events and mortality are less clear.7 In addition to the high cost of recombinant Factor VIIa, the original formulation must be stored at 2−8 °C, and the newer formulation may only be stored at 25 °C for 6 h, neither of which is compatible with the battlefield setting.8 Thus, there is a great need for the development of a novel therapeutic that can stop noncompressible hemorrhage.
H
emorrhage is the leading cause of mortality in battlefield trauma and the second leading cause of death in civilian trauma.1,2 Although therapies have been developed for compressible hemorrhage, noncompressible torso hemorrhage, or bleeding that cannot be controlled by external compression, is a major problem in trauma. Noncompressible torso hemorrhage is defined as a pulmonary injury, a grade IV or greater injury to a solid organ, injury to a named axial vessel, or a pelvic fracture with ring disruption.3 These are common injuries, representing 13% of all battlefield injuries in Iraq and Afghanistan, with 18% of those injured presenting in shock and requiring immediate control of hemorrhage.4 Current options for noncompressible torso hemorrhage in the far-forward setting are limited. Blood products such as fresh frozen plasma, platelets, and cryoprecipitate can improve coagulopathy,5 but these products © XXXX American Chemical Society
Received: September 24, 2015 Accepted: December 23, 2015
A
DOI: 10.1021/acsnano.5b06025 ACS Nano XXXX, XXX, XXX−XXX
Article
www.acsnano.org
Article
ACS Nano Utilizing the platform of targeted therapeutics allows the development of an injectable therapy that can be targeted specifically to the site of hemorrhage to stop bleeding, thereby avoiding off-target side effects. For this therapeutic, we chose to target tissue factor (TF) since it plays an important role in initiating the coagulation cascade. Tissue factor is an integral membrane protein found in the adventitial cells of all blood vessels larger than capillaries and is only exposed to the circulating blood elements within the intravascular space following disruption of the vessel, making it an ideal target for a therapy to control hemorrhage.9 In order to develop a targeted therapeutic for hemorrhage, we chose a peptide amphiphile (PA) delivery vehicle. PA monomersconsisting of a hydrophobic alkyl tail, a β-sheet forming peptide domain, and a bioactive peptide segmentcan self-assemble into supramolecular nanostructures.10−13 In particular, our previous PA platform that is targeted against collagen IV forms nanofibers that are typically 6−8 nm (nm) in diameter.14 PA nanofibers are ideal for developing a therapeutic for clinical use given that they are infinitely modifiable, biocompatible, biodegradable into excreted products, can encapsulate drugs in their hydrophobic core, and can be redosed as necessary.14−16 The overall goal of this study was to develop a nanotherapeutic that specifically targets TF to reduce hemorrhage. We hypothesized that a TF-targeting peptide would specifically bind to TF in vitro, and that TFtargeted PA nanofibers would home to the site of organ injury in vivo and reduce hemorrhage.
Figure 1. Structural models of the interaction of TF with Factor VII. Protein models were produced using PyMol software (Schrodinger, Cambridge, MA), and the sequence was deposited in the National Center for Biotechnology Information Protein Data Bank for tissue factor (TF, blue) and Factor VII (yellow) (PDB # 1DAN).19 TF binding sequences derived from linear sequences of Factor VII are shown in red. (A) EGRNCETHKDDQL (EGR) found in the EGF-2 domain of Factor VII; (B) RLMTQDCLQQRSK (RLM) found in the heavy chain of Factor VII; and (C) RTLAFVRFK (RTL) found in the heavy chain of Factor VII.
the micrometer range. For the RTL-targeted PA, as the percentage of the RTL targeting peptide increased, the fibers increased in average diameter from 11 ± 2 to 16 ± 3 nm (Figure S1). In addition, shorter assemblies appeared (400− 900 nm in length), displaying ribbon-like character. Higher targeting epitope densities potentially favor epitope−epitope interactions that could affect the overall supramolecular structure. Circular dichroism performed on each targeted and nontargeted nanofiber showed that the targeted nanofibers had increased β-sheet character over the nontargeted nanofiber (Figure 2G). In addition, increasing the percent of the targeting ligand increased the β-sheet character compared to the nontargeted nanofibers, as evidenced by the blue shift of the circular dichroism (CD) minima closer to 220 nm (Figure 2H). RLM and RTL TF-Targeting Peptides Bind to Recombinant TF in Vitro. To determine which of the three binding peptide sequences bind TF the best, a binding assay was performed. Both the RLM and RTL peptide sequences had the highest absorbance (0.29 ± 0.02 and 0.28 ± 0.03, respectively, *p < 0.05) compared with the negative control (0.18 ± 0.02), nonspecific peptide (0.2 ± 0.03), and EGR peptide sequence (0.19 ± 0.3, Figure 3A). In addition, the RLM and RTL peptides were found to have absorbance readings similar to the recombinant TF-coated positive control group (0.32 ± 0.01), suggesting that the RLM and RTL peptides achieved maximal binding to recombinant TF. To further delineate the nature of the binding, a dose−response assay was performed in which the RLM and RTL peptides were coated on 96-well plates and incubated with increasing concentrations of recombinant TF. Colorimetric immunodetection demonstrated increased absorbance with increasing concentrations of recombinant TF (Figure 3B), further validating the binding of RLM and RTL to TF in vitro. TF-Targeted Nanofibers Bind to Areas of Liver Injury in Vivo. To determine if the targeted nanofibers bind to TF in vivo, rats underwent the liver punch injury model, and binding of the different targeted nanofibers was assessed using fluorescent imaging of the liver. As can be seen in Figure 4A,
RESULTS Targeted PAs Form Nanofibers and Display Increased β-Sheet Character That Increases with Ligand Density. TF binding peptides were selected by examining the interaction between TF and Factor VII (FVII). Studies on the crystal structure of TF and FVII, as well as mutational studies, have identified several areas important for binding.9,17,18 We chose three linear binding sequences from FVII. The first sequence is from the EGF-2 domain and contains arginine 79, which has been shown to be important for TF binding by crystal structure interactions19 and through mutational studies.17,18 This sequence, consisting of the amino acids EGRNCETHKDDQL, is referred to as the EGR sequence (Figure 1A). The next sequence is located on the heavy chain of FVII, forms an αhelix, and contains the amino acids methionine 306 and aspartate 309, both of which have been shown to be important in the binding of FVII to TF.18,19 This sequence, consisting of the amino acids RLMTQDCLQQRSK, is referred to as the RLM sequence (Figure 1B). The third sequence is also located on the heavy chain of FVII, forms an α-helix, and contains arginine 277, which has been found to be important in the binding of FVII to TF.18,19 This sequence, consisting of the amino acids RTLAFVRFK, is referred to as the RTL sequence (Figure 1C). In addition, a nonspecific peptide, consisting of the filler peptide A2V2K2−C12, was selected as a nontargeted control. Each of the three targeted nanofibers (EGR, RLM, and RTL, Figure 2A−C) was synthesized and coassembled in a ratio of 25% targeted PA (EGR PA, RLM PA, and RTL PA) to 75% nontargeted PA (K2A2V2 PA filler). For RTL, this range was further expanded to include 10 to 100% RTL PA. Cryotransmission electron microscopy (cryo-TEM) confirmed fiber formation (Figure 2D−F) and found that the fibers were between 10 and 19 nm in diameter (Figure S1) with lengths in B
DOI: 10.1021/acsnano.5b06025 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano
Figure 2. Targeted PAs form nanofibers and display increased β-sheet character, which increases with ligand density. Chemical structures of the (A) EGR PA, (B) RLM PA, and (C) RTL PA. Cryo-transmission electron microscopy of the (D) EGR PA, (E) RLM PA, and (F) RTL PA nanofibers confirms the fiber shape. Circular dichroism demonstrates (G) increased β-sheet formation with the EGR PA, RLM PA, and RTL PA nanofibers as compared to the nontargeted nanofibers (K2A2V2 PA filler), and (H) increased β-sheet formation with increasing RTL targeting ligand density.
no fluorescence was observed in the injured liver from shamtreated rats or rats that received nonspecific PA nanofibers (Figure 4A). Fluorescence was observed near the vasculature in the injured liver from rats that received the EGR PA, RLM PA, and RTL PA targeted nanofibers (Figure 4A), indicating binding of the targeted nanofibers to the site of injury. No fluorescence was observed in liver remote to the area of injury (i.e., the uninjured liver) in any of the treatment groups. RTL TF-Targeted Nanofibers Reduce Hemorrhage in a Rat Model of Liver Injury. Next, to determine if the targeted nanofibers impact hemorrhage, blood loss during the liver punch model was assessed for all treatment groups. Cumulative blood loss increased over time for sham-treated animals. Surprisingly, all three targeted nanofibers, as well as the nonspecific PA nanofiber, had a time-dependent effect of reducing blood loss compared to the sham-treated rats (*p < 0.001, Figure 4B). However, only the RTL PA-targeted nanofiber reduced blood loss over time compared to both sham and nonspecific PA nanofiber-treated rats (τp < 0.05, Figure 4B). To assess the effect of treatment group independent of time, total blood loss was analyzed. The RTL
PA nanofiber was the only sequence that significantly reduced total blood loss compared to sham-treated rats (53% reduction, *p < 0.05, Figure 4C). No statistically significant differences were noted in total blood loss between the other treatment groups and sham-treated rats: nonspecific PA (21 ± 2%), EGR PA (19 ± 3%), and RLM PA (21 ± 1%). RTL TF-Targeted Nanofibers with Increasing Ligand Density Bind to Injured Liver. Since both the RLM and RTL binding sequences exhibited superior binding compared to the EGR sequence in the in vitro binding assay, with only the RTL PA-targeted nanofiber exhibiting less blood loss in the animal model of hemorrhage, the RTL PA was selected for further evaluation. The effect of targeting ligand density was evaluated on both binding and response to hemorrhage by coassembling different ratios of the targeted PA (10−100%) with the nontargeted PA (K2A2V2 PA filler). Similar to previous experiments, animals underwent tail vein injection of the nanofiber followed by immediate liver injury. Fluorescent microscopy of injured liver from rats that received the nonspecific or RTL PA nanofiber with different targeting densities is shown in Figure 5. No fluorescence was observed in C
DOI: 10.1021/acsnano.5b06025 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano
compared to blood loss over time from animals that received the nonspecific PA nanofiber (τp < 0.001). To assess the effect of RTL ligand density independently of time, we evaluated total blood loss (Figure 6B). There was a significant ligand densitydependent reduction on total blood loss (p = 0.0081). The 100% RTL PA nanofiber resulted in 60% less blood loss compared to sham-treated animals (12 ± 1 vs 30 ± 4%, respectively, **p < 0.05). The 100% RTL TF-targeted nanofiber also reduced blood loss compared to the animals treated with nonspecific PA nanofiber (21 ± 2%, **p < 0.05). Blood losses for the 75% RTL (14 ± 2%), 50% RTL (16 ± 2%), 25% RTL (14 ± 2%), and 10% RTL (18 ± 3%) were each reduced compared to sham (*p < 0.05). To assess the possible mechanism of action of the RTL PA nanofiber on reducing hemorrhage, we evaluated platelet aggregation and measured free fibrinogen levels, as we hypothesize that the targeted nanofiber potentiates coagulation by physically binding to platelets and fibrinogen at the site of hemorrhage. First, we analyzed platelet adhesion to a RTL PA nanofiber network using SEM. We found that platelets adhere to surfaces coated with RTL PA nanofiber, whereas they do not adhere to uncoated tissue culture surfaces (Figure 7). We also assessed the effect of RTL PA on free fibrinogen levels in whole blood. After collecting rat whole blood, plasma was incubated ex vivo with RTL PA (80 μM) or with saline. The free fibrinogen level in control plasma was 202 ± 4 mg/dL, whereas in RTL PA nanofiber-treated plasma, it was below the limit of detection of the assay (
