Engineering a Two‐Component Hemostat for the Treatment of Internal Bleeding through Wound‐Targeted Crosslinking

Abstract

Primary hemostasis (platelet plug formation) and secondary hemostasis (fibrin clot formation) are intertwined processes that occur upon vascular injury. Researchers have sought to target wounds by leveraging cues specific to these processes, such as using peptides that bind activated platelets or fibrin. While these materials have shown success in various injury models, they are commonly designed for the purpose of treating solely primary or secondary hemostasis. In this work, a two‐component system consisting of a targeting component (azide/GRGDS PEG‐PLGA nanoparticles) and a crosslinking component (multifunctional DBCO) is developed to treat internal bleeding. The system leverages increased injury accumulation to achieve crosslinking above a critical concentration, addressing both primary and secondary hemostasis by amplifying platelet recruitment and mitigating plasminolysis for greater clot stability. Nanoparticle aggregation is measured to validate concentration‐dependent crosslinking, while a 1:3 azide/GRGDS ratio is found to increase platelet recruitment, decrease clot degradation in hemodiluted environments, and decrease complement activation. Finally, this approach significantly increases survival relative to the particle‐only control in a liver resection model. In light of prior successes with the particle‐only system, these results emphasize the potential of this technology in aiding hemostasis and the importance of a holistic approach in engineering new treatments for hemorrhage.

In this work, a two‐component system consisting of a targeting component (azide/glycine‐arginine‐glycine‐aspartic acid‐serinepolyethylene glycol (PEG)‐poly(lactic‐co‐glycolic) (PLGA) nanoparticles) and a crosslinking component (multifunctional dibenzylcyclooctyne (DBCO)) is developed to treat internal bleeding. The system leverages increased injury accumulation to achieve crosslinking above a critical concentration and is demonstrated to amplify platelet recruitment, decrease clot degradation, and increase survival in a lethal liver injury model.

1. Introduction

Traumatic hemorrhage contributes to upwards of 2.5 million casualties per year,^{[ 1 ]} resulting in a significant number of deaths in both civilian and military populations.^{[ 2 ]} The process of achieving hemostasis in response to traumatic bleeding involves several steps:^{[ 3 ]} the activation of platelets via exposed collagen, the binding of activated platelets to von Willebrand factor, collagen, and fibrinogen to form a platelet plug (primary hemostasis), the intrinsic and extrinsic pathways of coagulation that culminate in thrombin generation, and the cleavage of fibrinogen via thrombin into self‐polymerizable fibrin^{[ 4 ]} (secondary hemostasis). These steps are all mechanistically interconnected: for instance, fibrinogen‐platelet binding aids in the formation of a platelet plug, while an activated platelet provides the surface on which prothrombin is converted to thrombin, which leads to fibrin crosslinking.^{[ 5 ]}

Several solutions in literature have been proposed to accelerate hemostasis through interactions with these wound‐specific components. For topical wounds, these include materials such as positively charged chitosan, which can cause nonspecific aggregation of platelets and red blood cells,^{[ 6 ]} or self‐assembling peptide coatings such as RADA16 that mimic the structure of crosslinked fibrin.^{[ 7 ]} For internal wounds, these include intravenously‐delivered linear polymers conjugated with fibrin‐specific or von‐Willebrand‐binding peptides,^{[ 8 ]} as well as polymer and liposomal nanoparticles functionalized with platelet‐aggregating peptides.^{[ 9 ]} Though certain challenges remain, such as complement activation in response to nanoparticle infusion^{[ 10 ]} and nonspecific accumulation in lung and clearance organs,^{[ 11 ]} these materials have been demonstrated to significantly increase survival in a wide variety of animal models and present valuable options for the treatment of internal hemorrhage.

As noted, hemostasis is achieved through a multitude of interactions, involving both protein and cellular components of the wound microenvironment. However, the aforementioned intravenous hemostats have either focused on enhancing the platelet‐aggregating aspect of primary hemostasis, increasing the stability of the fibrin clot formed in secondary hemostasis, or recovering thrombin generation, neglecting key components of the process when fibrinogen concentration, clot stability, platelet function, and platelet availability are adversely impacted during traumatic blood loss and subsequent fluid resuscitation. For instance, while fibrin‐crosslinking polymers may slow the degradation of clots via plasminolysis,^{[ 8b ]} platelets have also been demonstrated to be instrumental in overall clot strength;^{[ 12 ]} similarly, while increased platelet number has been associated with higher rates of clot formation and thrombin generation,^{[ 13 ]} it fails to recover fibrinogen depleted through blood loss, even though fibrinogen deteriorates below critical levels ahead of other coagulation factors.^{[ 14 ]} Moreover, the lethal triad of trauma triggered by acute hemorrhage— hypothermia, coagulopathy, and acidosis—exacerbates fibrinogen degradation^{[ 15 ]} and decreases platelet activation,^{[ 16 ]} which may be difficult to counter solely by enhancing a discrete portion of hemostasis. These points illustrate the multifarious effects of traumatic bleeding on the coagulation system, underscoring the need for a more comprehensive solution.

To address this, a two‐component system consisting of wound‐targeted nanoparticles with crosslinkable groups and polymeric crosslinkers was designed to both recruit platelets and counter fibrinogen depletion, enhancing both primary and secondary hemostasis and slowing clot degradation. Polyethylene glycol (PEG)‐b‐poly(lactic‐co‐glycolic) (PLGA) nanoparticles (NPs) were functionalized with GRGDS and azide biorthogonal click groups, while the second component was functionalized with the corresponding copper‐free click moiety dibenzylcyclooctyne (DBCO), with the aim of crosslinking the targeted nanoparticles through a fibrin‐independent mechanism to achieve wound‐targeted clot stabilization. This process is illustrated below in Scheme  1 . Two versions of this second component were tested—a multiarm PEG‐DBCO crosslinker, or a DBCO‐functionalized nanoparticle crosslinker. Crosslinking of the two components leading to nanoparticle aggregation was first confirmed to occur only at high concentrations and not at the concentration circulating in the bloodstream. The effect of this system on platelet recruitment over multiple incubations was then assessed and demonstrated to be superior to the nanoparticle‐only system, and subsequent assays with different doses of nanoparticle revealed that a lower dose of the two‐component system could be used to achieve similar degrees of platelet recruitment to the nanoparticle‐only system, resulting also in lower complement activation. The system was also able to recover platelet recruitment in diluted plasma to levels comparable to undiluted plasma with a nanoparticle‐only treatment, as well as significantly improve clot formation relative to the nanoparticle‐only group. When challenged with a liver resection injury model over the course of three hours, the two‐component system resulted in a significant increase in survival relative to both saline and the particle‐only control groups, corroborating the enhanced performance observed in vitro and ex vivo and presenting new avenues for the development of multicomponent hemostats.

Scheme 1.

Two‐component system designed for wound‐targeted crosslinking. a) Single‐component system with activated‐platelet‐binding nanoparticles and b) two‐component system with click‐functionalized nanoparticles and crosslinking groups for wound‐targeted biorthogonal crosslinking.

2. Results and Discussion

2.1. Synthesis and Characterization of the Two‐Component System

PEG‐b‐PLGA copolymers functionalized with platelet‐binding peptide GRGDS, DBCO, Cyanine 7, or azide groups were synthesized via ring‐opening polymerization as previously described,^{[ 2d ]} with minor adjustments in solvent and reaction time made to accommodate different functional groups. These polymers were used to generate nanoparticles (NPs) of approximately 180 nm, as previous studies had demonstrated that this size was optimal for the recruitment of activated platelets and accumulation at the injury site.^{[ 17 ]} With the exception of pure azide NPs (only used in kinetic studies), all sizes fell within the range of 140–220 nm. The size of the nanoparticle crosslinker, also within the range of 140–220 nm, was chosen to match its biodistribution and pharmacokinetic properties to the targeting component, maximizing the timeframe over which the two components would coexist within the bloodstream. Larger nanoparticles were not considered due to a high degree of nonspecific pulmonary accumulation.^{[ 17} , ^{18 ]}

Cyanine 7 (Cy7)‐labeled nanoparticles were synthesized via mixing PEG‐b‐PLGA‐Cy7 with GRGDS‐PEG‐b‐PLGA and using the polymer solution for nanoprecipitation, while mixed azide‐GRGDS nanoparticles (GNPP) were likewise synthesized through mixing of the two polymers and subsequent nanoprecipitation. DBCO‐PEG‐b‐PLGA nanoparticles were also synthesized through nanoprecipitation, and 4‐arm‐DBCO PEG was purchased and used as‐is. For ease of comparison to the single‐component system, the concentration of the two‐component system refers to nanoparticle concentration unless otherwise specified. Figure  1 above provides the synthetic scheme of all functionalized PEG‐b‐PLGA polymers and representative nanoparticle size distributions as measured via dynamic light scattering (DLS). All other characterization (nuclear magnetic resonance, gel permeation chromatography, NP diameters, PDI values, and size stability) has been listed in the Supporting Information (Figures S1–S3 and Table S1, Supporting Information).

Figure 1.

Polymer synthesis and nanoparticle formulation and characterization. a) Synthesis scheme of GRGDS‐functionalized PEG‐b‐PLGA, azide‐PEG‐b‐PLGA, and DBCO‐PEG‐b‐PLGA. b) Particle size distributions of unlabeled particles, with z‐averages of each size range noted. c) Formulations of GRGDS, azide/mixed and pure DBCO nanoparticles (NPs).

2.2. Optimal Ratio of Peptide to Azide Functionality

The optimal ratio of peptide to azide functionality on mixed nanoparticles was first determined by screening the platelet recruitment ability of five different GRGDS: azide ratios: pure GRGDS‐functionalized nanoparticles, 5:1, 3:1, 1:1, and pure azide‐functionalized nanoparticles. This was performed through a lactate dehydrogenase (LDH) assay to gauge whether or not the inclusion of a crosslinkable moiety resulted in increased platelet accumulation relative to the nanoparticle‐only control, and to ensure that the click‐functionalized groups and the lower percentage of platelet‐targeting peptide did not adversely impact the ability of the nanoparticles to recruit platelets. Multiple rounds of incubation in platelet‐rich‐plasma (PRP) were performed to mimic the flow of fresh blood over the wound site, and this experiment was repeated with both four‐arm‐PEG‐DBCO (4ADP) and DBCO‐PEG‐PLGA nanoparticles (DPP). All measurements were normalized to nonspecific binding to quiescent platelets through the LDH assay.

Increased platelet recruitment was observed at several ratios with the two‐component system, occurring at the 4^th incubation for the combination of mixed GRGDS‐azide nanoparticle (GNPP) + 4ADP (Figure  2b) and at the second incubation for the combination of GNPP + DPP (Figure 2d). Some degree of nonspecific binding or saturation of platelet binding was observed, in particular during the later incubation stages following multiple additions of platelet‐rich‐plasma and nanoparticle solution, phenomena that have been previously described in light transmission‐based platelet aggregation assays and lactate‐dehydrogenase‐based binding assays.^{[ 2} , ^{18 ]} Increased nanoparticle recruitment was also observed for mixed nanoparticles (5:1 and 3:1 for GNPP + 4ADP and 3:1 and 1:1 for GNPP + DPP). For both types of crosslinkers, a GRGDS‐to‐azide ratio of 3: 1 resulted in significantly increased platelet recruitment relative to the nanoparticle‐only control—as such, all future mixed nanoparticles were synthesized according to this ratio.

Figure 2.

Nanoparticle (NP) and platelet recruitment over multiple rounds of incubation as determined via the LDH assay. a) NP accumulation when using a polymer crosslinker; b) Platelet accumulation when using a polymer crosslinker. c,d) NP and platelet accumulation when using a NP crosslinker, respectively. The original NP system has been marked out in green. For all tests, ns: p > 0.05; *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001 as determined through 2‐way ANOVA with Bonferroni post‐tests.

2.3. Crosslinking Behavior of the Two‐Component System

As wound‐targeting nanoparticles were observed in several studies to result in significantly higher accumulation at the wound site,^{[ 9} , ¹⁷ , ^{18 ]} the aim of the two‐component system was to leverage this phenomenon and promote significant crosslinking selectively at higher concentrations. The formation of a gel occurs only when polymer concentrations exceed the critical gelation concentration (CGC),^{[ 19 ]} which has been reported at concentrations of 130 mg mL⁻¹ for PEG‐PLGA^{[ 20 ]} and 260 mg mL⁻¹ for PEG‐PLA gels.^{[ 21 ]} Representative inversion tests of the nanoparticle solutions at 150 mg mL⁻¹ at 1:1 stoichiometry have been provided in Figure S4 (Supporting Information), confirming gelation within this range.

To ensure safe intravenous delivery of the two‐component system, nanoparticle concentrations ranging from 2 mg mL⁻¹ up to 50 mg mL⁻¹ were incubated with crosslinkers at a 1:1 stoichiometric ratio, between the circulation concentration of 0.5‐1 mg/mL and the CGC. A 1:1 stoichiometric ratio was shown to result in maximum crosslinking for both polymeric and nanoparticle crosslinkers, with significantly lower crosslinking observed at either insufficient or excess crosslinker (Figures S3 and S5, Supporting Information). Using dynamic light scattering to monitor crosslinking behavior, the functionality of the clickable groups was first evaluated using pure azide nanoparticles (Figure  3a), before mixed nanoparticles were also tested at these concentrations (Figure 3b). Control groups of unfunctionalized nanoparticle (MPP) and unfunctionalized four‐arm‐PEG (4AUP) were included to confirm that any observed aggregation did not occur due to high nanoparticle concentration and subsequent sedimentation.

Figure 3.

Two‐component system crosslinking behavior at various concentrations with a) Pure azide nanoparticles and b) Azide‐GRGDS nanoparticles. NPP—N₃ ‐PEG‐PLGA; DPP—DBCO‐PEG‐PLGA; MPP —methoxy‐PEG‐PLGA; 4ADP —4‐arm‐DBCO‐PEG; 4AUP —4‐arm‐PEG; GNPP— mixed GRGDS‐ N₃ ‐PEG‐PLGA; DNPP—mixed GRGDS‐DBCO‐PEG‐PLGA.

Pure azide‐functionalized nanoparticles (Figure 3a) demonstrated visible increases in size within the first two hours of the experiment when incubated with both nanoparticle and multiarm polymer crosslinkers at 50 mg mL⁻¹. No significant increases in sizes were observed at 5–10 times circulating concentration (0.5–1 mg mL⁻¹) for the nanoparticle + nanoparticle crosslinker combination (NPP + DPP), while a slight increase in size was observed in the nanoparticle + multiarm polymer system (NPP + 4ADP). No size increases were observed with the control groups of unfunctionalized nanoparticle and unfunctionalized multiarm polymer. Mixed nanoparticles (GNPP) were also observed to increase in size at higher concentrations, though the kinetics appeared to be delayed in comparison to the pure azide nanoparticles, likely due to the lower percentage of crosslinkable functionalities (Figure 3b). Two types of DBCO‐functionalized nanoparticle crosslinkers were tested in this experiment—one of pure DBCO‐PEG‐b‐PLGA (DPP), and one mixed with GRGDS‐PEG‐b‐PLGA (GDPP). Nanoparticle aggregation behavior was observed to be much more pronounced in the GNPP + DPP solution despite the same amount of DBCO‐functionalized polymer, a phenomenon that could be potentially attributed to the higher functionality per molecule of pure DBCO‐functionalized nanoparticles. As a result, all further experiments were conducted with DPP instead of GDPP, to ensure crosslinking could still occur upon accumulation at the injury site. Only minimal increases in nanoparticle size were observed for the GNPP + 4ADP combination at circulating concentration, and no increases in size were observed with unfunctionalized nanoparticles. Overall, the two‐component system with either nanoparticle or multiarm polymer crosslinker led to significant nanoparticle aggregation at higher concentrations but not at circulating concentration.

To confirm that the targeted crosslinking would proceed in vivo without unintended thrombosis from platelet aggregation, two additional experiments were conducted: flow cytometry of platelets when incubated with the two‐component system and crosslinking in a solution containing serum proteins. Crosslinking in the presence of serum proteins resulted in a similar crosslinking behavior to results obtained in deionized water (Figure S5b, Supporting Information). A slightly lower size increase was observed with the polymeric crosslinker, though the difference was not significant (P = 0.4168 and 0.3736 for 50 and 20 mg mL⁻¹, respectively). This could potentially be attributed to the similar sizes scales of the polymer crosslinker and serum proteins, which would make it difficult to differentiate size increases from the polymer crosslinker reacting with the nanoparticle or from protein absorption onto the nanoparticle. Notably, such an effect was not observed with the nanoparticle crosslinker, which is significantly larger in size compared to major serum proteins or the polymer crosslinker. This indicates that the click reaction itself is not significantly affected by the addition of serum proteins.

Finally, the two‐component system was incubated with quiescent platelets at increasing concentrations to confirm that no spontaneous platelet activation would occur (Figure S6, Supporting Information). This was quantified through the use of anti‐CD41, a platelet‐specific antibody, and then through anti‐CD62P, an activation marker; representative gating plots are as shown in Figure S6a–d (Supporting Information). The two‐component system appeared to have no evident effect on increasing platelet activation, resulting in similar populations of CD62P⁺/CD41⁺ when compared to PRP‐only controls (Figure S6m, Supporting Information), with no trend with increasing concentration up to 10 mg mL⁻¹ (20× circulating concentration). These results suggest that unintended thrombosis due to platelet targeting of the two‐component system is unlikely in the range of concentrations tested.

2.4. Platelet Recruitment at Decreased Doses of Hemostatic Nanoparticle

The dosage of hemostatic nanoparticles was then decreased to confirm if a decreased dosage of the two‐component system would achieve similar levels of platelet recruitment as a normal dose of nanoparticle‐only treatment (Figure  4a). Complement activation levels were likewise evaluated to gauge if decreasing the nanoparticle concentration could lead to lower complement production, as high levels of complement have been observed to exacerbate hemorrhage and can cause severe adverse side effects such as shock or even death.^{[ 22 ]} As can be seen, the two‐component system with both polymeric and nanoparticle crosslinker results in average platelet recruitment above that of nanoparticle‐only groups at both lower concentrations tested. This was in part due to significantly lower nonspecific binding of the two‐component system relative to the particle‐only group, as evidenced by binding to quiescent platelets (Figure S7, Supporting Information). Additionally, decreasing the dosage to 0.5 mg mL⁻¹ of nanoparticle in the two‐component system with multiarm polymer crosslinker resulted in an almost five‐fold decrease (10.7% vs 51.2%) in Complement 5a (C5a) production compared to 1 mg mL⁻¹ of the particle‐only treatment, suggesting that this system has the potential to decrease immune‐mediated side effects while retaining the hemostatic effects of particle‐only treatments. In contrast, the two‐component system with nanoparticle treatment resulted in an increase in C5a concentration, an effect that could potentially be attributed to the formation of isolated aggregates^{[ 23 ]} that were less prevalent in the nanoparticle‐polymer‐crosslinker group.

Figure 4.

Decreased doses of two‐component system result in increased specific platelet accumulation relative to nanoparticle‐only systems. a) Schematic of dilutions; b) Changes in complement levels; c) Specific platelet recruitment at two incubations and at d) four incubations. For C5a levels, n = 5, ns: p > 0.05; *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001 as determined through one‐way ANOVA with Tukey's post‐test for trial groups (without Zymosan); for all other tests, n = 6, ns: p > 0.05; *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001 as determined through two‐way ANOVA with Tukey's post‐test.

2.5. Platelet Recruitment under Dilutional Coagulopathy Conditions

Subsequently, the effect of the two‐component system on recovering platelet recruitment in hemodiluted environments was evaluated. In brief, PRP was diluted by 20% and 40% with isotonic saline, simulating Grade II–Grade IV blood loss with subsequent fluid resuscitation (Figure  5a). The two‐component system was observed to significantly enhance platelet recruitment in comparison to the single‐component system at these diluted concentrations for both two incubations and four incubations. In addition to that, the system was observed to recover platelet recruitment to a level equivalent to 20% less dilution (e.g., 20% dilution versus undiluted plasma or 40% dilution versus 20% dilution), demonstrating the improved functionality of the two‐component system relative to the single‐particle system even in hemodiluted environments with decreased fibrinogen, platelet, and clotting factors.

Figure 5.

Two‐component system recovers platelet recruitment in hemodiluted conditions. a) Schematic of the procedure; The two‐component system is capable of recovering platelet recruitment in diluted PRP relative to the single component system at higher plasma concentrations for b) two incubations and c) four incubations. For all tests, n = 6, ns: p > 0.05; *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001 as determined through two‐way ANOVA with Tukey's post‐test.

2.6. Fibrin Clot Formation and Degradation under Severe Dilutional Coagulopathic Conditions

The two‐component system was found to significantly improve clot formation and decrease clot degradation under severe dilutional coagulopathic conditions. In brief, PRP was isolated from citrated whole blood and diluted by 40% in isotonic saline, whereupon it was incubated with trial groups of saline, GRGDS‐NPs, and both polymeric and nanoparticle versions of the two‐component at a concentration of 5 mg mL⁻¹ nanoparticle. This concentration—5 to 10‐fold the circulating concentration of 0.5–1 mg mL⁻¹—was used in the following experiments, as prior studies had demonstrated that the hemostatic nanoparticles accumulated at the injured vessel section at approximately 5 to 15 times relative to uninjured vessels,^{[ 17 ]} though it is possible that concentrations localized at the point of hemorrhage could exceed even that. The absorbance of these solutions following the addition of CaCl₂ was then monitored to yield the overall coagulation potential (OCP) and change in transmission as a measure of fibrin polymerization (Figure  6a,c).

Figure 6.

Two‐component system increases fibrin crosslinking and decreases clot plasminolysis by ≈40% in dilutional coagulopathic clotting conditions. a) Representative absorbance curves for fibrinogen coagulation and b)the overall coagulation potential is significantly increased for the two‐component system; c) Representative transmission curves for fibrinogen coagulation. d) Fibrin crosslinking is significantly increased with the two‐component system relative to the single‐component system and the control. e) Lightening and movement of interface in control clot (top: saline; bottom: plasmin); f) Degradation profiles of fibrin clot, nanoparticle + fibrin clot, and two‐component system + fibrin clot. For a–d) n = 12, ns: p > 0.05; *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001 as determined through one‐way ANOVA with Tukey's post‐test; for (e,f), n = 5, ns: p > 0.05; *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001 as determined through two‐way ANOVA with Tukey's post‐test.

As can be seen, the two‐component system is capable of significantly increasing fibrin polymerization by up to 149% even in severe dilutional coagulopathic conditions relative to the saline control (Figure 6b,d). In particular, the two‐component system with polymeric crosslinker also significantly enhanced clot formation relative to the nanoparticle‐only control (132% and 125% for OCP and change in transmission respectively). While similar anticoagulated or coagulopathic models have been proposed in literature to evaluate the effect of synthetic hemostats on clot formation, they have generally only depleted or inhibited specific components of hemostasis, such as platelets, coagulation factors, or fibrinogen.^{[ 8} , ^{24 ]} Direct dilution of platelet‐rich‐plasma with saline instead of platelet‐poor‐plasma results in the decrease of all plasma components, which is observed upon fluid resuscitation in response to massive hemorrhage.^{[ 25 ]} Overall, the two‐component system shows promise in recovering fibrin polymerization, which is critical to achieving hemostasis in traumatic bleeding.

Subsequently, fibrin clot degradation was assessed using a 1D degradation assay in a glass capillary, as fibrinolysis is likewise exacerbated in dilutional coagulopathy.^{[ 26 ]} Fluorescently labeled fibrinogen was gelled with thrombin at 1.25 mg mL⁻¹ in a capillary, this time mimicking severely coagulopathic conditions below critical fibrinogen thresholds with complete depletion of platelets and coagulation factors.^{[ 8b ]} This was then filled with plasmin using a gel loading tip, and fibrinogen degradation was measured via diffusion of the labeled fibrinogen into the plasmin solution, as illustrated in Figure 6e.

The inclusion of the single‐component hemostatic nanoparticles led to a slight decrease (≈25%) in the level of fibrinogen diffusion into the plasmin chamber over the course of 24 h, while the addition of the two‐component system led to approximately a 40% decrease in fibrinogen loss. Figure 6f illustrates the degradation profiles of fibrin clots in nanoparticle‐free, hemostatic nanoparticle, and two‐component system trial groups, where fibrinolysis has been mitigated over time. In conjunction with the enhancements in fibrin polymerization and platelet recruitment, these results demonstrate the significant benefit of wound‐targeted crosslinking to promote clot formation under coagulopathic conditions.

2.7. Biodistribution, Pharmacokinetics, and Biocompatibility in an Uninjured Mouse Model

To ascertain the safety of the system, the biodistribution, pharmacokinetics, and biocompatibility in an uninjured mouse model were first measured using Cy7‐labeled fluorescent nanoparticles. A study on injection order of the nanoparticles and crosslinkers was also conducted, where it was ascertained that this had no significant effect on biodistribution or pharmacokinetics (Figure S8, Supporting Information). The effect of DBCO crosslinker concentration on retention and blood circulation lifetime was also assessed. As shown below in Figure  7 , there appears to be a slight decrease in retention time for higher concentrations of polymeric crosslinker, though this was not found to be significant. Notably, this was not observed in the two‐component system with nanoparticle crosslinker. No differences in nanoparticle circulating concentration could be detected among the trial groups (Figure 7b), and no visible differences in accumulation could be observed via IVIS images (Figure S9, Supporting Information). These mice were monitored for two weeks post‐injection for any long‐term inflammation due to the injection through the measurement of TNFα and IL‐6 levels, cytokines commonly used to indicate the presence of inflammation.^{[ 27 ]} No significant differences were observed between trial groups and mice that had not been injected (Figure S10, Supporting Information), indicating that the two‐component system does not pose any additional risk for long‐term inflammation despite its additional crosslinking ability. This may be because at low concentrations any reaction between the two components is incapable of forming large aggregates, as shown in vitro in Figure 3.

Figure 7.

Biodistribution and pharmacokinetics of the two‐component system at various crosslinker concentrations relative to nanoparticle‐only groups. a) Retention of the two‐component system and controls. b) Blood circulation concentration of the two‐component system and controls. Biodistribution not normalized to mass c) and normalized to mass d) for the two‐component system and controls.

Similarly, organ biodistribution was found to remain similar over all groups tested, with the highest concentrations of DBCO crosslinker selected from the previous study to visualize any potential differences between groups. This indicates that despite the higher degree of kidney clearance of polymers (relative to nanoparticles),^{[ 28 ]} the non‐targeted crosslinkers did not significantly skew the biodistribution of the peptide‐functionalized, targeted nanoparticles. Additionally, no significant increases in pulmonary accumulation were observed in the two‐component system, corroborating the prior in vitro results that demonstrated no significant nanoparticle aggregation at circulating conditions. Histology results one week post‐injection were also obtained (Figure S11, Supporting Information), where gross examination of organ cross‐sections did not reveal notable differences from the isotonic glucose control.

2.8. Hemostatic Efficacy of the Two‐Component System

The hemostatic efficacy of the two‐component system was first evaluated in a closed liver resection model, with an observation period of three hours. This model was selected due to the prevalence of liver injuries in traumatic hemorrhage, specifically as the most commonly injured organ in blunt abdominal trauma and the second‐most injured organ in penetrating abdominal trauma. Additionally, liver hemorrhage models have been established in literature as one of the primary methods in which intravenous hemostats have been evaluated.^{[ 9} , ¹⁸ , ^{29 ]} The timeframe was selected based on the average time for prehospital care, including prehospital time and on‐scene time. For civilians, median values range from 47 min to 61 min,^{[ 30 ]} while for field applications this range is extended to 73 to 130 min^{[ 31 ]} and in some cases exceed a day.^{[ 32 ]} This is reflected in literature, where all intravenous hemostats tested in a lethal injury model to date have been evaluated over a timeframe of 1–3 h for their ability to prolong survival.^{[ 8} , ¹⁷ , ¹⁸ , ^{33 ]} BALB/c mice were dosed with prophylactic injections of the two‐component system, as vasoconstriction due to acute hemorrhage led to challenges in successful tail‐vein injections. This prophylactic approach is considered standard in the dosing of intravenous hemostats via the tail vein in mouse injury models.^{[ 8} , ²⁹ , ^{34 ]}Two doses of nanoparticle‐only treatment were administered: the original concentration at 1 mg mL⁻¹, which was previously observed to increase complement activation by ≈50% relative to the particle‐free concentration (Figure 5), and the decreased dose at 0.5 mg mL⁻¹ for comparison with the two‐component system at the same nanoparticle concentration. Three mice per group were injected with Cy7‐labeled nanoparticles and dissected to obtain organ biodistribution.

As can be seen in Figure  8 , both the two‐component systems with polymeric and nanoparticle crosslinkers resulted in significantly increased survival (100% survival at P = 0.0099** and 87.5% survival at P = 0.0316*, respectively). Though prophylactic treatment may have contributed to increased efficacy, these results are relative to the nanoparticle‐only group at the same dose (40% survival), as well as the control. This is especially significant, as the nanoparticle‐only group (GRGDS‐NP) has been demonstrated to prolong survival in a treatment model of a lethal inferior vena cava puncture in a prior study. Overall, these results indicate that the two‐component system's ability to prolong survival exceeds that of the original system, which has been shown to be successful in treatment models. While no significant differences in blood loss were observed between these trial groups, this phenomenon is consistent with similar studies on uncontrolled, lethal hemorrhage models.^{[ 9} , ¹¹ , ^{18 ]}

Figure 8.

Hemostatic efficacy of the two‐component system in a lethal mouse liver resection model. a) Schematic of the procedure. b) Accumulation in resected versus remnant liver. c) Blood loss. d) Survival over three hours; ns: p > 0.05; *: p ≤ 0.05; **: p ≤ 0.01; ***: p ≤ 0.001 as determined through Kaplan‐Meier analysis and two‐way ANOVA.

The two‐component system also resulted in a significant difference in remnant versus resected liver accumulation, which was not observed in the nanoparticle‐only system, potentially indicating increased accumulation at the injury site. The two‐component system with nanoparticle crosslinker likewise exhibited a similar trend in enhanced liver accumulation. As no significant differences in blood fluorescence levels were detected (Figure S8, Supporting Information), this suggests that the difference in accumulation is not entirely due to a higher level of hepatic clearance over time, as the concentration of nanoparticles in blood should have also decreased. This corroborates ex vivo results obtained in Figure 2, where increased nanoparticle and platelet accumulation was observed upon the inclusion of a crosslinking component to the nanoparticle‐only system. The biodistribution profiles of other organs have been provided in Figure S12 (Supporting Information).

Finally, a pilot tail transection experiment was included to gauge the effect of the two‐component system on short‐term bleeding time. BALB/c mice were dosed with prophylactic injections of the two‐component system or the isotonic control, before 5 mm of the tail tip was transected and immersed in a tube of saline. The time required for cessation of bleeding ranged from 3.5 to 11.2 minutes for the two‐component system and from 7.08 to 25.6 minutes for the control (≈2.3× as long), indicating the ability of the two‐component system to expedite hemostasis. This is also reflected in blood loss (quantified via hemoglobin content), with the two‐component system group experiencing approximately one third of the blood loss suffered by the control (Figure S13, Supporting Information).

The results of the injury model corroborate conclusions drawn from prior in vitro and ex vivo experiments regarding the efficacy of the two‐component system. Specifically, the two‐component system with polymer crosslinker overall achieved the most significant improvement in platelet recruitment both under normal and hemodiluted conditions, in decreasing complement activation, and in fibrin polymerization relative to the nanoparticle‐only treatment group. The comparatively poorer performance of the two‐component system with nanoparticle (though still significantly higher than that of the nanoparticle‐only group) may be attributed to a number of factors, including but not limited to increased nonspecific binding (Figure 2), slower crosslinking kinetics (Figure 3), and increased complement activation. While these differences do not significantly influence survival in the liver resection model, they may factor in the selection of trial groups in testing for larger animal models and serve to validate the efficacy of the in vitro assays in screening synthetic hemostats.

To this date, the transfusion of blood products (whole blood, coagulation factors, and blood components) remains the preferred treatment for internal bleeding in the clinic, though this is not always feasible due to the perishability and advanced storage needs of these materials.^{[ 2} , ¹⁷ , ¹⁸ , ^{24 ]} Researchers have thus sought to address this by engineering synthetic hemostats that mimic blood components, such as peptide‐decorated polymers, nanoparticles and liposomes, and have demonstrated the efficacy of these materials across a wide variety of injury models.^{[ 17} , ³³ , ^{35 ]} Prior to this study, one approach that has yet to be investigated in nanoparticle‐based hemostats is the strengthening of the clot through fibrin‐independent mechanisms. Fibrin is critical to the formation of a clot—however, it is also the first of the coagulation factors to fall below critical levels upon blood loss and hemodilution,^{[ 36 ]} ahead of other blood proteins such as thrombin and FVII.^{[ 36} , ^{37 ]} In addition, it is sensitive to the effects of the trauma triad beyond coagulopathy: acidosis increases the rate of clot lysis,^{[ 38 ]} while hypothermia further decreases fibrinogen synthesis.^{[ 39 ]} In the absence of stable clot formation—which may occur as part of trauma‐induced coagulopathy^{[ 40 ]}—there is a significant need to target the site of hemorrhage and strengthen the clot in a fibrin‐independent manner.

The two‐component system proposed herein provides a potential solution to these points. By leveraging increased accumulation of activated‐platelet‐targeted materials at the wound site—a phenomenon widely reported in literature^{[ 17} , ^{41 ]}—this system applies the concept of the critical gelation concentration to achieve wound‐targeted clot strengthening without increased lung accumulation or systemic toxicity. Its targeting mechanism through activated platelets and bioorthogonal crosslinking through azide‐DBCO reactions mitigate the effect of fibrin depletion in severe hemorrhage, as demonstrated through enhanced platelet recruitment and overall coagulation potential even in fully hemodiluted environments. This is particularly significant, as the effect of nanoparticle hemostats on clot stability under simultaneously decreased levels of clotting factors, platelets, and fibrinogen has yet to be reported in literature. The decreased therapeutic dose, nonspecific binding, and complement activation of the two‐component system relative to the NP‐only treatment also suggest that this approach may result in fewer side effects upon translation to more advanced animal models, while its increased survival affirms the effectiveness of this treatment in the context of already highly successful NP‐only systems.^{[ 17} , ¹⁸ , ^{41 ]} Future studies may expand towards engaging specific factors of the coagulation cascade^{[ 24a ]} or to larger animal models in order to further optimize and gauge the translational potential of this system. Ultimately, the results of this study demonstrate the promise of a multi‐pronged approach toward treating internal hemorrhage, and introduce the use of enhanced injury accumulation and bioorthogonal crosslinking to establish wound‐targeted hemostatic effects.

3. Conclusions

In this work, a two‐component hemostat for targeted, biorthogonal crosslinking was developed for the treatment of internal bleeding by functionalization of PEG‐b‐PLGA nanoparticles with GRGDS and clickable azide moieties. Nanoparticles and multiarm polymers with corresponding copper‐free crosslinking groups (DBCO) were delivered as a second component. The two‐component system was demonstrated to only result in particle aggregation/significant concentration at several times circulating concentrations, and to enhance both platelet and nanoparticle recruitment at a peptide : azide ratio of 3: 1. In addition to that, the system appeared to slow plasmin lysis of fibrin clots as monitored through a 1D capillary assay, recover platelet recruitment and fibrin polymerization in hemodiluted environments, and require a lower therapeutic dose that resulted in decreased complement activation. Biodistribution, pharmacokinetics, and biocompatibility were evaluated in an uninjured model, and no signs of long‐term inflammation or skewed biodistribution from the targeted nanoparticles were observed, confirming the safety of the system for systemic delivery. Finally, the two‐component system was demonstrated to result in significantly improved survival relative to the nanoparticle‐only group in a mouse liver resection model. These findings underscore the importance of supplementing the native hemostasis process in a comprehensive manner, providing guiding principles for the future development of artificial blood replacements.

4. Experimental Section

Materials

All chemicals and proteins were purchased from Sigma Aldrich unless otherwise specified. GRGDS peptide (95%) was purchased from China Peptides. RIPA Lysis Buffer, CyQUANT LDH Cytotoxicity Assays, Complement C5a Human ELISA kits were purchased from ThermoFisher Scientific; TNFα and IL‐6 kits were purchased from Abcam. 4‐arm‐PEG‐DBCO (20 kDa) was purchased from CreativePEGworks, and 4‐arm‐PEG‐OH (20 kDa) was purchased from JenKem; both polymers were used as‐is. Cyanine 7 free acid (Cy7) and N‐hydroxysuccinimide‐DBCO were purchased from Lumiprobe. Deuterated solvents (CDCl₃ and DMSO‐d₆) were purchased from Cambridge Isotope Laboratories. Citrated whole human blood was acquired from Research Blood Components (Watertown, MA) as a commercial, de‐identified specimen, and donor consent was obtained via the company. Female BALB/c mice were purchased from Taconic.

Characterization Methods

All synthesized materials were characterized as previously described.^{[ 17 ]} In brief, polymers were characterized via nuclear magnetic resonance (NMR) on a Bruker Advance III DPX 400 spectrometer in deuterated DMSO at a concentration of 5 mg mL⁻¹. The ratio of lactide to glycolide was quantified via lactide protons (≈5.2 ppm) and glycolide protons (≈4.8 ppm), and the molecular weight of the polymer was calculated using the PEG macroinitiator as a standard (≈3.5 ppm). Mnova software was used to perform the analysis.

Gel permeation chromatography (GPC) was carried out at 1 mg mL⁻¹ using an Agilent 1260 GPC system with three DMF ResiPore columns, a Wyatt Mini‐DAWN TREOS 3‐angle static light scattering detector, and a Wyatt Optilab T‐rEX refractive index detector, with DMF + 0.01 m LiBr used as the eluent. Amino acid analysis was performed on an Agilent 1260 Infinity Quaternary LC System to detect peptide conjugation efficiency. Column conditions and elution gradients used were as previously described.^{[ 18a ]} The polymer sample was dissolved in 6 n HCl at a concentration of 10 mg mL⁻¹ and hydrolyzed at 105 °C for 24 hours. Subsequently, excess HCl was removed via vacuum, and the residue was redissolved in 0.1 n HCl for analysis. A standard curve was generated for aspartic acid over the range of 0.01–1 × 10⁻³ m and used to calculate the amount of peptide conjugated to polymer (21 µmol GRGDSg⁻¹ of PEG‐b‐PLGA).

Dynamic light scattering measurements were performed on Malvern Zetasizer Nano ZS90 in deionized water at 0.25 mg mL⁻¹. The laser wavelength on the machine was 633 nm. The acquisition angle was 90°. Acquisition times were set to automatic on the machine and averaged 60–70 s per run.

Polymer and Nanoparticle Synthesis: Azide‐Poly(ethylene glycol)‐b‐Poly(d,l‐lactide‐co‐glycolide (N₃‐PEG‐b‐PLGA)

450 mg of azide‐PEG‐OH (5 kDa, purchased from Sigma‐Aldrich) was dissolved in 5 mL of dimethylformamide (DMF) and 5 mL of tetrahydrofuran (THF) and left overnight on activated molecular sieves. The solution was then transported to the glovebox, where 1828 mg of racemic lactide was added to the solution and dissolved. The flask was then sealed with a rubber septum. In a separate glass vial, 1120 mg of glycolide was dissolved in 4.5 mg of DMF and taken up in a syringe. In another glass vial, 22 µL of 1,8‐diazabicyclo(5.4.0)undec‐7‐ene (DBU) was mixed with 2 mL of DMF and also taken up in a syringe. The DBU solution was injected into the lactide and PEG solution; immediately afterward, the glycolide solution was slowly infused into the flask via syringe pump at a rate of 60.4 µL per min. The reaction was allowed to proceed for 90 min, before it was terminated via the addition of excess benzoic acid. The polymer was precipitated into isopropanol twice and diethyl ether once, resulting in a yield of 84% (≈2.5 g).

DBCO‐Poly(ethylene glycol)‐b‐Poly(d,l‐lactide‐co‐glycolide (DBCO‐PEG‐b‐PLGA)

DBCO‐functionalized PEG‐b‐PLGA was synthesized via the use of DBCO‐PEG macroinitiator. 100 mg of DBCO‐ N‐hydroxysuccinimide ester (DBCO‐NHS) was added to 1 g of NH₂‐PEG‐OH (5 kDa) at a slight molar excess (≈1.16 equiv.), dissolved in 10 mL of anhydrous DCM, and stirred at RT overnight to react. The polymer solution was then concentrated to 5–6 mL and precipitated in ether three times before it was dried under vacuum, resulting in a yield of 90% (90 mg). This macroinitiator was used in the same manner described in the original NHBoc‐PEG‐b‐PLGA ring‐opening polymerization reaction, with no additional changes to solvent conditions or reaction time (80 min).

GRGDS‐Poly(ethylene oxide)‐b‐Poly(d,l‐lactide‐co‐glycolide) (GRGDS‐PEG‐PLGA)

GRGDS‐PEG‐b‐PLGA was synthesized as previously described^{[ 17 ]}—a summary of the protocol has been provided below. In brief, OH‐PEG‐NH₂ was protected with Boc₂O by incubating the polymer with Boc anhydride for three hours at room temperature and precipitating three times in ether. Following this step, the dried polymer was dissolved in dichloromethane (DCM) with lactide, while a glycolide solution in dimethylformamide (DMF) was prepared in a syringe, with a target molecular weight of ≈30 kDa PLGA. The reaction mixture was catalyzed via DBU and allowed to react for 80 min, over which the glycolide‐DMF solution was infused via syringe pump. The polymer was precipitated twice in isopropanol and once in diethyl ether before it was dried under vacuum. Deprotection of the Boc protecting group was performed in 50% trifluoroacetic acid (TFA)and dichloromethane (DCM) overnight at room temperature, and the polymer was precipitated twice in diethyl ether after vacuum removal of TFA and DCM. Conjugation of the peptide was accomplished through a two‐step process: activation of the amine terminus via carbonyldiimidazole (CDI) in dioxane at 37 °C, followed immediately by incubation with a five‐fold excess of GRGDS peptide over 48 h. Excess peptide was then removed by dialyzing the solution for 72 h in deionized water, after which the polymer was freeze‐dried and stored at ‐20 °C.

Poly(ethylene oxide)‐b‐Poly(d,l‐lactide‐co‐glycolide)‐Cyanine 7 (PEG‐PLGA‐Cy7)

PEG‐PLGA‐Cy7 was synthesized as previously described^{[ 17 ]}—a summary of the protocol has been provided below. In brief, one equivalent of Cyanine 7 free acid was added to PEG‐b‐PLGA‐OH (synthesized using a methoxy‐PEG macroinitiator) and dissolved in DCM with 0.4 equivalent of 4‐dimethylaminopyridine (DMAP). N,N'‐dicyclohexylcarbodiimide (DCC) was dissolved separately in DCM and added dropwise to the polymer solution under stir to achieve a final concentration of 50 mg polymer mL⁻¹ solvent. The solution was stirred overnight at RT and precipitated in isopropanol and diethyl ether.

Nanoparticle Synthesis

All nanoparticles were synthesized through nanoprecipitation. 40 mg of PEG‐b‐PLGA polymer and 5 mg of Resomer 503H pure PLGA were dissolved in 1.5 mL dimethylformamide (DMF) and tetrahydrofuran (THF), stirred overnight, and sonicated until the solution was clear (≈15 min). The solution was then added dropwise into deionized water and stirred at 720 rpm, before the resulting nanoparticle solution was washed repeatedly in Amicon ultracentrifugal filters with more deionized water to remove excess solvent and concentrate the solution. All nanoparticles were used within three days of purification, a timeframe over which they remained stable as measured via dynamic light scattering.

Characterization of Nanoparticle Crosslinking via Dynamic Light Scattering

The crosslinking stability of click‐functionalized nanoparticles was evaluated at concentrations of 2, 5, 10, 20, and 50 mg mL⁻¹. Circulating concentrations in animals were approximately 1–2 mg mL⁻¹. In brief, azide‐functionalized nanoparticles were formulated using the synthesis conditions for the intermediate nanoparticle size, though the eventual size of the nanoparticles was slightly smaller—approximately 120 nm in diameter. DBCO‐functionalized nanoparticles and methoxy‐functionalized nanoparticles were similarly synthesized, as were mixed nanoparticles with varying ratios of GRGDS peptide to azide functionality. Four‐arm‐DBCO‐functionalized PEG (4ADP, MW = 20 kDa) and four‐arm unfunctionalized PEG (4AP, MW = 20 kDa) was dissolved to stoichiometric concentrations calculated from the molecular weights of the azide‐functionalized polymer. For a purely azide‐functionalized polymer (≈27 kDa) nanoparticle at 1 mg mL⁻¹ in deionized water, this would be equivalent to 0.1852 mg mL⁻¹ 4ADP (20 kDa). The formula for obtaining this is shown below:

$${\mathrm{Concentration}}_{4\mathrm{ADP}}=\frac{{\mathrm{Concentration}}_{\mathrm{azide}}}{{\mathrm{MW}}_{\mathrm{azide}-\mathrm{PEG}-\mathrm{PLGA}}}\times \frac{{\mathrm{MW}}_{4\mathrm{ADP}}}{4}$$ (1)

For each concentration previously specified, 25 µL of azide‐functionalized nanoparticle at twice the desired concentration was added to a PCR tube. 25 µL of DBCO‐functionalized nanoparticle was added to this tube. This step was repeated with 25 µL of unfunctionalized nanoparticle, 25 µL of four‐arm‐PEG‐DBCO, and 25 µL of unfunctionalized four‐arm‐PEG in three different PCR tubes. These tubes were incubated at 37 °Cwith a water basin to prevent evaporation, and these were measured at the start of the experiment to record the size at the start of the experiment. Samples were taken at 2, 4, and 24 h post‐incubation to track size changes in the nanoparticles, using the unfunctionalized combinations as negative controls. A separate experiment with different stoichiometric ratios (azide: DBCO = 0.75 or 1.25) was performed to confirm that a 1:1 stoichiometric ratio indeed resulted in maximum crosslinking, and the effect of a protein corona on crosslinking was evaluated by repeating the crosslinking stability experiment in cell culture media with 10% fetal bovine serum (FBS).

Blood Components Preparation

Citrated whole blood was separated into hematocrit and platelet‐rich plasma (PRP) upon receipt. In brief, blood was centrifuged at 200 rcf in a 50 mL conical tube for 20 min with no brake at room temperature. The upper fraction, PRP, was then used in subsequent ex vivo assays (platelet recruitment and transmission‐based fibrin polymerization assays).

In Vitro/Ex Vivo Evaluation of Hemostatic Efficacy: Determining the Optimal Peptide to Azide Ratio Using a Platelet Recruitment Assay

The effect of nanoparticle size on the total amount of platelet aggregation/binding to a surface was evaluated by a lactate dehydrogenase (LDH) assay. In brief, a mixture of platelet‐rich plasma and nanoparticle solution was added to black cell‐treated wells (54 µL PRP to 6 µL nanoparticle stock solution at 11 mg mL⁻¹). Half the wells received 6 µL adenosine diphosphate stock solution (ADP) as a platelet activation agonist, while the other half received saline. Four‐arm‐PEG‐DBCO and DBCO‐PEG‐PLGA nanoparticles were added in a 6 µL dose at stoichiometric equivalence to the azide‐functionalized nanoparticles. The mixture was allowed to incubate for an hour, before it was washed three times with saline (100 µL per well). The prior few steps (addition of PRP, nanoparticle stock, crosslinker/saline, and agonist/saline) were then repeated for following incubation steps. After the desired number of incubations, the wells were again washed three times with saline before 60 µL of Pierce RIPA lysis buffer was added. These samples were then diluted threefold and then used in the CyQuant LDH assay according to manufacturer's instructions in a clear‐bottom well plate. The plate's absorbance was then measured at 490 nm and 680 nm, with the final results calculated by subtracting the readings at 680 nm for those at 490 nm. Six incubations were completed for the four‐arm‐PEG‐DBCO system, and four incubations were completed for the DBCO‐PEG‐PLGA system. Each trial group included six replicates with the exception of outliers as excluded via Grubbs’ test.

Platelet Flow Cytometry

Flow cytometry was conducted using a FACS LSR II instrument (BD Biosciences) with FACSDIVA software and 405, 488, 561, and 640 nm lasers in a modified protocol from the literature.^{[ 17 ]} In brief, a diluted solution of activated PRP was prepared by combining 85 µL of FACS buffer (0.9 wt % saline + 5% bovine serum albumin), 15 µL of 100 × 10⁻³ m calcium chloride, 100 µL of thrombin at 100 U mL⁻¹, 150 µL of 15 × 10⁻³ m GPRP‐NH₂, and 150 µL PRP. Inactivated PRP was prepared similarly by substituting thrombin and calcium chloride with FACS buffer. Nanoparticle solutions were concentrated to 100, 50, 10, 5, and 1 mg mL⁻¹ and added to PRP in a 1:10 ratio (yielding 10, 5, 1, 0.5, and 0.1 mg mL⁻¹ solutions), and crosslinkers were added in stoichiometric proportions. PRP‐only positive and negative controls received an equivalent amount of FACS buffer. All solutions were incubated for half an hour under agitation at 37 °C. The two antibodies (anti‐CD41‐AF488 and anti‐CD62P‐BV421) were diluted at a 1:5 ratio from the concentration provided by the manufacturer, and 5 µL of each antibody was added to the mixture. For unstained and single‐color controls, the deficit in volume was compensated with an equivalent amount of FACS buffer. The samples were incubated at RT for half an hour in the dark, before they were fixed with 600 µL of 1% paraformaldehyde in saline. Samples were gated for CD41‐positive events to separate platelets from other cells or cell debris, then gated for CD62P‐positive events to identify activated platelet populations (CD62P⁺/CD41⁺). Representative dot plots and analysis of the acquired data was completed with FlowJo and are displayed in Figure S6 (Supporting Information).

Evaluating Platelet Recruitment at Lower Nanoparticle Doses

A variation of the prior assay using different nanoparticle doses was later completed with four incubations. Nanoparticle stock solutions were formulated at 11 , 5.5, and 2.75 mg mL⁻¹ and diluted in PRP as described to attain final concentrations of 1, 0.5, and 0.25 mg mL⁻¹, respectively. Only 0.5 and 0.25 mg mL⁻¹ concentrations of DBCO‐NP crosslinkers were tested to keep total nanoparticle concentration below 1 mg mL⁻¹, as complement activation was observed at higher nanoparticle concentrations. All subsequent steps are as described in the prior section. Each trial group included six replicates with the exception of outliers as excluded via Grubbs’ test.

Evaluating Complement Activation of Hemostatic Nanoparticles

Complement activation of hemostatic nanoparticles was measured following a protocol modified from literature.^{[ 22 ]} In brief, PRP was incubated with nanoparticle solutions or zymosan (positive control, 0.5 mg mL⁻¹) at concentrations of 0.5–1 mg mL⁻¹ in isotonic glucose for 45 min at 37 °C under agitation. They were then spun down at 3000×g to remove nanoparticles and platelets. The resulting clear serum was then assayed via Invitrogen's Complement C5a Human ELISA Kit according to manufacturer's instructions at a dilution of 1:25 (kit included pre‐coated ELISA wells). All reported results were normalized to C5a concentrations in a blank isotonic glucose control.

Evaluating Platelet Recruitment in Hemodiluted Conditions

A variation of the platelet recruitment assay was again performed, this time with various dilutions of plasma. In brief, plasma was diluted to 60% and 80% its original concentration by combining 30 mL of PRP with 20 mL of isotonic saline or 40 mL of PRP with 10 mL of isotonic saline, respectively. These were denoted as ‐40%, ‐20%, and 0% dilution, respectively. The PRP dilutions and undiluted PRP were all tested with the two‐component system and nanoparticle‐only groups at 0.5 mg polymer per mL PRP. All subsequent steps are as described in the section “Determining the optimal peptide to azide ratio using a platelet recruitment assay.” Each trial group included six replicates.

Evaluating Fibrin Crosslinking in Hemodiluted/Dilutional Coagulopathy Conditions

Fibrin crosslinking under dilutional coagulopathy conditions was evaluated using platelet‐rich plasma diluted to 60% its original concentration to simulate Grade IV hemorrhage. In brief, 7.5 µL of nanoparticle solutions at 44 mg mL⁻¹ or deionized water, 6 µL of CaCl_2, and 52.5 µL of diluted PRP was added to each well of a clear flat‐bottom non‐coated 96‐well plate to obtain a final concentration of 5 mg nanoparticle mL⁻¹ PRP. The change in absorbance over time was then monitored at 650 nm at one‐minute intervals for one hour. All absorbance or transmission measurements were normalized to initial values at t = 0 min to account for variations due to nanoparticle scattering. The overall coagulation potential, OCP, was calculated by plotting the normalized absorbance curves and computing the area underneath each curve; the change in transmission was calculated by converting absorbance to transmission via the following equation: Transmission = 10^{2 − Absorbance} and subtracting the initial transmission value. Twelve replicates were collected for this measurement.

Evaluating Fibrin Clot Lysis in Hemodiluted/Dilutional Coagulopathy Conditions

A 1D capillary degradation assay was developed in order to assess the effect of the two‐component system on the stability of fibrin clots when exposed to plasma. 50 mg of human fibrinogen (plasminogen‐free) was dissolved in 5 mL HEPES buffered saline and labeled using AF488‐N‐hydroxysuccinimide (AF488‐NHS). This was washed multiple times using an Amicon ultracentrifugal filter or dialyzed overnight to remove unreacted dye, before it was diluted to a stock concentration of 5.5 mg mL⁻¹. Square glass capillaries of 1.0 mm inner diameter were ordered from VitroCom. 250 µL of fibrinogen solution was added to 250 µL nanoparticle stock solution and 100 µL HEPES buffered saline or 100 µL multiarm polymer/crosslinker solution and mixed well. 30 µL of thrombin (100 U mL⁻¹) and 500 µL HEPES buffered saline were added to a glass shell vial and mixed thoroughly. The fibrinogen mix (≈1.25 mg mL⁻¹) was then pipetted into the glass shell vial, and the capillaries were immediately immersed in the gelling mixture for 45 min. One end of the capillary was sealed with vacuum grease. Plasmin was then added using a gel loading pipette tip directly at the capillary interface until the capillary was full, at which point it was also sealed with vacuum grease. The capillaries were then imaged on a fluorescent microscope for a period of 24 h to track the degradation of the fibrin clot, which was measured by the increase in fluorescence due to fibrinogen loss on the plasmin end of the capillary.

In Vivo Experiments: Biodistribution, Biocompatibility, and Pharmacokinetics

Female BALB/c mice (12‐15 weeks old, 20–30 g) were used in accordance with procedures approved by the MIT Division of Comparative Medicine and the Institutional Animal Care and Use Committee (IACUC) of MIT (approval number 0821‐069‐24). Animals had ad‐libitum access to both food and water until the time of the procedure. Mice were injected via tail vein with the two‐component system (hemostatic nanoparticle and nanoparticle crosslinker), with the two injections spaced five minutes apart. In a separate study to compare injection order, mice were injected with the materials in a different order (first NP then crosslinker, or crosslinker then nanoparticle). The nanoparticles were labeled with Cy7 to facilitate imaging via In Vivo Imaging System (IVIS) at Ex. 745 nm/Em. 800 nm. Trial groups included nanoparticle + saline, unfunctionalized nanoparticle + hemostatic nanoparticle, various concentrations of crosslinking polymer, and a 1:1 stoichiometric ratio of crosslinking nanoparticle to hemostatic nanoparticle. These mice were imaged at set intervals over a period of 2.5–3.5 h, and then sacrificed to obtain the organ biodistribution of hemostatic nanoparticles. Blood was drawn from the cheek at the start and at the end of the experiment to evaluate the circulating concentration of the different trial groups. A small cohort of mice was monitored over a week post‐injection, then sacrificed to obtain histology data of their liver, spleen, heart, lungs, and kidneys. The organs were fixed in 10% neutral buffered formalin for 24 h, before they were transferred to 70% ethanol, embedded in paraffin, sliced cross‐section and stained via H&E. Finally, another small cohort of mice was monitored over the next few months for any adverse effects; cheek bleeding was performed at two weeks post‐injection. TNFα and IL‐6 levels were determined through ELISA according to manufacturer's instructions.

Lethal Liver Resection Model

Female BALB/c mice (20–30 g) were injected via tail vein with the two‐component system (hemostatic nanoparticle and crosslinker), with the two injections spaced five minutes apart. The mouse was anesthetized via inhaled isoflurane at 2.0–2.5%, and depth of anesthesia was assessed with toe pinch. The mouse was then positioned supine on a surgical board and its four limbs secured with tape, with a heating pad placed beneath the mouse. The abdomen and groin were shaved and wiped down with betadine. A tube containing 0.5 mL of PBS, two pre‐weighed gauze pieces, and one weigh boat was prepared. The liver was exposed via a ventral midline laparotomy incision, and one piece of gauze was placed on either side of the abdominal cavity. A ≈5 mm section of the left‐middle lobe of the liver was removed with scissors, and this section was placed in the tube with PBS. The incision was then closed with wound clips. The mouse was observed under anesthesia for a period of <3 h and euthanized at 3 h via cardiac puncture or isoflurane overdose (e.g., 5% isoflurane for 5 min and vital organ removal/exsanguination.) Blood loss was determined by weighing gauze immediately following the conclusion of the experiment, and organs were weighed and imaged via IVIS to obtain organ biodistribution from fluorescently labeled nanoparticles.

Statistical Analysis

All normalized values were calculated from calibration curves for the specific study (ex: hemoglobin, Cy7, etc.) . The sample sizes for nanoparticle binding experiments for 2, 4, 6 washes were n = 15, 10, 5 respectively, and for platelet binding experiments they ranged from n = 5–6 as detailed in the figure legends. Crosslinking studies were completed with n = 3 samples. For light‐transmission‐based clotting tests, n = 12, and for plasminolysis tests n = 5. All uninjured or minor injury models were completed with n = 3 mice, while for the lethal injury model the two‐component systems had n = 8 mice, the NP‐only group had n = 10 mice, and the controls had n = 5 mice. Statistical analyses were carried out using GraphPad Prism 5. Data were analyzed with one or two‐way analysis of variance (ANOVA) with Bonferroni or Tukey post‐tests as recommended via Prism software and presented as mean ± SEM.

Conflict of Interest

The authors declare no conflict of interest.

Author Contributions

C.H. and Y.H. designed the experiments and contributed equally to this work. C.H. synthesized the polymer materials. C.H. and P.A.B. performed all in vitro and ex vivo studies; Y.H. and C.H. performed uninjured murine studies. Y.H. and C.H. performed lethal injury surgeries at MIT, and C.H. and P.A.B. performed necropsies for organ biodistribution. C.H., Y.H., and corresponding authors wrote the manuscript.

Supporting information

Supporting Information

Hong C., He Y., Bowen P. A., Belcher A. M., Olsen B. D., Hammond P. T., Engineering a Two‐Component Hemostat for the Treatment of Internal Bleeding through Wound‐Targeted Crosslinking. Adv. Healthcare Mater. 2023, 12, 2202756. 10.1002/adhm.202202756

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.