Biomaterials 86 (2016) 42e55

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Reversible hemostatic properties of sulfabetaine/quaternary ammonium modified hyperbranched polyglycerol Jiying Wen a, b, Marie Weinhart a, c, Benjamin Lai a, d, Jayachandran Kizhakkedathu a, d, Donald E. Brooks a, b, d, * a

Centre for Blood Research, 2350 Health Sciences Mall, University of British Columbia, Vancouver V6T 1Z3, Canada Department of Chemistry, University of BC, 2036 Main Mall, Vancouver V6T 1Z1, Canada €t Berlin, Takustrasse 3, 14195 Berlin, Germany Institute of Chemistry and Biochemistry, Freie Universita d Department of Pathology & Laboratory Medicine, University of BC, 2211 Wesbrook Mall, Vancouver V6T 2B5, Canada b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 November 2015 Received in revised form 31 December 2015 Accepted 29 January 2016 Available online 2 February 2016

A library of hyperbranched polyglycerols (HPGs) functionalized with different mole fractions of zwitterionic sulfabetaine and cationic quaternary ammonium ligands was synthesized and characterized. A post-polymerization method was employed that utilized double bond moieties on the dendritic HPG for the coupling of thiol-terminated ligands via UV initiated thiol-ene “click” chemistry. The proportions of different ligands were precisely controlled by varying the ligand concentration during the irradiation process. The effect of the polymer library on hemostasis was investigated using whole human blood. It was found that polymer with
40% of alkenes converted to positive charges and the remainder to sulfabetaines caused hemagglutination at
1 mg/mL, without causing red blood cell lysis. The quaternary ammonium groups can interact with the negative charged sites on the membranes of erythrocytes, which provides the bioadhesion. The zwitterionic sulfabetaine evidently provides a hydration layer to partially mask the adverse effects that are likely to be caused by cationic moieties. The polymer was also found able to enhance platelet aggregation and activation in a concentration and positive charge densitydependent manner, which would contribute to initiating hemostasis. In a variety of other assays the material was found to be largely biocompatible. The polymer-induced hemostasis is obtained by a process independent of the normal blood clotting cascade but dependent on red blood cell agglutination, where the polymers promote hemostasis by linking erythrocytes together to form a lattice to entrap the cells. © 2016 Elsevier Ltd. All rights reserved.

Keywords: Hyperbranched polyglycerol Betaines Polycations Hemostasis Bioadhesion

1. Introduction Enhancement of hemostasis at the site of a wound is a very attractive method to limit bleeding and to reduce the need for blood transfusion support. The ideal hemostatic agent would be easy to use, highly efficacious, nonantigenic, fully absorbable and inexpensive [1,2]. Another factor for consideration is to have the hemostat disassemble or “reversible” to avoid undesired ongoing clotting. Synthetic hemostatic agents such as cyanoacrylates and glutaraldehyde cross-linked albumin are reported to exhibit toxicity and potential mutagenicity [3]. The application of widely

* Corresponding author. Centre for Blood Research, 2350 Health Sciences Mall, University of British Columbia, Vancouver V6T 1Z3, Canada. E-mail address: [email protected] (D.E. Brooks). http://dx.doi.org/10.1016/j.biomaterials.2016.01.067 0142-9612/© 2016 Elsevier Ltd. All rights reserved.

used zeolite-based QuickClot generates heat that can cause burn injuries [4]. Fibrin bandages overcome these problems, yet they still have some limitations such as high cost and scarcity of supply [5]. Thus despite much research in the field of hemostasis and a plethora of hemostatic agents, the need for an improved hemostatic material that can be easily and rapidly applied to severed vessels to reduce perioperative blood loss at all scales still exists. Cationic polymers aggressively bind to a variety of biological cells because of their ability to interact strongly with negatively charged cell membranes; they bind more strongly than neutral and anionic substances due to the interaction between opposite charges [6]. Such ionic, multivalent interactions of cationic polymers with cell membranes can cause aggregation of cells in a concentration and surface charge density dependent manner [7], events which also offer other biomedical functions, i.e., hemostatic and antimicrobial [8]. The hemostatic potential of polycations, for example

J. Wen et al. / Biomaterials 86 (2016) 42e55

chitosan [9] and poly-L-lysine [10], has been investigated since the 1950s. Polycations have achieved levels of success in fields such as gene delivery [11], drug delivery [12] and tissue engineering [13,14], but they have also been faced with the challenge of biocompatibility. Their relatively high cytotoxicity often is caused by the disruption of the cell membrane [15,16], which has limited their therapeutic applications. Therefore, methods to achieve defined and adjustable polymer/biomembrane interactions to control the strength of cell adhesiveness and enhance biocompatibility are needed. Zwitterionic materials with a balance of positively and negatively charged moieties within the same segment have been considered as “stealth” materials in the biomedical community due to their very weak interaction with bioentities, such as proteins and cells [17,18]. Almost all of the zwitterionic materials that have been reported are known to have excellent non-fouling properties, except methyl decorated choline phosphate, which displays a strong affinity for biological membranes [19e21]. The high degree of hydration of zwitterionic materials, which exceeds the hydration capacity of polyethylene glycol (PEG) [22], is thought to be responsible for the observed stealth effect of surfaces decorated with dipolar groups [23]. Several types of zwitterionic entities, including sulfobetaine [24,25], carboxybetaine [26,27] and phosphorylcholine [28,29], have been used as biocompatible coating materials for various nanoparticles, but to the best of our knowledge, sulfabetaine-modified surfaces have only been reported as lipid head groups [30], not associated with polymers. Unlike sulfobetaine, where the sulfur atom is directly linked to carbon characteristic of a sulfonate group, the carbon atom in sulfabetaine is linked to sulfur through an oxygen atom resembling a sulfate group. The presence of an additional heteroatom in the form of oxygen can introduce more hydrophilicity to the polymer, and also increase the size of the anionic moiety resulting in a lower charge density [31]. Hyperbranched polyglycerol (HPG) has a glycerol-based macromolecular architecture that is structurally related to PEG, but with a high degree of branching. Probably due to its globular nature, the hydration capacity of HPG is higher than that of PEG [32]. Moreover the terminal hydroxyl groups can be directly used as linker functionalities, allowing the coupling of multiple copies of bioactive molecules and chemical entities, enhancing the design potential for multifunctional nanostructures. Due to the low degree of molecular weight dispersity, good chemical stability, high hydrophilicity, inertness under biological conditions, and flexible design [33], HPG has emerged as a scaffold with a broad range of possible structural designs in biomedical applications, including drug conjugates [34,35], erythrocyte surface camouflage [32,36,37], antidotes for anticoagulation [38], and presentation of antibacterial peptides [39]. In this study we describe a novel class of multivalent water soluble HPG-based polymeric materials that combine several functions by incorporating multiple anti-fouling and/or adhesive bioactive agents. These bioactive agents, whether the zwitterionic moieties or the quaternary amine groups, are intended to function synergistically to form the basis of a biocompatible hemostatic wound dressing agent when coupled to a suitable flexible base material. We anticipate the resulting bioactive material to possess hemostatic functions with high efficacy since cationic moieties can strongly bind red blood cells and enhance the formation of a blood plug while both the zwitterionic ligands and HPG foundation can impart hemocompatibility to the structure to mask the cytotoxicity of polycations and also absorb wound fluids due to their strong hydration properties.

43

2. Materials and methods 2.1. Materials All chemicals were purchased from SigmaeAldrich (Oakville, ON) and all solvents were HPLC grade from Fisher Scientific (Ottawa, ON) and used without further purification unless otherwise stated. 1,1,1-Tris(hydroxymethyl)propane and potassium methylate solution (25 wt% in methanol) were purchased from Fluka (Oakville, ON). Dialysis membranes with different molecular weight cut-offs (Spectra/Por Biotech regenerated cellulose dialysis membranes) were obtained from Spectrum Laboratories, Inc. (Rancho Dominguez, CA). Glycidol (96%) was purified by distillation under reduced pressure and stored at 4  C. Anti-CD62PE and goat anti-mouse PE antibodies were purchased from Immunotech Inc. (Vaudreuil-Dorion, QC). GVB2þ (0.1% gelatin, 5 mM Veronal, 145 mM NaCl, 0.025% NaN3 with 0.15 mM CaCl2 and 0.5 mM MgCl2, pH 7.3) and antibody-sensitized sheep erythrocytes were purchased from CompTech (Tyler, TX). Human umbilical vein endothelial cells (HUVECs) were obtained from Lorus Pharmaceuticals (Allendale, NJ). 2.2. Synthesis of sulfabetaine (SB) ligand All air and moisture sensitive reactions were carried out in flame-dried glassware under an argon inert gas atmosphere. Bis(3dimethylaminopropyl)disulfide was obtained from the corresponding HCl salt via extraction of an aqueous sodium hydroxide solution with diethylether, phase separation, drying over Na2SO4 and concentration under reduced pressure. Bis(3-dimethylaminopropyl)disulfide (2.36 g, 0.01 mol, 1 equiv.) was dissolved in anhydrous acetone (15 ml) and cooled to 0  C in an ice bath. 1,3,2-Dioxathiolane 2,2-dioxide (2.73 g, 0.22 mol, 2.2 equiv.) dissolved in anhydrous acetone, (20 mL) was added to the chilled bis(3-dimethylaminopropyl)disulfide solution under vigorous stirring in a dropwise manner via a syringe. The reaction was continued for 1 h at 0  C then allowed to warm to room temperature overnight. A white precipitate was obtained during the reaction, which was filtered, washed thoroughly with acetone and dried under vacuum to obtain pure sulfabetaine disulfide as a colorless, very hygroscopic solid in 91% yield (4.4 g). The corresponding sulfabetaine thiol (3.5 g, 80%) was obtained by reduction of the corresponding disulfide (2.20 g, 0.0045 mol, 1 eq) with triphenylphosphine (12 g, 0.045 mol, 10 eq) under emulsifying conditions in a solvent mixture of dichloromethane (DCM), trifluoroethanol (TFE) and water at room temperature for 48 h. The aqueous phase was separated and TFE was removed from it by rotatory evaporation. The remaining aqueous phase was freezedried to yield 3.5 g (80%) of the final thiol product as a white, hygroscopic solid, which was stored under an inert gas atmosphere. 1 H NMR (d: ppm, 300 MHz, D2O) 4.50 (t, 2H, eCH2SO4), 3.76 (t, 2H, SO4eCH2CH2Ne), 3.55 (t, 2H, HSCH2CH2CH2Ne), 3.20 (s, 6H, eN(CH3)2), 2.65 (t, 2H, HSCH2e), 2.14 (m, 2H, HSCH2CH2CH2N). 13 C NMR (d: ppm, 75 MHz, D2O) 64.1 (eCH2SO 4 ), 62.5, 61.9 (2 CH2Ne), 51.8 (eN(CH3)2), 26.2 (HSCH2e), 20.5 (HSCH2CH2CH2e). 2.3. Synthesis of quaternary ammonium (QA) ligand Sodium thiosulfate pentahydrate (5.46 g, 0.022 mol, 1.1 equiv.) was dissolved in 20 mL deionized water and added into a flask with (3-bromopropyl)trimethylammonium bromide (5.20 g, 0.02 mol, 1 equiv.). The mixture was heated to 100  C and kept under reflux and stirring for a total of 12 h. The resulting Bunte salt solution was then hydrolyzed without isolation to the corresponding thiol by adding concentrated H2SO4 to a final acid concentration of 1 M and

44

J. Wen et al. / Biomaterials 86 (2016) 42e55

refluxing for further 24 h. Afterwards the reaction was neutralized by addition of aqueous sodium hydroxide solution and the crude product was isolated by freeze-drying. The crude product was separated from inorganic salt byproducts by dissolution with methanol. The methanolic solution was filtered from remaining insoluble, inorganic salts and concentrated in vacuum. To ensure 100% thiol functionalities on the QA ligand eventually formed, disulfides during isolation were reduced with triphenylphosphine at room temperature as described for the SB ligand to yield pure 3mercapto N,N,N-trimethylpropan-1-ammonium bromide as a colorless solid in 82% yield (3.5 g). 1 H NMR (d: ppm, 300 MHz, D2O) 3.47 (t, 2H, eCH2CH2Ne), 3.16 (s, 9H, eN(CH3)3), 2.65 (t, 2H, HS-CH2e), 2.22 (m, 2H, eCH2CH2CH2e). 13 C NMR (d: ppm, 75 MHz, D2O) 65.3 (eCH2CH2Ne), 53.1 (eN(CH3)3), 26.6 (HSeCH2e), 20.6 (eCH2CH2CH2e). 2.4. Synthesis of Allyl functionalized HPG HPG was synthesized by ring opening multi-branching anionic polymerization using a procedure described previously [40]. HPG (Mn ¼ 20 K, PDI ¼ 1.3, 6.5 g containing 8.78  102 mol hydroxyl groups), powdered NaOH (17.56 g, 0.439 mol, 5 mol per mole of hydroxyl group) and distilled water (17.56 g, the same mass as NaOH), tetrabutylammonium bromide (TBAB) (0.8491 g, 2.63  103 mol, 3 mol% on each mole of hydroxyl groups), were added into a two-neck round-bottom flask equipped with a condenser and a dropping funnel. The mixture was heated to 60  C and stirred at this temperature for 1 h to get a viscous slurry. TBAB was used as a phase-transfer catalyst. Next, allyl chloride (33.59 g, 0.439 mol, 5 mol per mole of hydroxyl group) was transferred into the dropping funnel and added in drops to the mixture over a period of 2 h. The mixture was kept under reflux with vigorous stirring. After 12 h of reaction, dichloromethane (200 mL) was added to the mixture. The organic phase was then separated, washed with water, dried over Na2SO4, filtered, and concentrated by rotary evaporator to yield the crude product, which was further purified by dialysis against acetone. The final product was obtained as a colorless oil after drying under vacuum (~9 g). For long-term storage, it was necessary to keep the allyl ether product under an inert gas atmosphere, i.e. argon, at 20  C to prevent slow crosslinking of the allyl groups. Allyl functionalized HPG was characterized by NMR analysis. The number of allyl groups on the polymer was calculated from 1H NMR integration values. 1 H NMR (d: ppm, 300 MHz, CDCl3) 5.84 (m, CH2]CHCH2e), 5.24 (s, CH2¼C), 5.19 (s, CH2¼Ce), 5.10 (t, CH2¼Ce), 4.10 (s, eOCH2CH] Ce), 3.95 (s, eOCH2CH]Ce), 3.30e3.80 (m, eOCH2e). 13 C NMR (d: ppm, 75 MHz, CDCl3) 135.5, 135.1 (CH2]Ce), 116.9 (CH2]Ce), 77.7, 77.2, 76.8, 72.4, 72.1, 72.0, 71.8, 71.4, 70.4 (polymer backbone, CH2]CHeCH2e). 2.5. Synthesis of SB/QA functionalized HPG The functionalization of allyl-HPG was performed using a UVinitiated radical-mediated thiol-ene “click” reaction. Briefly, allylHPG (100 mg, 3.43  106 mol, 7.41  104 mol allyl groups), a mixture of SB and QA ligands in various molar ratios (2 mol per mole of allyl groups), 2-hydroxy-40 -(2-hydroxyethoxy)-2methylpropiophenone (34 mg, 1.53  104 mol, 0.2 mol per mole of allyl groups), methanol/TFE (5 mL, methanol: TFE ¼ 1:4, v/v) were transferred to a 20 mL glass reaction vial and stirred until the mixture was clear. The reaction was carried out under the irradiation of UV light with a wavelength of 365 nm for 45 min. The lamp used was a BlueWave™ 50AS UV light curing spot lamp from DYMAX Corporation (Torrington, CT), which emits 365 nm light at

50 W. The reaction vessel was checked every 5 min by gradually adding drops of deionized water to ensure the homogeneity of the mixture. Excessive zwitterionic ligands and the photoinitiator were removed by dialysis. The mixture was first dialyzed against 0.15 M aqueous sodium chloride solution and then deionized water with a cellulose dialysis membrane (MWCO ¼ 3500). The number of QA and SB ligands on the polymer and the molecular weight (Mn) of the resulting functionalized polymers were calculated from the 1H NMR integration values. 1 H NMR (d: ppm, 400 MHz, D2O) 4.50 (t, eCH2SO4 on SB ligand), 3.7e3.9 (HPG backbone), 3.76 (t, SO4eCH2CH2Ne on SB ligand), 3.47 (t, eCH2CH2Ne on QA ligand) 3.23 (s, eN(CH3)2 on SB ligand), 3.17 (s, eN(CH3)3 on QA ligand), 2.70 (m, eCH2SCH2e), 2.13 (m, eCH2CH2CH2e), 1.93(m, eCH2CH2CH2e) 13 C NMR (d: ppm, 100 MHz, D2O) 65.6 (eCH2CH2Ne on QA ligand), 70e80 (HPG backbone), 64.0 (eCH2SO 4 on SB ligand), 62.4 (SO4eCH2CH2Ne on SB ligand), 61.7 (eCH2CH2CH2Ne on SB ligand), 53.2 (eN(CH3)3 on QA ligand), 52.0 (eN(CH3)2 on SB ligand), 29.0 (eCH2CH2CH2e), 28.0 (eCH2CH2Se), 22.6 (eCH2CH2CH2e) 2.6. Quantification of bound water The amount of structured water associated with the polymers was quantified using a Q2000 differential scanning calorimeter (DSC) from TA Instrument (New Castle, DE). Polymer was first dissolved in deionized water with a 10 (w/w)% polymer concentration. Around 15 mL of polymer solution was pipetted into a Tzero pan to evaluate the ice-water fusion enthalpy. An empty pan was used as the reference. Scanning was conducted from 30  C to 30  C at a rate of 2  C/min, the power (mW) was recorded as a function of temperature and the fusion enthalpy was evaluated from the peak integral at negative temperatures around 0  C using the TA Universal Analysis Software. The number of water molecules bound to each polymer molecule, N, was calculated from the following equation (see Supplementary Information (SI) for derivation):


DH0 1  Cp  DHp N ¼ $Mn 18:02DH0 Cp

[1]

Where DH0 is the fusion enthalpy of pure water, DHp is the fusion enthalpy of free water in the polymer solution, Cp is the mass fraction of polymer in solution, and Mn is the molecular weight of the polymer calculated using 1H NMR integration. 2.7. Blood analysis 2.7.1. Blood preparation Blood was drawn from healthy consented volunteers into a 3.8% sodium citrated tube (BD Vacutainer™) with a blood/coagulant ratio of 9:1 at the Centre for Blood Research, by a protocol approved by the University of British Columbia Clinical Ethics Committee. For serum preparation, blood was collected in serum tubes without anticoagulant. After 20 min, the clotted whole blood sample was centrifuged at 1200  g for 30 min in an Allegra X-22R centrifuge (Beckman Coulter, Canada). The supernatant serum was transferred into a plastic tube and stored at room temperature until use. Platelet poor plasma (PPP) was isolated as the supernatant from centrifugation of citrated whole blood samples at 3000  g for 15 min at room temperature. Platelet rich plasma (PRP) was isolated as the supernatant from centrifugation of the citrated blood at 150  g for 15 min. Packed RBCs were separated by centrifugation of citrated whole blood at 1000  g for 5 min. The plasma and buffy coat were removed, and RBCs were further washed three times

J. Wen et al. / Biomaterials 86 (2016) 42e55

with sterile PBS solution by centrifugation. RBC suspensions were prepared by resuspending one part of packed RBCs in seven parts of the corresponding buffer (PBS or saline) to yield a hematocrit of approximately 10%. 2.7.2. Red blood cell aggregation Red blood cell aggregation studies were performed using both whole blood and washed RBC suspensions. RBC aggregation in whole blood was assessed according to a previous publication from this laboratory [41]. Ten microliters of stock solution (50 mg/mL or 10 mg/mL) of polymer in PBS was first transferred to an Eppendorf tube, and then 90 mL of whole blood was added into the polymer solution with a micropipette to achieve a final concentration of 5 mg/mL or 1 mg/mL of polymer in the blood. To prevent high local concentration of the polymer, the mixture was homogenized by several up-and-down aspirations and then incubated for 5 min at 37  C. Whole blood incubated with PBS was used as the negative control. After incubation, the RBCs and plasma/polymer were separated by centrifugation at 3000  g for 3 min. Aliquots of 2 mL of the red blood cells were resuspended in 10 mL of plasma/polymer and examined by bright field light microscopy (Zeiss Axioskop 2plus) using wet mounted slides. Images were captured with a digital microscope camera (AxioCam ICc 1, Carl Zeiss Microimaging Inc.) at 40  magnification of ten different observation fields on the same chip for statistically sound results. RBC aggregation with washed RBCs was performed by mixing one part of polymer stock solution with nine parts of 10% hematocrit washed RBC suspension; aliquots of 5 mL of the mixture were used for microscope examination. 2.7.3. Platelet activation P-selectin (CD62P) surface expression is commonly used as a marker to quantify the level of platelet activation. For a typical experiment, 90 mL of PRP was mixed with 10 mL of stock polymer samples (10 mg/mL or 1 mg/mL) to achieve final polymer concentrations of 1 mg/mL or 0.1 mg/mL respectively. After 1 h incubation at 37  C, aliquots of the incubation mixtures were removed for assessment of the platelet activation state. Aliquots of 5 mL postincubation platelet/polymer mixture were added to 45 mL PBS buffer, and further incubated for 20 min in the dark at room temperature with 5 mL of fluorescently labelled monoclonal anti-CD62PE. The incubation was then stopped by adding 300 mL of PBS buffer. The level of platelet activation was analyzed in a BD FACSCanto II flow cytometer (Becton Dickinson) by gating platelet specific events based on their light scattering profile. Activation of platelets was expressed as the percentage of platelet activation marker CD62-PE fluorescence detected in the 10,000 total events counted. Duplicate measurements were performed for each donor, and three different donors were used. The results were reported as mean with standard deviation. 2.7.4. Platelet aggregation Aggregation of platelets was studied in a flat-bottomed 96-well plate; the method is modified from an experimental procedure that has been previously published [42]. Aliquots of 90 mL of PRP were pipetted per well into the microplate and 10 mL of polymer stock solution was added into the appropriate wells. Adenosine diphosphate (ADP) with a concentration of 5 mM and PBS were used as a positive and negative control respectively. For calibration, 100 mL of PRP and 100 mL of PPP were placed in wells. After the addition of agonists, kinetic reading of the plate was started immediately. The light transmittance (LT) was then measured at 595 nm at an interval of 15 s for 20 min at 37  C using a Spectramax PLUS plate reader (Molecular Devices, Sunnyvale, CA). During the run time the plate was mixed every 3 s using the automix function of the reader.

45

The percentage of aggregation was calculated using the formula:

% of aggregation ¼

LT595;PRP  LT595;test sample  100% LT595;PRP  LT595;PRP

[2]

2.7.5. Hemolysis assay The membrane disruption of RBCs was estimated to determine the hemocompatible nature of the polymers. The hemolysis assay was performed according to a procedure modified from previous references [43]. A 40 mL aliquot of the stock polymer solution prepared in PBS (50 mg/mL) was added to 360 mL of the 10% w/v RBCs suspension to make the final polymer concentration 5 mg/mL and then incubated at 37  C for 1 h. After incubation, the mixture was centrifuged at 3000  g and the supernatants transferred to a 96well plate (150 mL per well). Release of hemoglobin was measured by spectrophotometric analysis of the supernatant at 540 nm. Deionized water and PBS were used as positive and negative controls, respectively. The percent of hemolysis was calculated as follows:

Hemolysis% ¼

Abs540;test sample  Abs540;PBS  100% Abs540;water  Abs540;PBS

[3]

2.7.6. Complement activation The level of complement activation was measured by the CH50 sheep erythrocyte complement lysis assay in human serum, according to a previous publication from our group [44]. Stock solutions of polymer at concentrations of 10 mg/mL or 1 mg/mL were prepared. Aliquots of 10 mL stock solutions were mixed with 90 mL of GVB2þ diluted human serum to obtain a final polymer concentration of 1 mg/mL or 0.1 mg/mL respectively. Heat-aggregated human IgG and PBS mixed with diluted human serum were used as positive and negative controls respectively. After 1 h incubation at 37  C, 75 mL of the polymer/serum mixture was further mixed with 75 mL antibody-sensitized (Ab-sensitized) sheep erythrocyte and then incubated for another hour. Then, 300 mL of cold GVBEDTA buffer was added to each sample. Ab-sensitized sheep erythrocytes incubated with distilled H2O were regarded as the 100% lysis control. Intact Ab-sensitized sheep erythrocytes were then centrifuged at 3000  g for 3 min and the supernatants were sampled in a 96-well plate (100 mL per well). Percentage sheep erythrocyte lysis was calculated using average absorbance values at 540 nm as follows:

%lysis ¼

Abs540;test sample  Abs540;PBS  100 Abs540;100%lysis  Abs540;PBS

[4]

Percentage of complement consumed by the polymer was expressed as 100 e %lysis. 2.7.7. Blood coagulation 2.7.7.1. Activated partial thromboplastin time (aPTT) and prothrombin time (PT). Activated partial thromboplastin time (aPTT) and prothrombin time (PT) were measured using a coagulation analyzer using mechanical end point determination (ST4, Diagnostica Stago) according to the procedure published previously [44]. The effect of each polymer on the intrinsic pathway of coagulation (aPTT) was studied after mixing PPP with the polymer stock solution (9:1 v/v polymer: plasma) in the cuvette-strips at 37  C for 3 min before adding the partial thromboplastin reagent (Actin® reagent) and calcium chloride (CaCl2) solution. Actin® reagent contains phospholipids, ellagic acid, stabilizers and

46

J. Wen et al. / Biomaterials 86 (2016) 42e55

preservative. For PT determination, stock solution of the polymer was first mixed with PPP (9:1 v/v PPP: polymer), and the extrinsic coagulation pathway was activated by incubating plasma with innovin® reagent and the clotting time then measured. Innovin® is a lyophilized reagent consisting of recombinant human tissue factor and synthetic phospholipids (thrombosplastin), calcium ions, a heparin-neutralizing compound, buffers and stabilizers (bovine serum albumin). A control experiment was done by adding identical volume of PBS buffer to plasma. Each measurement was performed in triplicate, using plasma from at least three different donors.

2.7.7.2. Thromboelastography. The effect of polymers on blood clotting was examined using thromboelastography (TEG) analysis. TEG is used to assess the viscoelastic changes in whole blood clotting under low shear conditions after adding a specific coagulation activator [45]. In a typical experiment, 40 mL of stock polymer solution was mixed with 360 mL of freshly drawn citrated whole blood (within 5e10 min of venipuncture). Aliquots of 340 mL of the polymer/blood mixture were loaded into TEG cups in the analyzer. The coagulation analysis began when the citrated whole blood was re-calcified with 20 mL of CaCl2 solution (0.20 M). PBS (40 mL) mixed with citrated whole blood (360 mL) served as the normal control for experimental study. Each experiment was performed at least three times with different donors. All TEG experiments were performed

at 37  C and ended after 2 h. The detailed procedure was published previously [46].

2.8. Cytotoxicity measurements The cytotoxicity of the polymers was studied by using the propidium iodide live cell staining assay [47]. The cells were seeded on 96-well plates at a density of approximately 5  103 cells per well. After 24 h of cell-seeding incubation at 37  C, polymer stock solution (dissolved in cell culture media and filtered) was added to the appropriate wells to achieve a final concentration of 0.25, 0.5, 1 and 2 mg/mL and incubated for 48 h. A positive control was included using 50% DMSO in six wells per plate to induce cell death. Twelve wells in each plate were also reserved as a negative control with no chemical addition. After chemical exposure, 100 mL of staining solution (1 mL of 20 mg/mL Hoechst, 100 mL of 50 mg/mL PI, 10 mL pre-warned medium) was added directly into the wells without aspirating the treatment medium and incubated for 30 min at 37  C. The final concentrations of Hoechst and PI in each well were both 0.5 mg/mL. Cells were then imaged using the Cellomics™ Arrayscan VTI (Thermo Scientific, Pittsburgh, PA) high content imaging system. Cell viability was calculated as the proportion of all cells not stained with propidium iodide. The untreated wells were regarded as reference during the scan and the data output % responder

Scheme 1. Synthesis of hemostatic polymers and controls. Hydroxyl groups of HPG were modified with different amounts of zwitterionic sulfabetaine (SB) and cationic quaternary ammonium QA) ligands.

J. Wen et al. / Biomaterials 86 (2016) 42e55

47

Fig. 1. 1H NMR spectra HPG-SB and HPG-SB-QAx where x indicates the percentage of QA.

indicated as a percentage of cells with average intensity in cytoplasmic greater than the average intensity in the untreated cells. 2.9. Statistics Data were expressed as means ± standard deviation (SD). Measurements were conducted in at least triplicate if not stated. All statistical analysis was performed using the Prism software program (GraphPad, San Diego, CA). Data were statistically analyzed using one-way analysis of variance. To compare the effect of polymer treatment and buffer control, Dunnett's multiple comparison test was performed. Reported p-values were adjusted to achieve an overall confidence level of 95%. A p value < 0.05 was considered as statistically significant. 3. Results and discussions 3.1. Synthesis and characterization of hemostatic polymers In this study, hyperbranched polyglycerols were used as the starting material to be modified by allyl chloride to introduce pendant alkene groups in the structure, which were further modified with different amounts of thiol-terminated zwitterionic

Table 1 Physicochemical properties of macromolecules used in this study. Polymer HPG HPG-SB HPG-SB-QA15 HPG-SB-QA20 HPG-SB-QA30 HPG-SB-QA40 HPG-SB-QA45 Control HPG-QA

# SBa NA 216 184 173 150 130 120 NA

# QAa NA 0 32 43 66 86 96 130

Mna NA 8.2  7.8  7.7  7.4  7.2  7.1  5.4 

M nb 104 104 104 104 104 104 104

2.0  8.0  8.4  7.8  8.9  1.0  1.1  NAd

4

10 104 104 104 104 105 105

PDIb

Rh (nm)c

1.3 1.4 1.5 1.6 1.5 1.7 1.8 NAd

3.0 4.5 4.7 4.7 4.9 5.8 6.2 NAd

sulfate betaine and cationic quaternary ammonium ligands via UV initiated thiol-ene “click” chemistry (Scheme 1). The feeding ratios of QA/SB ranging from 0 to 1 were used to control the degree of modification. An excess of thiol-containing components was used in the reaction mixture to suppress the undesired crosslinking reactions between the double bonds on the backbone. A photo-initiator was used to increase the radical concentration and thus increase the rate of reaction. The residual thiol-containing molecules and photo-initiator were removed by subsequent dialysis. The control polymer was synthesized in a similar procedure, except that the zwitterionic ligands were replaced by equal moles of 1-thiolglycerol. These HPG-SB/QA polymers are referred to as hemostatic polymers, anticipating the results described below. The functionalized ratio of QA/SB on the HPG backbone can be obtained from 1H NMR spectra (Fig. 1). Compared to the 1H NMR of HPG-allyl (Fig. S1), the new peak at 3.23 ppm represents the protons on the two methyl groups connected to the nitrogen of sulfabetaine, and the peak at 3.17 ppm represents the protons on the three methyl groups connected to the nitrogen of the quaternary amine. It is very clear that as the proportion of QA increases, the peak at 3.17 ppm goes up while the peak at 3.23 ppm goes down.

Table 2 The ice-to-water-like transition temperatures (Tc) and enthalpy changes (DH) of water as measured by DSC for aqueous solutions of different polymers. Pure water was used as a normal control. Polymer

Ca(g/g)

DH (J/g)

Nb

Nc

Nd

Ne

Pure H2O HPG HPG-SB-QA20 HPG-SB-QA30 HPG-SB-QA40 HPG-SB-QA45 HPG-QA Control polymer

NA 0.096 0.089 0.090 0.088 0.086 0.098

362 289 302 301 303 304 298

NA 1200 3700 3600 3300 3200 2400

~5

~14

~4

a

a

b

b

Calculated from 1H NMR integration with certain uncertainties. Determined by GPC at 25  C in 1 N NaNO3. c Determined by dynamic light scattering (QELS) at 25  C in 1 N NaNO3. d The GPC measurement of the control polymer characteristics failed due to the positive charge.

c d e

The The The The The

polymer concentration in aqueous solution determined by TGA. number of moles of water bound per mole of polymer. number water molecules bound per glycerol unit. number of water molecules bound per SB unit. number of water molecules bound per QA unit.

48

J. Wen et al. / Biomaterials 86 (2016) 42e55

Fig. 2. (A)e(E): Optical microscopic images of human whole blood incubated with or without hemostatic polymers; (F) reversal of aggregation caused HPG-SB-QA45 by LPG sulfate. All images are at 40  magnification. (A) PBS control, (B) HPG-SB 1 mg/mL, (C) HPG-SB-QA30 1 mg/mL, (D) HPG-SB-QA40 1 mg/mL, (E) HPG-SB-QA45 1 mg/mL, (F) whole blood first incubated with 1 mg/mL (final) HPG-SB-QA45 for 5 min, then LPG-sulfate (1 mg/ml final concentration) was added and incubated for another 5 min.

These two signals of the resulting end groups do not interfere with other signals in the spectrum and therefore permit reliable integration of the 1H NMR spectra for the determination of the ratio between SB and QA ligands on the polymer as well as the resulting molecular weight. There are other new peaks at 2.70 ppm corresponding to the methylene groups next to the thio-ether linkage after full reaction of the alkene groups, indicating the successful conjugation of thiol head groups with the peripheral double bonds. Simultaneously, the original two peaks located at 5.1 ppm and 5.8 ppm representing the protons of the allyl group at the periphery of the polymer disappeared, indicating the alkene side groups were quantitatively consumed during the thiol-ene addition. The quantification of the ratio of SB and QA on the HPG backbone was represented by the relative mole fraction of QA content, as summarized in Table 1. The relative proportion of the ligands in the reaction (SB/QA ratio of 100/0, 90/10, 75/25, 70/30, 60/40, and 50/50) was close to those of the modified HPG with relative SB/QA

ratios of 100/0, 85/15, 80/20, 70/30, 60/40, and 55/45, respectively. This indicated that the relative proportion of zwitterions and cations was effectively controlled. The hydrodynamic radii of the series of hyperbranched polymers ranged from 4.5 to 6.2 nm, while the hydrodynamic radius of the core HPG is ~3 nm. For HPG-SB, the contribution of the ligands to the overall hydrodynamic radius is ~1.5 nm, which is reasonably close to that estimated from the fully extended geometric length of the sulfabetaine ligands (~1.7 nm). As the proportion of the QA moieties increases, the hydrodynamic size as well as the polydispersities of the polymer get larger. The former may represent electrostatically driven swelling as the charge density increases, albeit in a high ionic strength medium. The increased PDI could be due to formation of a tiny fraction of aggregates in the solution, which may also explain why the Mn determined by GPC is generally higher than that calculated from 1H NMR spectra for the HPG-SBQAX series.

J. Wen et al. / Biomaterials 86 (2016) 42e55

49

Fig. 3. Optical microscopic images of washed RBCs incubated with or without hemostatic polymers. All images are at 40  magnification. (A) PBS, (B) HPG-SB, (C) HPG-SB-QA40, (D) HPG-SB-QA45; polymer concentration: 1 mg/mL.

Fig. 4. Platelet aggregation and activation by hemostatic polymers. (A) Maximal percentage aggregation (mean ± S.D.) of PRP measured by light transmission with a microplate reader; (B) Platelet activation in PRP measured by the expression of platelet activation marker CD62P using anti-CD62P antibody by flow cytometry. Statistical testing was performed to compare the effect of the polymers with PBS buffer. Asterisks indicate significant differences: (A) ****p < 0.0001; for the four not significant (ns) comparisons, p values are 0.9996, 0.4729, 0.5393, 0.5908 respectively; (B)**p value ¼ 0.0015; for the four not significant (ns) comparisons, p values are > 0.9999, 0.9997, 0.9984, 0.4187 respectively.

3.2. Quantification of structured water associated with the hemostatic polymers The amount of structured or bound water associated with native and modified HPGs was analyzed by DSC, and the results are summarized in Table 2. Thermal transitions were observed for all of the native and modified HPGs, indicating varying amounts of free water exists with the binding sites of polymers being essentially

saturated by water molecules; the free water has a transition over a temperature range similar to that of the ice-to-water transition for bulk water. The non-freezing water is generally assumed to be bound by the polymers but models which describe the H-bonding potential of materials like PEG tested with spectroscopic methods show that layers of structured water must exist around some biocompatible materials that do not interact directly with the polymer and no generally satisfactory understanding of the

50

J. Wen et al. / Biomaterials 86 (2016) 42e55

Fig. 5. (A) Hemolysis assay for SB/QA-modified HPG of different degrees. Results indicated that with the presence of zwitterionic SB, the percentage of RBC lysis is less than 5%. Polymer concentration is 5 mg/mL. Relative hemolysis is expressed considering the hemolysis of PBS as 0% and hemolysis of water as 100%. Error bars correspond to standard deviation from three separate measurements. Statistical testing was performed to compare the effect of HPG-SB-QA and the control polymer with HPG-SB. Asterisks indicate a significant difference, ****p < 0.0001; for the four not significant (ns) comparisons, p values are >0.9999, 0.9998, 0.9983, 0.9694 respectively. (B) Images of RBCs incubated with polymers (5 mg/mL final concentration) at 37  C for 1 h.

situation has yet been reached, even for linear polymers [48,49]. Much less is known about water interactions with hyperbranched polymers [32], let alone those carrying charged and zwitterionic groups. We will refer to non-freezing water as bound hereafter, recognizing this is likely an inexact description. The hydration of a material plays a key role in its biocompatibility. The functionalization of HPG with SB/QA ligands greatly increased its ability to bind water molecules. In addition to the water molecules bound to the dendritic HPG via hydrogen bonding, the charged ligands can generate a strong hydration layer via ionic solvation. For the cationic QA ligand, water molecules should bind to the positively charged group (Nþ(CH3)3). For the zwitterionic SB ligand, water molecules simultaneously would have bound to both

Fig. 6. Complement activation determined in human serum using sheep erythrocyte based complement consumption assay. Final polymer concentration: 1 mg/mL. IgG and PBS were used as positive and negative controls respectively. Statistical testing was performed to compare the effect of the polymers with PBS buffer. Asterisks indicate a significant difference, ****p < 0.0001; for the four not significant (ns) comparisons, p values are 0.7351, 0.2874, 0.6938, 0.3014 respectively.

the positively charged group (Nþ(CH3)2) and the negatively charged group (SO 4 ). A zwitterionic ligand can hold more water molecules compared to a cationic ligand because of these dual ions. This is confirmed by the data shown in Table 2; as the percentage of cationic ligand increases, the moles of water bound per mole of polymer decreases. There are on average 5 water molecules bound per glycerol unit in the HPG structure, which is comparable with results from a previous publication [32]. Subtracting that from the data for HPG-SB-QA and applying multiple linear regression [50] gives 14 water molecules bound to one SB ligand, and 4 water molecules bound to one QA ligand. 3.3. Investigation of the interaction of hemostatic polymer and red blood cells The role of red cells in hemostasis has recently attracted renewed attention. Recent studies showed the aggregation of RBCs can change them to a polyhedron shape that can form a perfect seal [51]. Fig. 2(A) to (E) includes representative images of RBC morphology in whole blood observed in the presence or absence of hemostatic polymers. In the absence of polymer, native RBCs were stacked together in roughly linear groups of 4e10 cells and formed typical rouleaux (Fig. 2(A)). Similar features were observed upon the addition of HPG-SB solution (Fig. 2(B)), indicating that the charge neutral polymer would not induce hemagglutination. As the percentage of positively charged QA on the polymer increased, native RBCs clumped together and led to irregularly formed aggregates. As shown in Fig. 2(C) to (E), the higher percentage of cationic groups, the greater the size of the aggregates. The strongest aggregation activity was observed with HPG-SB-QA45 (Fig. 2(E)), where RBCs lost their original biconcave morphology and showed an unusual affinity towards each other, with almost no monodispersed cells. We also investigated the aggregation of washed erythrocytes induced by the hemostatic polymer (Fig. 3) in the absence of serum proteins. As expected there was no rouleaux formation in washed erythrocytes, and the RBCs retained their original biconcave shape when dispersed in PBS buffer. In the absence of potentially interfering proteins aggregation occurs at lower concentration in washed RBCs. The aggregates in Fig. 3(C) and (D) were much easier

J. Wen et al. / Biomaterials 86 (2016) 42e55

51

Fig. 7. Effect of hemostatic polymers on coagulation (A) prothrombin time (PT); (B) activated partial thromboplastin time (aPTT). Final polymer concentration is 1 mg/mL. Statistical testing was performed to compare the effect of the polymers with PBS buffer. Asterisk indicates significant difference, ****p < 0.0001; for the four not significant comparisons in (A), p values are 0.7640, 0.2494, 0.7640, 0.3633 respectively; the p value for the not significant comparison in (B) is 0.9992.

to identify compared to those in Fig. 2(D) and (E). This can be explained by a partial neutralization of the cationic ligands through their interactions with plasma proteins, which are mostly negatively charged as well. Thus these plasma proteins likely compete with RBCs for the cationic ligands. 3.4. Investigation of hemostatic polymer effects on platelet aggregation and activation Platelets are key blood components that are involved in hemostasis. They are capable of reacting sensitively to the presence of foreign surfaces, giving rise to their activation and aggregation. We hypothesize that the positive charge on the surface of HPG-SB-QA could bind to platelets and enhance their adhesion and activation, which would result in the formation of a cell embolus and promote blood clotting. Platelet aggregation caused by HPG-SB-QA in human PRP, as shown in Fig. 4(A), was significant compared to the PBS control. The ability of HPG-SB-QA45 to induce platelet aggregation was as strong as that induced by the physiological activator ADP. HPG-SB bearing no positively charged moieties could not cause platelet aggregation nor did HPG-SB-QA15, whose ability to induce platelet aggregation was not significantly different from PBS buffer, indicating only high concentration of cationic groups per polymer, not neutral or low concentration of cationic groups per molecule, would induce the aggregation of human platelets in plasma. The percentage of aggregation was clearly related to the relative concentration of quaternary amines on the polymer. Study of platelet activation provides an independent and sensitive method for testing platelet interaction with added materials that is relavent to understanding the hemostatic potential of HPGSB-QA polymers. Platelet activation by polymers with proportions of cationic QA ligands higher than 20% in human PRP, as determined by measuring the expression of the platelet activation marker CD62P, was significantly different compared to the PBS control, while the neutral polymer HPG-SB and the polymer with 15% cationic ligands again showed no difference compared to PBS control (Fig. 4(B)) in agreement with the results from platelet aggregation. Decreasing the QA moieties on the HPG reduced the number of activated platelets. It was shown that almost no activated platelets were caused by HPG-SB with overall charge neutrality.

3.5. Blood compatibility of hemostatic polymers 3.5.1. Hemolysis Having demonstrated the aggregation of erythrocytes induced by the HPG-SB-QA polymers, we were interested in evaluating the extent of RBC membrane damage. For this purpose, a hemolysis assay was performed. Hemolysis is the disruption of the RBC membrane, hence causing the release of hemoglobin, which can be detected by absorbance at 540 nm. Hemolysis was measured and compared with the PBS control, and the result is summarized in Fig. 5. A control polycationic polymer with no zwitterionic groups was included. Although the polymers, namely HPG-SB-QA40 and HPG-SB-QA45, showed strong RBC aggregation in PBS, the hemolytic activity of polymers carrying the zwitterions is lower than 2%, and the release of hemoglobin was not significantly different from the control. This could be explained if cell aggregation involved only the outer layer of the RBC membrane, as the release of hemoglobin would require disturbance of the inner part of the membrane structures sufficiently to affect cell permeability. This suggests the polymer does not penetrate through the RBC membrane which would likely result in cell lysis; instead it seems to be limited to being adsorbed on the RBC surface, keeping a more open conformation with flexible cationic ligands able to bridge adjacent RBCs. The zwitterionic SB ligands might either block membrane penetration or reduce the cross-linking action mediated by the cationic QA ligands. The control cationic polymer carrying no stabilizing zwitterions was highly hemolytic, however, as seen from the red supernatant in Fig. 5. The excellent compatibility with RBCs indicates that the HPG-SB-QA polymers could be promising candidates for applications in vivo. 3.5.2. Complement activation Complement activation is a highly relevant parameter to evaluate since it has been shown that polycations can activate complement [52]. Biomaterial-induced complement activation is an undesirable effect that should be avoided as it may provoke rapid clearance of the biomaterials and inflammatory reactions [53]. The 50% hemolytic complement assay (CH50 assay) is a screening assay for the activation of the classical complement pathway [54]. CH50 screens for the activation of the complement pathway by detecting the hemolysis of antibody sensitized sheep erythrocytes. The effects of the HPG-SB-QAX polymers and control polymer HPG-QA on

52

J. Wen et al. / Biomaterials 86 (2016) 42e55

Fig. 8. (A) Normal TEG trace; (B) Representative TEG traces for hemostatic polymers at concentration of 1 mg/mL (final). Black line shows the PBS buffer. Green shows HPG-SB-QA30. Pink shows HPG-SB-QA40. Brown shows HPG-SB-QA45; (C) Representative TEG traces for polymers with concentration of 0.1 mg/mL. Black line shows the PBS buffer. Green shows HPG-SB-QA30. Pink shows HPG-SB-QA45.

J. Wen et al. / Biomaterials 86 (2016) 42e55 Table 3 Clotting kinetics parameter values of human whole blood mixed with polymer solutions. Polymer (mg/mL)

R (min)

K (min)

a (deg)

MA (mm)

Reference range [60] PBS control HPG-SB-QA30, 0.1 HPG-SB-QA30, 1 HPG-SB-QA45, 0.1 HPG-SB-QA40, 1 HPG-SB-QA45, 1

3e22 12.1 ± 11.9 ± 14.9 ± 13.8 ± 23.9 ± 62.0 ±

2e7 3.9 ± 3.2 ± 3.4 ± 3.6 ± 8.2 ± NA

47e78 50.7 ± 9.4 50.4 ± 7.2 47.7 ± 10.4 47.3 ± 6.9 19.1 ± 5.5 5.4 ± 0.8

48e70 55.4 ± 4.1 53.1 ± 8.6 54.4 ± 6.9 49.6 ± 1.2 47.7 ± 0.4 20.9 ± 2.2

1.8 1.0 3.6 3.1 4.1 5.9

1.4 0.8 1.2 1.0 1.9

complement activation in human serum at 37  C were investigated. Immunoglobulin G (IgG) and PBS buffer are used as positive and negative controls respectively. The results are presented in Fig. 6. The results show that the SB-QA polymers did not activate complement pathways compared to a PBS control. In contrast, the control QA polymer and IgG produced strong complement consumption, as expected.

3.6. Effects on coagulation A change in plasma coagulation properties upon incubation with the polymer is an indication of its interactions with blood components. Accepted in vitro blood coagulation tests include prothrombin time (PT) and activated partial thromboplastin time (aPTT). PT is used to evaluate the extrinsic and common coagulation pathway whereas aPTT is used to evaluate the intrinsic pathway. PT measures the amount of time it takes factor VIIa to form a complex with tissue factor and proceed to clot formation [55], with 11e14 s as the normal range. As shown in Fig. 7(A), the presence of HPG-SB-QA polymers in PPP did not produce any statistically significant difference in the time taken for clots to form compared to the PBS buffer control, suggesting that the extrinsic pathway of coagulation is not affected by these polymers. However, the control polycationic polymer almost doubled the time and fell outside the normal range. The results of aPTT are expressed in seconds and measure the speed of the intrinsic pathway [55]. As shown in Fig. 7(B), the neutrally charged HPG-SB did not affect aPTT but all positively charged HPG-SB-QAX polymers increased aPTT significantly compared to the PBS control, indicating that the presence of HPG-SB-QA somehow affects the intrinsic coagulation pathway by an unknown mechanism. The increases put these samples' aPTT values into the range typical of heparinized patients, however

53

(2e3.5  control times; 60 s e 105 s in our case [56]) so are not particularly dangerous clinically. It has been shown that polycations do reduce thrombin activity and thereby decrease blood coagulation rates [57] however further studies are needed to elucidate the details of how the current polymers interact with the coagulation cascades. The aPTT of the control polycationic polymer fell outside the measurement range of the instrument (The measurement stopped automatically after 500 s). PT and aPTT alone cannot always reflect overall biomaterialinduced blood clotting activity as platelets and RBCs are not present. Further studies are required to better understand these effects. TEG characterizes the formation and strength of a blood clot in anticoagulated fresh human whole blood after Ca2þ is added to initiate coagulation as a function of time, providing more detailed information on blood coagulation compared to aPTT and PT. TEG is used to investigate the following parameters in the blood clotting process [58] related to the viscoelastic properties of the clot formation and breakdown, shown in the standard TEG curve in Fig. 8(A). During the TEG measurement the outer cup oscillates through a small angle while the inner element registers the stress transmitted to the inner surface through the coagulating sample. The TEG trace represents the maximum stress detected, in each direction, through the series of oscillations applied. R represents the time until the first evidence of a clot is detected after calcium is added back into the citrated whole blood and is measured at the point that the two lines have separated 2 mm. K represents the clot propagation rate and is measured as the time from R until the curves are 20 mm apart. The angle a represents the tangential angle of the curve with the horizontal and also represents clot propagation. MA is the maximal amplitude represented as the distance in millimeters between the two curves once they are parallel and represents peak clot rigidity. LY30 is the rate of amplitude reduction 30 min after maximum amplitude, and assesses the clot breakdown. Here the influences of SB/QA modified HPGs on the whole blood clotting process were measured, using PBS buffer as a normal control. The TEG data is listed in Table 3. Additionally, representative TEG traces are given in Fig. 8(B) and (C). TEG traces for PBS control, and HPG-SB-QA30 (0.1 mg/mL) did not differ. For HPG-SBQA30 (1 mg/mL) and HPG-SB-QA45 (0.1 mg/mL), the initial clotting time (R) was prolonged by 2e3 min, but still considered within normal range. TEG traces did not lose their characteristic shape but the blood clotting time was prolonged significantly when positive charge density and concentration of the conjugate both increased, which agreed with the aPTT result. Upon the treatment of HPG-SB-

Fig. 9. Dose-response curves for cell viability of HUVECs (A) and CHO cells (B) after being treated 48 h with hemostatic polymers by the propidium iodide live cell staining assay. Results are means ± SD (n ¼ 6).

54

J. Wen et al. / Biomaterials 86 (2016) 42e55

QA40/45, the initial “artificial” clot formed by the aggregation of RBCs and platelets would help plug the wound; the coagulation system is initiated although it takes a longer time. 3.7. Cytotoxicity

adjusting the molar ratio of these two ligands. The polymers exhibit abilities to aggregate erythrocytes and to aggregate and activate platelets in a concentration and positive charge density-dependent manner. The hemostatic polymers induce the aggregation of erythrocytes likely by interaction with erythrocyte surface charges and promote hemostasis by activation and aggregation of platelets. Their effects on coagulation are negligible or mild, depending on the fraction of polycation and polymer concentration. The presence of zwitterionic sulfabetaine moieties on the hyperbranched polymer partially masks the adverse effects that are normally caused by polycationic moieties. As a result, the current hemostatic polymer does not induce any hemolysis or endothelial cell toxicity. Furthermore, the polymer does not affect the complement system. It is also notable that the RBC aggregation effect could be reversed by using another polymer bearing a higher negative charge density, which is a desired feature for the application as blood vessel sealant. Overall, this class of macromolecules appears promising as a new hemostatic material.

Cell membrane damage is an important aspect of nucleated cell response to polymers upon exposure. The in vitro cytotoxicity of HPG-SB-QA polymers was measured as a function of polymer concentration using the propidium iodide live cell staining assay. When cells have membrane damage, the propidum iodide in the assay solution passively diffuses into the cytoplasm and binds with intracellular RNA or DNA. By quantitating the percentage of propidium iodide positive cells, the percentage of cells in the total cell population experiencing membrane damage can be estimated. As endothelial cells line the entire vascular system, the influence of HPG-SB-QA on the membranes of human umbilical vein endothelial cells (HUVECs) was investigated. The results are summarized in Fig. 9(A), showing that for the control HPG-QA polymer, cytotoxicity increases with increasing polymer concentration. Greater than 30% of the cell population lost viability at the highest polymer concentration (2 mg/mL). In contrast, viability of cells incubated with HPG-SB-QA polymers only slightly decreased with increasing concentration, retaining greater than 80% of their viability even when the concentration used was up to 2 mg/mL. To provide some breadth to the conclusion we also repeated the testing with a Chinese hamster ovary (CHO) cell line with the results shown in Fig. 9(B). Again, the safety of the HPG-SB-QA polymers even at the concentration of 2 mg/mL was evident, suggesting that combining positively charged surface quaternized amine groups with zwitterionic moieties is effective in reducing the interaction force between the polymer and cell membranes, thus reducing the toxicity.

Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.biomaterials.2016.01.067.

3.8. Reversal of erythrocyte aggregation

References

As discussed above, the hemostatic application envisaged is driven by the aggregation of erythrocytes via electrostatic interaction between the positively charged HPG-SB-QA and the negatively charged cell membrane. This also suggests a mechanism to reverse the aggregation, which is via a polymeric species bearing a higher density of negative charges than the negatively charged membrane components that can preferentially bind to the positive moieties of the polymer and thereby disengage the positive ligands from the cells. We demonstrate such reversal of erythrocytes aggregation using a sulfated linear polyglycerol [59] (LPG-sulfate, molecular characteristics shown in Table S2). As the highest aggregation activity was observed with the polymer HPG-SB-QA45, it was used as a model polymer for this study. Fig. 2(F) shows the result of adding LPG-sulfate (1 mg/ml final concentration) to whole blood aggregates formed by adding 1 mg/ mL HPG-SB-QA45. From the visual observations, it was clear that the aggregates could be immediately reversed by this low molecular weight negatively charged LPG-sulfate. As indicated by ITC studies (Fig. S2, Table S1), the small molecule LPG-sulfate bound to HPGSB-QA45 with a binding affinity in the order of 105 M1, which is strong enough to cause the positive ligand on the polymer to reequilibrate and bind to the LPG-sulfate instead of the RBC membrane. As a result, the adjacent cells are no longer connected by polymers and are thus able to flow freely. 4. Conclusions Double bond-carrying hyperbranched polyglycerols were surface engineered via a UV initiated thiol-ene “click” reaction. The proportion of SB/QA ligands can be well controlled through

Acknowledgments This research is supported by Natural Sciences and Engineering Council of Canada (NSERC) (RPGIN 238852-11) and the Canadian Blood Services (201209-293). JNK is a recipient of a Michael Smith Foundation for Health Research Career Scholar Award. MW is grateful to the German Research Foundation (WE 5377/1-1) (DFG) for financial support. Appendix A. Supplementary data

[1] H.E. Achneck, B. Sileshi, R.M. Jamiolkowski, D.M. Albala, M.L. Shapiro, J.H. Lawson, A comprehensive review of topical hemostatic agents: efficacy and recommendations for use, Ann. Surg. 251 (2) (2010) 217e228. [2] F. di Lena, Hemostatic polymers: the concept, state of the art and perspectives, J. Mater. Chem. B 2 (23) (2014) 3567e3577. [3] D.H. Park, S.B. Kim, K.-D. Ahn, E.Y. Kim, Y.J. Kim, D.K. Han, In vitro degradation and cytotoxicity of alkyl 2-cyanoacrylate polymers for application to tissue adhesives, J. Appl. Polym. Sci. 89 (12) (2003) 3272e3278. [4] P. Rhee, C. Brown, M. Martin, A. Salim, D. Plurad, D. Green, L. Chambers, D. Demetriades, G. Velmahos, H. Alam, QuikClot use in trauma for hemorrhage control: case series of 103 documented uses, J. Trauma 64 (4) (2008) 1093e1099. [5] R.F. Reiss, M.C. Oz, Autogous fibrin glue: production and clinical use, Transfus. Med. Rev. 10 (2) (1996) 85e92. [6] S.K. Samal, M. Dash, S. Van Vlierberghe, D.L. Kaplan, E. Chiellini, C. van Blitterswijk, L. Moroni, P. Dubruel, Cationic polymers and their therapeutic potential, Chem. Soc. Rev. 41 (21) (2012) 7147e7194. [7] C. Li, D. Zhong, Y. Zhang, W. Tuo, N. Li, Q. Wang, Z. Liu, W. Xue, The effect of the gene carrier material polyethyleneimine on the structure and function of human red blood cells in vitro, J. Mater. Chem. B 1 (14) (2013) 1885e1893. [8] A.M. Carmona-Ribeiro, L.D. de Melo Carrasco, Cationic antimicrobial polymers and their assemblies, Int. J. Mol. Sci. 14 (5) (2013) 9906e9946. [9] W.G. Malette, H.J. Quigley, R.D. Gaines, N.D. Johnson, W.G. Rainer, Chitosan: a new hemostatic, Ann. Thorac. Surg. 36 (1) (1983) 55e58. [10] N. Biezunski, E. Shafrir, A. De Vries, E. Katchalski, The action of polylysine on the conversion of fibrinogen into fibrin by coagulase thrombin, Biochem. J. 59 (1) (1955) 55e58. [11] Y.W. Cho, J.-D. Kim, K. Park, Polycation gene delivery systems: escape from endosomes to cytosol, J. Pharm. Pharmacol. 55 (6) (2003) 721e734. [12] R. Esfand, D.A. Tomalia, Poly(amidoamine) (PAMAM) dendrimers: from biomimicry to drug delivery and biomedical applications, Drug Discov. Today 6 (8) (2000) 427e436. [13] J.-K.F. Suh, H.W.T. Matthew, Application of chitosan-based polysaccharide biomaterials in cartilage tissue engineering: a review, Biomaterials 21 (24) (2000) 2589e2598. [14] N. Khanam, C. Mikoryak, R.K. Draper, K.J. Balkus Jr., Electrospun linear polyethyleneimine scaffolds for cell growth, Acta Biomater. 3 (6) (2007) 1050e1059. [15] S. Hong, P.R. Leroueil, E.K. Janus, J.L. Peters, M.-M. Kober, M.T. Islam, B.G. Orr,

J. Wen et al. / Biomaterials 86 (2016) 42e55

[16]

[17]

[18] [19]

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

J.R. Baker, M.M. Banaszak Holl, Interactions of polycationic polymers with supported lipid bilayers and cells: nanoscale hole formation and enhanced membrane permeability, Bioconjugate Chem. 17 (3) (2006) 728e734. D. Fischer, Y. Li, B. Ahlemeyer, J. Krieglstein, T. Kissel, In vitro cytotoxicity testing of polycations: influence of polymer structure on cell viability and hemolysis, Biomaterials 24 (7) (2003) 1121e1131. K. Pombo Garcia, K. Zarschler, L. Barbaro, J.A. Barreto, W. O'Malley, L. Spiccia, H. Stephan, B. Graham, Zwitterionic-coated “stealth” nanoparticles for biomedical applications: recent advances in countering biomolecular corona formation and uptake by the mononuclear phagocyte system, Small 10 (13) (2014) 2516e2529. M.-C. Sin, S.-H. Chen, Y. Chang, Hemocompatibility of zwitterionic interfaces and membranes, Polym. J. 46 (8) (2014) 436e443. X. Yu, Z. Liu, J. Janzen, I. Chafeeva, S. Horte, W. Chen, R.K. Kainthan, J.N. Kizhakkedathu, D.E. Brooks, Polyvalent choline phosphate as a universal biomembrane adhesive, Nat. Mater. 11 (5) (2012) 468e476. X. Yu, X. Yang, S. Horte, J.N. Kizhakkedathu, D.E. Brooks, ATRP synthesis of poly(2-(methacryloyloxy)ethyl choline phosphate): a multivalent universal biomembrane adhesive, Chem. Commun. 49 (61) (2013) 6831e6833. X. Chen, T. Chen, Z. Lin, X. Li, W. Wu, J. Li, Choline phosphate functionalized surface: protein-resistant but cell-adhesive zwitterionic surface potential for tissue engineering, Chem. Commun. 51 (3) (2015) 487e490. J. Wu, W. Lin, Z. Wang, S. Chen, Y. Chang, Investigation of the hydration of nonfouling material poly(sulfobetaine methacrylate) by low-field nuclear magnetic resonance, Langmuir 28 (19) (2012) 7436e7441. S. Chen, L. Li, C. Zhao, J. Zheng, Surface hydration: principles and applications toward low-fouling/nonfouling biomaterials, Polymer 51 (23) (2010) 5283e5293. E. Muro, T. Pons, N. Lequeux, A. Fragola, N. Sanson, Z. Lenkei, B. Dubertret, Small and stable sulfobetaine zwitterionic quantum dots for functiona livecell imaging, J. Am. Chem. Soc. 132 (2010) 4556e4557. N. Zhan, G. Palui, M. Safi, X. Ji, H. Mattoussi, Multidentate zwitterionic ligands provide compact and highly biocompatible quantum dots, J. Am. Chem. Soc. 135 (37) (2013) 13786e13795. L. Wang, Z. Wang, G. Ma, W. Lin, S. Chen, Reducing the cytotoxity of poly(amidoamine) dendrimers by modification of a single layer of carboxybetaine, Langmuir 29 (28) (2013) 8914e8921. W. Yang, L. Zhang, S. Wang, A.D. White, S. Jiang, Functionalizable and ultra stable nanoparticles coated with zwitterionic poly(carboxybetaine) in undiluted blood serum, Biomaterials 30 (29) (2009) 5617e5621. W. Zhou, J. Shao, Q. Jin, Q. Wei, J. Tang, J. Ji, Zwitterionic phosphorylcholine as a better ligand for gold nanorods cell uptake and selective photothermal ablation of cancer cells, Chem. Commun. 46 (9) (2010) 1479e1481. K. Hirota, K. Murakami, K. Nemoto, Y. Miyake, Coating of a surface with 2methacryloyloxyethyl phosphorylcholine (MPC) co-polymer significantly reduces retention of human pathogenic microorganisms, FEMS Microbiol. Lett. 248 (1) (2005) 37e45. V.J. Venditto, A. Dolor, A. Kohli, S. Salentinig, B.J. Boyd, J. Szoka, C. Francis, Sulfated quaternary amine lipids: a new class of inverse charge zwitterlipids, Chem. Commun. 50 (65) (2014) 9109e9111. V.A. Vasantha, S. Jana, A. Parthiban, J.G. Vancso, Halophilic polysulfabetaines e synthesis and study of gelation and thermoresponsive behavior, RSC Adv. 4 (43) (2014) 22596e22600. R. Chapanian, I. Constantinescu, N.A. Rossi, N. Medvedev, D.E. Brooks, M.D. Scott, J.N. Kizhakkedathu, Influence of polymer architecture on antigens camouflage, CD47 protection and complement mediated lysis of surface grafted red blood cells, Biomaterials 33 (31) (2012) 7871e7883. D. Wilms, S.-E. Stiriba, H. Frey, Hyperbranched polyglyerols: from the controlled synthesis of biocompatible polyether polyols to multipurpose applications, Acc. Chem. Res. 43 (1) (2010) 129e141. L. Ye, K. Letchford, M. Heller, R. Liggins, D. Guan, J.N. Kizhakkedathu, D.E. Brooks, J.K. Jackson, H.M. Burt, Synthesis and characterization of carboxylic acid conjugated, hydrophobically derivatezed, hyperbranched polyglycerols as nanoparticulate drug carriers for cisplatin, Biomacromolecules 12 (1) (2011) 145e155. C. Mugabe, B.A. Hadaschik, R.K. Kainthan, D.E. Brooks, A.I. So, M.E. Gleave, H.M. Burt, Paclitaxel incorporated in hydrophobically derivatized hyperbranched polyglycerols for intravesical bladder cancer therapy, BJU Int. 103 (7) (2009) 978e986. R. Chapanian, D.H. Kwan, I. Constantinescu, F.A. Shaikh, N.A. Rossi, S.G. Withers, J.N. Kizhakkedathu, Enhancement of biological reactions on cell

55

surfaces via macromolecular crowding, Nat. Commun. 5 (2014) 4683. [37] R. Chapanian, I. Constantinescu, D.E. Brooks, M.D. Scott, J.N. Kizhakkedathu, In vivo circulation, clearance, and biodistribution of polyglycerol grafted functional red blood cells, Biomaterials 33 (10) (2012) 3047e3057. [38] R.A. Shenoi, M.T. Kalathottukaren, R.J. Travers, B.F.L. Lai, A.L. Creagh, D. Lange, K. Yu, M. Weinhart, B.H. Chew, C. Du, D.E. Brooks, C.J. Carter, J.H. Morrissey, C.A. Haynes, J.N. Kizhakkedathu, Affinity-based design of a synthetic universal reversal agent for heparin anticoagulants, Sci. Transl. Med. 6 (260) (2014), 260ra150. [39] P. Kumar, R.A. Shenoi, B.F. Lai, M. Nguyen, J.N. Kizhakkedathu, S.K. Straus, Conjugation of aurein 2.2 to HPG yields an antimicrobial with better properties, Biomacromolecules 16 (3) (2015) 913e923. [40] R.K. Kainthan, J. Janzen, E. Levin, D.V. Devine, D.E. Brooks, Biocompatibility testing of branched and linear polyglycidol, Biomacromolecules 7 (3) (2006) 703e709. [41] R.K. Kainthan, S.R. Hester, E. Levin, D.V. Devine, D.E. Brooks, In vitro biological evaluation of high molecular weight hyperbranched polyglycerols, Biomaterials 28 (31) (2007) 4581e4590. [42] S. Krause, T. Scholz, U. Temmler, W. Losche, Monitoring the effects of platelet glycoprotein IIb/IIIa antagonists with a microtiter plate method for detection of platelet aggregation, Platelets 12 (7) (2001) 423e430. [43] T. Yu, A. Malugin, H. Ghandehari, Impact of silica nanoparticle design on cellular toxicity and hemolytic activity, ACS Nano 5 (7) (2011) 5717e5728. [44] M. Imran ul-haq, B.F. Lai, R. Chapanian, J.N. Kizhakkedathu, Influence of architecture of high molecular weight linear and branched polyglycerols on their biocompatibility and biodistribution, Biomaterials 33 (35) (2012) 9135e9147. [45] D. Bolliger, M.D. Seeberger, K.A. Tanaka, Principles and practice of thromboelastography in clinical coagulation management and transfusion practice, Transfus. Med. Rev. 26 (1) (2012) 1e13. [46] A. Zambruni, U. Thalheimer, G. Leandro, D. Perry, A.K. Burroughs, Thromboelastography with citrated blood: comparability with native blood, stability of citrate storage and effect of repeated sampling, Blood Coagul. Fibrinolysis 15 (1) (2004) 103e107. [47] J.M. Breier, N.M. Radio, W.R. Mundy, T.J. Shafer, Development of a highthroughput screening assay for chemical effects on proliferation and viability of immortalized human neural progenitor cells, Toxicol. Sci. 105 (1) (2008) 119e133. [48] M. Tanaka, T. Hayashi, S. Morita, The roles of water molecules at the biointerface of medical polymers, Polym. J. 45 (7) (2013) 701e710. [49] Q. Shao, S. Jiang, Molecular understanding and design of zwitterionic materials, Adv. Mater. 27 (1) (2015) 15e26. [50] T. Hothorn, B.S. Everitt, A Handbook of Statistical Analyses Using R, 2nd ed., Chapman and Hall/CRC, 2009. [51] D.B. Cines, T. Lebedeva, C. Nagaswami, V. Hayes, W. Massefski, R.I. Litvinov, L. Rauova, T.J. Lowery, J.W. Weisel, Clot contraction: compression of erythrocytes into tightly packed polyhedra and redistribution of platelets and fibrin, Blood 123 (10) (2014) 1596e1603. [52] C. Plank, K. Mechtler, F.C. Szoka, E. Wagner, Activation of the complement system by synthetic DNA complexes: a potential barrier for intravenous gene delivery, Hum. Gene Ther. 7 (12) (1996) 1437e1446. [53] B. Nilsson, K.N. Ekdahl, T.E. Mollnes, J.D. Lambris, The role of complement in biomaterial-induced inflammation, Mol. Immunol. 44 (1e3) (2007) 82e94. [54] CH50 testL.W. Moreland (Ed.), Rheumatology and Immunology Therapy, Springer Berlin Heidelberg, 2004, pp. 205e206. [55] T.G. DeLoughery, Tests of hemostasis and thrombosis, in: T.G. DeLoughery (Ed.), Hemostasis and Thrombosis, 3 ed., Springer International Publishing, 2015, pp. 9e14. [56] J.W. Eikelboom, J. Hirsh, Monitoring unfractionated heparin with the aPTT: time for a fresh look, Thromb. Haemost. 96 (5) (2006) 547e552. [57] F. Ni Ainle, R.J.S. Preston, P.V. Jenkins, H.J. Nel, J.A. Johnson, O.P. Smith, B. White, P.G. Fallon, J.S. O'Donnell, Protamine sulfate down-regulates thrombin generation by inhibiting factor V activation, Blood 114 (8) (2009) 1658e1665. [58] S. Vig, A. Chitolie, H. Bevan, A. Halliday, J. Dormandy, Thromboelastography: a reliable test? Blood Coagul. Fibrinolysis 12 (7) (2001) 555e561. €ger, B.F. Lai, D.E. Brooks, J.N. Kizhakkedathu, [59] M. Weinhart, R.A. Shenoi, D. Gro R. Haag, Anionic Linear Polyglycerol Derivatives, a Method for Manufacturing and Applications, deutsche Patentanmeldung, 2015. EP 15151794. [60] C.L. Davis, W.L. Chandler, Thromboelastography for the prediction of bleeding after transplant renal biopsy, J. Am. Soc. Nephrol. 6 (4) (1995) 1250e1255.