computed tomography thrombus imaging

Biomaterials 150 (2018) 125e136

Contents lists available at ScienceDirect

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

Thrombin-activatable fluorescent peptide incorporated gold nanoparticles for dual optical/computed tomography thrombus imaging Sung-Pil Kwon a, b, 1, Sangmin Jeon a, c, 1, Sung-Hoon Lee a, b, Hong Yeol Yoon a, Ju Hee Ryu a, Dayil Choi a, Jeong-Yeon Kim d, Jiwon Kim d, Jae Hyung Park c, Dong-Eog Kim d, Ick Chan Kwon a, e, Kwangmeyung Kim a, e, *, Cheol-Hee Ahn b, ** a Center for Theragnosis, Biomedical Research Institute, Korea Institute of Science and Technology, 39-1 Hawolgok-dong, Seongbuk-gu, Seoul, 136-791, Republic of Korea b Research Institute of Advanced Materials, Department of Materials Science and Engineering, Seoul National University, Seoul, 08826, Republic of Korea c School of Chemical Engineering, Sungkyunkwan University, Suwon, 440-746, Republic of Korea d Molecular Imaging and Neurovascular Research Laboratory, Dongguk University College of Medicine, Goyang, 10326, Republic of Korea e KU-KIST Graduate School of Converging Science and Technology, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul, 02841, Republic of Korea

a r t i c l e i n f o

a b s t r a c t

Article history: Received 3 May 2017 Received in revised form 3 September 2017 Accepted 8 October 2017 Available online 9 October 2017

Thrombosis is an important pathophysiologic phenomenon in various cardiovascular diseases, which can lead to oxygen deprivation and infarction of tissues by generation of a thrombus. Thus, direct thrombus imaging can provide beneficial in diagnosis and therapy of thrombosis. Herein, we developed thrombinactivatable fluorescent peptide (TAP) incorporated silica-coated gold nanoparticles (TAP-SiO2@AuNPs) for direct imaging of thrombus by dual near-infrared fluorescence (NIRF) and micro-computed tomography (micro-CT) imaging, wherein TAP molecules were used as targeted thrombin-activatable peptide probes for thrombin-specific NIRF imaging. The freshly prepared TAP-SiO2@AuNPs had an average diameter of 39.8 ± 2.55 nm and they showed the quenched NIRF signal in aqueous condition, due to the excellent quenching effect of TAP molecules on the silica-gold nanoparticle surface. However, 30.31-fold higher NIRF intensity was rapidly recovered in the presence of thrombin in vitro, due to the thrombinspecific cleavage of quenched TAP molecules on the gold particle surface. Furthermore, TAPSiO2@AuNPs were successfully accumulated in thrombus by their particle size-dependent capturing property, and they presented a potential X-ray absorption property in a dose-dependent manner. Finally, thrombotic lesion was clearly distinguished from peripheral tissues by dual NIRF/micro-CT imaging after intravenous injection of TAP-SiO2@AuNPs in the in situ thrombotic mouse model, simultaneously. This study showed that thrombin-activatable fluorescent peptide incorporated silica-coated gold nanoparticles can be potentially used as a dual imaging probe for direct thrombus imaging and therapy in clinical applications. © 2017 Elsevier Ltd. All rights reserved.

Keywords: Silica-coated gold nanoparticle Thrombin-activatable fluorescent probe Thrombus imaging Dual optical/CT imaging

1. Introduction

* Corresponding author. Center for Theragnosis, Biomedical Research Institute, Korea Institute of Science and Technology, 39-1 Hawolgok-dong, Seongbuk-gu, Seoul, 136-791, Republic of Korea. ** Corresponding author. Research Institute of Advanced Materials, Department of Materials Science and Engineering, Seoul National University, Seoul, 08826, Republic of Korea. E-mail addresses: [email protected] (K. Kim), [email protected] (C.-H. Ahn). 1 These authors have equally contributed. https://doi.org/10.1016/j.biomaterials.2017.10.017 0142-9612/© 2017 Elsevier Ltd. All rights reserved.

Thrombosis is pathophysiological phenomena, which are the formation of thrombus in a blood vessel. The thrombus formation can obstruct the flow of blood, resulting in tissue oxygen deprivation and infarction. Therefore, angiostenosis and the oxygen deprivation by thrombosis are a major cause of morbidity and death [1]. When the vessel wall or endothelium is damaged, thrombus formation is induced by the interaction between collagen and tissue factor in the blood [2]. During the thrombus formation,

126

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

tissue factor which exposed to the blood activates the fibrin formation as well as the thrombin generation. Importantly, thrombin is a serine protease which proteolytically cleaved from of coagulation factor II (prothrombin) [3]. In addition, thrombin is one of the principal mediators that converts fibrinogen into fibrin during the thrombus formation. Although the exact mechanism of thrombin activation and thrombus formation in vivo has been well investigated in vitro, direct diagnosis of thrombosis has been limited by the lack of the appropriate imaging probes in vivo. Since 1974, conventional imaging probes of the thrombus have been developed to fibrin- and platelet-related imaging [4]. Furthermore, conventional thrombus imaging probes were limited to monitor aged clots or clots after anticoagulation treatment [5]. Until now, magnetic resonance (MR) and computed tomography (CT) images with various venography techniques have been used to directly image thrombus in the clinical field [6]. In particular, CT venography with X-ray irradiation is the most commonly used for imaging of thrombus because it provides prompt, accurate and detailed information of thrombotic lesion [7]. For example, cerebral venous sinus thrombosis (CVT) and deep vein thrombosis (DVT) were often scanned using non-contrast CT as a first-line investigation in the emergency in the clinical field [8]. However, noncontrast CT is difficult to obtain a precise distribution of thrombus, because that density of the thrombus in blood vessels is similar to adjacent blood [9]. In this regard, development of thrombus-specific contrast agents which can directly visualize the thrombus and thrombin activity in the blood vessels will contribute to a prompt and accurate diagnosis of thrombosis. Thrombosis also may commonly be confronted by various clinicians due to the diversity of causes, including surgery, trauma, cancer, pregnancy, and inflammatory bowel disease [10e12]. In addition, the symptoms and signs of thrombosis are not enough to thrombus-specific diagnosis. It can cause a fatal problem such as pulmonary embolism, heart attack, or stroke [11]. Once the diagnosis is performed, adequate therapeutic treatment which combined anticoagulants with symptomatic treatment started as soon as possible to decrease post-thrombotic morbidity [13]. Therefore, early and precise diagnosis of thrombosis is important to restore venous patency and preserves valvular function, resulting in increasing survival rate of a patient. It has been reported that gold nanoparticles (AuNPs) have been studied extensively as CT contrast agents, due to their high Xray absorption coefficient property [9,14]. Furthermore, bismuth (Bi), ytterbium (Yb) or gold (Au)-based nanoparticles which have Kedge values in the X-ray energy band have developed as a novel imaging agent for the imaging of thrombosis by CT [9,15e17]. In addition, various types of nanoplatforms such as thrombin inhibitor-conjugated nanoparticle and fibrin-target peptide conjugated nanoparticles have developed for early and precise diagnosis of thrombosis using MRI or CT [18e21]. Although these thrombosis imaging nanoplatforms can provide thrombus-specific imaging with high spatiotemporal resolution, imaging of thrombosisrelated enzyme activity is still challenging. For example, we reported that tumor [22,23] or thrombosis [9,24] specific-CT images were successfully carried out using biocompatible polymer coated AuNPs in various animal models. In particular, biocompatible glycol chitosan coated AuNPs showed the substantial accumulation in the thrombus by their particle sizedependent capturing property, allowing the high-resolution CT visualization of primary and recurrent thrombus in mouse carotid artery [9]. Moreover, fibrin targeting peptide modified AuNPs showed the high-resolution in vivo CT images of thrombus in carotid artery [24]. Although AuNP-based nanoparticles were welldeveloped for CT imaging of thrombus, it is difficult to direct imaging of thrombin activity that is one of the key enzymes during the thrombus formation. It is because that direct imaging of thrombin

activity might provide an opportunity for rapid and precise diagnosis of thrombotic lesion from peripheral tissue. However, AuNPbased CT contrast agents that directly image the fibrin activity in the thrombus have not been developed so far. Importantly, multimodal imaging based on the nanotechnology and molecular imaging have provided opportunities for the development of activatable probes for detecting of biological changes in living body. In this point of view, AuNPs have attracted a great deal of attention as a multimodal platform due to their intrinsic advantages such as easy fabrication, controllable size and shapes and biocompatibility, etc. [25]. Moreover, the excellent surface area-to-volume ratio and fluorescence quenching properties of AuNPs lead to development of various fluorescence-activatable optical imaging probes [26,27]. Herein, we have developed a novel fluorescence/micro-CT dual imaging probe wherein the optical imaging probe of near-infrared fluorescence (NIRF) dye, Cy5.5, conjugated thrombin-activatable peptide (Cy5.5-Gly-D-Phe-Pip-Arg-Ser-Gly-Gly-Gly-Gly-Lys-Cys), in resulting thrombin-activatable fluorescent peptide (TAP) that may be specifically cleavable by thrombin. And then TAP molecules were directly incorporated to silica caped AuNPs (SiO2@AuNPs), in resulting TAP-SiO2@AuNPs (Scheme 1a). The freshly prepared dual imaging probe of TAP-SiO2@AuNPs may present the excellent quenched NIRF signal in a normal blood vessel, because of distancedependent quenching effect of excited states of Cy5.5 on the surface of TAP-SiO2@AuNPs [25,28]. However, the quenched fluorescence of Cy5.5 molecules on the surface of TAP-SiO2@AuNPs can be strongly activated in the presence of thrombin at the thrombotic lesion, due to the rapid cleavage from thrombin specific-cleavable peptide on the surface TAP-SiO2@AuNPs. In addition, silica caped AuNPs also can be accumulated in thrombus by their particle sizedependent capturing property, and they allow a high-resolution micro-CT imaging of thrombus at the targeted thrombotic lesion, due to their high X-ray absorption coefficient property (Scheme 1b). Finally, for the first time, we demonstrated that activatable fluorescence/micro-CT dual imaging using TAP-SiO2@AuNPs provided the successful dual imaging of thrombin activity and anatomical information of thrombotic lesion with high sensitivity in a live thrombotic animal model, simultaneously.

2. Material and methods 2.1. Materials Gold(III) chloride trihydrate (HAuCl4$3H2O, 99.9%), trisodium citrate dihydrate, N-(3-dimethylaminopropyl)-N0 -ethylcarbodiimide hydrochloride (EDC; 98%), N-hydroxy-succinimide (NHS; 98%) tetraethylorthosilicate (TEOS, 99%), ammonium hydroxide solution (NH4OH, 28e30%), (3-aminopropyl)triethoxysilane (APTES; 99%), thrombin (form human plasma) and calcium chloride anhydrous were purchased from Sigma-Aldrich (St. Louis, MO, USA). Methoxy polyethylene glycol sulfhydryl 5000 (mPEG-SH, MW ¼ 5000 Da) was obtained from SunBio (Anyang, Korea). Thrombin-specific cleavable peptide (Gly-D-Phe-Pip-Arg-Ser-GlyGly-Gly-Gly-Lys(BOC)-Cys-NH2; where Pip indicates pipecolic acid) was purchased from Peptron (Dajeon, Korea). Cy5.5 monofunctional N-hydroxysuccinimide ester (Cy5.5-NHS) (lex/lem ¼ 675/ 695 nm) was purchased from GE Healthcare (Little Chalfont, United Kingdom). Rabbit polyclonal anti-thrombin (ab92621) antibody and Alexa Fluor® 594 conjugated goat anti-rabbit IgG antibody (ab150080) were purchased from abcam (Cambridge, United Kingdom). All other chemicals were purchased as reagent grade and used without further purification. Ten-weeks-old male C57BL/ 6 mice were obtained from Orient Bio (Seoul, Korea).

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

127

Scheme 1. (a) Synthetic illustration of TAP-SiO2@AuNPs. (b) Schematic illustration of thrombin-specific near-infrared fluorescence (NIRF) and micro-CT imaging using TAPSiO2@AuNPs.

2.2. Preparation of silica-coated AuNP (SiO2@AuNPs) Gold nanoparticles (AuNPs, 15e20 nm) was prepared by sodium citrate reduction method as previous reports [29e32]. Briefly, 50 mg of HAuCl4$3H2O (0.5 mM) was dissolved in 94 ml of deionized water (DIW) and heated to 100  C. Then, 6 ml of trisodium citrate dehydrate (1% w/w) were added to a HAuCl4$3H2O solution. The reaction mixture was maintained at the boiling temperature for further 10 min before allowing to cool to room temperature (RT). To surface modification of AuNPs with silica, 7.0  107 mol of mPEG-SH in the DIW was dropwise into 10 ml of AuNPs (0.5 mM of Au) solution. The mixture was stirred for 24 h at €ber method RT to obtain mPEG coated AuNP (PEG@AuNPs) by Sto [32,33]. After that, PEG@AuNPs were centrifuged (9000 rpm, 60 min) twice with DIW to remove uncoated mPEG-SH, and redispersed in 2 ml of ethanol for silica coating. Finally, 8 ml of TEOS ethanol solution (1.1 mM) was added to PEG@AuNPs in ethanol followed by further shaking for 24 h, resulting in silica coated PEG@AuNPs (SiO2@AuNPs) and then SiO2@AuNPs was centrifuged (11000 rpm, 20 min) three times with ethanol to remove unreacted TEOS, and then re-dispersed in ethanol (0.5 mM of Au). 2.3. Preparation of thrombin-activatable fluorescent peptide (TAP) incorporated SiO2@AuNPs (TAP- SiO2@AuNPs) Fluorophore conjugated thrombin-specific cleavable peptide (TAP) probe was synthesized by conjugating the N-terminal glycine (Gly) of thrombin substrate peptide (Gly-D-Phe-Pip-Arg-Ser-GlyGly-Gly-Gly-Lys(BOC)-Cys-NH2; the substrate site is italicized and the cleavage site is between Arg and Ser) to Cy5.5-NHS [34]. In brief,

Cy5.5-NHS (1.9  106 mol) was incubated with 1 equivalent of the thrombin-specific cleavable peptide (1.9  106 mol) in 200 ml DMSO for 4 h at RT. After then, 20 ml trifluoroacetic acid (TFA)/DIW/ anisole (95:2.5:2.5, v/v) solution was added to deprotect Boc group attached to Lys. The peptide was purified by reverse-phase high performance liquid chromatography (RP-HPLC) equipped with C18 column as the following conditions: 30%e100% acetonitrile (0.1% (v/v) TFA): DIW (0.1% (v/v) TFA) over 14 min at a flow rate of 1.0 ml/ min. The appropriate fractions (>95% purity) were collected and lyophilized to obtain purified thrombin-activatable fluorescent peptide (TAP). The structural properties of TAP were analyzed using FT-IR (is 10, ThermoFisher Scientific, USA) and 600 MHz 1H NMR (DD2 600 MHz FT NMR, Agilent Technologies, USA). The molecular weight of TAP was measured using matrix-assisted laser desorption/ionization (MALDI) analysis (AB Sciex TOF/TOF 5800 System, USA) with 2,5-dihydroxybenzoic acid (DHB) matrix. The scan parameters in the MS mode were a mass range of 1900 to 2200 Da. For the adsorption of TAP to the SiO2@AuNPs, surface aminated SiO2@AuNPs (NH2-SiO2@AuNPs) were prepared as previous reports [35]. In brief, APTES ethanol solution (1 ml, 4.3 mM) was added to 9 ml of SiO2@AuNPs solution (0.5 mM of Au). Then, the reaction mixture was sonicated for 2 h at RT and stirring for 1 h at 60  C. After the reaction, surface aminated SiO2@AuNPs was centrifuged (11000 rpm, 20 min) three times with ethanol to remove unreacted APTES and re-dispersed in 2 ml of ethanol for further reaction. Finally, for the incorporating TAP to the NH2-SiO2@AuNPs, TAP (1.2  107 mol) in 200 ml DMSO was mixed with NH2-SiO2@AuNPs (2 mg/ml in DIW), and the mixture was stirred for 12 h at RT to obtain TAP incorporated SiO2@AuNPs (TAP-SiO2@AuNPs). The freshly prepared TAP-SiO2@AuNPs was centrifuged (11000 rpm, 20 min) three times with DIW to remove any residues and stored at

128

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

4  C before use. The surface properties of TAP-SiO2@AuNPs were analyzed using FT-IR (is 10, ThermoFisher Scientific, USA). Quantification of the number of TAP molecules on the SiO2@AuNPs was calculated based on absorbance between fed Cy5.5 and unreacted Cy5.5 by UVeVis spectrometer. In brief, TAP (1.2  107 mol) in 200 ml DMSO was mixed with NH2-SiO2@AuNPs (2 mg/ml in DIW) for the incorporating TAP to the NH2-SiO2@AuNPs, and the mixture was stirred for 12 h at RT to obtain TAP incorporated SiO2@AuNPs (TAP-SiO2@AuNPs). Then, TAP-SiO2@AuNPs was centrifuged (11000 rpm, 20 min) three times with DIW. After centrifugation, absorbance at 680 nm in all supernatants was measured by UVeVis spectrometer to monitor unreacted TAP molecules. The amount of unreacted TAP molecules were calculated based on the fluorescence standard curve of TAP, which measured at 680 nm. 2.4. In vitro characterization The size and surface charge of AuNPs, SiO2@AuNPs and TAPSiO2@AuNPs were measured by dynamic laser scattering (DLS, Nano ZS, Malvern, UK) at 25  C. UV absorption spectrum of AuNPs, SiO2@AuNPs and TAP-SiO2@AuNPs were measured from 400 to 800 nm using UVeVis spectrophotometer (G1103A, Agilent, USA). The morphology of TAP-SiO2@AuNPs were observed by field emission transmission electron microscopy (FE-TEM, JEOL 200CX, Japan). Micro-CT phantom image of TAP-SiO2@AuNPs was acquired using micro-CT scanner (NFR Polaris-G90, NanoFocusRay, Korea) at the various concentration of TAP-SiO2@AuNPs (5, 10, 25, 50 mg/ml). In addition, in vitro X-ray absorption property of TAP-SiO2@AuNPs was analyzed using micro-CT phantom image, compared to the clinically used Ultavist® (50 mg/ml). Relative signal intensity of micro-CT phantom image was analyzed using Image J software (NIH, Bethesda, USA). 2.5. In vitro thrombin-specific fluorescent activation of TAPSiO2@AuNPs The thrombin-specific fluorescent activation of TAPSiO2@AuNPs was evaluated by measuring the changes in nearinfrared fluorescence (NIRF) intensity after incubation of 0.5 mg/ ml of TAP-SiO2@AuNPs with or without 13.4 nM of thrombin in 50 mM Tris-buffered saline (TBS, 150 mM of NaCl and 2 M of CaCl2). After post-incubation with thrombin, the changes in NIRF intensity were monitored using an Fluorescence Spectrophotometer (F7000, Hitachi, Japan) at 0, 1, 10, 30 and 60 min at 37  C (lex/ lem ¼ 675/695 nm). In addition, thrombin concentrationdependent NIRF intensities were acquired by incubating 0.5 m/ml of TAP-SiO2@AuNPs with various concentrations (0, 2.7, 6.7, 13.4 nM) of thrombin for 20 min or 120 min. The NIRF images of TAP-SiO2@AuNPs were also acquired using a 4000 MM Kodak image station (Kodak Co., Rochester, NY), equipped with a Cy5.5 bandpass emission filter (680e720 nm) and a CCD camera. For the inhibition experiment, thrombin-specific Cy5.5 releasing from TAPSiO2@AuNPs was monitored UVeVis spectrophotometer. In brief, 0.5 mg/ml of TAP-SiO2@AuNPs was incubated with 13.4 nM of thrombin in the presence of Hirudin (4 unit), as a thrombin inhibitor. UV absorbance of the Cy5.5 which released from TAPSiO2@AuNPs in supernatant was measured after centrifugation (10000 rpm, 5 min). Fluorescence stability of TAP-SiO2@AuNPs was evaluated by measuring the increment of NIRF intensity that resulted from releasing of TAP. To mimic physiological conditions in the blood, we incubated TAP-SiO2@AuNPs in 50% fetal bovine serum (FBS) or 13.4 nM of thrombin for 60 min at 37  C. In addition, fluorescence stability of TAP-SiO2@AuNPs was evaluated using the whole blood. In brief, 200 ml of TAPSiO2@AuNPs (0.5 mg/ml) and Cy5.5 (equal fluoresce intensity) were

incubated for 90 min at 37  C with 1 ml of the heparin-treated mouse whole blood. The fluorescence stability was monitored by measuring the changes in NIRF intensity using the IVIS Spectrum imaging system (PerkinElmer, USA). The NIRF intensities of Cy5.5 and TAP-SiO2@AuNPs in blood were analyzed by Living Image software (PerkinElmer, USA), respectivley. Lastly, non-specific absorption of released Cy5.5 molecules from NH2-SiO2@AuNPs was monitored by measuring the changes in NIRF intensity. 10 mg of free Cy5.5 was mixed with 200 ml of NH2-SiO2@AuNPs (0.5 mg/ml) to mimic Cy5.5-releasing condition. As a control condition, 10 mg of Cy5.5 was directly conjugated with 200 ml of NH2-SiO2@AuNPs (0.5 mg/ml) in the presence of EDC and NHS. Both Cy5.5-mixed NH2-SiO2@AuNPs and Cy5.5-conjugated NH2-SiO2@AuNPs were incubated for 24 h at 37  C. And then the changes in the NIRF intensity were measured by the IVIS Spectrum imaging system (PerkinElmer, USA). 2.6. In vivo thrombus imaging in mouse in situ thrombotic model All in vivo animal experiments were approved by the Institutional Animal Care and Use Committee of the Dongkuk University and the Korea Institute of Science and Technology. Mouse in situ thrombotic model was established as previously described [17]. For the in vivo thrombus imaging, ten-weeks-old C57BL/6 mice were anaesthetized with avertin (250 mg/kg) and in situ thrombotic models were induced using FeCl3 soaked filter paper on the exposed left distal common carotid artery (CCA) for 5 min. After 30 min of thrombus formation, 100 mg/kg of TAP-SiO2@AuNPs or Cy5.5 with equal fluorescence intensity was directly injected via mouse tail vein and the time-dependent in vivo fluorescence images of thrombus were acquired using the IVIS Spectrum imaging system (PerkinElmer, USA). In addition, for the high resolution NIRF imaging of CCA, thrombotic CCA was dissected from the mouse after 30 min post-injection of both Cy5.5 and TAP-SiO2@AuNPstreated. The high resolution NIRF images of dissected CCA were obtained by using the Olympus OV-100 Whole Mouse Imaging System (Olympus Corp., Japan). Also, the NIRF intensities of thrombotic lesion and non-thrombotic vessel were analyzed with Image J software (NIH, Maryland, USA). Furthermore, in vivo microCT images of CCA from TAP-SiO2@AuNPs-treated mice were acquired using micro-CT scanner (NFR Polaris-G90, NanoFocusRay, Korea) with standard imaging protocol (65 kVpp, 60 mA, 26.7  26.7 mm field of view, 0.053  0.053  0.054 mm3 voxel size, 500 ms per frame, 360 views, 512  512 reconstruction matrix, and 600 slices). The acquired micro-CT images were analyzed quantitatively using 3D-rendering software (Lucion, MediSYS, Korea). As a control experiment, micro-CT images of Cy5.5 and SiO2@AuNPs-treated thrombotic model and TAP-SiO2@AuNPstreated non-thrombotic model were acquired standard imaging protocol, as described above. 2.7. Histological analysis of CCA from mouse in situ thrombotic model For the histological analysis, Cy5.5-treated and TAPSiO2@AuNPs-treated CCAs of in situ thrombotic model were dissected and fixed with 4% paraformaldehyde solution. Then, the fixed CCA tissues were embedded in paraffin. The sliced CCAs (6 mm) were stained by hematoxylin and eosin (H&E) and observed by optical microscopy (BX 51, Olympus, Japan). The NIRF signals of Cy5.5-treated and TAP-SiO2@AuNPs-treated CCAs were observed after immunofluorescence staining by rabbit polyclonal antithrombin antibody. In brief, the dissected CCAs from in situ thrombotic model was fixed in optimal cutting temperature compound (OCT, Sakura Finetek, Tokyo, Japan) and freezed for 24 h at

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

deep freezer. And then, the CCA tissue blocks were sectioned at 10 mm thickness. For thrombin staining, the CCA tissue slide was washed with PBS (pH 7.4) for twice and blocking solution was treated for 1 h at RT. Then, the CCA tissue slide was washed with PBS for twice and incubated with rabbit polyclonal anti-thrombin antibody (0.15 mg/ml, 0.3% BSA contained PBS (pH7.4)) for 2 h at room temperature. And the CCA tissue slide was washed with PBS for twice. To visualization of thrombin, Alexa Fluor® 594 conjugated goat anti-rabbit IgG antibody (1:1000, 0.3% BSA contained PBS (pH7.4)) was treated to the CCA tissue slide for 1 h at RT. Finally, the CCA tissue slide was washed with PBS for twice and mounted using cover glass. The fluorescence in the CCA tissue was observed using an inverted fluorescence microscope (IX71, Olympus, Japan). The CCA tissue was also observed through an inverted fluorescence microscope equipped with a dark-field condenser. For the CryoTEM analysis, ultrathin slices of the CCA were prepared as follows. Briefly, the CCA was fixed in 0.1 M cacodylate buffer containing 5 wt% of glutaraldehyde and 2 wt% of paraformaldehyde for 12 h. Then, it was post-fixed in 0.1 M cacodylate buffer with 2 wt% of osmium tetroxide for 90 min. The fixed CCA was washed and dehydrated using ethanol and infiltrated with EPON epoxy resin. For the polymerization of resin, it was incubated for 12 h at 38  C and for 48 h at 60  C, respectively. The CCA block was sliced by ultra-microtome (RMC MT-XL; Boeckler Instruments, Tucson, AZ) and ultrathin slices (60 nm) were collected on copper grid. The slices were stained with 4% uranyl acetate and 4% lead citrate. Finally, they were observed using Cryo-TEM at 80 kV (JEOL-1400 TEM, Japan). 2.8. In vivo toxicological analysis For the in vivo toxicological analysis of TAP-SiO2@AuNPs, major organs (liver, lung, spleen, kidney, heart) were dissected from mice at 90 min after I.V injection of TAP-SiO2@AuNPs (100 mg/kg, n ¼ 3). And then, they were fixed with 4% paraformaldehyde solution and embedded in paraffin. The sliced organs (6 mm) were stained by Hematoxylin and Eosin (H&E) and observed by optical microscope (BX 51, Olympus, USA). 2.9. Statistical analysis The differences between experimental and control groups were analyzed using one-way ANOVA and considered statistically significant (marked with an asterisk (*) in figure). 3. Results and discussion 3.1. Rational design and synthesis of thrombin-activatable fluorescent peptide (TAP) incorporated silica-coated AuNPs (TAPSiO2@AuNPs) For a thrombin-specific dual imaging in thrombus, firstly, we prepared the thrombin activatable peptide (TAP) probe by direct conjugating Cy5.5-NHS to thrombin substrate peptide (Gly-D-PhePip-Arg-Ser-Gly-Gly-Gly-Gly-Lys-Cys), wherein the thrombin specific cleavable site is between Arg and Ser [34]. To confirm of chemical structures of TAP, it was freshly dissolved in DMSO-d6 and characteristic peaks were measured using 1H NMR. The chemical structure of TAP was analyzed using characteristic peaks at 2.22e2.3 ppm (eCH2 at Cy5.5), 4.18e4.47 ppm (eCH at peptide), 4.94e5.0 ppm (-OH at Ser), 5.75e5.76 ppm and 8.41e8.51 ppm (-CH at Cy5.5) (Fig. S1a). The chemical structure of TAP was also confirmed by FT-IR measurement (Fig. S1b). Both amide I band (1620 - 1650 cm1) of TAP and eSO-3 ion of Cy5.5 band (910e990 cm1) were clearly observed in FT-IR spectrum,

129

indicating that Cy5.5 was successfully conjugated to the thrombinspecific cleavable peptide. The molecular weight of TAP was further confirmed using matrix-assisted laser desorption/ionization (MALDI) analysis with 2,5-dihydroxybenzoic acid (DHB) matrix. The exact mass of TAP was found to be 2133.54 [Mþ 3Kþþ2Naþþ5Hþ] and 2155.53 [Mþ 3Kþþ3Naþþ4Hþ] m/z, respectively (Fig. S1c). Next, in order to make stable gold nanoparticles, AuNPs were prepared by sodium citrate reduction method [29e32] and then mPEG-SH was deposited on the surface of AuNPs as a coating agent for improving colloidal stability during the further silica capping reaction. The change of surface charge from 20.6 ± 1.2 mV of AuNPs to 6.1 ± 1.4 mV of mPEG@AuNPs, resulted from the PEG coating on the AuNP surface. To improve biocompatibility and colloidal stability of our mPEG@AuNPs, silicon dioxygen (SiO2) was used as capping agent of mPEG@AuNPs. After the mPEG treatment, silica (SiO2) was firstly incorporated on the €ber procedure [32]. surface of mPEG@AuNPs via the standard Sto Then, amine functionalized silica capped AuNPs (NH2-SiO2@AuNPs) was prepared by incorporating of (3-aminopropyl)triethoxysilane (APTES) on SiO2@AuNPs for further incorporation of TAP molecules onto NH2-SiO2@AuNPs surface through charge-charge interactions. Each single AuNP was capped with a uniform silica shell thickness 8.3 ± 0.9 nm of silica shell (Fig. S2a). The surface charge of SiO2@AuNPs (20.2 ± 0.8 mV) was changed positively to 42.1 ± 0.1 mV of NH2-SiO2@AuNPs, due to the surface-amination of SiO2@AuNPs. Finally, TAP molecules were directly incorporated onto NH2-SiO2@AuNPs via physical adsorption method through charge-charge interactions between TPA molecules and NH2SiO2@AuNPs. The freshly prepared TAP incorporated NH2-SiO2@AuNPs (TAPSiO2@AuNPs) showed negative surface charge (14.9 ± 2.0 mV), suggesting the charge neutralization of primary amine groups on NH2-SiO2@AuNPs after physical adsorption of TAP molecules. Based on the UVeVis spectrum, 2300 molecules of TAP molecules were incorporated onto the single particle of NH2-SiO2@AuNPs. It has been extensively reported that the physical adsorption mechanism of peptide molecules on the surface of NH2-SiO2@AuNPs has been revealed with various mechanisms, such as hydrogen bonds formation, hydrophobic interactions, and van-der Waals interactions between peptide molecules and amine functionalized SiO2 surface [36,37]. Furthermore, negatively charged TPA molecules may be also strongly incorporated on the positively charged surface of NH2SiO2@AuNPs through electrostatic interactions, resulting in stable TPA incorporated SiO2@AuNPs (TPA-SiO2@AuNPs) [38,39]. The hydrodynamic diameters and surface charges of different AuNPs which from each reaction step were summarized in Fig. 1a. As we expected, the hydrodynamic diameters of bare AuNPs slightly increased from 24.3 ± 8.8 nm to 39.8 ± 2.55 nm according to each reaction step, confirmed using dynamic light scattering (DLS) in distilled water. Furthermore, the zeta potential of negatively charged bare AuNPs (20.6 ± 1.2 mV) was changed to 20.2 ± 0.8 mV (SiO2@AuNPs), 42.1 ± 0.1 mV (NH2-SiO2@AuNPs) and 14.9 ± 2.0 mV (TAP-SiO2@AuNPs), respectively, indicating different reaction steps clearly affected on the surface charges of AuNPs. Also, as the silica layer and TAP molecules were continuously incorporated on the surface of AuNPs, the plasmon resonance (SPR) peaks of AuNPs were slightly red-shifted from 523 nm to 529 nm (SiO2@AuNPs) and 532 nm (TAP-SiO2@AuNPs), respectively, due to the red-shifts by a local increase of refractive index of each material (Fig. 1b) [40]. Moreover, generation of absorption peak of Cy5.5 at 675 nm further supported that TAP molecules were successfully incorporated on the NH2-SiO2@AuNPs surfaces. Importantly, the TEM images of TAP-SiO2@AuNPs showed the approximately 33.1 ± 2.1 nm of spherical nanoparticle structure (Fig. 1c), which consist of 16.7 ± 2.8 nm of AuNPs core and

130

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

Fig. 1. (a) The hydrodynamic diameters and surface charges of different AuNPs which from each reaction step measured by dynamic laser scattering (DLS). (b) UVeVis spectrums of AuNPs, SiO2@AuNPs and TAP- SiO2@AuNPs in distilled water. (c) FE-TEM image of TAP-SiO2@AuNPs in distilled water (1 mg/ml). (d) Micro-CT phantom image of TAP-SiO2@AuNPs at the various concentrations (5e50 mg/ml). (e) Relative signal intensity of micro-CT phantom images of TAP-SiO2@AuNPs in distilled water. The relative signal intensity was measured using Image J software. (*) indicates difference at the p < 0.01 significance level.

8.3 ± 0.9 nm of homogeneous silica shell on the surface of AuNPs, respectively (Fig. S2a). The formation of TAP-SiO2@AuNPs was further confirmed by FT-IR measurement (Fig. S2b). As we expected, characteristic bands of amide I (1620 - 1650 cm1 of TAP), Si-O-Si (1020e1100 cm1 of SiO2@AuNPs) and Si-OH band (930e960 cm1 of SiO2@AuNPs) were clearly observed in FT-IR spectrum of TAP-SiO2@AuNPs, indicating TAP was successfully incorporated onto the SiO2@AuNPs. As a new micro-CT imaging probe, in vitro X-ray absorption property of TAP-SiO2@AuNPs was evaluated with micro-CT phantom images at the various concentrations of TAP-SiO2@AuNPs (5e50 mg/ml) (Fig. 1d). The micro-CT contrast effect was closely dependent on the concentration of TAP-SiO2@AuNPs. Relative signal intensity also gradually increased by the TAP-SiO2@AuNPs concentration. It was 2.5 fold higher than the water signal and comparable with the commercial control contrast agent of Ultravist®300 (300 mg/ml of iodine) at the 50 mg/ ml of TAP-SiO2@AuNPs, due to the higher X-ray absorption of gold

particles (Fig. 1e).

3.2. In vitro thrombin-specific fluorescence activation of TAPSiO2@AuNPs To analyze the thrombin-specific fluorescence activation of TAPSiO2@AuNPs as a new fluorescent probe, the selective fluorescence activation without or with thrombin was first measured in vitro. The freshly prepared TAP-SiO2@AuNPs showed the complete quenched NIRF signal in 50 mM Tris-buffered saline (150 mM of NaCl and 2 M of CaCl2) as well as bare AuNPs and SiO2@AuNPs, due to the selfquenching effect between TAP molecules on the gold particle surface [41]. However, the strong NIRF signal of TAP-SiO2@AuNPs was rapidly recovered after expose to 13.4 nM of thrombin within 1 h post-incubation in the reaction buffer (Fig. 2a). Next, we observed Cy5.5-release properties of TAP-SiO2@AuNPs with or without thrombin inhibitor in 50 mM Tris-buffered saline (150 mM of NaCl

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

131

Fig. 2. (a) 96-well plate optical and NIRF images of AuNPs, SiO2@AuNPs and TAP-SiO2@AuNPs (with or without 13.4 nM of thrombin) in 50 mM Tris-buffered saline (TBS, 150 mM of NaCl and 2 M of CaCl2). (b) UVeVis spectrum and optical images of supernatant, which from thrombin treated TAP-SiO2@AuNPs with/without Hirudin (4 unit). (c) Time-dependent NIRF activation and 96-well plate NIRF image of TAP-SiO2@AuNPs with 13.4 nM of thrombin. (d) Thrombin concentration-dependent NIRF activation and 96-well plate NIRF image of TAP-SiO2@AuNPs.

and 2 M of CaCl2) (Fig. 2b). The absorbance of Cy5.5 at 675 nm was dramatically increased in the presence of thrombin. However, the change of absorbance at 675 nm showed negligible increment when it treated with Hirudin (4 unit) as a thrombin inhibitor. This is because TAP can be specifically cleaved by thrombin, resulting in thrombin-specific releasing of Cy5.5 molecules which can emit strong fluorescence from TAP-SiO2@AuNPs. In addition, timedependent fluorescence intensity was observed by incubating a fixed concentration of the TAP-SiO2@AuNPs (0.5 mg/ml) in a quartz cuvette containing 13.4 nM of thrombin in the reaction buffer at  37 C for 1 h (Fig. 2c). The NIRF signal of thrombin-treated TAPSiO2@AuNPs rapidly increased only after 2 min post-incubation and dramatically increased about 30.31 fold higher than TAPSiO2@AuNPs after 1 h post-incubation with thrombin. It clearly demonstrated that significant time-dependent activation of the NIRF signals were closely related to the specific cleavage reaction between TAP molecules and thrombin in the reaction buffer, whereas the NIRF signals did not increase without thrombin. Furthermore, the fluorescence stability of TAP-SiO2@AuNPs confirmed by measuring of NIRF intensity after incubating for 1 h in the fetal bovine serum (FBS, 50% v/v) containing 50 mM Tris buffered saline (150 mM of NaCl and 2 M of CaCl2) at 37 C. Importantly, TAP-SiO2@AuNPs did not show an increased NIRF  signal in the reaction buffer up to 1 h post-incubation at 37 C (Fig. S3a). This is because that TAP-SiO2@AuNPs can form stable nanoprobe in the buffer condition. Furthermore, the fluorescence stability of TAP-SiO2@AuNPs in the whole blood also evaluated by measuring changes in the NIRF intensity (Figs. S3b and S3c). The NIRF intensity of TAP-SiO2@AuNPs was slightly increased in the

whole blood. It means that small amount of TAP molecules may be released out from SiO2@AuNPs in the whole blood. However, the fluorescence intensity of TAP-SiO2@AuNPs was still quenched for 90 min in the whole blood, compared to Cy5.5. Therefore, we expected that TAP-SiO2@AuNPs formed stable nanoprobe to detect the thrombin activity, in vivo condition. However, the fluorescence intensity of TAP-SiO2@AuNPs increased 1.9 and 12.5 fold higher than that of TAP-SiO2@AuNPs in Tris-buffered saline when it was incubated with 50% of FBS or 13.4 nM of thrombin, respectively. It means that small amount of TAP molecules may be released out from SiO2@AuNPs in the presence of FBS. However, the NIRF signal of TAP-SiO2@AuNPs in the presence of thrombin greatly increased 6.6 times higher than that of TAP-SiO2@AuNPs in only FBS, indicating that the specific activity of thrombin in blood can be visualized using TAPSiO2@AuNPs. To further evaluation of relationship between thrombin concentration and the NIRF activation, TAP-SiO2@AuNPs were incubated with various concentrations of thrombin (0, 1.3, 2.7, 6.7 and 13.4 nM) for 20 min or 2 h (Fig. 2d). The NIRF images of the TAP-SiO2@AuNPs showed the thrombin-dose dependent NIRF intensities and the NIRF intensity also increased according to the incubation time of 20 min and 2 h in the presence of thrombin. These in vitro thrombin-specific fluorescence activation data clearly demonstrated that TAP-SiO2@AuNPs could present the strong fluorescence signals in the presence of the targeted enzyme of thrombin in vitro wherein the NIRF intensity was closely related to the thrombin concentration and incubation time. Finally, nonspecific absorption of Cy5.5 onto NH2-SiO2@AuNPs was confirmed by measuring changes in NIRF intensity (Fig. S3d). When

132

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

Cy5.5 was mixed with NH2-SiO2@AuNPs, the fluorescence from Cy5.5 did not change, compared to the free Cy5.5 without NH2SiO2@AuNPs. However, there was strong quenching effect when Cy5.5 was directly conjugated on the NH2-SiO2@AuNPs surface. Therefore, we expected that released Cy5.5 molecules could not be electrostatically absorbed on the NH2-SiO2@AuNPs surface.

3.3. NIRF fluorescence imaging of in situ thrombotic animal model To observe thrombin-specific NIRF imaging of TAP-SiO2@AuNPs in vivo, in situ thrombotic models were induced using FeCl3 soaked filter paper on the exposed left distal common carotid artery (CCA) for 5 min [9,24]. After 30 min post-thrombus formation, TAPSiO2@AuNPs (100 mg/kg) or Cy5.5 (the equal amount of Cy5.5 in TAP-SiO2@AuNPs) were intravenously injected to in situ thrombotic mice models. And then, the NIRF signals in thrombotic lesion were monitored at 10, 20, 30, 40, 50, 60 and 90 min using the IVIS Spectrum imaging system (Fig. 3a). There is no strong NIRF signal in thrombotic lesion (yellow circle) after 10 min post-injection. However, the strong NIRF signals in thrombotic lesion which treated with TAP-SiO2@AuNPs could be clearly observed within 20 min postinjection, indicating the strong NIRF signals were recovered in the presence of thrombin overexpressed at the thrombotic lesion. Importantly, thrombotic lesion was easily distinguished from normal vessels and tissues with the strongest NIRF signal only after 30 min post-injection, which was maintained up for 2 h. However, it was difficult to distinguish thrombotic lesion from normal tissues when

the Cy5.5 was treated with the same animal model, indicating the lower targeting efficiency of Cy5.5 molecules in thrombotic lesion. Also, the NIRF intensities of TAP-SiO2@AuNPs treated thrombotic lesion were 2e3 times higher than that of free Cy5.5-treated thrombotic lesion after 30 min post-injection both free Cy5.5 and TAP-SiO2@AuNPs (Fig. S4a). The NIRF intensity of TAP-SiO2@AuNPsinjected thrombotic models maintained 17.04, 30.53, 33.6, 33.4, 33.8, 31 and 29.5 fold higher than non-thrombotic region in CCA at 10, 20, 30, 40, 50, 60 and 90 min, respectively. However, NIRF signals of Cy5.5 treated thrombotic model were rapidly decreased with the time, due to the rapid clearance of small molecular Cy5.5 in the blood stream [42]. These results suggest that nano-sized TAP-SiO2@AuNPs can accumulate at the thrombotic lesion by the structural properties of thrombus, which formed network structure with fibrin and red blood cells, wherein the captured TAP-SiO2@AuNPs can be rapidly activated to present strong NIRF signals in the presence of thrombin [43,44]. High resolution ex vivo NIRF images of CCA clearly supported that TAP-SiO2@AuNPs accumulated and present the strong NIRF signal at thrombus (yellow dotted area) (Fig. 3b). More clearly, TAPSiO2@AuNPs-treated CCA showed the strong NIRF signal at the thrombus, which was co-localized with the border of dark yellow thrombus in CCA. However, negligible NIRF signal was observed in thrombus as well as non-thrombotic vessel when it treated with Cy5.5, as a control. The relative NIRF signals between thrombotic lesion and non-thrombotic vessel were 2.3- and 1.1-fold higher when it treated with TAP-SiO2@AuNPs or Cy5.5, respectively (Fig. S4b).

Fig. 3. (a) Time-dependent NIRF images of in situ thrombotic model treated with TAP-SiO2@AuNPs or Cy5.5 alone. The yellow circles indicate thrombotic lesion. (b) High resolution NIRF images of TAP-SiO2@AuNPs or Cy5.5-treated CCA. The CCAs were dissected from mouse in situ thrombotic model at 30 min post-injection. The yellow lines indicate thrombotic lesion in CCA. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

3.4. Micro-CT imaging of in situ thrombotic animal models Next, we investigated that TAP-SiO2@AuNPs could be used as micro-CT contrast agent for direct imaging of thrombus in mouse in situ thrombotic model. Carotid thrombus was induced using FeCl3

133

solutions as described above and the CCA changed dark yellow after thrombus formation (Fig. 4a). After 30 min post-injection of TAPSiO2@AuNPs (100 mg/kg) into the thrombotic model, micro-CT images were acquired using micro-CT scanner with standard imaging protocol (65 kVpp, 60 mA, 26.7  26.7 mm field of view,

Fig. 4. (a) Micro-CT axial slice, sagittal slice and 3D-reconstruction images of thrombotic lesion from TAP-SiO2@AuNPs, Cy5.5 and SiO2@AuNPs-treated thrombotic model and TAPSiO2@AuNPs-treated non-thrombotic model. The yellow rectangles indicate CCA in mouse in situ thrombotic model. (b) Relative brightness in thrombotic lesion which was denoted yellow rectangles in (a) (*) indicates difference at the p < 0.001 significance level. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

134

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

0.053  0.053  0.054 mm3 voxel size, 500 ms per frame, 360 views, 512  512 reconstruction matrix, and 600 slices). As a control experiment, micro-CT images of Cy5.5 or SiO2@AuNPs-treated thrombotic model and TAP-SiO2@AuNPs-treated non-thrombotic model were acquired using micro-CT scanner with standard imaging protocol. Importantly, the thrombotic lesion (yellow rectangle) in CCA showed the stronger micro-CT attenuation than peripheral tissues in axial and sagittal micro-CT slices, when it treated with TAP-SiO2@AuNPs or SiO2@AuNPs after thrombus formation. In addition, three-dimensional (3D) reconstruction image showed more clearly the thrombotic lesion in in situ thrombotic

model. However, the thrombotic lesion could not be identified through axial and sagittal micro-CT slices as well as 3D-reconstruction image in both Cy5.5-treated thrombotic model and TAPSiO2@AuNPs-treated non-thrombotic model. The relative brightness of thrombotic lesion in axial slices was 2.2 and 1.9 fold brighter than that of peripheral tissues after treatment with TAPSiO2@AuNPs or SiO2@AuNPs, respectively (Fig. 4b). However, the relative brightness of thrombotic lesion was not significantly different from peripheral tissues in both Cy5.5-treated thrombotic model and TAP-SiO2@AuNPs-treated non-thrombotic model. From the micro-CT images, intravenously injected TAP-SiO2@AuNPs were

Fig. 5. (a) NIRF and micro-CT dual imaging of in situ thrombotic model treated with TAP-SiO2@AuNPs. The white arrows indicating that thrombus induced lesion. (b) H&E stained dissected-CCA from TAP-SiO2@AuNPs or Cy5.5-treated in situ thrombotic model. (c) Dark field images of dissected-CCA from TAP-SiO2@AuNPs or Cy5.5-treated in situ thrombotic model. The scale bar indicates 150 mm. (d) Fluorescence images of immunofluorescence stained dissected-CCA from mouse in situ thrombotic model. (v) and (T) indicate vessel endothelium and thrombus, respectively. Channel: thrombin ¼ green, TAP-SiO2@AuNPs or Cy5.5 ¼ red. The scale bar indicates 150 mm. (e) Electron microscopy image of dissectedCCA from mouse in situ thrombotic model. The scale bar indicates 300 nm. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

successfully localized at thrombotic lesion by their particle sizedependent capturing property in thrombus, and their strong Xray absorbance allowed clear visualization of thrombus using micro-CT imaging. 3.5. Dual NIRF and micro-CT imaging of thrombus in in situ thrombotic animal models Encouraged by NIRF and micro-CT images in mouse in situ thrombotic model, finally, we investigated that TAP-SiO2@AuNPs could be used as a new NIRF and micro-CT dual contrast agent for direct dual-imaging of thrombus in mouse in situ thrombotic model. Carotid thrombus was induced using FeCl3 solutions as described above. After 30 min post-injection of TAP-SiO2@AuNPs (100 mg/kg) into the thrombotic model, NIRF image and micro-CT image were simultaneously acquired using the IVIS Spectrum imaging system and micro-CT scanner (NFR Polaris-G90) at the same time. Firstly, the left NIRF image in Fig. 5a implying that higher thrombin activity based on the strong NIRF activation at thrombotic lesion (middle bright field image). In addition, the right 3Dreconstruction micro-CT image showed highlighted thrombus with high spatial resolution based on X-ray absorption property of TAPSiO2@AuNPs. Interestingly, the signals from NIRF and micro-CT images of TAP-SiO2@AuNPs were co-localized at the thrombotic lesion in in situ thrombotic model. For the precise analysis, H&E-stained sections of CCA from in situ thrombotic mouse model showed a distinct thrombus in CCA tissues following FeCl3 exposure (Fig. 5b). The dark field images of TAP-SiO2@AuNPs-treated CCA tissue showed strong scattering bright signals of gold particles in thrombus as that indicated by H&E staining, whereas Cy5.5-treated CCA tissue showed negligible scatter in thrombus (Fig. 5c). Next, we performed immunofluorescence stain of CCA tissues to prove that the NIRF signal was originated from the specificity of TAP-SiO2@AuNPs against thrombin activity. The dissected CCA tissues were immunofluorescence (IF) stained using thrombin antibody to visualize of thrombin in thrombotic lesion (Fig. 5d). As expected, the strong green fluorescence signals from thrombin antibody were clearly observed in thrombus, and they were co-localized with NIRF signals from TAP-SiO2@AuNPs, indicating the thrombin-specific NIRF image of TAP molecules on the gold particle surface. However, NIRF signals from Cy5.5-treated tissue were negligible compared to that of TAP-SiO2@AuNPs-treated tissue. Also, the strong NIRF signals from TAP-SiO2@AuNPs showed not only thrombus but endothelium. It is because that dysfunctional endothelium can be also promoted up-regulation of thrombogenic proteins, resulting in activation of thrombin in thrombotic lesion [45]. Interestingly, electron microscopy image showed that thrombus has been formed network structure and numerous TAP-SiO2@AuNPs were successfully localized within thrombus by their particle size-dependent capturing property (Fig. 5e). These ex vivo data of TAPSiO2@AuNPs-treated thrombotic lesion decisively supported that NIRF signals and micro-CT signals at the carotid thrombotic lesion were certainly originated from the thrombus-specific localization and activation of TAP-SiO2@AuNPs in in situ thrombotic animal models. Histopathological analysis of major organs to determine biocompatibility of TAP-SiO2@AuNPs in vivo (Fig. S5). H&E staining images of major organs were exhibited no signs of histopathological abnormalities such as tissue damage or inflammation after administration of TAP-SiO2@AuNPs. This is because TAPSiO2@AuNPs consisted of biocompatible materials such as silicacoated gold nanoparticle, thrombin-specific peptide substrate and fluorescent dye, Cy5.5. However, a high dose-dependent toxicity and long-term biocompatibility of TAP-SiO2@AuNPs should be

135

carefully studied for future clinical applications. 4. Conclusions In this study, we have developed a novel dual-imaging nanoplatform based on thrombin-activatable fluorescent peptide (TAP) incorporated silica capped gold nanoparticles (TAP-SiO2@AuNPs) for direct optical/micro-CT imaging of thrombus in thrombolytic animal models. The TAP-SiO2@AuNPs could monitor of thrombin activity in vitro and in vivo via thrombin-specific NIRF activation properties, due to the thrombin-specific cleavage of quenched TAP molecules from TAP-SiO2@AuNPs. Furthermore, the nano-sized TAP-SiO2@AuNPs could rapidly be accumulated in thrombus in an in situ thrombotic mouse model, due to the structural properties of thrombus which formed network structure with fibrin and red blood cells. Importantly, the thrombin-specific strong NIRF image of TAP-SiO2@AuNPs provided directly discrimination thrombotic lesion from peripheral tissues. In addition, TAP-SiO2@AuNPs could provide the accurate anatomical X-ray image of the thrombus using micro-CT, due to the high x-ray absorption property of TAPSiO2@AuNPs. Therefore, our TAP-SiO2@AuNPs showed their potential as a thrombin-specific dual NIRF/micro-CT imaging probe for direct thrombus dual imaging in vivo. Although various types of multi-modal imaging nanoplatforms have been developed for imaging of thrombosis, the challenges are still remaining for acute thrombosis imaging in clinical applications. This is because that the passive accumulation property of nanoplatforms to the thrombus which has formed network structure at late stage of thrombosis is utilizing for the thrombosis imaging. In this point of view, the combination of thrombin-activatable peptide probe (TAP) and SiO2@AuNPs can be utilized for imaging of acute thrombosis by NIRF/micro-CT dual-imaging of enzyme activity and thrombus formation. In addition, development of such a dual-imaging nanoplatform might be able to utilize for providing therapy strategy of thrombosis. Finally, our study suggests that thrombinactivatable fluorescent peptide (TAP) incorporated silica capped gold nanoparticles will provide rapid and direct dual optical/micoCT imaging of thrombosis with thrombin-dependent biological information and high spatial resolution for direct thrombus diagnosis and therapy. Acknowledgments This work was supported by the Basic Science Research Program (2010-0027955) (MEST), Global Research Lab (GRL) Program (NRF2015K1A1A2028228, NRF-2013K1A1A2A02050115), the KU-KIST School Project and the Intramural Research Program of KIST. Appendix A. Supplementary data Supplementary data related to this article can be found at https://doi.org/10.1016/j.biomaterials.2017.10.017. References [1] S. Falati, P. Gross, G. Merrill-Skoloff, B.C. Furie, B. Furie, Real-time in vivo imaging of platelets, tissue factor and fibrin during arterial thrombus formation in the mouse, Nat. Med. 8 (10) (2002) 1175e1180. [2] B. Furie, B.C. Furie, Mechanisms of thrombus formation, N. Engl. J. Med. 359 (9) (2008) 938e949. [3] C.T. Esmon, Basic mechanisms and pathogenesis of venous thrombosis, Blood Rev. 23 (5) (2009) 225e229. [4] G.N.W. Kerrigan, M.R. Buchanan, J.F. Cade, E. Regoeczi, J. Hirsh, Investigation of the mechanism of false positive 125I-labelled fibrinogen scans, Br. J. Haematol. 26 (3) (1974) 469e473. [5] G.M. Lanza, G. Cui, A.H. Schmieder, H. Zhang, J.S. Allen, M.J. Scott, T. Williams, X. Yang, An unmet clinical need: the history of thrombus imaging, J. Nucl. Cardiol. Offi. Publ. Am. Soc. Nucl. Cardiol. (2017), https://doi.org/10.1007/

136

S.-P. Kwon et al. / Biomaterials 150 (2018) 125e136

s12350-017-0942-8 (Epub ahead of print). [6] J.L. Leach, R.B. Fortuna, B.V. Jones, M.F. Gaskill-Shipley, Imaging of cerebral venous thrombosis: current techniques, spectrum of findings, and diagnostic pitfalls, RadioGraphics 26 (suppl_1) (2006) S19eS41. [7] D.F. Yankelevitz, G. Gamsu, A. Shah, J. Rademaker, D. Shaham, N. Buckshee, M.D. Cham, C.I. Henschke, Optimization of combined CT pulmonary angiography with lower extremity CT venography, Am. J. Roentgenol. 174 (1) (2000) 67e69. [8] J. Linn, T. Pfefferkorn, K. Ivanicova, S. Müller-Schunk, S. Hartz, M. Wiesmann, M. Dichgans, H. Brückmann, Noncontrast CT in deep cerebral venous thrombosis and sinus thrombosis: comparison of its diagnostic value for both entities, Am. J. Neuroradiol. 30 (4) (2009) 728e735. [9] D.-E. Kim, J.-Y. Kim, I.-C. Sun, D. Schellingerhout, S.-K. Lee, C.-H. Ahn, I.C. Kwon, K. Kim, Hyperacute direct thrombus imaging using computed tomography and gold nanoparticles, Ann. Neurol. 73 (5) (2013) 617e625. [10] J.A. Heit, M.D. Silverstein, D.N. Mohr, T.M. Petterson, W. O'Fallon, L. Melton, Iii, Risk factors for deep vein thrombosis and pulmonary embolism: a populationbased case-control study, Arch. Intern. Med. 160 (6) (2000) 809e815. [11] A.A. Khorana, C.W. Francis, E. Culakova, R.I. Fisher, N.M. Kuderer, G.H. Lyman, Thromboembolism in hospitalized neutropenic cancer patients, J. Clin. Oncol. Offi. J. Am. Soc. Clin. Oncol. 24 (3) (2006) 484e490. [12] A.A. Khorana, C.W. Francis, E. Culakova, G.H. Lyman, Risk factors for chemotherapy-associated venous thromboembolism in a prospective observational study, Cancer 104 (12) (2005) 2822e2829. [13] M.H. Meissner, P. Gloviczki, A.J. Comerota, M.C. Dalsing, B.G. Eklof, D.L. Gillespie, J.M. Lohr, R.B. McLafferty, M.H. Murad, F. Padberg, P. Pappas, J.D. Raffetto, T.W. Wakefield, Early thrombus removal strategies for acute deep venous thrombosis: clinical practice guidelines of the society for vascular surgery and the american venous forum, J. Vasc. Surg. 55 (5) (2012) 1449e1462. [14] I.-C. Sun, D.-K. Eun, H. Koo, C.-Y. Ko, H.-S. Kim, D.K. Yi, K. Choi, I.C. Kwon, K. Kim, C.-H. Ahn, Tumor-targeting gold particles for dual computed tomography/optical cancer imaging, Angew. Chem. Int. Ed. 50 (40) (2011) 9348e9351. [15] D. Pan, C.O. Schirra, A. Senpan, A.H. Schmieder, A.J. Stacy, E. Roessl, A. Thran, S.A. Wickline, R. Proska, G.M. Lanza, An early investigation of ytterbium nanocolloids for selective and quantitative “multicolor” spectral CT imaging, ACS Nano 6 (4) (2012) 3364e3370. [16] D. Pan, E. Roessl, J.-P. Schlomka, S.D. Caruthers, A. Senpan, M.J. Scott, J.S. Allen, H. Zhang, G. Hu, P.J. Gaffney, E.T. Choi, V. Rasche, S.A. Wickline, R. Proksa, G.M. Lanza, Computed tomography in color: NanoK-enhanced spectral CT molecular imaging, Angew. Chem. Int. Ed. 49 (50) (2010) 9635e9639. [17] J.Y. Kim, J.H. Ryu, D. Schellingerhout, I.C. Sun, S.K. Lee, S. Jeon, J. Kim, I.C. Kwon, M. Nahrendorf, C.H. Ahn, K. Kim, D.E. Kim, Direct imaging of cerebral thromboemboli using computed tomography and fibrin-targeted gold nanoparticles, Theranostics 5 (10) (2015) 1098e1114. [18] J. Myerson, L. He, G. Lanza, D. Tollefsen, S. Wickline, Thrombin-inhibiting perfluorocarbon nanoparticles provide a novel strategy for the treatment and magnetic resonance imaging of acute thrombosis, J. Thromb. Haemost. 9 (7) (2011) 1292e1300. [19] M. Sirol, V. Fuster, J.J. Badimon, J.T. Fallon, J.-F. Toussaint, Z.A. Fayad, Chronic thrombus detection with in vivo magnetic resonance imaging and a fibrintargeted contrast agent, Circulation 112 (11) (2005) 1594e1600. [20] M.E. Andia, P. Saha, J. Jenkins, B. Modarai, A.J. Wiethoff, A. Phinikaridou, S.P. Grover, A.S. Patel, T. Schaeffter, A. Smith, Fibrin-targeted magnetic resonance imaging allows in vivo quantification of thrombus fibrin content and identifies thrombi amenable for thrombolysis, Arterioscler. Thromb. Vasc. Biol. 34 (6) (2014) 1193e1198. [21] R.U. Palekar, J.W. Myerson, P.H. Schlesinger, J.E. Sadler, H. Pan, S.A. Wickline, Thrombin-targeted liposomes establish a sustained localized anticlotting barrier against acute thrombosis, Mol. Pharm. 10 (11) (2013) 4168e4175. [22] I.C. Sun, D.K. Eun, H. Koo, C.Y. Ko, H.S. Kim, D.K. Yi, K. Choi, I.C. Kwon, K. Kim, C.H. Ahn, Tumor-targeting gold particles for dual computed tomography/optical cancer imaging, Angew. Chem. Int. Ed. Engl. 50 (40) (2011) 9348e9351. [23] I.-C. Sun, J.H. Na, S.Y. Jeong, D.-E. Kim, I.C. Kwon, K. Choi, C.-H. Ahn, K. Kim, Biocompatible glycol chitosan-coated gold nanoparticles for tumor-targeting CT imaging, Pharm. Res. 31 (6) (2013) 1418e1425.

[24] J.-Y. Kim, J.H. Ryu, D. Schellingerhout, I.-C. Sun, S.-K. Lee, S. Jeon, J. Kim, I.C. Kwon, M. Nahrendorf, C.-H. Ahn, K. Kim, D.-E. Kim, Direct imaging of cerebral thromboemboli using computed tomography and fibrin-targeted gold nanoparticles, Theranostics 5 (10) (2015) 1098e1114. [25] M. Swierczewska, S. Lee, X. Chen, The design and application of fluorophoregold nanoparticle activatable probes, Phys. Chem. Chem. Phys. 13 (21) (2011) 9929e9941. [26] D. Liu, X. Huang, Z. Wang, A. Jin, X. Sun, L. Zhu, F. Wang, Y. Ma, G. Niu, A.R. Hight Walker, X. Chen, Gold nanoparticle-based activatable probe for sensing ultralow levels of prostate-specific antigen, ACS Nano 7 (6) (2013) 5568e5576. [27] F. Degliangeli, P. Kshirsagar, V. Brunetti, P.P. Pompa, R. Fiammengo, Absolute and direct MicroRNA quantification using DNAegold nanoparticle probes, J. Am. Chem. Soc. 136 (6) (2014) 2264e2267. [28] K.G. Thomas, P.V. Kamat, Chromophore-functionalized gold nanoparticles, Acc. Chem. Res. 36 (12) (2003) 888e898. [29] G. Frens, Controlled nucleation for the regulation of the particle size in monodisperse gold suspensions, Nature 241 (105) (1973) 20e22. [30] J. Turkevich, P.C. Stevenson, J. Hillier, The formation of colloidal gold, J. Phys. Chem. 57 (7) (1953) 670e673. nez, F.M. Romero, N.G. Bastús, V. Puntes, Small gold nanoparticles [31] I. Ojea-Jime synthesized with sodium citrate and heavy water: insights into the reaction mechanism, J. Phys. Chem. C 114 (4) (2010) 1800e1804. ndez-Lo  pez, C. Mateo-Mateo, R.A. Alvarez-Puebla, J. Pe rez-Juste, [32] C. Ferna n, Highly controlled silica coating of PEGI. Pastoriza-Santos, L.M. Liz-Marza capped metal nanoparticles and preparation of SERS-encoded particlesy, Langmuir 25 (24) (2009) 13894e13899. €ber, A. Fink, E. Bohn, Controlled growth of monodisperse silica spheres [33] W. Sto in the micron size range, J. Colloid Interface Sci. 26 (1) (1968) 62e69. [34] F.A. Jaffer, C.-H. Tung, R.E. Gerszten, R. Weissleder, In vivo imaging of thrombin activity in experimental thrombi with thrombin-sensitive nearinfrared molecular probe, Arterioscler. Thromb. Vasc. Biol. 22 (11) (2002) 1929e1935. [35] H. Wang, M. Peng, J. Zheng, P. Li, Encapsulation of silica nanoparticles by redox-initiated graft polymerization from the surface of silica nanoparticles, J. Colloid Interface Sci. 326 (1) (2008) 151e157. [36] V. Puddu, C.C. Perry, Peptide adsorption on silica nanoparticles: evidence of hydrophobic interactions, ACS Nano 6 (7) (2012) 6356e6363. [37] S. Joshi, I. Ghosh, S. Pokhrel, L. M€ adler, W.M. Nau, Interactions of amino acids and polypeptides with metal oxide nanoparticles probed by fluorescent indicator adsorption and displacement, ACS Nano 6 (6) (2012) 5668e5679. [38] S.V. Patwardhan, F.S. Emami, R.J. Berry, S.E. Jones, R.R. Naik, O. Deschaume, H. Heinz, C.C. Perry, Chemistry of aqueous silica nanoparticle surfaces and the mechanism of selective peptide adsorption, J. Am. Chem. Soc. 134 (14) (2012) 6244e6256. [39] V. Puddu, C.C. Perry, Peptide adsorption on silica nanoparticles: evidence of hydrophobic interactions, ACS Nano 6 (7) (2012) 6356e6363. n, Direct coating of [40] E. Mine, A. Yamada, Y. Kobayashi, M. Konno, L.M. Liz-Marza gold nanoparticles with silica by a seeded polymerization technique, J. Colloid Interface Sci. 264 (2) (2003) 385e390. [41] S. Lee, E.-J. Cha, K. Park, S.-Y. Lee, J.-K. Hong, I.-C. Sun, S.Y. Kim, K. Choi, I.C. Kwon, K. Kim, C.-H. Ahn, A near-infrared-fluorescence-quenched goldnanoparticle imaging probe for in vivo drug screening and protease activity determination, Angew. Chem. 120 (15) (2008) 2846e2849. [42] K. Kim, J.H. Kim, H. Park, Y.-S. Kim, K. Park, H. Nam, S. Lee, J.H. Park, R.-W. Park, I.-S. Kim, K. Choi, S.Y. Kim, K. Park, I.C. Kwon, Tumor-homing multifunctional nanoparticles for cancer theragnosis: simultaneous diagnosis, drug delivery, and therapeutic monitoring, J. Control. Release 146 (2) (2010) 219e227. €ns, Fibrin clot structure and function: a role in the [43] A. Undas, R.A.S. Arie pathophysiology of arterial and venous thromboembolic diseases, Arterioscler. Thromb. Vasc. Biol. 31 (12) (2011) e88ee99. [44] A.S. Wolberg, Thrombin generation and fibrin clot structure, Blood Rev. 21 (3) (2007) 131e142. [45] L.A. Harker, S.R. Hanson, M.S. Runge, Thrombin hypothesis of thrombus generation and vascular lesion formation, Am. J. Cardiol. 75 (6, Supplement 1) (1995) 12Be17B.