Molecular Imaging of Factor XIIIa Activity in Thrombosis Using a Novel, Near-Infrared Fluorescent Contrast Agent That Covalently Links to Thrombi
Background— Activated factor XIII (FXIIIa) mediates fibrinolytic resistance and is a hallmark of newly formed thrombi. In vivo imaging of FXIIIa activity could further elucidate the role of this molecule in thrombosis and other biological processes and aid in the clinical detection of acute thrombi.
Methods and Results— An FXIIIa-sensitive near-infrared fluorescence imaging agent (A15) was engineered by conjugating a near-infrared fluorochrome to a peptide ligand derived from the amino terminus of α2-antiplasmin. To evaluate the molecular specificity of A15 for FXIIIa, a control agent (C15) was also synthesized by modifying a single key glutamine residue in A15. Fluorescence imaging experiments with A15 demonstrated stronger thrombosis enhancement in human plasma clots in vitro (P<0.001 versus C15 clots and other controls). A15 was found to be highly specific for the active site of FXIIIa and was covalently bound to fibrin. In vivo murine experiments with A15 demonstrated significant signal enhancement in acute intravascular thrombi (P<0.05 versus C15 group). Minimal A15 enhancement was seen in older aged thrombi (>24 hours), consistent with an expected decline of FXIIIa activity over time. Imaging results were confirmed on correlative histopathology and fluorescence microscopy.
Conclusions— A15 is a novel optical imaging agent that is specifically crosslinked to fibrin by FXIIIa, permitting detection of FXIIIa activity in experimental thrombi in vivo. This agent should permit assessment of FXIIIa activity in a broad range of biological processes and could aid in the clinical diagnosis of acute thrombi.
Received December 18, 2003; de novo received February 13, 2004; revision received March 23, 2004; accepted March 24, 2004.
Thrombosis is an important pathological hallmark of several vascular diseases, including acute coronary syndromes, stroke, venous thromboembolism, and peripheral arterial disease. Because thrombi become resistant to fibrinolytic therapy within hours,1 immediate recognition and treatment of these disorders can improve clinical outcome. Experimentally, activated factor XIII (FXIIIa), a thrombin-activated tetrameric transglutaminase, is an important mediator of fibrinolytic resistance.2 FXIIIa augments thrombus stability by crosslinking fibrin γ-chains and α-chains, increasing their tensile strength, and by covalently binding molecules that impede plasmin activity, such as α2-antiplasmin.3 In thrombosis models, therapy with anti-FXIIIa compounds hastens fibrinolysis and reduces thrombus weight.4,5 More recently, FXIIIa has been found to rapidly crosslink α2-antiplasmin into thrombi with a catalytic half-life of ≈20 minutes in vivo and is therefore a marker of biologically acute thrombi.6
In vivo imaging of FXIIIa activity could further elucidate its role in thrombosis and fibrinolytic resistance and could aid in the clinical diagnosis of acute thrombi, such as those found in acute coronary and cerebrovascular syndromes. Specific diagnosis of biologically active thrombi7,8 could guide antithrombotic therapy with fibrinolytic or anti-FXIIIa agents, direct or indirect thrombin inhibitors, or new compounds. In addition, specific imaging of FXIIIa activity could further define the role of FXIIIa in other biological processes in which this enzyme has been implicated, such as surgical wound healing,9 vitreoretinopathies,10 maintenance of normal pregnancy,11 platelet activation,12 and genetic polymorphisms of factor XIII.13
To image FXIIIa activity, we recently synthesized a FXIIIa-specific near-infrared fluorescent (NIRF) imaging agent derived from α2-antiplasmin and characterized its properties in vitro.14 In this report, we show that this agent (1) specifically images FXIIIa activity in human plasma clots, (2) covalently binds to thrombi, and (3) detects FXIIIa activity in experimental thrombi in vivo.
Synthesis of the Factor XIIIa (A15) and Control (C15) NIRF Imaging Agents
The chemical synthesis of the FXIIIa-sensitive NIRF imaging agent A15 has been described in detail.14 Briefly, an FXIIIa peptide substrate, N13QEQVSPLTLLK24,6 based on the N-terminus of α2-antiplasmin, was synthesized by solid-phase peptide synthesis and then reacted with an NIR fluorochrome (Alexa Fluor 680C2 maleimide [AF680]; Molecular Probes) via its cysteine side chain. The product was purified by use of high-performance liquid chromatography, resulting in a final compound, A15, that possessed excitation and emission wavelengths of 679 and 702 nm, respectively. To evaluate the molecular specificity of A15 for the active site of FXIIIa, a control imaging agent (C15) was engineered similarly but contained an alanine residue in place of a pivotal glutamine residue at position 14 (Table), greatly reducing its affinity for FXIIIa.6
In Vitro Imaging of Activated FXIII Activity in Human Plasma Clots
To confirm that A15 was specific for human FXIIIa, the A15 probe was evaluated in human fresh-frozen plasma (FFP). Clots were created by mixing 90 μL FFP with 3 μL of A15, C15, AF680 (free NIR fluorochrome), or normal saline. All imaging agents had the same concentration of 100 μmol/L based on the fluorochrome AF680. In 1 of the 5 experimental groups, iodoacetamide (IAA, 2 μL, 10 mg/mL), an alkylating agent capable of inactivating FXIIIa,3 was added to an FFP aliquot for 5 minutes before addition of A15. In a 96-well plate, a final aliquot of 93 μL for each of the 5 groups was then mixed with 5 μL of thrombin (1 U/10 μL) and 5 μL of 0.4 mol/L CaCl2. Experiments were performed in quadruplicate. Clots were incubated at 37°C for 90 minutes and then washed in 1× Tris-buffered saline (TBS), compressed, and resuspended in TBS. After incubation, clots were imaged on a custom-built NIRF reflectance imaging system.15 The well plate was placed in the center of the imaging field, and light and NIRF images were obtained as digital 16-bit images. After imaging, clots were snap-frozen at −80°C.
Western Blot Protein Analysis of Human Plasma Clots
To determine whether A15 thrombosis signal enhancement was caused by FXIIIa-mediated covalent crosslinking of A15 to fibrin, clots were formed as above. After washing, the clots were solubilized sequentially in 9 mol/L urea and then sample buffer under reducing conditions as previously described.6 The solubilized proteins were electrophoresed on SDS-polyacrylamide gels, electrophoretically transferred to membranes, and immunoblotted with a polyclonal antibody against the N-terminus of α2-antiplasmin.
In Vivo Imaging of FXIIIa Activity in Murine Intravascular Thrombi
Ferric Chloride Thrombosis Model
To assess whether A15 could detect FXIIIa activity in vivo, an intravascular model of thrombosis was created using topical ferric chloride (FeCl3).8,16 BALB/c mice were obtained and cared for according to our institution’s animal facility guidelines. The animal studies committee of our hospital approved all protocols. Mice were anesthetized with an intraperitoneal mixture of ketamine 100 mg/kg and xylazine 10 mg/kg. After a midline skin incision, the superficial femoral vessels were exposed by blunt dissection. Subsequently, a 2×5-mm strip of Whatman #1 filter paper (Whatman Inc) was soaked in 10% FeCl3 and applied to the superficial femoral bundle for 3 minutes. At discrete time points, each mouse received an intracardiac injection of 5 nmol of either A15 or C15.
Intravital Fluorescence Microscopy
After injection of imaging agent, the mouse was placed onto an inverted epifluorescence microscope (Zeiss Axiovert) with a cooled CCD camera (Sensys Photometrics) interfaced to a Macintosh computer. Femoral vessels were viewed in phase-contrast or NIRF mode (680 nm long-pass emission filter, Omega Optical). Control fluorescence images were also obtained using the fluorescein isothiocyanate (FITC) channel (emission bandpass filter, 515 to 565 nm; OmegaOptical). Serial light and fluorescence digital 16-bit images (FITC exposure time 1 second; NIR exposure time 10 to 30 seconds) were obtained.
Correlative Histopathological and Fluorescence Microscopy
Thrombosed segments of the femoral vessels were identified visually, and specific branches were marked to allow precise localization during histological processing. The vessels were harvested, soaked in 30% sucrose at 4°C, and embedded in OCT on dry ice. Sections 10 μm thick were obtained by use of a cryostat. FITC and NIRF fluorescence microscopic images were then obtained. The same sections were then stained with hematoxylin and eosin to identify intravascular thrombi.
Fluorescence Reflectance Image Analysis
Images were analyzed with a commercial software package (Kodak Digital Science 1D software). For each thrombus, the NIRF signal intensity was measured with a manual region-of-interest (ROI) tracing tool. Images were further processed by use of Photoshop 7.0 software (Adobe Systems).
Intravital Fluorescence Microscopy Image Analysis
Intravital fluorescence microscopy (IVFM) images were analyzed with a manual ROI tracing program on IP Laboratory Spectrum software, version 3.6.3 (Scanalytics). The NIRF signal was measured in ROIs of intravascular thrombi and their adjacent perivascular background. The contrast-to-noise ratio was defined as 100%×(thrombus signal−adjacent background signal)/(adjacent background signal).
In a time course experiment of A15-enhanced thrombi, serial NIRF images from the same anatomic area were analyzed with CMIR Image, an interactive image analysis tool developed in the Interactive Data Language (IDL, Research Systems). Images were manually registered by 2D rotation and translation until the femoral vessels were coincident. Thereafter, the NIRF signal intensity was measured in each of the registered images by use of the same ROIs of thrombus and background.
Data are presented as mean±SD. For differences between multiple groups, 1-way ANOVA followed by a post hoc Tukey test for multiple comparisons was used. For differences between 2 groups, a Student t test was used. A value of P<0.05 was considered statistically significant.
A15 Specifically Images FXIIIa Activity in Human Plasma Clots
The ability of A15 to be crosslinked by human FXIIIa was first tested in vitro in plasma clots. After incubation of plasma clots with individual imaging agents, the NIRF fluorescence signal in each group was as follows: A15, 1562.3±193.8 arbitrary units (AU); A15+IAA, 111.5±28.2 AU; C15, 247.5±65.6 AU; AF680, 328.0±72.9 AU; normal saline, 4.5±0.6 AU (P<0.001 A15 versus each other group; Figure 1). Therefore, thrombi in the A15 group had a 347-fold NIRF signal increase over the saline group and a 4.8- to 6.3-fold increase over the other control groups. Preincubation of FFP thrombi with IAA, a known inhibitor of FXIIIa, reduced the A15 thrombus enhancement by 92.9% (P<0.001), further confirming the specificity of A15 for FXIIIa.
FXIIIa Covalently Crosslinks A15 to Fibrin in Clots
To determine whether A15 was crosslinked to fibrin by FXIIIa, plasma clots were denatured and subjected to immunoblotting. A polyclonal antibody was used that binds to the N-terminal peptide of α2-antiplasmin,6 which forms the backbone of the A15 agent. Clots formed with A15 (with A15 having a mass of <3 kDa) showed a band at ≈70 kDa that was strongly immunoreactive with the antibody to the N-terminus of α2-antiplasmin (Figure 2). This process required active FXIIIa, because the immunoreactive band was markedly diminished in the presence of an FXIIIa inhibitor (IAA), and no band was detected in the absence of A15 (saline control). In addition, no band was detected in the clots formed with C15, an agent containing a single amino acid change that blocks crosslinking by FXIIIa. These data demonstrated that A15 was covalently crosslinked to fibrin in the plasma clot by FXIIIa in a highly specific manner.
A15 Specifically Images FXIIIa Activity in Experimental Intravascular Thrombi
A total of 11 mice were studied. Eight mice received an imaging agent within 2 hours of FeCl3 application (acute thrombus group). Specifically, 6 mice received an injection of either the A15 (n=3) or C15 (n=3) agent 30 minutes after FeCl3, and 2 mice received the A15 agent at either 0 or 120 minutes after FeCl3. Three additional mice received the A15 agent 28 hours after FeCl3 application (subacute thrombus group). All animal experiments were well tolerated; no adverse outcomes were noted from intracardiac or imaging agent injection.
Intravital fluorescence images demonstrated thrombosis NIR signal enhancement in all animals with acute thrombi that received the A15 agent (Figure 3). Enhancement of multiple segments of femoral artery and vein were noted within 30 to 45 minutes and persisted up to completion of the imaging experiment. In animals that received an imaging agent 30 minutes after FeCl3 application, the peak contrast-to-noise ratio between thrombus and background by 90 minutes was 15.0±6.7% (versus −1.4±2.5% in the C15 group, P=0.039). Greater NIR thrombus signal enhancement was noted at the vessel wall–thrombus interface or at the leading thrombus edges, presumably in areas of greater thrombin and consequent FXIIIa activity. In the subacute thrombus group (28 hours after FeCl3) injected with A15, there was no thrombosis enhancement in the main femoral vessels (Figure 3E); occasional minor enhancement was noted in a small branch vessel. This finding is consistent with an expected temporal decline in FXIIIa activity.6
Remarkably, C15, differing from A15 by only a single amino acid residue, produced no significant thrombus enhancement in vivo (Figure 3F), despite the presence of intravascular thrombi on histopathological sections in the C15 group (sections not shown). This indicates that covalent crosslinking of A15 by active FXIIIa, which did not occur with C15 (Figure 2), was required for thrombosis signal enhancement.
In a time-course experiment in which A15 was injected immediately after FeCl3 injury, the contrast between the A15-enhanced thrombus and adjacent background increased over time (Figure 4). This finding suggests that NIR signal washed out of the intravascular and tissue background faster than from thrombi, consistent with the covalent binding of A15 to thrombi.
A15 Intravascular NIRF Signal Enhancement Colocalizes With Thrombus
Identification of vascular branches on the intravital images allowed precise registration of corresponding frozen and histopathological sections (Figures 3B and 5⇓). Fluorescence microscopy of A15-enhanced microthrombi revealed strong focal NIR signal from within arterial and venous thrombi identified on the corresponding intravital images. As in the in vivo images, stronger signal enhancement was noted at the thrombus–vascular wall interface or at the outer margin of the thrombus (Figure 5). This may reflect higher levels of FXIIIa activity in these biologically active areas or, less likely because of the small size of A15 (<3 kDa), a concentration gradient in the penetration of A15 into the thrombus center. FITC autofluorescence was not seen within thrombi, confirming that the observed NIR signal was not caused by bleed-through from background autofluorescence. Histology demonstrated that FeCl3 induced venous nonocclusive and/or occlusive thrombi in all animals and nonocclusive arterial thrombi in most animals.
In this report, we show that A15, an FXIIIa-specific NIRF imaging agent, can image FXIIIa activity in experimental thrombi. In vitro, A15 increased the thrombus NIR signal by nearly 350-fold over background and 5- to 6-fold over an engineered control probe (C15) and other control groups. Western blot analysis demonstrated that A15 was specifically and covalently bound to fibrin via FXIIIa crosslinking. In vivo, A15 significantly enhanced intravascular murine thrombi compared with C15, demonstrating that specific FXIIIa activity was required for thrombosis enhancement. Contrast between thrombi and background increased over time, consistent with the covalent binding of A15 to thrombi in vivo. These results demonstrate that A15 is specific for FXIIIa, covalently binds to fibrin, and permits in vivo detection of FXIIIa activity in newly formed experimental thrombi.
Enhancement of human and murine thrombi by A15 is probably a result of a combination of synergistic factors. First, the core peptide backbone of A15, N13QEQVSPLTLLK24, is based on the amino terminus of α2-antiplasmin, an avid and specific substrate for FXIIIa.6 As demonstrated on immunoblot analysis, A15 also capitalizes on the ability of FXIIIa to covalently bind its substrates to fibrin, permitting higher and more durable thrombus contrast. This favorable property of A15 is a form of biological signal amplification, a term describing imaging agents that can be locally concentrated at their targets, resulting in higher target-to-background ratios.17 Crosslinking of A15 to fibrin probably explains how thrombus contrast increased over time and suggests that higher contrast might be realized at time points >90 minutes in the absence of fibrinolysis or embolization.
NIRF imaging of A15 therefore offers a new method to image the presence and activity of activated coagulation factor XIII in arterial and venous thrombi in vivo. By use of a standardized thrombosis model, the A15 agent could allow further insight into thrombus aging and permit assessment of direct or indirect (eg, thrombin inhibitors) anti-FXIIIa pharmacological agents. Furthermore, because A15 is covalently bound, NIR thrombosis signal decay reflects a combination of fibrinolysis and washout of the circulating unbound fraction of A15. Additional time-course experiments and modeling of the NIR thrombosis signal decay could yield insight into the kinetics of fibrinolysis. In addition to thrombosis, A15 may allow further insight into the role of FXIIIa in other aspects of biology, including surgical wound healing,9 vitreoretinopathies,10 pregnancy,11 platelet activation,12 and genetic polymorphisms of factor XIII.13
Indocyanine green, an NIRF contrast agent, has been used in thousands of patients and possesses an excellent safety profile;18 thus, the A15 agent could allow clinical imaging of human thrombi in vivo. Potential advantages of A15 as a clinical thrombosis imaging agent include its ability to covalently bind to fibrin, permitting biological amplification of the thrombus signal; its ability to detect biologically “active” thrombi7,8 that may be more amenable to medical treatment; and its small size, permitting rapid thrombus enhancement. From a clinical imaging standpoint, NIRF intraoperative19 and tomographic optical imaging20 systems are under development for patient use and, in conjunction with A15, could detect recent cerebral, intracardiac, peripheral arterial, and pulmonary thromboemboli. In addition, this agent could be used with a clinical coronary catheter NIRF imaging system to detect biologically acute thrombi on atherosclerotic lesions.
Compared with a recently described thrombin-activatable NIR agent used for acute thrombus detection, the FXIIIa agent is covalently bound and therefore is not subject to signal loss from washout of the cleaved thrombin agent.8 A relative advantage of the thrombin agent is that it uses a chemical amplification strategy, which serves to enhance target-to-background ratios more dramatically than the biological amplification strategy used by the FXIIIa agent.17 A common strength of both the FXIIIa and thrombin NIRF agents is that a single biologically active enzyme can process multiple contrast agent molecules as substrates, increasing their efficiency in NIRF signal amplification. For thrombosis detection, a limitation of both agents is their lower sensitivity for older (less biologically active) thrombi.
There are limitations to our feasibility study designed to image FXIIIa activity. First, in vivo thrombus enhancement was demonstrated invasively using IVFM, a surface-weighted method that is only semiquantitative. Second, adjustments in the dose and/or injection time of A15 could also produce greater thrombus enhancement than reported here.
In summary, we have shown that A15 is a novel, covalently bound NIRF imaging agent that specifically detects factor XIIIa activity in experimental thrombi in vivo. This agent should permit further insight into the role of FXIIIa in thrombosis and fibrinolysis and other biological and disease processes. With further development of NIRF imaging systems, A15 could also aid in the clinical diagnosis of biologically acute thrombi.
This study was supported by National Institutes of Health grants R01-CA99385 (Dr Tung), P50-CA86355 (Dr Weissleder), R24-CA92782 (Dr Weissleder), and HL-64057 (Dr Reed), the Donald W. Reynolds Foundation, and the Center for Molecular Imaging Research Development Fund (Dr Jaffer). We gratefully acknowledge Edward E. Graves, PhD, for assistance with image analysis and Qing Zeng, MD, for assistance with experimental thrombosis models.
Fibrinolytic Therapy Trialists’ (FTT) Collaborative Group. Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomised trials of more than 1000 patients. Lancet. 1994; 343: 311–322.
Reed GL, Houng AK. The contribution of activated factor XIII to fibrinolytic resistance in experimental pulmonary embolism. Circulation. 1999; 99: 299–304.
Shebuski RJ, Sitko GR, Claremon DA, et al. Inhibition of factor XIIIa in a canine model of coronary thrombosis: effect on reperfusion and acute reocclusion after recombinant tissue-type plasminogen activator. Blood. 1990; 75: 1455–1459.
Robinson BR, Houng AK, Reed GL. Catalytic life of activated factor XIII in thrombi: implications for fibrinolytic resistance and thrombus aging. Circulation. 2000; 102: 1151–1157.
Badimon JJ, Fuster V. Can we image the “active” thrombus? Arterioscler Thromb Vasc Biol. 2002; 22: 1753–1754.
Jaffer FA, Tung CH, Gerszten RE, et al. In vivo imaging of thrombin activity in experimental thrombi with thrombin-sensitive near-infrared molecular probe. Arterioscler Thromb Vasc Biol. 2002; 22: 1929–1935.
Ariens RA, Lai TS, Weisel JW, et al. Role of factor XIII in fibrin clot formation and effects of genetic polymorphisms. Blood. 2002; 100: 743–754.
Farrehi PM, Ozaki CK, Carmeliet P, et al. Regulation of arterial thrombolysis by plasminogen activator inhibitor-1 in mice. Circulation. 1998; 97: 1002–1008.
Jaffer FA, Weissleder R. Seeing within: molecular imaging of the cardiovascular system. Circ Res. 2004; 94: 433–445.