Left Ventricular Thrombus Enhancement After Intravenous Injection of Echogenic Immunoliposomes
Studies in a New Experimental Model
Background— Targeted echogenic immunoliposomes (ELIPs) for ultrasound enhancement of atheroma components have been developed. To date, ELIP delivery has been intra-arterial. To determine whether ELIPs can be given intravenously with enhancement of systemic structures, a left ventricular thrombus (LVT) model was developed.
Methods and Results— In 6 animals plus 1 dose-ranging animal, the apical coronary arteries were ligated, and an LVT was produced by injecting Hemaseel fibrin adhesive through the apical myocardium. The thrombus was imaged epicardially and transthoracically at 0, 1, 5, and 10 minutes after anti-fibrinogen ELIP injections. The dose of ELIPs was varied. PBS and unconjugated ELIPs were controls. The apical thrombi were easily reproduced and clearly visible with epicardial and transthoracic ultrasound. Enhancement occurred with 2 mg anti-fibrinogen ELIPs and increased with dose. With 8 mg ELIPs, enhancement was different from control within 10 minutes (P<0.05). Rhodamine-labeled anti-fibrinogen ELIPs were seen with fluorescence microscopy of the LVT. Blinded viewing detected enhancement by 10 minutes in all animals after anti-fibrinogen ELIPs.
Conclusions— We describe an easily reproducible LVT model. Anti-fibrinogen ELIPs delivered intravenously, as a single-step process, rapidly enhance the ultrasound image of a systemic target. This allows for future development of ELIPs as a targeted ultrasound contrast agent.
Received February 28, 2002; accepted March 21, 2002.
Nontargeted ultrasound contrast agents are a routine tool for clinical ultrasound image enhancement. Targeted ultrasound contrast agents, however, have not yet been developed to the level that allows clinical utility. One reason for this has been the complexity associated with conjugation, echogenicity, size, and reticuloendothelial system survival of targeted contrast agents. Whereas nontargeted contrast agents opacify cavities through high concentrations of intrinsically acoustically reflective particles, targeted contrast agents rely, in addition, on surface binding.
Lanza et al1 first demonstrated that a multistep acoustic biotinylated, lipid-coated, perfluorocarbon nanoemulsion as an ultrasound contrast agent could be successfully targeted to dog thrombi in vivo. Since then, Lanza et al2,3⇓ have developed this contrast agent using acoustic microscopy to measure backscatter power in imaging avidin-conjugated thrombi and nitrocellulose membranes. They have also demonstrated that this agent can infiltrate arterial walls and localize tissue factor expression.4
Our laboratory has developed targeted echogenic immunoliposomes (ELIPs) for intravascular ultrasound imaging of atherosclerosis. ELIPs are small (<1 μm) ultrasound contrast agents, produced in the absence of added air, that can be used to evaluate molecular components of vasoactive and pathological endothelium/atherosclerosis.7,8⇓ With conjugation to anti-ICAM, anti-fibrinogen, anti-fibrin, and anti-VCAM antibodies (Ab/Abs), these liposomes have been shown in preliminary studies to target and highlight early and later atheroma components by use of intravascular and transvascular ultrasound.9,10⇓ Once atheroma can be staged, appropriate therapy may potentially be targeted. In addition, ELIPs can be used as targeted therapeutic delivery systems, because they are nontoxic and can carry either water-soluble compounds in the aqueous compartment or insoluble compounds within the phospholipid bilayer. For such purposes, liposomes have been evaluated as drug11–13⇓⇓ and gene-delivery14 systems.
The purpose of this study was to demonstrate the systemic ultrasound enhancement of left ventricular (LV) thrombus after the intravenous injection of ELIPs conjugated to anti-fibrinogen.
Chemicals used were anti-fibrinogen 313R from American Diagnostica; Hemaseel APR fibrin sealant Kit from Hemacure Corp; and component phospholipids and all other chemicals from Sigma and Avanti.
Component phospholipids (phosphatidylcholine [PC], phosphatidylglycerol [PG], maleimido-4(p-phenylbutyrate)-phosphatidylethanolamine [MPB-PE], and cholesterol in a molar ratio of 69:8:8:15) were dissolved in chloroform and mixed; the chloroform was then evaporated under argon.15 The resulting lipid film was placed under vacuum for 12 hours at 10 mm Hg pressure. The dry lipid film was rehydrated with deionized water and sonicated in a water bath for 5 minutes; then the liposome size was assessed according to light absorbance.16 Before lyophilization, the size was 60 to 100 nm. To increase the multilamellar structure, mannitol was added, and the lipid mixture was lyophilized for 24 hours. Anti-fibrinogen ELIPs (animal 3) and unconjugated ELIPs (animal 6) were labeled with fluorescent rhodamine.
For conjugation, 6 mg of rabbit anti-human fibrinogen was reacted with 3-(2-pyridyldithio)propionic acid-N-hydroxysuccinimide ester (SPDP) at a SPDP-to-protein molar ratio of 15:1 for 30 minutes at room temperature. Protein was separated from unreacted SPDP by gel chromatography on a 50-mL Sephadex G-50 column equilibrated with 0.05 mol/L sodium citrate, 0.05 mol/L sodium phosphate, and 0.05 mol/L sodium chloride, pH 5.5. Protein fractions were identified by optical absorbance at 280 nm (A280), pooled, and concentrated to <2 mL with Centricon YM-10 centrifugal filter units. The PDP-protein was reduced in 25 mmol/L dithiothreitol for 30 minutes at room temperature. The thiolated protein was isolated by use of the G-50 column, equilibrated, and eluted with pH 6.7 citrate-phosphate buffer. Protein-containing fractions were pooled and concentrated. The thiolated protein was reacted with reconstituted MPB liposomes (10 mg lipid/mL 0.1 mol/L phosphate buffer, pH 6.62) under nitrogen overnight at room temperature. Anionic liposomes were separated from free protein and low-molecular-weight products by dialysis in 5 mL Spectra/Por cellulose ester DispoDialyzer units (molecular weight cutoff 300 kDa) versus 100 volumes of 0.92 mol/L PBS, pH 7.4, with 2 changes of the same buffer in 24 hours. The retentates were lyophilized in the presence of 0.1 mol/L d-mannitol.
Thiolation of rabbit anti-fibrinogen, determined from A343, was 6.28 thiols per molecule. The conjugation efficiency was 4.2 μg IgG/mg lipid, determined by quantitative immunoblot assay17; 24.6% of anti-fibrinogen reacted with MPB liposomes for conjugation. The molar coupling ratio of rabbit anti-fibrinogen Abs conjugated to ELIPs is 18.4 μmol Ab/mol lipid. Because MPB-PE composes 8 mol%, the specific molar coupling ratio is 0.23 mmol Ab/mol MPB-PE (1 molecule of Ab coupled for every 4340 molecules of MPB-PE). On the basis of an Ab thiolation efficiency of 6.28, a maximum of 6 to 7 molecules of MPB-PE could be coupled per molecule of Ab. Previous work indicated that 13 500 molecules of rabbit anti-fibrin/fibrinogen were conjugated per liposome. Calculations based on immunoglobulin molecular dimensions revealed that 65% of the liposomal surface was covered with Abs, suggesting that optimal conjugation efficiency is dominated by steric considerations. Scatchard analysis indicated that 60% to 80% of liposomal Ab sites could be occupied by antigen (fibrinogen), suggesting that the preponderance of Abs is on the outer surface of the ELIPs.
This study was approved by the Animal Care and Use Committee of Northwestern University. In 7 mongrel dogs (20 to 30 kg), endotracheal intubation was performed after premedication with acepromazine (0.5 mg/kg) and sodium pentobarbital (29 mg/kg). Isoflurane anesthesia (1.5% to 3%) was then administered. Oxygen was delivered at 5 to 10 mL · kg−1 · min−1 with nitrous oxide (O2 to N2O ratio 1:1). Left femoral venous, pulmonary, and arterial lines were established for continuous monitoring.
After a midline thoracotomy, the apical coronary arteries were ligated to produce an apical myocardial infarction. An LV mural thrombus was created by use of the Hemaseel fibrin sealant; sealer protein concentrate was reconstituted with 2.0 mL bovine fibrinolysis inhibitor solution at 37°C. Thrombin was reconstituted with 2.0 mL of calcium chloride solution at 37°C. Two separate syringes used to withdraw the sealer protein concentrate and the thrombin solution were placed into the Hemaseel APR applicator. The fibrin sealant was injected in small increments (0.5 mL) directly into the LV cavity and then the myocardium as the needle was withdrawn. The size of the fibrin core was controlled by imaging immediately after each injection. The thrombus remained in situ for the duration of the experiment (Figure 1). Histological analysis of the thrombus demonstrated a fibrin core surrounded by fresh thrombus.
Anti-fibrinogen-conjugated ELIPs were injected through the left femoral vein. The dose of anti-fibrinogen-conjugated ELIPs was based on the weight of lipid. In animal 1, the dose of ELIPs started at 0.5 mg lipid and was incrementally increased through 1, 2, 4, and 8 mg to assess dose response. The maximum ELIP dose injected was 8 mg. For the other animals, the dose of ELIPs was kept constant at 8 mg. Saline and unconjugated ELIPs were used as controls.
Image Acquisition and Pixel Analysis
Epicardial ultrasound images were acquired onto videotape with an Acuson Sequoia 256. Imaging was performed with a 3.5-MHz probe in fundamental mode. The probe was placed on the apical anterior segment of the left ventricle, and imaging planes were oriented through the entire length of the thrombus to make images consistent. For each animal, images were acquired before intravenous injection and at 1, 5, and 10 minutes after injection. In 4 animals, the chest was then closed, and transthoracic imaging was performed with a 3.5-MHz probe with second harmonic imaging at 0, 1, 5, and 10 minutes after injection. The use of second harmonics, gain, compression, rejection, and dynamic range settings were tailored to maximize endocardial definition and kept constant throughout each experiment.
For each data set, 30 seconds of videotape was acquired. From the videotape, a blinded observer digitized 20 end-diastolic images to 640×480 pixel spatial resolution (≈0.2 mm/pixel) and 8-bit (256 gray levels) amplitude resolution. For each image, an area of interest (AOI) over the fresh thrombus/blood interface was created by subtracting all gray-scale values >230 from the analysis, which represented the central echogenic fibrin core. A gray-scale histogram was obtained, and the mean gray scale of the remaining fresh thrombus was calculated with Image-Pro Plus software (V4.1, Media Cybernetics). An AOI was placed over the LV cavity, and the mean gray scale was calculated. The LV thrombus mean gray scale minus the LV cavity mean gray scale was calculated for each image. An AOI was placed over the entire imaging sector to calculate the mean gray scale for the entire image.
Anti-fibrinogen-conjugated ELIPs and unconjugated ELIPs, both labeled with rhodamine, were injected into 2 separate animals, and in each, the thrombus was harvested postmortem. Frozen sections of the thrombus and adjacent normal myocardium were viewed by fluorescence microscopy.
A blinded observer was shown 30-second videotaped ultrasound images of the LV thrombus at random time intervals (1, 5, or 10 minutes) after each injection of anti-fibrinogen ELIPs, unconjugated ELIPs, or saline and compared thrombus brightness with the baseline image (0 minutes). The observer identified whether the thrombus was brighter (+), the same (−), or uncertain (±).
Data are presented as percentage change of LV thrombus minus cavity gray scale (%). Data were analyzed by use of SigmaStat version 2.03 statistical software (SPSS Inc). Results are expressed as mean±SEM. Differences between groups were analyzed with a 1-way ANOVA Student-Newman-Keuls method; a value of P<0.05 was considered statistically significant.
After the anti-fibrinogen-conjugated acoustic liposome injection, the thrombus became visibly more enhanced, and this enhancement increased over a number of minutes (Figure 1).
A fluorescent microscopic image of a thrombus is shown in Figure 2. This demonstrates fluorescent rhodamine-labeled anti-fibrinogen ELIPs attached to the thrombus surface with possible penetration into the immediate thrombus interior. No fluorescence was detectable on the normal myocardium or on the thrombus surface after injection with unconjugated rhodamine-labeled ELIPs in another animal.
Cumulative Dose Response
The effect of cumulative dose on the LV thrombus minus background mean gray scale, after each ELIP injection, is shown in Figure 3. The LV thrombus minus background mean gray scale for combined time points (1, 5, and 10 minutes) is different from control (0 minutes) at 0.5, 2, 4, and 8 mg (n=60, P<0.05). A cumulative dose-response effect is demonstrated.
Change in LV Thrombus Minus Background Mean Gray Scale
The effect of 8-mg doses of anti-fibrinogen ELIPs on LV thrombus minus background mean gray scale is displayed in Figure 4 and Table 1. Epicardial ultrasound enhancement was visibly detectable at 5 minutes. The percentage change in LV thrombus minus background mean gray scale was different from saline and unconjugated ELIPs at 10 minutes (P<0.05). The percentage change in LV thrombus minus background mean gray scale was 23±8.6%. There was no difference between unconjugated ELIPs and saline at all time points (P=NS).
The effect of 8-mg doses of anti-fibrinogen ELIPs on LV thrombus minus background mean gray scale is displayed in Figure 5 and Table 2. The percentage change in LV thrombus minus background mean gray scale was different from saline and unconjugated ELIPs at 10 minutes (P<0.05). The percentage change in LV thrombus minus background mean gray scale was 31±14.5%.
No hemodynamic changes were observed during the procedure. After each ELIP injection, there was no difference in mean arterial pressure, heart rate, or mean pulmonary artery pressure (P=NS).
Variability of Data
Intraobserver variability of the image analysis was calculated by reanalyzing the 12 separate time frames, each with 20 images, before and after ELIPs, saline, and unconjugated ELIP injections. The results were compared with the first analysis performed by the same operator. The difference between the means of the 2 groups was 2.3%.
The image variability attributable to transducer angulation, sonographer movement, ventilator respiration, and cardiac motion was measured by assessing the total mean gray scale of every digitized image and the thrombus mean gray scale data. The total intra-time point image variability for the epicardial images was 6±2.8%, and for transthoracic images, 5±2.6% (n=20 images per time point). The intra-time point variability of the thrombus image alone was 8±3.5% for the epicardial images and 9±3.8% for the transthoracic images (n=20 images per time point).
After anti-fibrinogen ELIPs, the blinded observer was able to detect an increase in the echogenicity of the LV thrombus in all 6 animals at 10 minutes and in 3 animals at 5 minutes. No enhancement was identified in any of the controls (Table 3). The blinded observer was able to detect enhancement in the dose-ranging animal at 4 mg.
After anti-fibrinogen ELIPs, the blinded observer was able to detect an increase in the echogenicity of the LV thrombus in all 4 animals at 10 minutes and in 1 animal at 1 and 5 minutes. In animal 7, the observer was not sure of thrombus enhancement at 10 minutes after saline and unconjugated injections (Table 3).
Our discussion will focus on the rationale for this methodology, implications, factors influencing our results, and potential applications.
ELIPs are composed of 4 primary lipids: PC, MPB-PE, PG, and cholesterol, and are made by a lyophilization process.7,8,15⇓⇓ We have been able to quantitate the ELIP intravascular ultrasound enhancement of fibrin in vitro9,18,19⇓⇓ and qualitatively demonstrate acoustic image enhancement of atheroma with anti-fibrinogen ELIPs delivered locally and imaged by transvascular and intravascular ultrasound.10 Tiukinhoy et al14 showed comparable echogenicity per particle of ELIPs compared with Albunex and other liposomal preparations. Through manipulation of component lipids, our laboratory recently increased the echogenicity of ELIPs by ≈300%. By manipulating the negatively charged lipids, we reduced aggregation to allow surface layering. We demonstrated in an in vitro model that interface thickness and brightness are the components that make up total interface enhancement.19
In this study, we initially used 1 animal to demonstrate the dose-response curve of the ELIPs. Using this animal, we found that an 8-mg dose (known to be nontoxic in other experimental studies) provided adequate enhancement of the thrombus. We did not explore the upper end of the dose-response relationship.
Rapid enhancement occurs after intravenous injection, and this demonstrates that acoustic liposomes successfully cross the pulmonary vasculature without uptake by the pulmonary reticuloendothelial system. Although some images were visibly brighter at 5 minutes, it was not until 10 minutes that all images were different from baseline. Although this is prolonged compared with previous ELIP intra-arterial injection studies and in vitro flow chamber studies,10,19⇓ some of the increased time to enhancement may be a result of delayed pulmonary circulation. It is known that nontargeted intravenous ultrasound contrast agents take a number of seconds/minutes to cross the lungs, and this effect continues over many minutes. Alternatively, ELIPs in diluted numbers, when administered systemically, may obey classic antibody-antigen kinetics that allow binding and enhancement to continue over time with recirculation. The contribution of lung clearance versus recirculation is unknown.
In their in vivo arterial thrombus study, Lanza et al1 describe a 3-step targeting process with an avidin-biotin-perfluorocarbon contrast agent that required 30 minutes for each step, which was then repeated before introduction of the perfluorocarbon contrast component. In comparison, the time to image enhancement by ELIPs is more rapid.
The degree of contrast enhancement of a systemic structure also depends on its intrinsic ultrasound reflectivity. If the structure is bright and reflects ultrasound, then it is difficult to increase reflected ultrasound energy. The thrombus model we developed has a reflective core of fibrin adhesive. By subtracting the fibrin core in our analysis, we were able to more closely approximate the detection of fresh thrombus, which is often difficult to discern clinically from underlying myocardium. The fibrin core yielded a large blooming artifact with gray intensities consistently >230. To exclude the fibrin core image data, all gray intensities >230 in the thrombus were removed. In addition to the fibrin core data, additional liposome enhancement at these high gray-scale levels may have been excluded. We would anticipate that greater ultrasound enhancement is possible when the targeted structure is more echolucent, as seen clinically.
Overall thrombus enhancement reported in this study was 23±8.6% and 31±14.5% for epicardial and transthoracic imaging, respectively. The human eye can detect 20 levels of gray-scale changes in brightness in a small area within a complex image.20 In detection of an intensity change at an interface, however, the human eye detects this change not in a linear fashion but in a log-linear relationship.20 This allows small changes in intensity to be detected more easily at an interface. The videodensitometric method used quantifies the degree of overall thrombus enhancement but does not assess the effect of edge enhancement, a component the blinded observer uses to help detect a difference from control. Thus, the videodensitometric method may underestimate visual enhancement.
The anti-fibrinogen Ab used in the study is equally reactive with human fibrin and fibrinogen. Using an ELISA, we demonstrated that this Ab reacts with canine fibrin. To what extent these conjugated Abs bind to circulating fibrinogen is unknown; however, there is good evidence of Ab uptake by the thrombus, as determined histologically by the rhodamine-labeled ELIPs. We know that conjugated liposomes display specific and nonspecific binding. After saturation of fibrin binding sites, some degree of nonspecific binding may occur.
Because a thrombus is not a smooth structure and liposomes attach to the external edges with signal augmentation, the ELIPs can appear to be slightly internally incorporated. The rhodamine histology demonstrates external or close to external binding of the liposomes to the thrombus. These images suggest a small degree of internalization of the liposomes. To what degree this occurs has not been determined.
After intravenous injection of the liposomes, we demonstrated that there were no hemodynamic effects on mean arterial blood pressure, heart rate, or mean pulmonary arterial pressures. Liposomes, especially those with a particularly high negative charge, have been shown in animals to cause a pseudoallergic response through uptake by pulmonary macrophages and activation of the complement system with hypotension and tachycardia.21 The ELIPs used in this experiment have a small negative charge that prevents aggregation.22 Many nontargeted contrast agents have been developed in a canine model23 as an immunological model, because it has a pulmonary macrophage population similar to that of humans.24 Both the model and the small anionic charge deter adverse reactions of the ELIPs. These data support the preliminary safety of our echogenic liposomes.
We chose to take a cardiac mural thrombus model from the distant literature25,26⇓ and modify this to create a useful thrombus model. Our methodology, in particular, seems to create an easily imaged and reproducible thrombus similar to those in pathological conditions.
Clinical ultrasound equipment settings may not have been optimized to demonstrate maximal enhancement. The images were obtained to optimize endocardial definition and simulate the clinical situation. In open-chest animals, nonharmonic imaging was used, because harmonic imaging tended to bloom the baseline images. In closed-chest animals, harmonic imaging was required for optimal endocardial definition. This resulted in more uniform image display between the open- and closed-chest animals. Greater acoustic enhancement was seen with the transthoracic image data. Our liposomes do contain some very small trapped-air components. The extent to which air contributes to their acoustic properties and the influence that harmonic imaging has on liposomal accentuation is being evaluated in our laboratory. We chose to image continuously for 10 minutes after liposomal injection, because this was thought to be clinically appropriate, given hand-held transducer technology. Although variability in image data did occur, it was in the range of 5% for the total image and 8% to 9% for the thrombus alone, well within accepted variability limits. According to classic antibody-antigen kinetics, the Ab binding to the thrombi should occur for up to 3 hours. The extent to which added acoustic enhancement occurred as a result of increased imaging times was not addressed.
This study describes an easily reproducible LV thrombus model. It demonstrates that anti-fibrinogen-labeled ELIPs delivered intravenously as a single-step process successfully traverse the lungs and rapidly enhance epicardial and transthoracic ultrasound images of an LV apical thrombus without hemodynamic effects. It extends the application of these contrast agents beyond local/arterial injection and broadens their use for clinical applications.
This study was supported in part by grant NHLBI HL-59586 from the National Institutes of Health and by the Feinberg Cardiovascular Research Institute. Dr Hamilton is a research scholar of the Cardiac Society of Australia and New Zealand.
- ↵Lanza GM, Wallace KD, Scott MJ, et al. A novel site-targeted ultrasonic contrast agent with broad biomedical application. Circulation. 1996; 94: 3334–3340.
- ↵Lanza GM. Acoustically reflective liposomes and methods to make and use the same. US patent 5 612 057, 2000. University of Illinois and Northwestern University.
- ↵Spragg DD, Alford DR, Greferath R, et al. Immunotargeting of liposomes to activated vascular endothelial cells: a strategy for site-selective delivery in the cardiovascular system. Proc Natl Acad Sci U S A. 1997; 94: 8795–8800.
- ↵MacDonald RC, MacDonald RI. Applications of freezing and thawing in liposome technology. In: Gregoriadis G, ed. Liposome Technology. Boca Raton, Fla: CRC Press; 1993: chap 13.
- ↵Umbaugh SE. Computer Vision and Image Processing: A Practical Approach Using CVIP Tools. Upper Saddle River, NJ: Prentice Hall PTR; 1998: 13–22.
- ↵Szebeni J, Fontana JL, Wassef NM, et al. Hemodynamic changes induced by liposomes and liposome-encapsulated hemoglobin in pigs: a model for pseudoallergic cardiopulmonary reactions to liposomes: role of complement and inhibition by soluble CR1 and anti-C5a antibody. Circulation. 1999; 99: 2302–2309.