Investigation of a Potential Scintigraphic Tracer for Imaging Apoptosis: Radioiodinated Annexin V-Kunitz Protease Inhibitor Fusion Protein

Radiolabeled annexin V (ANV) has been widely used for imaging cell apoptosis. Recently, a novel ANV-Kunitz-type protease inhibitor fusion protein, ANV-6L15, was found to be a promising probe for improved apoptosis detection based on its higher affinity to phosphatidylserine (PS) compared to native ANV. The present paper investigates the feasibility of apoptosis detection using radioiodinated ANV-6L15. Native ANV and ANV-6L15 were labeled with iodine-123 and iodine-125 using Iodogen method. The binding between the radioiodinated proteins and erythrocyte ghosts or chemical-induced apoptotic cells was examined. ANV-6L15 can be radioiodinated with high yield (40%−60%) and excellent radiochemical purity (>95%). 123I-ANV-6L15 exhibited a higher binding ratio to erythrocyte ghosts and apoptotic cells compared to 123I-ANV. The biodistribution of 123I-ANV-6L15 in mice was also characterized. 123I-ANV-6L15 was rapidly cleared from the blood. High uptake in the liver and the kidneys may limit the evaluation of apoptosis in abdominal regions. Our data suggest that radiolabled ANV-6L15 may be a better scintigraphic tracer than native ANV for apoptosis detection.


Introduction
Apoptosis (programmed cell death) plays an important role in the maintenance of physiological homeostasis as well as in the pathogenesis of a number of disorders including cerebral and myocardial ischemia, autoimmune diseases, and neurodegeneration [1][2][3]. Apoptosis also plays a crucial role in tumor response to radiation, chemotherapy, and photodynamic therapy (PDT) [4]. The extent and time frame of cancer cell apoptosis induced by anticancer treatments provide crucial clinical information on both the disease status and the therapeutic efficacy.
The plasma membrane phospholipids of mammalian cells are normally asymmetrically distributed, in which the phosphatidylserine (PS) and phosphatidylethanolamine (PE) are segregated to the internal leaflet whereas the phosphatidylcholine (PC) and sphingomyelin (SM) reside on the outer leaflet [5,6]. Once apoptosis has been initiated, the caspase-mediated signaling cascade of cells is activated. One of the earliest events occurred in apoptotic cells is the externalization of PS to the outer leaflet of the plasma membrane.
Annexin V (ANV), an endogenous human protein with a molecular weight of 35.8 kDa, binds to membrane-bound 2 Journal of Biomedicine and Biotechnology PS in a Ca 2+ -dependent manner with a high affinity (K d = 0.5-7 nM). Since externalization of PS occurs in the early stage of apoptosis, fluorescein-and radionuclide-labeled ANV have been used for detection of apoptosis in vitro and under development for in vivo imaging as well [3,7]. However, the physiological concentration of Ca 2+ is lower than that required for optimal binding of native ANV to PS [8], rendering a suboptimal binding condition for apoptosis detection in vivo. In addition, excessive uptake by the liver and the kidney further limits the application of native ANV for in vivo apoptosis imaging [9]. In light of these drawbacks of native ANV concerning the in vivo detection of apoptosis, the development of derivatives of ANV with an improved binding profile to PS and pharmacokinetic properties has been under intensive investigation recently.
ANV possesses anticoagulant and antithrombotic activity by forming 2-dimensional arrays on anionic membrane surfaces, and thus making the anionic phospholipids unavailable for assembly of coagulation enzyme complexes [10]. In attempts to developing potent thrombogenesis inhibitors, a series of recombinant anticoagulant fusion proteins, consisting of an ANV moiety and a Kunitz protease inhibitor (KPI) domain that binds to various coagulation factors with high affinity and specificity, were reported recently [11]. One of these constructs, ANV-6L15, was found to possess a higher binding affinity to PS compared to native ANV at physiological Ca 2+ concentrations; the apparent dissociation constant for ANV-6L15 binding to erythrocyte ghosts was approximately 4-fold lower than that of ANV at 1.2-2.5 mM Ca 2+ [12]. The previous study reveals that ANV-6L15 may provide improved detection of PS exposed on the membrane surfaces of pathological cells in vitro and in vivo. The present study investigates the use of radioiodinated ANV-6L15 as an imaging agent for apoptosis detection.

Materials and
Reagents. 123 I-NH 4 I was produced by a compact cyclotron (Ebco TR30/15, Vancouver, Canada) at the Institute of Nuclear Energy Research (INER, Longtan, Taoyuan, Taiwan). 125 I-NaI was supplied by IZO-TOP Institute of Isotopes (Budapest, Hungary). Iodogen (1,3,4,6-tetrachloro-3,6-diphenylglycoluril) precoated iodination tubes were purchased from Pierce Biotechnology (Rockford, IL, USA). ANV-6L15 was produced by expression in Escherichia coli as described previously [9]. ANV and FITC-ANV apoptosis detection kit were obtained from Becton Dickinson (Franklin Lakes, NJ, USA). Nanosep 10 K Centrifugal Devices were purchased from Pall Life Science (Ann Arbor, MI, USA). Stabilized 5C Cell Control was a generous gift from Beckman Coulter (Fullerton, CA, USA). Camptothecin (CPT) was purchased from Sigma (Cambridge, MA, USA). Acute lymphoblastic cell line of human Jurkat T-cell was obtained from Bioresource Collection and Research Center (Hsinchu, Taiwan). Cell culture materials were obtained from Gibco BRL (Grand Island, NY, USA). All other chemicals were purchased from Merck (Darmstadt, Germany).

Radioiodination of ANV and ANV-6L15.
The Iodogen method [13,14] was adopted for radiolabeling of ANV and ANV-6L15 with 123 I or 125 I. Briefly, ten μg of ANV and ANV-6L15 were dissolved in 15 μL of 0.1 M KH 2 PO 4 (pH 8) solution and reacted with radioiodide (370 MBq 123 I-NH 4 I or 185 MBq 125 I-NaI) in a tube precoated with 50 μg of Iodogen. After gentle agitation of the tube at room temperature for 10 min, the reaction mixture and the rinse of the tube (with 400 μL 0.01 M KH 2 PO 4 solution, pH 7.4) were loaded on a Nanosep 10 K Centrifugal Device prewashed with 400 μL 0.01 M KH 2 PO 4 solution (pH 7.4). After centrifugation of the Nanosep tube at 15,000×g for 5 min and washing twice with 400 μL 0.01 M KH 2 PO 4 (pH 7.4) solution, the radioiodinated protein on the membrane of Nanosep tube was harvested.

Radiochemical Analysis by High-Pressure Liquid
Chromatography. The radiochemical purities of radioiodinated ANV and ANV-6L15 were determined on an analytical HPLC system (Waters 600E, Milford, MA, USA) equipped with a size-exclusion column (Waters Ultrahydrogel 250 column, 7.8 × 300 mm, 6 μm) and a Waters Ultrahydrogel guard column (6.0 mm × 40 mm, 6 μm). After loading 10 μL of diluted radiolabeled protein, the column was eluted with 0.01 M KH 2 PO 4 solution (pH 7.4) at a flow rate of 0.8 mL/min. The radioactivity of the eluate was monitored online by a flow detector (FC-1000, Bioscan, Washington, DC, USA).

Erythrocyte Ghost Binding
Assay. Stabilized erythrocyte ghosts were prepared by hypotonic treatment of 5C Cell Control (Beckman Coulter) to expose more PS sites. The cell control erythrocytes were incubated alternatively with water and 10 mM HEPES buffer (containing 137 mM NaCl, 4 mM KCl, 0.5 mM MgCl 2 , 0.5 mM NaH 2 PO 4 , 0.1% D-glucose, and 0.1% BSA; pH 7.4) twice at 37 • C for 45 min. Following centrifugation at 16,000×g for 5 min, the pellet of stabilized erythrocytes was resuspended in 10 mM HEPES buffer. For the binding studies, erythrocyte ghosts (1.1 × 10 9 cells/mL) were incubated with 125 I-ANV and 125 I-ANV-6L15, respectively, in binding buffers (10 mM HEPES buffer, pH 7.4, with 1.2-10 mM calcium chloride solution) at room temperature for 30 min. Bound and free radioiodinated proteins were then separated by centrifugation at 16,000×g for 10 min. The radioactivities in the pellet and the supernatants were measured using an automatic gamma counter (Wizard 1470, PerkinElmer Wallac, Turku, Finland), and the percentage of radioactivity bound to erythrocyte ghosts was calculated. Binding of 125 I-NaI to erythrocyte ghosts served as a negative control.

Apoptotic Cell Binding Assay. Apoptotic
Jurkat T-cells were prepared by treatment with an anticancer drug camptothecin (CPT; 8 μM) for 24 hours [15,16]. In brief, cell culture flasks with RPMI 1640 medium (5 mL containing For biodistribution study, 18 eight-week-old male BALB/c mice (obtained from the National Animal Center, Taipei, Taiwan) were injected intravenously via the lateral tail vein with 444 ± 37 kBq 123 I-ANV-6L15 and sacrificed at 2, 10, 30, 60, 120, and 180 min after injection (three mice for each time point) under isoflurane anesthesia. Blockade of specific uptake of free iodide was achieved by intraperitoneal injection of KI (3.9 mg per mouse) 30 min before radiotracer injection. At the selected time, main organs and tissues were obtained, weighed, and counted for radioactivity on a Wallac 1470 gamma counter. The distribution data were expressed as percentage of the injected dose per organ (%ID/organ) and per gram of tissue (%ID/g).
For in vivo imaging study, one BALB/c mouse was injected via tail vein with 18.5 MBq of 123 I-ANV-6L15 in physiological saline (100 μL). Under isoflurane anesthesia, whole body scans were acquired on an X-SPECT (Gamma Medica, Northridge, CA, USA) equipped with an HRES (high resolution electronic system) collimator 60-210 min after injection.

Statistical
Analysis. Data were expressed as mean ± standard deviation (SD). The binding of 123 I-ANV and 123 I-ANV-6L15 in CPT-treated Jurkat T-cells were compared using Student's t-test, with P < .05 indicating statistical significance.
Due to limited availability of 123 I, we used 123 I mainly for animal studies (biodistribution and SPECT imaging) and related Jurkat T-cell binding assay. To prevent frequent shortage of 123 I during the course of study, we used relatively longer half-lived and commercially available 125 I for in vitro study on erythrocyte ghost cell. We assume similar binding characteristics between radiolabels of 123 I and 125 I.

Erythrocyte Ghost Binding Assay.
To determine whether ANV-6L15 maintained its biological activity after radioiodination, erythrocyte ghost binding assay was performed. The results showed that the binding of 125 I-ANV or 125 I-ANV-6L15 to erythrocyte ghosts was time-and Ca 2+ -dependent (Figures 2(a) and 2(b)). Maximal binding of either 125 I-ANV or 125 I-ANV-6L15 to erythrocyte ghosts occurred after 30 min incubation in the presence of 2.5 mM Ca 2+ ; the binding ratio was 16.97 ± 0.66% and 35.07 ± 0.53% for 125 I-ANV and 125 I-ANV-6L15, respectively (Figure 2(a)). The binding of either radioiodinated protein to erythrocyte ghosts steeply increased as Ca 2+ concentration increased from 1.2 mM to 10 mM, and the binding ratio reached 77.3 ± 0.45% and 84.7 ± 0.26% for 125 I-ANV and 125 I-ANV-6L15, respectively, at 10 mM Ca 2+ (Figure 2(b)). 125 I-ANV-6L15 exhibited higher binding ratios than 125 I-ANV (Figures  2(a) and 2(b)). Thus, radioiodinated ANV-6L15 appeared to possess a higher affinity to membrane PS-binding sites compared to radioiodinated ANV.
Lahorte et al. [13] reported that the optimal incubation time for maximal platelet binding was 20 min or longer in the presence of 5 mM Ca 2+ and reached a plateau at 20 mM Ca 2+ when incubated for 30 min. Our data showed that the maximal binding of either 125 I-ANV or 125 I-ANV-6L15 to erythrocyte ghosts appeared at 30 min in the presence of 2.5 mM Ca 2+ and reached a plateau at 8 mM Ca 2+ when incubated for 30 min.
The mechanism for the increased binding affinity of ANV-6L15 for erythrocyte membranes compared with ANV is currently unknown. The coexpression of phosphatidylethanolamine (PE) on the erythrocyte membranes was proposed to contribute importantly to the increased affinity of ANV-6L15 compared with ANV [12]. Figure 3 showed a typical flow cytometric dot plot for the untreated and apoptotic CPT-treated Jurkat T-cells. An increased proportion of apoptotic (ANV + PI − ) and necrotic (ANV + PI + ) cells was observed in CPT-treated cells. The binding of 123 I-ANV-6L15 to CPT-treated cells (23.3 ± 2.1 cpm/10 4 cells) was significantly higher than those of untreated-(8.2 ± 0.4 cpm/10 4 cells) and vehicle-(DMSO-) treated Jurkat Tcells (P < .03; Figure 4). In contrast, the binding of 123 I-ANV (approximately 4 cpm/10 4 cells) or 123 I-iodide (<1 cpm/10 4 cells) was far below those of 123 I-ANV-6L15 among different groups, and no increase of 123 I-ANV binding to apoptotic Jurkat T-cells was observed (Figure 4). The lack of increased 123 I-ANV binding to apoptotic Jurkat T-cells was possibly due to relatively low binding affinity of 123 I-ANV for PSexposed membranes at the physiological concentration Ca 2+ (1.2 mM).

Apoptotic Cell Binding Assays.
Journal of Biomedicine and Biotechnology

Biodistribution and Imaging of Radioiodinated ANV-6L15 in Mice.
The biodistribution of 123 I-ANV-6L15 in BALB/c mice are depicted in Table 2. After bolus injection via tail vein (approximately 25 ng protein/mouse), 123 I-ANV-6L15 showed an initial rapid clearance from the blood (40.01 ± 6.25% and 12.29 ± 1.22% ID/organ at 2 min and 60 min after injection, resp.). The liver had the highest radioactivity (11.40 ± 1.34% ID/organ) at 2 min after injection. The organ with the second highest radioactivity was the kidneys (8.61 ± 0.35% ID/organ at 10 min after injection). 123 I-ANV-6L15 fusion protein did not cross the blood-brain barrier as indicated by the low brain uptake (<0.5% ID/g in average). Low uptake in the thyroid reflected low level of radioiodide dissociation from the radiolabeled protein in vivo. The slight increase of thyroid uptake at 120 and 180 min could be due to in vivo 123 I dissociation from the radiolabeled protein.
As reported previously [3,14], 123 I-ANV was rapidly cleared from the blood following a biexponential decay and predominant uptake in the kidneys, liver, and gastrointestinal tract. The result of this study revealed that 123 I-ANV-6L15 also excreted via kidneys.
In vivo imaging of 123 I-ANV-6L15 distribution in a BALB/c mouse after bolus injection via tail vein (approximately 1 μg protein/mouse) showed high uptake of the tracer in the liver and the kidneys ( Figure 5). The uptake of 123 I-ANV-6L15 in these organs may limit the evaluation of apoptosis in abdominal regions. Further studies of biodistribution and SPECT imaging will be performed at later time points. We will conduct dosimetry calculation for 123 I-ANV-6L15 and compare with 123 I-ANV. The novel tracer will be evaluated in animal models with stress-induced apoptosis in the future. 6 Journal of Biomedicine and Biotechnology Journal of Biomedicine and Biotechnology

Conclusions
ANV-6L15 was successfully labeled with radioiodine using Iodogen method. Radioiodinated ANV-6L15 showed significantly higher binding to PS-exposed erythrocyte ghosts and camptothecin-induced apoptotic Jurkat T-cells at physiological concentration of Ca 2+ compared to that of radioiodinated ANV in vitro. Biodistribution study showed that 123 I-ANV-6L15 was rapidly cleared from the blood. Further imaging studies in animal models of apoptosis are warranted. Owing to higher binding affinity to PS, radioiodinated ANV-6L15 could be more sensitive than radioiodinated ANV to detect the apoptosis-associated treatment and human disorders, such as radiation/chemotherapy efficacy, myocardial ischemia or infarct, infectious diseases, and neurodegenerative diseases. Taken together, these results suggest that the radioiodinated ANV-6L15 may be a better scintigraphic tracer for apoptosis detection compared with ANV.