18F-PEG1-Vinyl Sulfone-Labeled Red Blood Cells as Positron Emission Tomography Agent to Image Intra-Abdominal Bleeding

18F-Labeled blood pool agents (BPAs) have attracted great attention for identifying bleeding sites. However, many BPAs are not sufficiently evaluated partially due to the limitations of labeling methods. In our previous work, we noticed that 18F-PEG1-vinyl sulfone (18F-VS) could efficiently label red blood cells (RBCs) ex vivo and in situ. However, its application as BPA is not fully evaluated. In this study, we systematically explored the feasibility of using 18F-VS-labeled RBCs as a positron emission tomography (PET) BPA for intra-abdominal bleeding diagnosis. In brief, we first optimized the labeling conditions, which lead to an 80% labeling yield of RBCs after incubating with 18F-VS in phosphate-buffered saline (PBS) at 37°C for 20 min. 18F-VS-labeled RBCs were found to be stable in vitro, which could simplify its transportation/storage for in vivo applications. In normal rat PET study, the cardiovascular system could be clearly imaged up to 5 h post injection (p.i.). An intra-abdominal hemorrhage rat model demonstrated that the 18F-VS-labeled RBCs clearly showed the dynamic changes of extravascular radioactivity due to intra-abdominal hemorrhage. Validation in the model of gastrointestinal bleeding clearly demonstrated the great potential of using 18F-VS-labeled RBCs as a BPA, which could be further evaluated in future studies.


INTRODUCTION
Blood pool imaging (BPI) is widely used in preclinical and clinical research including the detection of gastrointestinal bleeding (1), blood volume measurement (2), evaluation of cardiac function (3,4), localization of hemangiomas (5), cerebral blood flow (6,7), detection of infection (8), or lymphoma (9). Various kinds of BPI agents have been developed, including 18 F-FDG-, 111 In-oxine (8), or 99m Tc-HMPAO-labeled leukocytes (10); 68 Ga-NOTA-NEB- (9) and 111 In-oxine-labeled platelets (11); 99m Tc-PYP-labeled red blood cells (RBCs) (12); radionuclide-labeled peptides (13); and magnetic resonance angiography-based BPI agents (14). RBCs represent promising BPI agents due to their good stability and easy availability. In fact, radionuclide-labeled RBCs have been used to obtain functional information of the cardiovascular system through quantitative analysis of BPI. During the past few decades, radionuclide-labeled RBCs [such as 51 Cr-RBCs (15), 99m Tc-PYP-RBCs (16), and 68 Gaoxine-RBCs (17)] have shown great progress in preclinical and clinical research applications. They have been used for cardiac function evaluation, diagnosis of hemangioma and digestive tract bleeding, cerebral blood flow measurement, and spleen imaging and positioning. Despite the progress, limitations of these BPIagents include 1) unstable labeling yield: for example, the labeling rate of 99m Tc-PYP-RBCs was greatly affected by the type and dose of the drug applied (18) and 2) the resolution of single-photon emission nuclide 99m Tc often resulted in reduced image quality compared with the positron emission tomography (PET) nuclide (19). Therefore, researchers have been trying to develop BPI agents based on PET nuclides. 11 C, 13 N, and 68 Ga have been used to label RBCs, but their short half-lives or image resolution still limited their widespread clinical application to a certain extent (20,21).
The positron nuclide 18 F has the advantages of good image quality and suitable half-life (109.8 min) for commercialization and transportation compared with other positron nuclides. Moreover, the resulting carbon-fluorine bonds generally have reasonable stability, which could be advantageous for BPI (22). However, there are only a few 18 F-labeled RBCs reported as BPI agents partially due to the limitation of the labeling method (23). Therefore, there is a need to find new methods that could lead to easily prepared 18 F-labeled BPI agents for preclinical and clinical diagnosis applications.
Previously, we established a new method for site-specific labeling of thiol groups based on 18 F-labeled vinyl sulfone ( 18 F-VS). Both peptides and proteins were found to react with 18 F-VS through the Michael addition in an aqueous system (24). Moreover, the resulting conjugates are stable in the aqueous solution and would not be hydrolyzed in a neutral solution like maleimide conjugates (25). Recently, it was also found that 18 F-VS could react with amino groups in addition to thiol groups even though the reaction is slower (26). Interestingly, we observed that 18 F-VS could efficiently label RBCs in vitro and in vivo. Despite the observation, the conditions to label RBCs were not optimized, and its application as BPI was not studied.
In this study, we evaluate the use of 18 F-((2-(2fluoroethoxy)ethyl)sulfonyl)ethene ( 18 F-PEG1-vinyl sulfone [ 18 F-VS]) for RBC labeling, which are then applied as a new BPI agent for abdominal hemorrhage imaging in animal models.

MATERIALS AND METHODS
All chemicals involved in the synthesis were of reagent grade and purchased from Aladdin Bio-Chem Technology (Shanghai, China) or Sigma. The radiochemical purity was documented by high-performance liquid chromatography (LC-16). A gamma counter (CAPRAC-t, Huaruisen Technology Development Co., Ltd., Beijing, China) and a dose calibrator (CRC-15R, Capintec Inc., Florham Park, NJ) were used to measure the radioactivity of the samples. Mouse and rat data acquisitions were performed with a micro-PET/computed tomography (CT) scanner (Inveon, Siemens, Munich, Germany). Healthy Kunming mice (20 g ± 2 g) and Sprague-Dawley (SD) rats (120 g ± 12 g) were provided by the Animal Experimental Center of Southwestern Medical University (Animal License SCXK 2018- 17), and all studies were approved by the Ethics Committee of Southwest Medical University. The precursor 2-(2-(vinyl sulfonyl)ethoxy)ethyl 4nitrobenzene sulfonate was synthesized and characterized using the method described in the literature (24).

RBCs Preparation
The rats were anesthetized with isoflurane, and 4 ml of blood was collected from the heart using an injection needle. Heparin (1,000 IU/kg body weight) was used to prevent blood clotting. After centrifugation (400 g for 10 min at 20 • C), RBCs were located at the bottom of the tube. The plasma was separated from RBCs and stored for subsequent post-labeling stability studies. The buffy coating, which contains most of white blood cells and platelets, was removed, leaving the RBC layer undisturbed.

Optimization of Labeling Conditions
The separated RBCs and 18 F-VS solution (74 MBq, 600 µl) were mixed thoroughly and then divided into eight tubes and incubated separately at 0 • C and 37 • C (n = 4 in a group). The incubation time varied from 0 to 60 min, with 10-min intervals. At the corresponding incubation timepoint, 20 µl of the corresponding suspension was removed and mixed with an additional 100 µl PBS solution. Then, the mixture was subjected to centrifugation. Radioactivity of the supernatant and that of the erythrocyte sediment were measured separately. Optimal conditions for labeling RBCs with 18 F-VS were determined and used thereafter for further evaluation.

Post-Labeling Stability of 18 F-VS-RBCs
Post-labeling stability was evaluated by calculating the radioactivity released from 18 F-VS-RBCs. The optimized labeling condition was detailed below: the centrifugal-washed RBC suspension (3 ml) and 18 F-VS solution (74 MBq in 600 µl) were mixed thoroughly and incubated at 37 • C for 30 min. Then 18 F-VS-RBCs were washed three times with PBS (8 ml), and 18 F-VS-RBCs was isolated. To the purified 18 F-VS-RBCs, plasma solution was added and mixed. 18 F-VS-RBC suspension was divided into two tubes and incubated for 0 to 180 min (30-min intervals) at 0 • and 37 • C, respectively. After the corresponding incubation time, 18 F-VS-RBC suspension was cooled and uniformly resuspended. Then, the radioactivity of 18 F-VS-RBC suspension (10 µl) was measured (n = 4). The remaining 18 F-VS-RBC suspension was centrifuged at 450 g for 2 min, and the supernatant (10 µl) was sampled (n = 4). Radioactivity counts of the RBC suspension and supernatant were simultaneously measured. The release fraction was calculated according to the following formula: release fraction (%) = (radioactivity of the supernatant/radioactivity of initial RBC pellet) × 100%. The supernatant at 30 and 120 min timepoints were also analyzed by radio-HPLC. For the sample at the 120 min timepoint, elution from HPLC was collected per minute and counted by a gamma counter due to the low radioactivity.

Incubation of 19 F-VS With RBCs
19 F-VS was prepared with the previously reported method (26), which was then added to 5 µl RBCs in 2 ml saline to form a final concentration of 19 F-VS at 10 and 100 µM. No 19 F-VS was added in the blank control. All RBCs in the Petri dishes were incubated at 4 • C, and the shape of RBCs was observed at the 2, 6, 12, and 24 h timepoints.

PET Imaging of Normal Rats
Normal SD rats (120 ± 12 g) were used in this study. All rats were anesthetized by inhalation of isoflurane, which was maintained throughout the imaging procedure. The imaging study was performed using a small-animal PET system. SD rats were placed on a fixed plate in the supine position for scanning. 18 F-VS-RBCs (24.8 ± 2.6 MBq and 400 µl) was injected through the tail vein. At the same time, images were continuously acquired for 60 min using the list mode. The list mode data were reconstructed using a dynamic sequence (30 frames, 60 s). After reconstruction, regions of interest (ROIs, mm 3 ) of the heart, blood vessels, and spleen were obtained using the software provided by the supplier (Inveon Research Workplace 4.2, Siemens). The values were presented as dose per gram of organ (% ID/g).

Imaging Study of the Rat Intra-Abdominal Hemorrhage Model
A glycerin enema was injected into the colon through the anus of rats to promote defecation about 30 min before the image acquisition. Then, 22.5 MBq of tracer ( 18 F-VS-RBCs with 1,000 IU/kg body weight heparin for anticoagulation) was injected through the tail vein and the PET/CT dynamic acquisition was started simultaneously. List mode data were acquired for 60 min. Under steady-state BPI conditions (10-15 min after injection), a 12-gauge lumbar puncture needle was used to manually puncture the colon wall through the anus to cause abdominal bleeding.

Image Analysis
For the abdominal hemorrhage model, the data acquisition and reconstruction were performed using the procedures described above. The ROI was drawn in the corresponding bleeding area. Radioactivity was presented using %ID/g, which was then then used to obtain the corresponding time-activity curve. The post-bleeding image (58-60 min after injection) and prebleeding image (10-20 min after injection) were subtracted with the PMOD software (Zurich, Switzerland) to measure the radioactivity of the bleeding site in the abdominal hemorrhage model. Negative values of the pixels in the subtracted image were replaced with zero values. Then, the total radioactivity of the abdominal hemorrhage image was presented as the percentage of the injected dose after excluding bladder radioactivity.

Statistical Analysis
Quantitative data were presented as mean ± standard deviation. Statistical analyses were performed using the SPSS Statistics 20.0 software package (IBM, Chicago, IL). The significance level was set to 0.05.

RESULTS AND DISCUSSION 18 F-VS Preparation
Similar to previous reports, the 18 F-VS was obtained through a nucleophile substitution of VS-ONs with [ 18 F]TBAF. The reaction mixture was purified using HPLC (Supplementary Figure 1), and the resulting 18 F-VS was obtained in 25-40% yield with ≥99% radiochemical purity and a retention time of 12.2 min (Supplementary Figure 2).

Evaluation Labeling Efficiency of 18 F-VS-RBCs
As a new BPI agent, it is important to optimize the labeling conditions to maximize the yield of 18 F-VS-RBCs. It is also important to understand the release profile from RBCs under different conditions (Figure 1).

Evaluation in vitro Stability of 18 F-VS-RBCs
To evaluate the in vitro stability of 18 F-VS-RBCs, a post-labeling stability experiment involving different storage temperatures was performed (Figure 2). In this study, the released 18 Fcontaining fraction from 18 F-VS-RBCs was evaluated for 180 min. Samples were incubated at 37 • and 0 • C, respectively, to simulate the temperature of the human body and the potential transportation/storage conditions using ice packs. 18 F-VS-RBC suspension (10 µl) was then taken for analysis at different incubation timepoints.
Overall, the released fractions of 18 F-VS-RBCs in the in vitro study were rather low. Thus, we concluded that 18 F-VS-RBCs were relatively stable in vitro. Moreover, HPLC analysis of the supernatant indicated that the released fraction only had a small amount of 18 F-VS in addition to some unknown radioactive fractions (Supplementary Figure 3). This observation indicated that 18 F-VS likely reacted with RBC through a covalent bond instead of passive absorption. As the released radioactive fraction may be excreted in the urine or absorbed by extravascular tissue after injection, imaging should be performed at an early timepoint if possible.

Evaluate the Toxicity of 19 F-VS on RBCs
To evaluate the toxicity of the tracer, the 19 F-VS was prepared and incubated with RBCs at 4 • C in saline. The shape of RBCs was observed, and images of RBCs were recorded at each timepoint (Supplementary Figure 5). As shown in Supplementary Figure 5, the shape of RBCs stayed complete, and there was no obvious broken RBCs observed, indicating that 18 F-VS has no apparent toxicity on RBCs.

Evaluation of 18 F-VS-RBCs for PET Imaging
Under optimal conditions, the average time from drawing blood to intravenously injecting 18 F-VS-RBCs was ∼60 min. Microscopic examination showed that the morphology of 18 F-VS-labeled RBCs was normal without abnormal aggregation. The final 18 F-VS-RBC suspension (400 µl) had a radioactivity of 25.9 MBq (±3.7 MBq) and an LE of 70.09% (±0.61%) (n = 4).

PET Imaging of Normal Rats
In order to evaluate the distribution of 18 F-VS-RBCs in the cardiovascular system and the changes of radioactivity in the blood pool, we performed 18 F-VS-RBC imaging in normal rats. The maximum-intensity projection image and biodistribution of 18 F-VS-RBCs within 60 min after injection are shown in Figure 3 and Table 1, respectively. The cardiovascular system in normal rats had a strong uptake of 18 F-VS-RBCs. A high uptake was obtained within 20 min, which was maintained stably at late time-points. Compared with that at 10 min, cardiac radioactivity at 60 min only decreased by ∼0.7%. The radioactivity of most organs remained relatively constant within 60 min. The results indicated that 18 F-VS-RBCs had a good stability in vivo. Within 60 min after tracer injection, atrium and spleen radioactivity was higher than that in the liver and lung. Urine excretion was observed. The PET imaging of the cardiovascular system in rats was clearly visualized with a low background ratio using 18 F-VS-RBCs. As shown in Figure 3, 18 F-VS-RBCs mainly stayed in the blood pool, suggesting good stability in vivo. At a late timepoint, urine activity was observed. As shown in Supplementary Figure 3, a small percentage of activity was released to the supernatant, which contains a hydrophilic motif. This may lead to the observed urine activity. Nonetheless, additional characterization would be done in a future study to further confirm it. Overall, 18 F-VS-RBCs hold a great potential to imaging the cardiovascular system considering its slow clearance and stability profile.

Imaging Study of the Intra-Abdominal Hemorrhage Model on Rats
A dynamic PET scan was performed on rats to evaluate the feasibility of using this BPI agent for the diagnosis of intraabdominal hemorrhage. Dynamic PET imaging of the rat gastrointestinal bleeding model showed that shortly after the colon wall puncture, a high aggregation site appeared in the abdomen, indicating the tracer had extravasated due to bleeding (Figures 4A-C). Time-activity curves showed that the agent increased steadily at the bleeding site: radioactivity was stable within 10 min and continued to increase thereafter ( Figure 4D). 18 F-VS-labeled RBCs successfully found the location and direction of abdominal bleeding over time. Although this study did not quantify the amount of bleeding, it can provide a rough estimation based on the blood radioactivity curve. Clearly, PET imaging of the rat intra-abdominal hemorrhage model demonstrated that 18 F-VS-RBCs hold a great potential for applications in gastrointestinal bleeding. Furthermore, without dietary restrictions, the labeling procedure of 18 F-VS-RBCs is simpler and shorter than the reported 18 F-FDG-RBCs. We would also like to point out that the potential limitations of 18 F-VS-RBCs include the requirement of clean space due to blood collection and in vitro labeling procedures. The radiation exposure toward operators could be high. Further improvements may focus on simplifying the labeling step by reducing the number of washing and streamlining the production process. Nonetheless, 18 F-VS-RBC PET of normal animal and intra-abdominal hemorrhage model suggested that the agent could represent a new BPI agent.

CONCLUSION
In this study, 18 F-VS was synthesized and successfully labeled RBCs for BPI. The resulting 18 F-VS-RBCs clearly visualized the cardiovascular system and extravascular blood in the abdominal hemorrhage model. Compared with existing cardiac blood pool agents, 18 F-VS-RBCs could be prepared using a readily available precursor and simple procedure. The agent has prolonged retention time in cardiac blood pool with a high-quality image. The agent's uptake was not affected by blood sugar, which eliminated the need of fasting. Because 18 F-VS-RBCs cannot penetrate the blood-brain barrier, it would be interesting to test its application in detecting bleeding in the brain in future studies.

DATA AVAILABILITY STATEMENT
The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

ETHICS STATEMENT
The animal study was reviewed and approved by The Affiliated Hospital of Southwest Medical University.

AUTHOR CONTRIBUTIONS
XZ and LW contributed to the study design, the labeling process, the imaging scan, the data analysis and they contributed equally to this paper. XZ wrote the manuscript and LW revised the manuscript. WF, YF, CZ, TX, and JC were responsible for the integrity of the data and the accuracy of the data analysis. TZ and LZ contributed to precursor synthesis, the labeling and evaluation of the tracer stability during the revision of the work. The YC and ZL was responsible for revising for important intellectual content. All authors have read and approved the final manuscript.

FUNDING
This study was funded by the Nuclear Medicine Innovation Transformation Platform (2019JDPT0004) and the Science and Technology Project of Sichuan Province (2018JY0230).