Abstract
The use of radiolabeled leukocytes is considered the gold standard for scintigraphic imaging of inflammatory bowel disease. The disadvantages of 99mTc-hexamethylpropyleneamine oxime (HMPAO)-leukocytes, however, encourage the search for new imaging agents with at least similar diagnostic accuracy but without the laborious preparation and subsequent risk of contamination. In this study we investigated the imaging characteristics of a new imaging agent that specifically binds to the leukotriene B4 (LTB4) receptors expressed on neutrophils. Imaging characteristics of the 111In-labeled LTB4 antagonist (DPC11870) were compared with those of 18F-FDG and 99mTc-HMPAO-granulocytes in a rabbit model of experimental colitis. Methods: Acute colitis was induced in New Zealand White (NZW) rabbits by infusion of trinitrobenzene sulfonic acid in the descending colon. Forty-eight hours after induction of colitis, all animals were injected intravenously with 99mTc-granulocytes, 18F-FDG, or 111In-DPC11870. The pharmacokinetics and biodistribution were studied by serial scintigraphic imaging and by ex vivo counting of dissected tissues. Results: All 3 radiopharmaceuticals showed the inflamed colon as early as 1 h after injection. However, compared with 99mTc-granulocytes, both 111In-DPC11870 and 18F-FDG were superior in revealing the inflamed lesions. The biodistribution data showed that uptake of 111In-DPC11870 in the inflamed colon was highest (0.72 ± 0.18 percentage injected dose per gram [%ID/g]), followed by uptake of 99mTc-granulocytes (0.40 ± 0.11 %ID/g) and of 18F-FDG (0.16 ± 0.04 %ID/g). Because of low activity concentrations in the noninflamed colon, the radiolabeled LTB4 antagonist also revealed the highest ratio of affected colon to unaffected colon (11.6 for 111In-DPC11870, 5.5 for 99mTc-granulocytes, and 4.1 for 18F-FDG). Conclusion: The radiolabeled LTB4 antagonist DPC11870 clearly delineated acute colitis lesions in NZW rabbits within 1 h after injection. Because of high uptake in the inflamed lesions and a low activity concentration in the noninflamed colon, images acquired with 111In-DPC11870 were better than those acquired with 99mTc-granulocytes or 18F-FDG.
Several nuclear medicine imaging techniques are currently used as tools for diagnosis of inflammatory and infectious diseases. In general, scintigraphic imaging of inflammatory and infectious foci is performed using radiolabeled leukocytes. Especially in cases of inflammatory bowel disease, radiolabeled leukocytes are considered the best of the available scintigraphic imaging agents. 99mTc-Hexamethylpropyleneamine oxime (HMPAO)-leukocytes are a useful radiopharmaceutical for the management of patients with Crohn’s disease. Uptake of radioactivity has shown a correlation with endoscopy and histology (1). Despite the fact that 111In- or 99mTc-labeled leukocytes are considered adequate imaging agents, their laborious preparation and their requirement that potentially contaminated blood be handled has stimulated a search for an alternative radiopharmaceutical comprising the same imaging qualities.
Recent reports suggest that 18F-FDG may be such an alternative (2,3). 18F-FDG PET is a distinguished imaging tool in clinical oncology because of its ability to image the increased glucose uptake of tumor cells. Several studies have shown that increased glucose metabolism is not restricted to malignant cells (4). Increased 18F-FDG accumulation also occurs in inflammatory cells (5,6).
Besides 18F-FDG, chemokine receptor-binding agents may be used for the imaging of infection and inflammation (7,8). The use of radiolabeled chemotactic peptides (e.g., N-formyl-methionyl-leucyl-phenylalanine and chemokines) is based on the high-affinity interaction with their receptors specifically expressed on (activated) white blood cells. Because many of these leukocytes infiltrate at the site of inflammation, interaction with these receptors in vivo may lead to accumulation of the radiolabeled agent in the infectious or inflammatory foci. Compared with the radiopharmaceuticals currently used to image infectious and inflammatory lesions, these agents have additional advantages warranting their further investigation, including low molecular weight—leading to rapid penetration in the focus and fast clearance from the blood—and absence of the potential hazard associated with handling blood products (9).
Leukotriene B4 (LTB4) is a chemotactic molecule that could be studied for this purpose. LTB4 activates granulocytes and macrophages in reaction to an inflammatory response (10,11). Two types of LTB4 receptors have been identified (12,13). The first receptor type, BLT1, is a high-affinity receptor expressed mainly on human neutrophils. The second receptor type, BLT2, is a low-affinity receptor expressed more ubiquitously, with high expression in spleen, leukocytes, ovary, and liver (10,14). The 99mTc-labeled LTB4 antagonist RP517 was recently reported to bind to the neutrophil LTB4 receptor and was also found capable of imaging myocardial inflammation caused by coronary artery occlusion and reperfusion in a dog model (15). We found that RP517 can rapidly reveal intramuscular Escherichia coli abscesses in rabbits (16). The high lipophilicity of RP517 results in predominantly hepatobiliary excretion, making imaging of infectious and inflammatory foci in the abdomen difficult.
The LTB4 antagonist used in this study was DPC11870. DPC11870 is structurally related to RP517 but is made less lipophilic by the addition of pharmacokinetic modifying groups. We found that DPC11870 binds to human granulocytes with high affinity in vitro. Therefore, we hypothesized that in vivo interaction with granulocytes at the site of inflammation, combined with reduced uptake in the gastrointestinal tract, can lead to visualization of the inflamed colon. In this study we investigated the imaging characteristics of the 111In-labeled LTB4 antagonist DPC11870, 18F-FDG, and 99mTc-leukocytes in a rabbit model of acute, chemically induced colitis.
MATERIALS AND METHODS
DPC11870
DPC11870 is a bivalent LTB4 antagonist consisting of 2 identical LTB4 receptor-binding moieties, conjugated with DTPA to allow radiolabeling with 111In. The labeling of 20 μg DPC11870 and 75 MBq 111In was performed in metal-free 0.25 mol/L ammonium acetate buffer, pH 5.5, during 30 min at room temperature (17). Radiochemical purity was checked by instant thin-layer chromatography on silica gel strips (Gelman Sciences, Inc.) in 0.1 mol/L sodium citrate buffer, pH 6.0, and by reversed-phase high-performance liquid chromatography (model 1100 system; Agilent Technologies S.A./N.V.) on a C18 column (Zorbax Rx-C18, 4.6 mm × 25 cm; Agilent), using a linear gradient from 100% NH4Ac buffer, pH 7.0, to 100% acetonitrile in 50 min, at a flow rate of 1 mL/min. The specific activity of the 111In-DPC11870 was 3.7 MBq/μg (12 MBq/nmol), and the chemical purity exceeded 95%.
Radiolabeling of Granulocytes
Because radiolabeled leukocytes obtained from infected donors showed better imaging characteristics than did leukocytes obtained from healthy donors (18), blood was obtained from a donor rabbit in which acute colitis had been induced 48 h before. The collection of 50 mL of blood and the purification and labeling of granulocytes with 185 MBq of 99mTc-HMPAO was performed as described previously (19). Labeling efficiency exceeded 80%.
18F-FDG
18FDG was commercially obtained (DRN 9957; Tyco Healthcare). The activity concentration was 1.22 GBq/11.3 mL at calibration time.
Animal Model
Fifteen female New Zealand White (NZW) rabbits weighing 2.3–2.8 kg were kept in cages (1 rabbit per cage) and fed standard laboratory chow and water ad libitum. Thirty minutes before the induction of colitis and during the imaging experiment, the animals were sedated with a subcutaneous injection of 0.7 mL Hypnorm (fentanyl, 0.315 mg/mL, and fluanisone, 10 mg/mL; Janssen Pharmaceutical). Colitis was induced as described previously, with minor modifications (20,21). A flexible silicone tube (2.6-mm diameter, 40-cm length) was inserted into the colon, with the tip placed 20 cm from the anal sphincter. Through the tube, 1 mL of 50% ethanol, followed by 1 mL of ethanol containing 25 mg of trinitrobenzene sulfonic acid (Sigma Chemicals) and 3 mL of ethanol, was injected into the colon. After the induction, animals had free access to the normal amount of food and water. All animal experiments were approved by the local Animal Welfare Committee in accordance with the Dutch legislation and were performed in accordance with their guidelines.
Imaging and Biodistribution
Forty-eight hours after the induction of colitis, the animals received an intravenous injection of 11 MBq of 111In-DPC11870, 18.5 MBq of 99mTc-HMPAO-granulocytes, or 30 MBq of 18F-FDG. For comparison, a healthy rabbit also was injected with 11 MBq 111In-DPC11870. For scintigraphic imaging of 111In-DPC11870 and 99mTc-granulocytes, the rabbits were immobilized in a mold and placed prone on a γ-camera (Orbiter; Siemens) using a low-energy (99mTc-granulocytes) or medium-energy (111In-DPC11870) parallel-hole collimator. Images of 300,000 counts each were obtained at various times after injection and stored in a 256 × 256 matrix. All images were windowed identically, allowing a fair comparison among the experiments. For PET, the animals were imaged for 16 min starting 1 h after injection, using a rotated half-ring dedicated PET scanner (ECAT ART; Siemens/CTI) and a scan distance of 28.35 cm (38% transmission). The animals were euthanized after the last image was acquired. A blood sample was drawn by cardiac puncture. Tissues were dissected and weighed. The amount of radioactivity in the tissues was measured in a shielded well-type γ-counter (Wizard; Canberra Packard) together with the injection standards and was expressed as percentage injected dose per gram (%ID/g). The ratio of affected colon to unaffected colon was calculated for each inflamed colonic segment, using uptake of the compound in an unaffected segment at the proximal end as the reference value. The mean and SD of the individual ratios were calculated for each radiolabeled compound. The time at which biodistribution was performed was considered the optimal time point for each radiolabeled compound. Imaging beyond 1 h after injection of 18F-FDG is hardly feasible because of its short biologic and physical half-life. In the case of 99mTc-HMPAO-granulocytes, leakage of 99mTc-HMPAO and subsequent excretion to the gallbladder and intestine restricts the maximum imaging and evaluation time (19,22).
Statistical Analysis
All mean values are presented as mean ± SD. Statistical analysis was performed using ANOVA. The level of significance was set at 0.05.
RESULTS
Two days after the induction of colitis, all animals had diarrhea. The rabbits ate less food, but their water intake and behavior were normal. Scintigraphic images acquired at several different times after injection are shown in Figure 1. Each of the 3 radiopharmaceuticals showed the inflamed colon at 1 h after injection. Uptake of radiolabeled granulocytes at the site of the inflamed colon was moderate, especially when compared with uptake of 111In-DPC11870 at the same time points. Both 18F-FDG and 111In-DPC11870 clearly delineated the colitis and showed low uptake in noninflamed colon; however, 111In-DPC11870 revealed the inflamed lesions better than did 18F-FDG. The images acquired after injection of 111In-DPC11870 in the healthy animal showed no accumulation of radioactivity in the abdominal area at any time point.
Anterior scintigraphic images of rabbits that were placed prone on γ-camera. (A) Images acquired 0, 1, 2, and 3 h after injection of 18.5 MBq 99mTc-granulocytes in rabbit with acute colitis. (B) Images acquired 0, 1, 2, 4, and 6 h after injection of 11 MBq 111In-DPC11870 in rabbit with acute colitis. (C) 3-Dimensional representation of image acquired 1 h after injection of 30 MBq 18F-FDG in rabbit with acute colitis. (D) Images acquired 0, 1, 2, 4, and 6 h after injection of 11 MBq 111In-DPC11870 in healthy rabbit.
The general in vivo behavior of the 3 radiotracers was very different. Immediately after injection, 99mTc-granulocytes showed high uptake in lung and liver, and the activity concentration in both organs decreased with time. In all images, uptake of radioactivity was seen in the bone marrow. The pharmacokinetics of 111In-DPC11870 were different from those of 99mTc-granulocytes. A few minutes after injection, the radiolabeled LTB4 antagonist localized mainly in the circulation (heart), liver, and kidneys. At later times, radioactivity decreased in the circulation but accumulated in the bone marrow and inflamed colon. PET images acquired 1 h after 18F-FDG injection revealed a high radioactivity concentration in the kidneys. Before PET scanning, the bladder was emptied by catheterization. Figure 2 summarizes the biodistribution data of 18F-FDG at 1.5 h after injection, of 99mTc-granulocytes at 3 h after injection, and of 111In-DPC11870 at 6 h after injection. These times were considered optimal for the individual radiopharmaceuticals. Uptake in the inflamed colon was highest for 111In-DPC11870 and then for 99mTc-granulocytes (0.72 ± 0.18 %ID/g vs. 0.40 ± 0.11 %ID/g, respectively). Uptake of 18F-FDG in the inflamed colon was relatively low (0.16 ± 0.04 %ID/g). Furthermore, the total %ID in the inflamed colon was significantly higher for 111In-DPC11870 than for the other 2 radioactive tracers (P < 0.004), whereas there was no significant difference in colonic uptake between 99mTc-granulocytes and 18F-FDG (P > 0.6). Although the radioactivity concentration in the unaffected part of the colon was highest in rabbits injected with 111In-DPC11870, the ratio of affected colon to unaffected colon was nevertheless highest with this radiopharmaceutical (11.6 for 111In-DPC11870, 5.5 for 99mTc-granulocytes, and 4.1 for 18F-FDG; Fig. 3). After dissection of the tissues, activity in the colonic tissue and activity in the colonic content were measured separately. Each of the 3 radiolabeled compounds had low activity concentrations in the content of the colon, as shown in Figure 4. The radioactivity concentration in the colonic content did not significantly differ between the 3 agents (P > 0.6). For 99mTc-labeled granulocytes, the radioactivity concentration was high in spleen and bone marrow (0.98 ± 0.23 %ID/g and 1.06 ± 0.33 %ID/g, respectively). The 111In-DPC11870 biodistribution data revealed a similar high uptake in bone marrow (0.95 ± 0.17 %ID/g), whereas uptake in the spleen was considerably lower (0.45 ± 0.10 %ID/g). For this agent, considerable activity localized in the kidneys (0.33 ± 0.04 %ID/g). Biodistribution data from animals injected with 18F-FDG indicated that a large fraction of the injected dose had already been excreted from the body by 1 h after injection. Uptake of 18F-FDG in all organs (except inflamed colon) was very low, and activity concentrations in the bone marrow and spleen were low (0.12 ± 0.02 %ID/g and 0.11 ± 0.03 %ID/g, respectively).
Biodistribution data of rabbits that received intravenous injections of 111In-DPC11870, 99mTc-granulocytes, or 18F-FDG. Uptake in organs and tissues is expressed as %ID/g. Each bar represents mean ± SD. Ex vivo biodistribution was determined 6 h after injection for 111In-DPC11870, 3 h after injection for 99mTc-granulocytes, and 1 h after injection for 18F-FDG.
Mean ratios of affected colon to unaffected colon for 18F-FDG, 99mTc-granulocytes, and 111In-DPC11870 at 1, 3, and 6 h after injection, respectively. Ratios were derived from individual colonic segments taken at dissection.
Amount of activity (%ID) found in colonic wall and in content of affected colon for the 3 radiopharmaceuticals: 111In-DPC11870, 99mTc-granulocytes, and 18F-FDG.
DISCUSSION
The present study demonstrated that both 111In-DPC11870 and 18F-FDG are excellent agents for imaging experimental colitis in NZW rabbits. Both agents are superior to 99mTc-leukocytes, the standard imaging agent for scintigraphic evaluation of inflammatory bowel disease in humans. Images acquired after injection of 111In-DPC11870 showed the inflamed colon in as soon as 1 h. The activity concentration of radiolabeled DPC11870 in the colon at 6 h after injection was high, and visualization of inflamed lesions in the colon was better with 111In-DPC11870 than with 18F-FDG.
Numerous studies have focused on the role of leukotrienes and leukotriene antagonists as chemotactic compounds in the mediation of infection and inflammation (10,23). Because of the antiinflammatory effect of LTB4 antagonists, the mechanism of interaction and the receptor specificity of these compounds have been studied extensively (24). Yokomizo et al. (12,13) reported that 2 different LTB4 receptors could be identified. The high-affinity receptor BLT1 is expressed mainly on activated polymorphonuclear leukocytes at the site of infection or inflammation. Accumulation of 111In-DPC11870 in inflamed colon is most likely due to interaction with the BLT1 receptor expressed on these infiltrated cells. In vitro binding studies showed that DPC11870 binds to purified human granulocytes with high affinity (17).
Recently, several investigators reported the applicability of 18F-FDG for imaging of infection and inflammation. Animal models demonstrated that 18F-FDG accumulates in both acute and chronic infectious lesions (2,25). The accumulation of 18F-FDG in infectious and inflammatory lesions has been hypothesized to be a consequence of enhanced uptake of 18F-FDG in neutrophils, macrophages, causative microorganisms, or granulation tissue (5,26). Uptake of 18F-FDG in these cells is enhanced because of the increased glycolysis and the intensified stimulation of the hexose monophosphate shunt in these cells (27). In addition, activated inflammatory cells also have an increased expression of glucose transporters, and the affinity of glucose transporters (for deoxyglucose) is increased by several cytokines and growth factors (28,29). 18F-FDG PET for visualization of infectious and inflammatory foci is now a frequent subject of clinical studies. Hannah et al. (27) were among the first investigators to describe a patient with increased 18F-FDG uptake in the bowel, with the uptake probably being due to inflammation in the bowel wall. In a separate study of 5 patients diagnosed with colitis, Kresnik et al. (30) showed that 18F-FDG PET detected colitis at an early clinical stage.
The results of the present study on rabbits with acute colitis confirmed this increased 18F-FDG uptake in inflamed lesions in the colon. With 18F-FDG PET, the inflamed colon was rapidly revealed and, because of rapid and complete clearance of activity from the remainder of the body, was distinctly seen. Conventional 99mTc-granulocytes, a commonly used imaging agent in the management of patients with inflammatory bowel disease, showed that uptake of radioactivity in the colon was increased. However, because of relatively high background radioactivity, the inflamed colon was less evident than with 18F-FDG and 111In-DPC11870.
Differential counting of the colonic tissue and the colonic content indicated that each of the 3 agents almost exclusively localized in the inflamed colonic wall (91, 92, and 87 %ID for 99mTc-granulocytes, 18F-FDG, and 111In-DPC11870, respectively) rather than in the feces, indicating that radioactivity is cell associated. Occasionally, some local uptake in the stool was observed. Macroscopic examination of the colonic wall and the content of the colon indicated that this local uptake was observed mainly in regions with severe tissue damage and ulcers. This implies that occasional enhanced radioactivity in the feces may be caused by loss of cells from the lumen.
CONCLUSION
99mTc-granulocytes, 111In-labeled LTB4 antagonist DPC11870, and 18F-FDG reveal acute colitis in NZW rabbits early after injection. We found that both 111In-DPC11870 and 18F-FDG were superior to 99mTc-granulocytes in allowing visualization of inflamed lesions. At later times, the quality of the images obtained with 111In-DPC11870 continued improving because of ongoing accumulation of radioactivity at the site of inflammation and continuous clearance from the background.
Acknowledgments
The authors thank Gerry Grutters and Hennie Eijkholt (University of Nijmegen, Central Animal Laboratory) for technical support and Paul Jaegers and Peter Kok for assistance during the PET acquisitions.
Footnotes
Received Apr. 3, 2003; revision accepted Sep. 25, 2003.
For correspondence or reprints contact: Julliëtte E.M. van Eerd, BSc, Department of Nuclear Medicine, University Medical Center Nijmegen, P.O. Box 9101, 6500 HB Nijmegen, The Netherlands.
E-mail: J.vaneerd{at}nucmed.umcn.nl
