|Home | About | Journals | Submit | Contact Us | Français|
The aim of the study was to compare 68Ga-chloride with 2-[18F]fluoro-2-deoxy-d-glucose (FDG) for the imaging of pancreatic xenografts.
Rats with subcutaneous human pancreatic adenocarcinoma xenografts were evaluated in vivo by dynamic positron emission tomography (PET) and ex vivo by measuring radioactivity of excised tissues and by digital autoradiography of tumor cryosections.
Both tracers were capable of delineating all subcutaneous tumors from surrounding tissues by PET. The standardized uptake values of tumors by PET were 0.9±0.3 (mean±SD) for 68Ga-chloride (n=13) and 1.8±1.2 for FDG (n=11). Ex vivo studies showed tumor-to-muscle ratio of 4.0±0.3 for 68Ga-chloride (n=4) and 7.9±3.2 for FDG (n=4).
68Ga-chloride delineated subcutaneously implanted pancreatic adenocarcinoma xenografts by PET, but the uptake was lower than FDG. Further studies to clarify the value of 68Ga-chloride for PET imaging of tumors are warranted.
Gallium-67 citrate (67Ga; T1/2=78 h) has been used as an imaging agent in single-photon emission computed tomography (SPECT) for decades for the purpose of detecting certain tumors and infection/inflammation, in spite of its disadvantages resulting mainly from such unfavorable physical characteristics as high radiation exposure and long examination time. The first tumor uptake studies with a positron-emitting gallium-68 (68Ga; T1/2=68 min) were performed in the 1960s [1, 2]. For a long period of time, 68Ga was almost forgotten, but during the last years, it has gone through a renaissance. 68Ga has been applied for the labeling of peptides and oligonucleotides with the macrocyclic chelating agent 1,4,7,10-tetraazacyclododecane-N,N′,N″,N-tetraacetic acid (DOTA) in order to target tumors by positron emission tomography (PET) [3, 4]. Some 68Ga-labeled tumor targeting radiopharmaceuticals, e.g., 68Ga-DOTATOC, 68Ga-DOTANOC, and 68Ga-DOTA-Ianreotide binding to somatostatin receptors, 68Ga-DOTA-bombesin binding to bombesin receptors, and 68Ga-DOTA-D-Glu-gastrin binding to gastrin-releasing peptide receptors, have been reported to be potential imaging agents in clinical studies [5, 6]. Preclinical evaluation of interesting 68Ga-labeled biomolecules, e.g., 68Ga-deferoxamine-folate and 68Ga-DOTA-hEGF, has also been reported [7, 8].
2-[18F]Fluoro-2-deoxy-d-glucose (FDG) is the most widely used PET imaging agent. FDG is a glucose analog that reflects the level of glucose metabolism. Glucose consumption is increased in tumors because of rapid cell proliferation. Glucose and FDG uses GLUT-1 and GLUT-3 cell surface transporters, which are overexpressed in cancer cells . FDG, like glucose, is phosphorylated by hexokinases (particularly hexokinase II) into 2-[18F]fluoro-2-deoxy-d-glucose 6-phosphate (FDG 6-P). Subsequently, FDG 6-P is trapped inside the cell because it is not metabolized any further.
Very recently, we have been able to establish a relationship between the 68Ga-chloride uptake with PET at the site of local bone infection and underlying structural changes determined with peripheral quantitative computed tomography . In addition, when using 68Ga-chloride as a control in our 68Ga-oligonucleotide studies, we noticed uptake in tumors during 2 h PET imaging . The purpose of this study was to preliminarily evaluate if gallium-68 (68Ga) could be used for PET imaging of tumors in a similar fashion as 67Ga is used for SPECT. The advantage of 68Ga-chloride over FDG would be the fast and simple production, i.e., the cyclotron and labeling-free production. 68Ga-chloride is readily available at a PET laboratory. We tested 68Ga-chloride by PET imaging of experimental tumors in comparison with FDG. The results were also verified ex vivo. Compared with the use of 67Ga in SPECT, 68Ga-chloride PET would have the following advantages: better spatial resolution, higher sensitivity, possibility for quantitative assessment of tracer accumulation in tissues, and lower absorbed radiation dose for the study subject.
68Ga-chloride was eluted with 0.1 M HCl from a 68Ge/68Ga generator (Cyclotron Co., Obninsk, Russia). The radioactive elution peak, monitored online with a positron-sensitive photodiode detector (Hamamatsu S5591, Hamamatsu Photonics K.K. Solid State Division, Hamatsu City, Japan), was collected, and the 68Ga-chloride was neutralized with 1.0 M NaOH. FDG was prepared as described earlier .
The human pancreatic ductal adenocarcinoma cell line, BxPC-3, was purchased from American Type Culture Collection (Rockville, MD, USA). The cells were grown in RPMI 1640 culture medium supplemented (10%) with heat-inactivated fetal bovine serum, 2 mM l-glutamine, 100 U/mL penicillin, and streptomycin (100µg/mL) in a humidified atmosphere of 5% CO2 at 37°C. The cells were detached from the culture bottle by trypsin–ethylenediaminetetraacetic acid and resuspended (5×107cells/mL) in medium for implantation.
Twenty-nine athymic male Hsd/RH-rnu/rnu rats were obtained from Harlan (The Netherlands) at the age of 6 weeks. Tumor cells (107/rat) were injected subcutaneously into the neck, and tumors were allowed to grow to a size of 1 cm in diameter. All animal studies had been approved by the Lab-Animal Care & Use Committee of the University of Turku.
To detect the tumor uptake of tracers and to quantify their organ distribution and kinetics, 2 h dynamic PET imaging was performed. Nineteen rats with BxPC-3 xenografts (weight 336±50 g) were anesthetized intraperitoneally with sodium pentobarbital (60 mg/kg; Mebunat, Orion, Finland) or with a mixture of fentanyl citrate plus fluanisone (0.25 and 8.0 mg/kg, respectively; Hypnorm® = fentanyl citrate 0.315 mg/mL and fluanisone 10 mg/mL, Janssen Pharmaceutica, Beerse, Belgium) and midazolam (4.0 mg/kg; Dormicum 5 mg/mL, Roche, Espoo, Finland). The rats were placed in the center of the scanner gantry to obtain the best resolution for small objects and kept on a warm pallet during the imaging procedure. PET imaging was performed either with Advance scanner (GE Medical Systems, Milwaukee, WI, USA) or with HRRT camera (Siemens Medical Solutions, Knoxville, TN, USA) [12, 13].
After transmission scan for attenuation correction, a tracer was injected directly as a bolus intravenously via a tail vein. PET imaging started at the time of injection. The mean±SD dose was 16±3 MBq for 68Ga-chloride (n=13) and 16±9 MBq for FDG (n=11). Prior to tracer injection, rats fasted for at least 2 h. Of a total of 19 rats, five were imaged with both tracers on subsequent days. The acquisition times were 5×60 and 23×300 s, with the total imaging duration being 120 min.
PET data were iteratively reconstructed with the ordered-subsets expectation maximization (OS-EM) algorithm. All PET images were reviewed on computer screen in the transaxial, coronal, and sagittal planes along with maximum-intensity-projection images. Two reviewers aware of subcutaneously implanted tumor xenografts in the neck area independently evaluated tracer uptake both visually and semiquantitatively. The radiotracer uptake in the tumor was visually compared to the surrounding tissue (skin, fat, and muscle) uptake as negative, less-than-surrounding tissue uptake or equivalent-to-surrounding tissue uptake, and positive, uptake which is higher than the surrounding tissue uptake. In case of discrepancies between the two reviewers, a consensus was reached in a joint reading session. Semiquantitative analysis was performed by drawing a standardized circular region of interest (ROI) to the tumor (ROI Ø 5 mm), heart (ROI Ø 6 mm), kidney (ROI Ø 4 mm), liver (ROI Ø 6 mm), and urinary bladder (ROI Ø 4 mm) areas. The heart ROI was drawn to cover the whole heart including both myocardium and the blood pool radioactivity. The average radioactivity concentration in a ROI (kiloBecquerels per milliliter) was used for further analyses. The tracer uptake was reported as mean standardized uptake values (SUV), which was calculated as the radioactivity of the ROI divided by the injected dose per animal body weight. Biokinetic/time–activity curves (TAC), representing the relative radioactivity concentration in the organ of interest (SUV) vs. time after injection, were determined accordingly. TACs were decay corrected to the time of injection.
The uptake of 68Ga-chloride and FDG in tumors was studied ex vivo in tumor-bearing animals. Eight rats (weight 280±69 g) were anesthetized with a mixture of Hypnorm–Dormicum, as described above and intravenously administered with 12±3 MBq of 68Ga-chloride (n=4) or 15±3 MBq of FDG (n=4). Prior to tracer injection, rats fasted for at least 2 h. Ninety minutes after tracer distribution, a blood sample was obtained via intracardiac puncture, and the animals were killed by intracardiac administration of sodium pentobarbital (Mebunat®, Orion, Finland). The time point for ex vivo measurements was based on our previous studies . Samples of blood, tumor, liver, lung, muscle, and skin were excised, weighed, and measured for total radioactivity in an automated gamma counter (1480 Wizard 3″ Gamma Counter; EG & G Wallac, Turku, Finland) cross-calibrated with a dose calibrator (VDC-202, Veenstra Instruments, Joure, The Netherlands) and PET cameras. The tail was also measured for radioactive content to determine the accuracy of the injections. The radioactivity concentration was decay corrected to the time of injection, the radioactivity remaining in the tail was compensated, and the results were expressed as SUV (organ radioactivity/organ weight)/(total given radioactivity/rat body weight). The radioactivity ratios between the target (tumor) and nontarget (blood, liver, lung, muscle, and skin) organs were also calculated.
Two rats (weight 224 g and 213 g) were injected with 19 MBq of 68Ga-chloride or 24 MBq of FDG, respectively. After tracer distribution (90 min), the tumors were excised, frozen in dry ice, and cut with a cryomicrotome into 10–20-μm sections. Tumor sections were thaw-mounted onto microscope slides, briefly air dried, and exposed to an imaging plate (Fujifilm BAS TR, Fuji Photo Film Co, Japan) for two half-lives of radio-isotope in question. The distribution of radioactivity in the sections was digitally scanned using a Fuji BAS-5000 device (Fuji Tokyo, Japan) with the image resolution of 25μm.
After autoradiography, the same sections were stained with hematoxylin and eosin (HE) or using an immunohistochemical method for light microscopy to obtain corresponding histological information. In addition, some tumor samples were fixed with 4% formaldehyde, embedded in paraffin, and cut into 10-μm sections, and the sections were stained with HE. For immunohistochemical staining, DakoCytomation EnVision-system-HRP (K4001, Dako, Glostrup, Denmark) two-step immunohistochemical technique was used.
After 68Ga-chloride autoradiography, the sections were stained with mouse antirat CD68 monoclonal antibody (MCA341GA; AbD Serotec, Oxford, UK; optimal dilution 1:2,000) to examine if the radioactivity originates from macrophage uptake. Antibody was located with 3,3′-diaminobenzidine tetrahydrochloride (Liquid DAB Substrate, K3468; Dako, Glostrup, Denmark). Finally, immunohistochemical sections were slightly counterstained with Mayer's hematoxylin, washed, and mounted.
The digital autoradiographs were combined with digital histological and immunohistological images using GIMP 2.4.5 (GNU Image Manipulation Program, authored by Peter Mattis and Spencer Kimball; http://www.gimp.org/) and Hugin 0.7 beta 3 hugin (Hugin, authored by Andrew Mihal, Pablo d'Angelo, Max Lyons, Erik Krause, Konstantin Rotkvich, and Christoph Spiel; http://hugin.sourceforge.net/) softwares. The intratumoral tracer distribution was reviewed on computer screen by two observers.
All the results are expressed as mean±SD. After testing of normality and variance, an analysis of variance test was applied to study the significance of differences between the tracers. A t test was used for the comparison of PET and ex vivo data. A log transformation because of skeweness was used for all data, except for comparison of ex vivo measurements. A P value of less than 0.05 was considered statistically significant. Statistical analyses were conducted using SAS 9.1.3 statistical software (SAS Institute Inc., Cary, NC, USA).
Both tracers were capable of delineating all subcutaneous BxPC-3 tumor xenografts from surrounding tissues by PET (Fig. 1a, b). Based on 2 h dynamic PET imaging, the peak radioactivity of 68Ga-chloride in the tumor was reached faster in comparison to FDG. The uptake of 68Ga-chloride at the site of tumors showed very fast initial uptake and moderate decline with a plateau at 10 min after injection. In contrast, uptake of FDG accumulated slowly and reached a plateau at 60 min (Fig. 1c). According to PET imaging, the tumor uptake expressed as SUV was 0.9±0.3 for 68Ga-chloride (n=13) and 1.8±1.2 for FDG (n=11; Table 1). The FDG SUV of tumor was significantly higher than that of 68Ga-chloride (P=0.003; Fig. 2). The results obtained by using GE Advance scanner were in accordance with those obtained by using Siemens HRRT camera (P>0.05 for both tracers).
In case of 68Ga-chloride, the excess of radioactivity was distributed into heart, liver, and urinary bladder. The critical organs for FDG distribution were heart and urinary bladder (Fig. 1a, b). The corresponding tissue TACs are shown in Fig. 3.
The results of the ex vivo measurements are presented in Table 2, and they are in accordance with the results of in vivo PET imaging (P>0.05; Fig. 2). The tumor SUV was 0.9±0.2 for 68Ga-chloride and 2.2±1.2 for FDG, and the difference was statistically insignificant (P>0.05).
Target-to-nontarget ratios, i.e., tumor-to-blood, tumor-to-liver, tumor-to-lung, tumor-to-muscle, and tumor-to-skin ratios, are presented in Table 2 and Fig. 4. The tumor-to-blood ratio of 68Ga-chloride was significantly lower compared to FDG (P<0.0001). Both tracers gave tumor-to-muscle ratios high enough to differentiate the tumor from surrounding tissues (range 3.7–4.3 with 68Ga-chloride and 5.7–12.6 with FDG).
Representative digital autoradiographs of 68Ga-chloride and FDG distribution in the tumor tissue are shown in Fig. 5. The imaging agents showed different biodistribution within the tumor. The 68Ga radioactivity was quite homogeneously distributed, while distribution of FDG was more heterogenous.
The histological sections of BxPC-3 tumors showed mainly homogenous, viable tumor tissue without any major necrotic areas (Fig. 6). 68Ga-chloride uptake by the tumor was mostly by viable cancer cell islets. Immunohistochemical staining with mouse antirat CD68 recognizing rat macrophages revealed homogenous distribution in the tumor tissue (Fig. 7).
This experimental study was designed to investigate the feasibility of 68Ga-chloride for PET imaging of tumors by comparing it with FDG. Although FDG is a widely used imaging agent in PET and 67Ga-citrate in SPECT, comparison with 68Ga-chloride has not been performed earlier. Our results revealed that 68Ga-chloride was able to differentiate experimental tumors from surrounding tissues (Fig. 1). The tumor uptake of 68Ga-chloride was lower than that of FDG (Table 1; Fig. 2).
Human pancreatic adenocarcinoma xenograft in rat was chosen for an experimental tumor model, since imaging and early diagnosis of pancreatic cancer still remains a challenge [14, 15]. One problem that hinders early diagnosis is that only symptomatic patients with suspected malignancy are PET imaged. PET is not the primary method for diagnosis, even though it has been reported to be the most accurate one . Since the production of 68Ga-chloride is cyclotron-free, 68Ga-chloride PET might facilitate the availability of PET in centers where no cyclotron is available.
Both tested tracers were capable of visualizing subcutaneous tumors (Fig. 1). According to dynamic PET imaging, the tumor uptake of 68Ga-chloride was much faster than the uptake of FDG. Blood clearance of 68Ga radioactivity estimated from TAC of heart was slower compared with FDG (Fig. 3a), which is in accordance to the fact that Ga3+ binds to transferrin. The slow clearance was seen also as the low tumor-to-blood ratio of 68Ga-chloride as noticed by ex vivo studies (Table 2; Fig. 4a). In our previous studies, the 68Ga-chloride uptake to healthy rat pancreas was low (SUV 0.4±0.1 at 2 h after injection) . When tumor uptake was compared to the muscle, both 68Ga-chloride and FDG had ratio ≥4 (Table 2; Fig. 4d). However, there was a great variation in FDG accumulation in tumor xenografts between animals. We used skin as another background tissue because tumor cells were subcutaneously transplanted. In general, the determination of true skin radioactivity is difficult, since hair and subcutaneous fat may interfere with the results if not carefully removed from the skin sample. Tumor-to-lung and tumor-to-liver ratios are valuable when evaluating the potential of a new imaging agent to differentiate tumors/metastases from these tissues. Pancreatic cancer cells easily send metastases, in particular, to the liver after surgical operation , and many cancers, like breast cancer, are known to send metastases to the lungs . From the tested tracers, FDG seems to be clearly superior to differentiate tumors from the normal liver tissue (Table 2; Fig. 4b). FDG seems the best in visualizing tumors in the lungs also. Identification of tumors with 68Ga-chloride in the abdominal area might be challenging due to high blood pool and liver radioactivity, although use of magnetic resonance imaging (MRI) or CT as an anatomical reference to PET might help.
Despite technological advances, PET is characterized by a relatively low spatial resolution. In this work, we used clinical PET devices (GE Advance and Siemens HRRT) for the imaging of rat because at the time of the study, we did not have access to a dedicated small animal PET scanner. The animal was placed in the center of the scanner gantry giving the highest possible spatial resolution, i.e., 4.8 mm with GE Advance and 2.7 mm with Siemens HRRT. Unfortunately, the structures we were aiming at were quite small (tumor diameter approximately 10 mm), which might result in partial-volume effect (PVE) and possibly invalidate the in vivo quantitative PET data. PVE means that the apparent pixel values in PET images are influenced by the surrounding high pixel values . In the present study, the central 200-mm-diameter transaxial field of view was reconstructed using OS-EM algorithm . Using the 128×128 matrix, this leads to the pixel size of 1.56×1.56 mm, which is approximately one third (GE Advance) or half (Siemens HRRT) of the image resolution. Boellaard et al. have obtained superior image quality by using OS-EM reconstruction instead of filtered back-projection, resulting in lower PVE . In theory, PVE correction is possible if high-resolution structural imaging, such as magnetic resonance imaging, can be coupled with the information of the scanner resolution. However, lack of anatomic MRI of the rats prevented PVE correction in this study. To minimize the PVE, we used the average radioactive concentration within the ROI (kiloBecquerels per milliliter).
The majority of positron-emitting radionuclides are produced by expensive accelerators requiring qualified personnel. 68Ga offers a cyclotron-independent, convenient, and low-cost access to PET radiopharmaceuticals. 68Ga is readily available by elution from 68Ge/68Ga generator possessing 1–2 years life span depending on uploaded 68Ge radioactivity. Furthermore, 68Ga has convenient characteristics for PET detection: β+decay 89%, Eβ+max 1.9 MeV, and T1/2 68 min. It has been shown that the clinical image qualities of 68Ga-tracers and 18F-tracers (β+decay 97%, Eβ+max 0.64 MeV, T1/2 110 min) are not necessarily very much different from each other . Although the half-life of 68Ga is shorter compared to that of 18F, the significant advantage for 18F-tracers is the lower radiation dose due to their lower positron energy. The 68Ge/68Ga generator supplied by Cyclotron Co., which was used in this study, contains 68Ge attached to a column of an inorganic matrix-based titanium dioxide (TiO2). Several research groups that are currently using this generator type have reported that the 68Ge breakthrough is <0.01%, and concentration of other metal ions (Zn2+, Al3+, Ge4+, Ti4+, Cu2+) is <50nM in the 68Ga-eluate [3, 22–24].
When 68Ga is eluted from 68Ge/68Ga generator with 0.1 M HCl, it is in the form of 68GaCl3. At least some of the 68Ga is hydrolyzed after neutralization with NaOH in situ or after administration in vivo. In general, in aqueous solution, 68Ga is in the form of a soluble anion called gallate, , and/or insoluble neutral hydroxide colloids, 68Ga(OH)3, depending on pH and the concentration of 68Ga . The 68Ga radioactivity can migrate in the blood circulation as free 68Ga3+ or 68Ga3+ bound to transferrin, ferritin, or lactoferrin after rapid intravenous administration to the vascular system. 67Ga is injected as citrate for SPECT imaging. Since the citrate is only a weak chelator in vivo, 67Ga3+ cation is rapidly released, hydrolyzed, and/or bound to transferrin and other plasma proteins. Because the citrate is able to prevent precipitation of 67Ga(OH)3, it is assumed that only a soluble is formed in vivo. The use of 67Ga-citrate has been characterized by its slow accumulation in tumors, which makes the imaging of tumors by SPECT possible only after several hours, even days. In contrast, when using 68Ga-chloride for PET, we observed rapid localization of radioactivity in the target. The form of 68Ga, i.e., the ratio of 68Ga-gallate and 68Ga-hydroxide colloids, was not studied in this study.
The uptake mechanisms of radiogallium into tumors are not fully understood, and a wide variety of factors are involved. It is known that Ga3+ behaves in vivo similarly to Fe3+ by binding to transferrin and using the same transporters . Gallium seems to concentrate in tissues having high concentration of transferrin or transferrin receptors, lactoferrin, or ferritin. All nucleated cells of the body express transferrin receptors at different concentrations. Malignant cells generally express high levels of transferrin receptors because they divide and require iron for DNA synthesis. However, there is some evidence that Ga can also enter tumor cells by transferrin-independent mechanisms. This may be due to a very small soluble gallate ion, , or a neutral insoluble, Ga(OH)3, both hydrolysis products of Ga . In this study, 68Ga radioactivity seemed to be taken up mostly by viable cancer cell islets according to 68Ga autoradiography and HE staining (Fig. 5). The comparison of autoradiography and CD68 immunohistochemical staining (Fig. 7) suggests that the uptake might also be a result of macrophage phagocytosis. The resolution of autoradiography is not high enough to show the radioactivity uptake of individual cells, but it rather shows the accumulation in certain tissue areas, like tumor cell islets. It is possible that the 68Ga radioactivity is taken up by both tumor cells and infiltrating macrophages. Indeed, accumulation of Ga in macrophages, leukocytes, and bacteria has been reported [26, 27].
Fast tumor uptake of 68Ga-chloride would facilitate PET imaging as soon as 2 h after injection vs. long-lasting study procedure with 67Ga-citrate in SPECT imaging, although the slow blood clearance of 68Ga-chloride suggests for a later time point. At 2 h after injection, the high blood pool radioactivity may contribute to tumor uptake as well.
68Ga-chloride and FDG PET imaging were able to delineate subcutaneously implanted human pancreatic adenocarcinoma xenografts in rats. The tumor uptake of FDG was clearly superior, but there was great variation between the animals. 68Ga-chloride showed fast tumor uptake, high blood pool radioactivity, and low urinary excretion. Advantageously, 68Ga-chloride is readily available from a generator system and produces good PET image quality. However, further studies are needed to clarify the value of 68Ga-chloride for PET imaging of tumors.
We acknowledge Maija-Liisa Hoffren (SafetyCity Oy Ltd, Turku), Irina Lisinen (Research Centre of Applied and Preventive Cardiovascular Medicine, University of Turku), Jouko Sandholm (Cell Imaging Core, Turku Centre for Biotechnology), and Erica Nyman (Turku Centre for Disease Modeling, University of Turku) for excellent technical assistance. The study was conducted within the Finnish Centre of Excellence in Molecular Imaging in Cardiovascular and Metabolic Research supported by the Academy of Finland, University of Turku, Turku University Hospital, and Abo Academy University. In addition, study was financially supported by grants from the Academy of Finland (nos. 205757 and 103032), the Foundation for the Finnish Cancer Institute, the Instrumentarium Foundation, and the Eli Lilly Foundation. Tiina Ujula and Anu Autio are Ph.D. students supported by the Drug Discovery Graduate School of the University of Turku.
Open Access This article is distributed under the terms of the Creative Commons Attribution Noncommercial License which permits any noncommercial use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.
Significance: This is the first study to evaluate 68Ga-chloride for PET imaging of pancreatic xenografts in comparison with FDG.
Tiina Ujula, Phone: +358-40-5371750, Fax: +358-2-2318191, Email: if.utu@opnait.
Anne Roivainen, Phone: +358-2-3132862, Fax: +358-2-2318191, Email: email@example.com.