The pC1-tk-IRES-luc plasmid was constructed according to our previous study (24). F98 rat glioma tumor cells were transfected with this plasmid using cationic polymer, jetPEI (Polyplus Transfection, Strasbourg, France). F98 cells were cultured in a 6-well plate (2×10⁵ cells/well) after transfection. In a 1.5-ml Eppendorf vial, 3 μg of DNA plasmid was mixed with 6 μl jetPEI reagent in 100 μl of 150 mM NaCl followed by gentle vortexing, and then briefly spun down. The mixtures were incubated for 30 min at room temperature and were added to the well. Transfection was performed at 37°C for 24 h in a humidified incubator with 5% CO₂. Stable clones were selected with 600 μg/ml of G418 (Calbiochem, Darmstadt, Germany).

For the luc activity assay, the cell lysate was prepared by performing a freeze-thaw cycle in reporter lysis buffer (Promega, Madison, WI, USA). The bioluminescence was measured with a luminometer (Wallac 1420 Multilabel Counter; Life Sciences, Pittsburgh, PA, USA). The protein concentration of the cell lysate in each well was measured with a Bio-Rad protein assay kit (Bio-Rad Laboratories, Hercules, CA, USA) (24). F98 cells transfected with pC1-tk-IRES-luc vectors were cultured in a 6-cm diameter dish for 14 days, and 30 colonies were selected. Each colony was cultured in a 3-cm diameter dish for cell proliferation. Cells were treated with 200 μl lysis buffer, and were frozen at −80°C for 1 h. After thawing and centrifugation (1,200 rpm, Kubota Laboratory Centrifuge, model 5100; Kubota Corp., Tokyo, Japan), the supernatant was transferred to a black 96-well plate filled with D-luciferin. The bioluminescence was quantified as photons/sec/mg. For BLI imaging in vivo, male Fischer 344 rats and NOD/SCID mice were anaesthetized with 1–3% isofluorane, and then an i.p. injection with 300 μl D-luciferin (150 mg/kg in PBS) as the substrate. The IVIS 50 Imaging System (Xenogen, Alameda, CA, USA) was used for image acquisition. The time duration was 15 min for rats and 2 min for mice, followed by quantification with the Living Image software (Xenogen).

F98/tk-luc cells were cultured in 96-well plates (3,000 cells/200 μl/well) for 24 h, and then treated with 0–200 mM ganciclovir (GCV; F. Hoffmann-La Roche Ltd., Basel, Switzerland) for 120 h. 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma-Aldrich, St. Louis, MO, USA) colorimetric assay was used to determine cell survival (24).

[¹³¹I]FIAU was prepared according to our previous study (25). [¹⁸F]FBAU was synthesized according to a study reported by Alauddin et al (26). The purification of [¹⁸F]FBAU was followed according to our previous study (27).

[¹³¹I]FIAU (3.7×10⁴ Bq/μl)was added to each well (1×10⁶ cells/3 ml/well in 6-well culture plates), followed by incubation at 37°C for 15 and 30 min, 1, 2 and 4 h, respectively. Triplicates were performed at each time point.

F98 glioma cells (a kind gift from Dr Rolf F. Barth, The Ohio State University, USA) were transfected with the pC1-tk-IRES-luc plasmid (24), and renamed as F98/tk-luc glioma cells. These cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (containing 10% FBS, 100 U/ml penicillin and 100 μg/ml streptomycin) at 37°C in a humidified atmosphere of 5% CO₂. Male Fischer 344 rats (12–14 weeks of age, 250–280 g) were anesthetized intraperitoneally (i.p.) with a mixture of 110 mg/kg ketamine and 9 mg/kg xylazine. F98/tk-luc glioma cells (1×10⁵) in 10 μl Mg²⁺/Ca²⁺-free Hanks’ balanced salt solution (19) were then slowly (15–20 sec) injected into the left brain region of the rats. The syringe was held still for 2 min after injection before withdrawing, and the injection hole was sealed with bone wax. Finally, the wound was flushed with iodinated alcohol and sutured with a sterilized steel clip. Male NOD/SCID mice were inoculated with 5×10⁵ F98/tk-luc cells in 200 μl serum-free medium subcutaneously in the right flank. The tumor volume was measured with a digital caliper.

3T MR scanning was performed within 24 h post BLI imaging. A 40-mm diameter field-of-view (FOV) orbital surface coil was used as the receiver (28). Each rat was anesthetized with 3–4 ml 10% chloride hydrate. To obtain accurate position of the brain slice, the coronal and transverse plane scout images were scanned first. 2–4 coronal T1-weighted slices with a 1.5-mm thickness would cover the tumor. The T1-weighted images were acquired with a 256×256 matrix, FOV=89 mm, repetition time=435 msec and echo time=12 msec. For contrast-enhanced images, each animal received 0.2 mmol/kg contrast medium i.v. 5 min prior to T1-weighted sequence imaging. T2-weighted images were acquired with a 256×256 matrix, FOV=89 mm, repetition time=5,000 msec and echo time=104 msec. The DICOM images were collected, and the tumor boundary visualized in each slice was defined as the region-of-interest (ROI) using ImageJ (National Institutes of Health, Bethesda, MD, USA) for calculation. The sum of the ROI areas was processed by slice thickness to obtain the tumor volume.

After examination, the rats were sacrificed with CO₂. The brains were removed and fixed with 4% paraformaldehyde. The sections (5-μm) obtained from the paraffin-embedded samples were stained with hematoxylin and eosin (H&E).

Biodistribution of [¹⁸F]FBAU was determined on the 13th day post tumor implantation as previously described (19). Each rat was anesthetized with 3–4 ml of 10% chloride hydrate. The F98/tk-luc glioma-bearing rats were injected with 15 MBq/0.2 ml [¹⁸F]FBAU via lateral tail veins. At 0.5, 1, 1.5 and 2 h post injection, the rats (n=5 per group) were sacrificed with CO₂. Left and right brains, tumors, intestine, spleen, heart, blood, lung, liver, stomach, pancreas, kidneys, bone and muscles were removed, and the radioactivity was measured in a well-type multichannel analyzer (Raytest, Straubenhardt, Germany) (27). The uptake of [¹⁸F]FBAU in tissues or organs was expressed in counts per second (cps) corrected with the decay, and was normalized as the percentage injected dose per gram of tissue (%ID/g).

The PET images of F98/tk-luc glioma-bearing rats i.v. injected with 50 MBq/0.1 ml [¹⁸F]FBAU were acquired up to 160 min using the PET/CT system (GE Discovery LS PET-CT; GE Healthcare, Cleveland, OH, USA) on day 13 post tumor implantation. The imaging parameters were set as follows: 2-D brain model, FOV=15.52 cm, spatial resolution=5.46 mm, 128×128 matrix size. The reconstruction algorithm using ordered subset expectation maximization (OSEM) was similar to that of a previous study (19). The images were acquired with axial, coronal and sagittal planes, respectively. Each rat was anesthetized with 3–4 ml 10% chloride hydrate and i.v. injected with 0.1 ml of 50 MBq [¹⁸F]FBAU. Data acquisition by PET scanning was initiated at the first minute after drug injection. Dynamic coronal and sagittal images were acquired using 10 60-sec frames and 10 2-min frames, followed by 10-min frames up to 2 h. Standardized uptake values (SUVs) were calculated and normalized to body weight (BW) (29). %SUV was calculated using the following formula: %SUV = [(left brain SUV - right brain SUV)/right brain SUV] × 100%.

The ex vivo autoradiography was performed according to a previous report (20). Fischer 344 rats were anesthetized with 1–3% isofluorane. At 30, 60, 90 and 120 min post i.v. injection of [¹⁸F]FBAU (15–20 MBq/0.7 ml), the rats were sacrificed with CO₂. The brains were removed and embedded with OCT on a cryostat holder (diameter 2.7 cm), and were frozen immediately at −80°C. Coronal sections (20-μm thick) were performed using a cryomicrotome (Leica CM3050; Leica Microsystems, Bensheim, Germany). The imaging plates were assayed with an FLA5000 reader (Fuji Photo Film Co., Ltd., Tokyo, Japan) with the following settings: PMT=800 V, resolution=50 μm, gradation=16 bits, dynamic range=L5 and sensitivity=5,000 to acquire the phosphor images. ROIs were circled along the tumor contour of the F98/tk-luc glioma, and ROIs of equal size in the normal brain region at corresponding brain position were used as the reference. The intensity of photo-stimulated luminescence (PSL) of the tumor and normal brain tissue was measured with Image Gauge (version 4.0, Science Lab 2001; Fuji Photo Film).

The Studen’t t-test was used to analyze the significant difference between the control and drug-treated rats. All data are shown as the means ± standard error (SE). Differences between the means were considered significant at P≤0.05. The correlation between tumor volume and BLI photon flux was performed by Pearson correlation.

The pC1-tk-IRES-luc vectors as shown in Fig. 1A were transfected into the F98 cell line, which was renamed F98/tk-luc. The colony exhibiting luc expression of >1,500 counts per sec per mg (cps/mg) was selected for cell expansion as shown in Fig. 1B (colony no. 11). The bioluminescence of luc gene expression in the F98/tk-luc cells was 3.24 photons/sec/cell (Fig. 1C).

F98/tk-luc cells were treated with various concentrations of ganciclovir (GCV). The surviving fraction (SF) of F98/tk-luc cells was 18% at 15 μM and remained ~100% up to 200 μM for parental F98 cells after GCV treatment (Fig. 2A). In addition, the uptake of [¹³¹I]FIAU in the F98/tk-luc cells was increased with time, but was not found in the parental F98 cells (Fig. 2B).

F98/tk-luc cells were inoculated subcutaneously into the right flank of NOD/SCID mice. Two weeks later, tumor growth was monitored with BLI as shown in Fig. 3A. The quantification of bioluminescence (photons/sec × 10⁵) vs. tumor volume (mm³ × 10) from representative mice is plotted in Fig. 3B with linear regression (r²=0.90).

The BLI of F98/tk-luc glioma-bearing rats was obtained using the IVIS50 optical imaging system (Xenogen Corp.) on day 13 post inoculation of F98-tk-luc glioma cells into the left brain. The total photon flux rates of the tumors were 5.50×10⁵-4.14×10⁶ as shown in Fig. 4A (images were from representative rats). MR images of the representative rat brains were obtained from a 3T MR scanner including T1-, T2-weighted and contrast-enhanced images (Gd 0.2 mmol/kg) on day 14 (Fig. 4B).

In the present study, the tumor volume and its extent were not able to be demonstrated well by T1-weighted images but were greatly enhanced post contrast injection, which showed a clear boundary between the tumor and normal brain tissue. On the other hand, the tumor location could be shown by T2-weighted images whereas the boundary between the tumor and normal tissue was still indistinguishable. Fischer 344 rats after MR scanning were sacrificed, and the brain tissues were removed and preserved in 4% paraformaldehyde. H&E staining was subsequently performed. The results are shown in Fig. 4C, in which the tumor location and tumor volume were the same as those of the MR images shown in Fig. 4B.

The total photon fluxes of F98-tk-luc tumors obtained from BLI vs. tumor volumes calculated by 3T MR contrast-enhanced images are tabulated in Table I and plotted in Fig. 4D. The Pearson correlation coefficient (r) is 0.95.

The accumulation of [¹⁸F]FBAU in tissues of the F98/tk-luc tumor-bearing Fischer 344 rats (n=5) was measured at different time points after administration of 15 MBq/0.2 ml. The highest uptake of [¹⁸F]FBAU was found in kidneys, suggesting that the elimination route of [¹⁸F]FBAU is via renal clearance. The uptake of [¹⁸F]FBAU in the left and right normal brain tissues was very low in the range of 0.03–0.09 %ID/g (Fig. 5A and Table II). Notably, the uptake of [¹⁸F]FBAU in the F98/tk-luc glioma tumors reached up to 0.51±0.06 at 90 min post injection (Fig. 5B). The uptake of [¹⁸F]FBAU in the tumors was significantly higher than those of normal brain tissues (P<0.01). In addition, the uptake of [¹⁸F]FBAU in tumors in the 60 and 90 min post-injection groups was significantly higher than that of the 30 min post-injection group (P<0.05).

The coronal and sagittal PET images of the F98/tk-luc glioma-bearing rats were imaged with a clinical PET. The result showed that accumulation of [¹⁸F]FBAU was noted in F98/tk-luc glioma in the brain (arrows in Fig. 6A). The [¹⁸F]FBAU/PET images at 30–90 min post injection showed significantly higher %SUV values than that acquired at the initial time point (P<0.005; Fig. 6B).

Biodistribution of [¹⁸F]FBAU in the rat brain was also examined by ex vivo autoradiography. The result showed that the uptake of [¹⁸F]FBAU in the F98/tk-luc gliomas reached the maximum value at 90 min post injection (Fig. 7A). Autoradiogram of [¹⁸F]FBAU in the F98/tk-luc gliomas showed a similar result as that of the biodistribution study with the highest uptake at 90 min post injection. The quantification of the autoradiogram is depicted in Fig. 7B. The uptake of [¹⁸F]FBAU assayed by BLI in units of photo-stimulated luminescence (PSL)/mm² in the F98/tk-luc gliomas was significantly higher at each time point from 30 to 120 min compared to that of the contralateral normal brain tissue (Fig. 7B).