Ex vivo cell labeling with ⁶⁴Cu–pyruvaldehyde-bis(N⁴-methylthiosemicarbazone) for imaging cell trafficking in mice with positron-emission tomography

Radiotracer Synthesis.

⁶⁴Cu was produced at Washington University by cyclotron irradiation of an enriched ⁶⁴Ni target by using methods reported (10). ⁶⁴Cu–PTSM was prepared according to literature methods (9). In brief, a 10-μl aliquot of PTSM (1 mg/ml DMSO) was mixed with 150 μl of ⁶⁴Cu (1 M NaOAc, pH 5.5) and vortexed briefly. After 3–5 min, the mixture was added to an ethanol-preconditioned C-18 SepPak Light column. ⁶⁴Cu–PTSM was eluted off with 500 μl of ethanol after the initial 150-μl fraction. Approximately 1 mCi of ⁶⁴Cu–PTSM arrived at UCLA (specific activity, ≈4,000 Ci/mmol). Cell culture experiments, cell labeling for in vivo transfer, and imaging studies were initiated 1–3, 14–16, and 24–28 h after tracer arrival, respectively. FDG was synthesized at UCLA as described by Hamacher et al. (ref. 11; specific activity, ≈5,000 Ci/mmol).

Cell Preparation.

C6 cells were obtained from the American Type Culture Collection and cultured in high-glucose, deficient DMEM (def-DMEM). Unless otherwise noted, the growth medium was supplemented with penicillin (100 μg/ml), streptomycin (292 μg/ml), glutamine (100 mM), histidinol (27 μg/ml), and 5% FBS by volume. In all FDG studies, glucose-free medium was used starting 30 min before the experiment. C6 cells (10–18 × 10⁴ cells per ml) were seeded in 10-cm, 6-, 12-, or 24-well culture dishes the day before labeling study. PBS and medium volumes of 10, 3, 1, and 0.5 ml correspond to 10-cm, 6-, 12-, and 24-well culture plates, respectively. Lymphocytes for labeling were collected from spleens of Swiss–Webster mice. Splenocytes were dislodged from homogenized spleen tissue into PBS. Lymphocytes were isolated by density centrifugation with Histopaque-1077 (Sigma) and cultured in high-glucose DMEM supplemented similar to def-DMEM except with heat-inactivated FBS and without histidinol. The cells were always cultured in a humidified atmosphere at 37°C with 5% CO₂.

Analysis of ⁶⁴Cu–PTSM and FDG Uptake by C6 Cells.

Cells were labeled with ⁶⁴Cu–PTSM in medium (≈5 μCi/ml) supplemented with FBS for 5–480 min or without FBS for 5–90 min. To optimize labeling conditions, the uptake of ⁶⁴Cu–PTSM by C6 cells was evaluated also at 3, 6, and 11 h of incubation. In parallel experiments, cells were labeled for 5, 20, 35, and 90 min with 1.5–12.4 μCi of FDG in 1 ml of medium. Radiotracer uptake into cells was terminated by the removal of labeling medium followed by two PBS washes. Cells were lysed with 0.1% SDS, and protein was assayed with Bradford reagent (Bio-Rad). Aliquots of cell lysate and supernatant were counted for 1 min with a Cobra II γ counter (Packard, model 5003). The decay-corrected counts in lysate and supernatant totaled the initial labeling activity. Label uptake was expressed as radioactivity in lysate divided by total activity and normalized to micrograms of protein to correct for cell loss during an experiment.

Analysis of ⁶⁴Cu–PTSM and FDG Efflux from C6 Cells.

C6 cells growing in 12-well plates were labeled for 90 min with ≈5 μCi of ⁶⁴Cu–PTSM or ≈15 μCi of FDG. Labeled cells were washed twice (PBS) and then incubated in tracer-free medium. ⁶⁴Cu–PTSM efflux was assayed at 0, 0.5, 2, 5, 20, and 24 h, whereas FDG release was measured only at 0, 0.5, 2, and 4 h because of its short t_1/2. Afterward, efflux supernatant was removed, and cells were washed with PBS. Those cells remaining adherent to plate were lysed with 0.1% SDS. The counts in cell lysate and labeling and efflux supernatants were decay-corrected. The percentage of label retention was calculated by dividing the counts in lysate by total counts in lysate and efflux supernatant × 100.

Measurement of C6 Cell Viability and Proliferation.

The relative survival of cells exposed to different experimental conditions was evaluated by using trypan blue dye (0.4%) exclusion assay in preliminary experiments and colorimetric 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) reduction assays in later experiments (Sigma). Exclusion of trypan blue dye by viable C6 cells was evaluated under a light microscope by using a hemocytometer after 35- and 270-min exposures to ⁶⁴Cu–PTSM or FDG (5 μCi in 1 ml of medium). All other conditions were identical to uptake studies. Unexposed and UV-exposed cells were used as positive and negative controls, respectively. Trypan blue also was used to determine C6 cell viability 20 h after ⁶⁴Cu–PTSM labeling. The percentage viability equaled viable over total cell count × 100. The MTT assay is based on the ability of living cells to reduce MTT into a blue-colored formazan product. C6 cells preseeded into 24-well plates were incubated either with or without ≈10 μCi of ⁶⁴Cu–PTSM (1 × 10⁵ cells in 0.5 ml of medium) for 5 h. Cells exposed to ethanol were used as positive controls. After 0, 1, 5, 20, and 24 h of efflux, cells were washed with PBS. Cells then were incubated with 50 μl of MTT stock solution (5 mg/ml in PBS) in 0.5 ml of medium at 37°C, and 2 h later acid-isopropyl alcohol (0.5 ml, 0.04 M HCl in isopropyl alcohol) was added. After 0.5 h, wells were mixed to dissolve the blue formazan salt crystals. The intensity of the resulting color changes was measured by reading the spectrophotometric absorbance of 1-ml samples at 570 nm. The effect of ⁶⁴Cu–PTSM labeling on the C6 cell viability and proliferation rate was evaluated by comparing the absorbance of labeled to unlabeled cells over time. Trypan blue and MTT assays were done in 3 and 5 replicate wells, respectively. The two-tailed Student's t test was used for statistical analysis of differences in the viability and metabolic activity of ⁶⁴Cu–PTSM-labeled and unlabeled replicates for each of the indicated time points. The differences were considered significant at P < 0.05.

⁶⁴Cu–PTSM Labeling of Cells for in Vivo Transfer.

The C6 cell-labeling procedure for in vivo studies was similar to the protocol used to examine ⁶⁴Cu–PTSM efflux. In brief, cells were seeded in 6-well plates 1 day before labeling. Cells were labeled for 5 h with ⁶⁴Cu–PTSM (110 μCi in 3 ml of medium), washed twice (PBS), and incubated with tracer-free medium for 5 h. After the efflux period, labeled cells were PBS-washed and trypsinized (0.8 ml) at 37°C for 2–3 min. Detached cells were collected with medium (3.2 ml) and pelleted by centrifugation (5 min, 4°C, 1,600 × g). The pellet was resuspended in PBS (0.5 ml). Because lymphocytes require cytokine stimulation to prolong viability in culture, the labeling procedure was shortened to 1 h of uptake with no efflux period. Lymphocytes (28 × 10⁶ cells) were labeled with 72 μCi of ⁶⁴Cu–PTSM in 1 ml of medium for 55 min. The labeling period was terminated by the addition of tracer-free medium (3 ml). Lymphocyte suspension was pelleted by centrifugation (10 min at room temperature, 300 × g), and the pellet was resuspended in 0.5 ml of PBS. Before in vivo transfer, cell concentration and viability were measured by trypan blue exclusion. Typically, 80–90% of cells were viable.

MicroPET Studies and Data Analysis.

All animal handling was performed in accordance with UCLA Animal Research Committee guidelines. Male nu/nu were anesthetized by i.p. administration of ketamine (100 mg/kg) and xylazine (10 mg/kg) either before or after injection of radioactivity. Mice were injected via tail vein (±5–10% injection accuracy) with one of the following in ≈200 μl of PBS: (i) 4–12 × 10⁴ C6 cells, 0.05–4.27 μCi, (ii) ⁶⁴Cu–PTSM, 2–4 μCi, (iii) 7.2 × 10⁶ splenic lymphocytes, 11.3 μCi. All injected cells were labeled with ⁶⁴Cu–PTSM. Unless otherwise noted, day-1 scans were initiated within 1 h of i.v. injection, and day-2 scans at ≈20 h. All mice were whole body (WB)-scanned in a supine position by using a microPET scanner (11). All WB microPET scans consisted of 7 bed positions with 8 min of data acquisition per bed. Of the 10 mice injected with C6 cells, three were prepared for digital WB autoradiography (DWBA) immediately after the day-1 scan and two after the day-2 scan. The mice injected with ⁶⁴Cu–PTSM were used as control. Two control mice were scanned on days 1 and 2, and DWBA was performed on one of them after the second scan. The organs of the additional mice were harvested, weighed, and counted either on day 1 or 2 after the injection of C6 cells or ⁶⁴Cu–PTSM. The mouse injected with splenic lymphocytes was scanned at 0.12, 3.33, and 18.9 h and then prepared for DWBA. Figs. 4–6 show averaged coronal planes (each plane is ≈0.4 mm thick) from WB microPET images reconstructed with the MAP algorithm (12). Filtered back-projection reconstruction was used for image quantification. MicroPET scans were standardized to phantom scans but were not corrected for photon attenuation (11). The percentage of injected dose per gram of tissue (%ID/g) color scale reflects the signal intensity at a region of interest. A mouse-sized cylinder scan with a known concentration of ⁶⁴Cu in water (μCi/cm³) was used to obtain a calibration factor for converting regions of interest from counts/sec/voxel to μCi/cm³ or μCi/g assuming equal water (1 g/cm³) and tissue density. The %ID/g was calculated by normalizing converted regions of interest (μCi/g) to the injected dose (μCi) × 100. The %ID/g for each organ was expressed as the mean of three regions of interest ± SE.

DWBA.

Immediately after microPET scanning, mice were killed via i.v. injection of ketamine/xylazine (100 μl), and DWBA was performed by methods described previously (13). Coronal WB mouse sections (45 μm thick) were developed digitally with a final image resolution of 100 μm. In Figs. 4–6, a single autoradiogram section and its corresponding photo are shown for the adjacent microPET image (several averaged planes) to link radioactive signal with anatomy.

C6 Cell-Labeling Efficiency Is Greater with ⁶⁴Cu–PTSM Than with FDG.

The ⁶⁴Cu–PTSM extraction efficiency by C6 cells as a function of time, serum, and dose was evaluated to determine the optimal labeling conditions for in vivo studies. ⁶⁴Cu–PTSM passively diffuses into or out of cells by moving down its concentration gradient. Therefore, the C6 cell-labeling efficiency (percentage uptake/microgram of protein) improves with increasing ⁶⁴Cu–PTSM concentration but declines with increasing cell number (data not shown). The labeling procedure with ⁶⁴Cu–PTSM is very efficient (70–85% uptake at 5 h) for C6 cells. The results shown in Fig. 2 indicate that ⁶⁴Cu–PTSM is extracted rapidly into C6 cells during the first 3 h, and influx stabilizes between 3 and 8 h, suggesting a period of equal intra- and extracellular ⁶⁴Cu–PTSM concentration. Longer labeling studies under similar conditions revealed that uptake of ⁶⁴Cu–PTSM declines between 6 and 11 h of incubation (data not shown). This decrease in uptake could be caused by reversal of the ⁶⁴Cu–PTSM concentration gradient. Removing serum from medium increased C6 cell uptake (percentage per microgram of protein) of ⁶⁴Cu–PTSM by 32.8 ± 3.34% (data not shown), which is consistent with published data (14). For comparison, FDG cell-labeling efficiency was evaluated also; however, FDG studies were limited by the short t_1/2 of ¹⁸F. Fig. 2 clearly illustrates that the C6 cell-labeling efficiency (percentage uptake/microgram of protein) of ⁶⁴Cu–PTSM is greater than FDG. Because FDG is transported actively into cells, its labeling efficiency (percentage uptake/microgram of protein) depends on the number of available glucose transporters. As expected, FDG uptake by C6 cells was inhibited potently (≈100-fold lower) in high-glucose medium (data not shown). Therefore, glucose-free medium was used in FDG experiments. At the FDG concentrations tested (1.5–12.4 μCi/ml), labeling efficiency was not dose-dependent, suggesting saturation of transporters (data not shown).

Both ⁶⁴Cu–PTSM and FDG Efflux from C6 Cells.

To evaluate the efflux rate of ⁶⁴Cu–PTSM and FDG, leakage of radioactivity into tracer-free medium from labeled C6 cells was measured. As shown by the diminishing cell-associated activity in Fig. 3, both tracers rapidly efflux from C6 cells. Serum did not influence the rate of ⁶⁴Cu–PTSM release, but glucose accelerated FDG efflux. ⁶⁴Cu–PTSM retention was examined numerous times (n = 20) because of considerable variance after 5 h of efflux (±20% retention). Visual inspection of cells under a light microscope revealed that there was a substantial and progressive loss of cell adherence to plate over time. The cell detachment from plate may be a result of repeated washes, inadvertent aspiration, increasing confluence, and/or cell death and would overestimate ⁶⁴Cu–PTSM efflux and account for the experimental variance observed at later time points in Fig. 3.

⁶⁴Cu–PTSM Labeling Procedure Imposed Minimal Cytotoxicity.

The trypan blue dye exclusion and MTT assays evaluate living cells based on intact plasma membrane and metabolic activity, respectively. As measured by trypan blue, C6 cells labeled with ⁶⁴Cu–PTSM for 35 and 270 min, showed 85 and 80% viability, respectively, with no significant difference in viability of cells similarly labeled with FDG (data not shown). Unexposed controls had an average viability of 92–94%. In contrast, only 2–5% of cells exposed to UV light for 35 and 270 min remained viable. The percent viability of C6 cells at 20 h postlabeling was 80–85%. Living cells convert the tetrazolium component of the MTT dye (yellow) into a formazan product (blue). Because intensity of the resulting color change is proportional to the number of living cells, tetrazolium salt reduction can be used both as a measure of cell viability and proliferation. There was no significant difference in MTT reduction by ⁶⁴Cu–PTSM-labeled C6 cells and unlabeled controls, neither immediately nor up to 24 h after the 5-h labeling period, indicating ⁶⁴Cu–PTSM labeling does not alter cellular proliferation (data not shown).

⁶⁴Cu–PTSM-Labeled Cells Can Be Imaged Noninvasively in a Living Mouse by Using a MicroPET Scanner.

To demonstrate that the microPET scanner can detect the presence of cells in association with ⁶⁴Cu, mice were injected via tail vein with 94,500 ⁶⁴Cu–PTSM-labeled C6 cells (Fig. 4A). Because the location of activity in the microPET image (Fig. 4A) correlates with the autoradiogram (Fig. 4C), the anatomic origin of activity can be determined by matching the autoradiogram with its corresponding photo (Fig. 4B). Thus, Fig. 4 demonstrates that 0.45 h after injection of ⁶⁴Cu–PTSM-labeled C6 cells, most of the activity is in the lungs (97.2 ± 3.37%ID/g), and some is present also in the liver (6.96 ± 0.81%ID/g) and gut (1.31 ± 0.04%ID/g). MicroPET images of mice at 0.23 h after injection of labeled C6 cells show distribution of activity in the lungs (99.2 ± 6.31%ID/g) and the liver (4.82 ± 0.90%ID/g). By 18.8 h, activity redistributes from the lungs (4.78 ± 0.49%ID/g) to the liver (13.6 ± 1.69%ID/g) (data not shown). Gut activity increases slightly from 0.56 ± 0.02%ID/g (0.23 h) to 1.64 ± 0.65%ID/g (18.8 h). DWBA verified that the liver was the primary anatomic origin of the radioactivity at 18.8 h.

To compare the biodistribution of ⁶⁴Cu–PTSM-labeled C6 cells to free ⁶⁴Cu–PTSM, mice were imaged also after injection of ⁶⁴Cu–PTSM. Extensive studies on the tissue-specific mechanisms of Cu–PTSM extraction and retention have revealed that Cu–PTSM is predominantly metabolized and cleared by hepatocytes, which is reviewed in detail by Blower et al. (9). Accordingly, microPET images of mice i.v.-injected with free ⁶⁴Cu–PTSM showed primarily hepatic clearance of radiocopper at 0.10 h (10.7 ± 0.50%ID/g) and 20.3 h (10.3 ± 1.94%ID/g) (Fig. 5 A and B). In contrast to ⁶⁴Cu–PTSM-labeled C6 cells, injection of free ⁶⁴Cu–PTSM results in minimal activity in the lungs both at 0.10 h (3.30 ± 0.41%ID/g) and 20.3 h (2.34 ± 0.51%ID/g). Organ biodistribution measurements in additional mice injected with either free or C6 cell-labeled ⁶⁴Cu–PTSM corroborates the microPET results (data not shown).

As shown in Fig. 6, ⁶⁴Cu–PTSM also can be used to track the systemic migration of lymphocytes. Similar to C6 cells (Fig. 4), injection of ⁶⁴Cu–PTSM-labeled lymphocytes initially (0.12 and 3.33 h) leads to lung localization of activity (51.8 ± 1.15 and 13.1 ± 0.23%ID/g). By 18.9 h, very little activity remains in the lungs (2.41 ± 0.19%ID/g; Fig. 6). As the lung activity decreases, liver activity increases (9.94 ± 0.25, 18.4 ± 0.64, and 16.7 ± 0.95%ID/g at 0.12, 3.33, and 18.9 h, respectively). In contrast to C6 cells, after injection of lymphocytes activity clearly localizes to the spleen (11.8 ± 0.91, 32.9 ± 0.91, and 9.93 ± 0.42%ID/g at 0.12, 3.33, and 18.9 h, respectively). Activity also is present initially in the kidneys (4.27 ± 0.23 and 4.94 ± 0.18%ID/g at 0.12 and 3.33 h, respectively) but becomes indiscernible by 18.9 h. The gut activity progressively increases (0.44 ± 0.13, 1.85 ± 0.25, and 3.03 ± 1.90%ID/g at 0.12, 3.33, and 18.9 h, respectively). DWBA reveals that activity is present also in the bone marrow (Fig. 6E). This finding suggests that the unidentified signal, seen in the neck and thoracic regions of the third microPET scan, is likely caused by lymphocyte accumulation in the bone marrow.