We have assessed the potential of [18F]fluorothymidine positron emission tomography ([18F]FLT-PET) to measure early cytostasis and cytotoxicity induced by cisplatin treatment of radiation-induced fibrosarcoma 1 (RIF-1) tumor–bearing mice. Cisplatin-mediated arrest of tumor cell growth and induction of tumor shrinkage at 24 and 48 hours, respectively, were detectable by [18F]FLT-PET. At 24 and 48 hours, the normalized uptake at 60 minutes (tumor/liver radioactivity ratio at 60 minutes after radiotracer injection; NUV60) for [18F]FLT was 0.76 ± 0.08 (P = 0.03) and 0.51 ± 0.08 (P = 0.03), respectively, compared with controls (1.02 ± 0.12). The decrease in [18F]FLT uptake at 24 hours was associated with a decrease in cell proliferation assessed immunohistochemically (a decrease in proliferating cell nuclear antigen labeling index, LIPCNA, from 14.0 ± 2.0% to 6.2 ± 1.0%; P = 0.001), despite the lack of a change in tumor size. There were G1-S and G2-M phase arrests after cisplatin treatment, as determined by cell cycle analysis. For the quantitative measurement of tumor cell proliferation, [18F]FLT-PET was found to be superior to [18F]fluorodeoxglucose-PET (NUV60 versus LIPCNA: r = 0.89, P = 0.001 and r = 0.55, P = 0.06, respectively). At the biochemical level, we found that the changes in [18F]FLT and [18F]fluorodeoxglucose uptake were due to changes in levels of thymidine kinase 1 protein, hexokinase, and ATP. This work supports the further development of [18F]FLT-PET as a generic pharmacodynamic readout for early quantitative imaging of drug-induced changes in cell proliferation in vivo.
- [18F] FLT
- PET imaging
- chemotherapy and proliferation
Inhibition of cell proliferation is the pharmacologic objective of most anticancer agents. Defining pharmacodynamic end points of cell proliferation would, therefore, be advantageous in the development of most anticancer agents and pivotal in the development of tumor-targeted therapies. At present, the response of tumors to drug therapy is determined mainly by noninvasive computed tomography or magnetic resonance imaging ( 1– 5). These modalities measure tumor size and do not allow changes in proliferation rate to be determined until several weeks or months into the therapy, thus, preventing quick modification of treatment. This drawback has been overcome more recently through the use of imaging techniques like positron emission tomography (PET) that allow for functional and metabolic changes to be detected early in tumors in response to drug therapy. PET with fluorine-18–labeled fluoro-2-deoxy-d-glucose, [18F]fluorodeoxglucose ([18F]FDG), is now widely used as a noninvasive marker in many oncology and nuclear medicine centers around the world for diagnosis and staging of cancer ( 6– 8). The medium half-life of fluorine-18 (∼110 minutes) makes it ideally suitable for this purpose. Although [18F]FDG-PET has high detection sensitivity, some limitations in specificity, such as the difficulty to distinguish between proliferating tumor cells and inflammation, make it less suitable as a marker for studying drug response ( 9– 11). Shields et al. ( 12) first reported the potential utility of 3-deoxy-3[18F]-fluorothymidine ([18F]-FLT) for the noninvasive detection of cell proliferation by PET. [18F]FLT undergoes monophosphorylation to [18F]FLT-phosphate catalyzed by the cytosolic enzyme thymidine kinase 1 (TK1), which being cell cycle regulated, provides a surrogate measure of cells in S phase of the cell cycle ( 13, 14). Thus, changes in [18F]-FLT uptake can be correlated directly with cell proliferation.
cis-Dichlorodiammineplatinum (II), marketed as cisplatin, is one of the most widely used chemotherapeutic agents in the treatment of cancer, and has proved to be effective against cervical, lung, bladder, and prostate cancer ( 15– 17). Cisplatin inhibits cell proliferation and induces cell cycle arrest by forming interstrand and intrastrand DNA cross-links ( 18). However, the mechanism by which cisplatin regulates TK1 activity in relation to proliferation has not been previously investigated. The objective of the study, therefore, was to assess the effect of cisplatin on tumor [18F]FLT uptake and the TK1 enzyme system (TK1 protein and cofactor, ATP) in relation to cellular proliferation, and compare this to changes seen with [18F]FDG-PET.
Materials and Methods
Radiopharmaceuticals. [18F]FLT and [18F]FDG were produced on-site by Imaging Research Solutions Ltd. (MRC Cyclotron Building, Hammersmith Hospital, London, United Kingdom). [18F]FLT was prepared by radiofluoridation of the 2,3′-anhydro-5′-O-(4,4′-dimethoxytrityl)-thymidine precursor and has been described previously ( 19). [18F]FDG was produced by a stereospecific nucleophilic fluorination reaction followed by deprotection to obtain a no-carrier-added product as reported previously ( 20).
Animals, tumor model, and drug treatment. Six- to eight-week-old male C3J/Hej mice were obtained from Harlan United Kingdom Ltd. (Bicester, United Kingdom) and tumors were induced by inoculation of 5 × 105 radiation-induced fibrosarcoma 1 (RIF-1) cells s.c. on the back. RIF-1 tumor–bearing C3J/Hej mice (n = 8) were treated with vehicle (saline) or cisplatin at a single i.p. dose of 5 mg/kg. Animals were assessed before and at 24 and 48 hours posttreatment. Tumor dimensions were measured continuously using a caliper and tumor volumes were calculated using the equation: volume = (π/6) × a × b × c, where a, b, and c represent three orthogonal axes of the tumor. The treatment experiments were started at 14 to 18 days after inoculating of the RIF-1 cells. Drug treatment was commenced when the tumors had reached a volume of ∼100 mm3. [18F]FLT-PET and [18F]FDG-PET studies were done at 24 or 48 hours after drug treatment.
Immunohistochemical examination of radiation-induced fibrosarcoma 1 tumors. For histologic evaluation of the degree of tumor proliferation, vehicle-treated (n = 4) and cisplatin-treated tumors [24 hours (n = 4) and 48 hours (n = 4) posttreatment] were excised, fixed in formalin, embedded in paraffin, and cut into 5.0 μm sections. Adjacent sections were stained with H&E or with mouse monoclonal antibodies for PCNA (Novocastra, Newcastle upon Tyne, United Kingdom). The primary PCNA antibodies were subsequently detected with biotinylated goat anti-mouse immunoglobulin G (IgG) antibodies (Dako, Carpinteria, CA) and peroxidase-labeled streptavidin (Roche Diagnostics, Lewes, United Kingdom). Peroxidase-labeled cells were visualized by incubation with hydrogen peroxide and diaminobenzidine tetrahydrochloride (Chromogen, Sigma, Poole, United Kingdom). Sections from two different regions of each tumor, separated by at least 1.0 mm, were used for the analyses. The numbers of PCNA- and H&E-positive cells in adjacent sections were counted in eight randomly selected fields of view per section using a BX51 Olympus microscope (Olympus Optical, Tokyo, Japan) at ×400 magnification. The LIPCNA was calculated using the equation LIPCNA = (PCNA-positive cells / H&E-positive cells) × 100%.
Positron emission tomography studies. [18F]FLT-PET and [18F]FDG-PET studies were carried out on a second generation dedicated small animal (quad-HIDAC) PET scanner (Oxford Positron Systems, Weston-on-the-Green, United Kingdom). The features of this instrument have been previously described ( 21). Mice were pretreated with vehicle or cisplatin for 24 or 48 hours before they were scanned. For scanning, the tail veins of the mice were cannulated after induction of anaesthesia with isofluorane/O2/N2O. The animals were placed within a thermostatically controlled jig and positioned prone in the scanner. The jig was calibrated to provide a rectal temperature of ∼37°C. A bolus injection of [18F]FLT or [18F]FDG (80-100 μCi; 2.96-3.7 MBq) was administered i.v. via the tail vein cannula and scanning commenced. Dynamic emission scans were acquired in list-mode format over a 60-minute period. The acquired data were sorted into 0.5-mm sinogram bins and 19 time frames (0.5 × 0.5 × 0.5 mm voxels; 4 × 15, 4 × 60, and 11 × 300 seconds) for image reconstruction, which was done by filtered back projection using a two-dimensional Hamming filter (cutoff, 0.6). The image data sets obtained were transferred to a SUN workstation (Ultra 10, SUN Microsystems, Santa Clara, CA) and visualized using the Analyze software (version 6.0, Biomedical Imaging Resource, Mayo Clinic, Rochester, MN).
Image analysis and quantification of radiotracer uptake. Cumulative images comprising of 30 to 60 minutes of the dynamic data were used for visualization of radiotracer uptake and to define the regions of interest. Regions of interest were defined on five tumor and liver slices (each 0.5 mm thick). Dynamic data from these slices were averaged for each tissue and at each of the 19 time points to obtain a time versus radioactivity curve for these tissues. Tumor radioactivity was corrected for physical decay and normalized to that of liver to obtain the mean normalized uptake value (NUV) from the entire region of interest (as opposed to the maximum NUV). The NUV at 60 minutes postinjection (NUV60) was used for comparisons. The area under the NUV curve was calculated as the integral of NUV from 0 to 60 minutes. The fractional retention of tracer at 60 minutes relative to that at 2.5 minutes was also calculated.
Thymidine kinase 1 Western blot. RIF-1 tumors were pulverized in liquid nitrogen and homogenized in ice-cold Dulbecco's PBS. The samples were then centrifuged at 800 × g (for 30 minutes at 4°C) and supernatants analyzed for protein content using a commercial bicinchoninic acid (BCA) protein assay kit (Perbio Science, Cheshire, United Kingdom). Aliquots of the supernatants (containing 30 μg of protein) were mixed with equivalent volumes of native Tris-glycine sample buffer (Invitrogen Life Science, Paisley, United Kingdom) and separated on a precast Tris-glycine (4-20%) gel (Invitrogen, Life Sciences). Electrophoresis was done at a constant current of 12 mA. The separated proteins were transferred onto a nitrocellulose membrane (Hybond ECL, Amersham Pharmacia Biotech, Little Chalfont, United Kingdom) and subsequently blocked with TBS/casein-blocking buffer [20 mmol/L Tris, 0.5 mol/L NaCl, 1% casein (pH 7.4); Bio-Rad Laboratories, Hercules, CA]. TK1 was detected with a mouse antibody (monoclonal antibody no. 1D11) raised against a synthetic peptide (amino acids K211PGEAVAARKLFAPQ255) that corresponds to a part of the COOH terminus of rodent and human TK1 (Svanova Biotech, Uppsala, Sweden). The nitrocellulose membrane was incubated with the anti-TK1 monoclonal antibody [initial concentration 1 mg/mL; 1:25,000 dilution in TBS-T buffer (20 mmol/L Tris, 150 mmol/L NaCl, and Tween 20 0.1% v/v)] for at least 15 hours, washed with TBS-T, and incubated for 60 minutes with anti-mouse IgG horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, CA) at a dilution of 1:15,000 TBS-T buffer. The membrane was washed again and visualized using the enhanced chemiluminescence method (SuperSignal West Pico Chemiluminescent Substrate, Pierce, Rockford, IL). For quantification of band intensities, the films were scanned using a GS-710 Calibrated Imaging Densitometer (Bio-Rad Laboratories, Hertfordshire, United Kingdom) and analyzed with the Quantity One software (version 4.0.3, Bio-Rad Laboratories). Tumor samples were analyzed in triplicate/native gel, and the analysis was repeated thrice.
Determination of ATP levels in radiation-induced fibrosarcoma 1 tumors. The same supernatants used for the Western Blots above were used for determination of ATP levels. ATP levels were determined by a bioluminescence assay (ENLITEN ATP assay system, Promega Corporation, Southampton, United Kingdom). This assay measures the ATP-dependent emission of light at a wavelength of 560 nm (A560) from a luciferase/luciferin reaction. Light intensity, which is proportional to the ATP concentration, was measured using a microplate luminescence counter (TopCount NXT, Packard Instrument, Meriden, CT) and the results were normalized to the total cellular protein levels, which were determined by the BCA protein assay kit (Perbio Science). A calibration curve was prepared from ATP standards, which were provided with the ATP assay kit.
Hexokinase enzyme activity. Hexokinase enzyme activity was determined by a modification of the method described by Shimke and Grossbard ( 22). Briefly, tumors were ground in liquid nitrogen, homogenized in 50 mmol/L Tris-HCl buffer (pH 7.6) at 4°C, and centrifuged for 1 hour (4°C at 9,000 × g). The supernatant was retained and assayed for total hexokinase activity in a reaction mixture comprising 1 mmol/L NADP, 5 mmol/L glucose, 2 mmol/L MgATP, and 50 mmol/L Tris-HCl buffer (pH 7.6), at 20°C to 25°C, in the presence of 1 unit/mL of glucose-6-phosphate dehydrogenase. This reaction measures the release of NADPH by glucose-6-phosphate dehydrogenase after glucose activation by hexokinase; the rate of change in absorbance at 340 nm was determined over 20 minutes in a UV spectrometer. Results were expressed as nanomoles of NADPH released (by hexokinase-induced glucose-6-phosphate dehydrogenase activity) per minute per milligram of cellular protein.
Cell cycle analysis. The cell cycle distribution and DNA content of RIF-1 tumors were determined by fluorescence-activated cell sorting (FACS). Mice were treated with saline or cisplatin (5 mg/kg). At 24 or 48 hours after drug/vehicle treatment, the mice were injected with BrdUrd (100 mg/kg body weight i.p.). Tumors were excised at 1 hour after injection of BrdUrd, minced, and digested with a 5 mL solution of serum-free RPMI growth medium containing 0.4 mg/mL collagenase type IV (Sigma), 1.0 mg/mL DNase I (Sigma), and 2.5 mg/mL trypsin (Sigma) for 15 minutes in a 37°C shaking water bath. Single cells were harvested, cell density was determined by trypan blue exclusion, and the cells were washed and fixed with ethanol. A BrdUrd Flow kit (BD Biosciences, San Diego, CA) was used to label S-phase cells (anti-BrdUrd) and total DNA (7-AAD). Cell fluorescence was determined using a BD FACS Canto flow cytometer (BD Biosciences) and the cell cycle profiles were determined with the FACS Diva software (BD Biosciences).
Statistical analysis. Statistical analyses were done using the software GraphPad Prism, version 3.03 (GraphPad, San Diego, CA). Differences between vehicle-treated and cisplatin-treated groups with respect to radiotracers uptake, tumor volume changes, LIPCNA, TK1 protein levels, ATP, hexokinase activity, and cell cycle profiles were tested for significance using nonparametric Mann-Whitney test for unpaired data. Correlations between tumor radiotracer levels and LIPCNA were determined by linear regression analyses. A two-tailed P value of ≤0.05 was considered significant.
The effect of cisplatin on radiation-induced fibrosarcoma 1 tumor volume. The biological activity of cisplatin on RIF-1 tumor–bearing mice was determined by measuring changes in tumor volume. Figure 1 shows the growth characteristics of RIF-1 tumors in C3J/Hej mice injected with a single dose of cisplatin, 5 mg/kg i.p. No change in (mean ± SE) tumor size was observed at 24 hours after cisplatin treatment (164 ± 38.3 mm3 compared with 173 ± 39.9 mm3 for vehicle treatment). In contrast, a significant reduction in tumor size was seen at 48 hours after cisplatin treatment (97.3 ± 18.6 mm3, compared with vehicle treatment 189.7 ± 37.4 mm3; P = 0.03, n = 8).
The effect of cisplatin on LIPCNA of radiation-induced fibrosarcoma 1 tumors. Typical control and cisplatin-treated RIF-1 tumor sections immunohistochemically stained for PCNA are shown in Fig. 2A-C . The brown-stained cells represent proliferating tumor cells. The LIPCNA was determined by counting all the positive brown-stained cells in eight randomly selected fields of view and dividing this by the total number of H&E-positive cells in adjacent sections. Compared to vehicle-treated tumors (24 and 48 hours combined), the LIPCNA decreased significantly from 14.0 ± 2.0% to 6.2 ± 1.0% at 24 hours and 1.1 ± 0.1 % at 48 hours after cisplatin (P < 0.001) as shown in Fig. 2D.
Analysis of [18F]FLT-PET and [18F]FDG-PET. Figure 3A-C and D-F shows typical (0.5 mm) transverse slices of [18F]FLT PET and [18F]FDG PET images, respectively, taken from the late (30-60 minutes summed) time frame after i.v. injection of the respective radiotracers in vehicle- and drug-treated RIF-1 tumor–bearing mice. From the four-dimensional imaging data set, time versus radioactivity curves were derived from region-of-interest analysis of the tumor (normalized to liver). Figure 4A and B shows that the intensity of [18F]FDG uptake was higher than that of [18F]FLT. There was a reduction in [18F]FLT and [18F]FDG time versus radioactivity curves at 24 and 48 hours after cisplatin treatment compared with vehicle (24 hours, 48 hours, and vehicle-treated groups combined).
Parameters reflecting radiotracer retention, including NUV60, fractional retention of tracer, and area under the NUV curve, are summarized in Table 1 . In the [18F]FLT study, the mean ± SE NUV60 values decreased in the order vehicle >24 hours post-cisplatin >48 hours post-cisplatin (1.02 ± 0.12 versus 0.76 ± 0.08 versus 0.51 ± 0.08, respectively; P = 0.03). As seen in our previous study ( 21), pretreatment NUV60 values were significantly higher for [18F]FDG than [18F]FLT (P = 0.05). In the [18F]FDG study, NUV60 similarly decreased in the order vehicle >24 hours post-cisplatin >48 hours post-cisplatin (5.67 ± 0.85 versus 2.64 ± 0.36 versus 2.06 ± 0.19, respectively; P = 0.03). Unlike [18F]FLT, however, there was no significant difference in tumor [18F]FDG-derived NUV60 at 24 hours versus 48 hours post-cisplatin treatment. The variables area under the NUV curve and fractional retention of tracer were also found to be useful robust variables for discriminating between vehicle- and drug-treated tumors ( Table 1). Of note, the cytostatic effect of cisplatin (lack of a change in tumor size) at 24 hours posttreatment was detected by both immunohistochemistry and PET imaging.
Correlation between tumor [18F]FLT and [18F]FDG activity and LIPCNA. Given the differences in mechanism of tumor accumulation of the two radiotracers, TK1 activity for [18F]FLT and glucose metabolism for [18F]FDG, we sought to examine the degree of association of tumor radiotracer uptake with cell proliferation. This was done by comparing individual NUV60 values with corresponding LIPCNA. The results showed a linear correlation between tumor [18F]FLT NUV60 and LIPCNA (r = 0.89, P = 0.001; Fig. 5A ). A modest but statistically nonsignificant correlation was seen between tumor [18F]FDG NUV60 and LIPCNA (r = 0.55, P = 0.06); the results in this case were skewed by a few outliers ( Fig. 5B). In addition, there was a significant correlation between LIPCNA and tumor [18F]FLT area under the NUV curve (r = 0.8, P = 0.005), but not with [18F]FDG area under the NUV curve (r = 0.55, P = 0.06). These results suggest that [18F]FLT is superior to [18F]FLT for monitoring early changes in cell proliferation in response to cisplatin compared with [18F]FDG tumors.
Cisplatin treatment alters tumor thymidine kinase 1 enzyme and ATP levels. TK1 activity has been proposed as the most important determinant of cell cycle–regulated [18F]FLT uptake ( 21, 23) Here, we investigated the effect of cisplatin on tumor TK1 protein levels as well as the levels of ATP, which is a cofactor required for the catalytic activity of TK1 enzyme. Figure 6A shows a typical Western blot of TK1 protein in vehicle- and cisplatin-treated RIF-1 tumors. Analysis of the band intensities by densitometry showed that compared with the vehicle group, TK1 protein levels decreased significantly at 24 hours after cisplatin treatment to 24.4 ± 11.07% (P = 0.03), with recovery at 48 hours to levels that were not significantly different from that of control levels ( Fig. 6B). In contrast, the tumor ATP levels, as assessed by bioluminescence, showed a stepwise reduction in the order vehicle > 24 hours post-cisplatin > 48 hours post-cisplatin. ATP levels decreased from 100 ± 0.01% to 69.0 ± 14.8% and 51.5 ± 9.2%, at 24 and 48 hours, respectively (P = 0.03), compared with vehicle ( Fig. 6C).
Cisplatin reduced hexokinase activity in radiation-induced fibrosarcoma 1 tumors. [18F]FDG uptake is a function of the activities of glucose transporters and hexokinase activity ( 24, 25).The latter enzyme is responsible for phosphorylating [18F]FDG. We assessed whether cisplatin altered hexokinase activity in RIF-1 tumors. Compared with control tumors, there was a statistically significant reduction in hexokinase activity at 24 hours after cisplatin treatment (P = 0.01). There was no further change in hexokinase activity at 48 hours (in fact, a slight increase was seen) compared with 24 hours group ( Fig. 7 ). This finding indicates that cisplatin modulates glucose metabolism in RIF-1 tumors, at least in part, by decreasing hexokinase activity.
The effect of cisplatin on cell cycle arrest. We investigated whether the recoveries in TK1 protein and hexokinase activity were due to changes in cell cycle profile. Cell cycle analyses were done on excised RIF-1 tumors. Disaggregation of tumor cells for both control and treated tumors was successful as shown by trypan blue exclusion (data not shown). Figure 8 shows the DNA synthesizing population and cell phase distribution of vehicle- and cisplatin-treated (24 and 48 hours posttreatment) tumor cells as analyzed by FACS. The estimated percentages of cells in G0-G1, S, and G2-M phase phases are shown in Fig. 8B. Cisplatin caused a progressive time-dependent disruption of cell cycle phases with an increase in the G0-G1 population at 24 and 48 hours; 55.1 ± 2.4 % and 59.2 ± 1.1 % (P = 0.05), respectively, compared with control, 52.2 ± 2.2%. There was also a decrease in S-phase cells (P = 0.03) and an increase in G2-M-phase cells (P = 0.05) at 48 hours. Cell cycle arrest at 48 hours posttreatment was accompanied by a significant decrease in S-phase cells synthesizing DNA, as determined by BrdUrd incorporation (22.1 ± 0.3% relative to control, 27.8 ± 2.2%; P = 0.03). There were nonsignificant changes in BrdUrd incorporation at 24 hours post-cisplatin (P > 0.05) compared with control.
We have shown that [18F]FLT-PET can be used to provide an early assessment of chemosensitivity in RIF-1 tumor–bearing mice. It is often the practice to assess patients undergoing treatment with anticancer agents (including chemotherapy, radiation therapy, or biological therapy) by cross-sectional imaging (such as computed tomography, ultrasound, and magnetic resonance imaging) or, for superficial tumors, by caliper measurements at about 3 months after commencing treatment to see if they are responding to the agent. Earlier information on lack of chemosensitivity will enable oncologist to stop treatment and prevent undue toxicity associated with treatment, as well as reduce the cost of treatment. In this study, we evaluated the potential utility of [18F]FLT-PET to provide an early index of chemosensitivity. For this purpose, we employed RIF-1 tumor–bearing mice treated with a single dose of cisplatin (5 mg/kg, i.p.) as a model system. The dynamics of response in the RIF-1 tumors was characterized by tumor stasis at 24 hours posttreatment followed by tumor shrinkage at 48 hours posttreatment. The activity of cisplatin is thought to be a result of inter- and intrastrand DNA cross-links ( 26, 27). Cell death is often preceded by cell type–dependent induction of cell cycle–related proteins including p21WAF1/CIP1, p27KIP1, and p16INK4A, which can induce cell cycle arrest ( 28, 29). Cytostatic concentrations of cisplatin, therefore, result in cell arrest in the G1-S or G2-M phase in vitro ( 28– 30).
We showed that cisplatin-induced tumor stasis and shrinkage were both associated with a reduction in the number of proliferating cells as detected by immunohistochemistry. LIPCNA decreased by 56% at 24 hours posttreatment in RIF-1 tumors despite the lack of change in tumor size. Tumor shrinkage at 48 hours was associated with 93% decrease in LIPCNA compared with controls. Given that PCNA is a protein of which expression is associated with the late G1-S phase of the cell cycle ( 31, 32), we investigated the ability of [18F]FLT-PET to provide an index of proliferation and compared this with [18F]FDG-PET, which is the current most widely used radiotracer technique for clinical imaging of tumors. Of interest, the early cytostatic and cytotoxic effects of cisplatin in this model were detectable by PET imaging. For the two radiotracer, tumor radioactivity was normalized to liver radioactivity in each mouse. The rationale for using NUV was that the quad-HIDAC currently does not give fully quantitative data. Normalization minimizes small changes in detector sensitivity, body weight, and injected radioactivity. Normalization in this case was done by dividing radioactivity in the region of interest (tumor) to the corresponding radioactivity in a reference region. We chose liver as the reference region because (a) the localization of both FDG and FLT within mouse liver is low, albeit of different mechanisms [FDG uptake in liver is low due to high glucose-6-phosphatase levels ( 33); FLT uptake in mouse liver is low due to a lack of glucuronidation in the mouse, compared with humans ( 21)]; and (b) the variability in determining the radiotracer kinetics in liver tissue is low due to its size. We found empirically that normalization reduces the variability in the time versus radioactivity curve for each mouse. However, chemotherapy can potentially affect radiotracer metabolism in the liver. This was checked by analysis of radioactivity levels in tissues excised immediately after PET imaging using γ-counting to obtain an absolute variable, the percent injected activity per gram of tissue. There were no statistically significant differences in liver percent injected activity per gram of tissue for vehicle- and cisplatin-treated mice in both the FDG and FLT studies (data not shown), justifying the use of liver as a reference tissue in this particular study. As reported previously, using this tumor model ( 21) baseline tumor [18F]FDG-derived radioactivity levels were found to be significantly higher than that of [18F]FLT. One reason for the low uptake of FLT compared with FDG in mouse tumors is the high levels of thymidine phosphorylase in mouse plasma, about 9 to 15 times higher ( 34) in comparison with human plasma. In this regard, van Waarde et al. ( 9) have previously shown that treatment with thymidine phosphorylase improves the tumor-muscle ratio by 3.8-fold. In this study, [18F]FDG-PET was more sensitive for detecting RIF-1 tumors than [18F]FLT-PET. The cytostatic effect of cisplatin at 24 hours posttreatment was detectable by [18F]FLT-PET as a 25% decrease in NUV60; there was a 50% decrease at 48 hours when tumor shrinkage was seen. In an in vitro study, we also showed that 1 μg/mL cisplatin inhibited [3H]thymidine uptake in RIF-1 up to 60% as early as 2 hours in culture compared with controls (data not shown). These findings further support our hypothesis that cisplatin may function to inhibit cell proliferation very early before detecting any changes in tumor size. A decrease in NUV60 was also seen for [18F] FDG-PET at 24 hours after treatment. With [18F]FDG-PET, however, there was no significant change in tumor NUV60 at 48 hours posttreatment from that at 24 hours posttreatment. Overall, [18F]FLT-PET correlated better with cell proliferation (LIPCNA) than did [18F]FDG-PET. We have previously reported the superior specificity of [18F]FLT-PET compared with [18F]FDG-PET for detecting proliferation in RIF-1 tumors treated with 5-fluorouracil ( 21). Other investigators have similarly shown this phenomenon of superior specificity of [18F]FLT-PET in untreated patients with lung and gastrointestinal cancers where PET variables were compared with Ki-67 labeling index ( 35– 37).
The rationale for using [18F]FLT-PET for quantification of cell proliferation is based on its substrate specificity for the cell cycle regulated protein, TK1. Using TK1 −/+ and −/− L5178Y tumors, we have recently shown that TK1 and ATP levels determine the in vivo uptake of [18F]FLT-PET ( 38). TK1 activity is stimulated at the G1-S border through transcriptional regulation by E2F, as well as enhanced catalytic activity through the formation of a tetramer with ATP ( 13, 14, 39– 41). Posttranslation modifications ( 42, 43) and/or low ATP levels seen in vivo ( 21) may also determine the lower affinity of phosphorylated TK1 in late S-phase and G2-M arrested cells, despite the higher TK1 protein levels in these cells. Posttranslational modifications at the carboxyl terminus of TK1 provide the recognition signal for degradation of TK1 protein in late M phase, such that TK1 protein is low in newly divided G1 cells ( 42– 44). In this study we assessed the regulation of TK1 and ATP levels by cisplatin and found an initial decrease in TK1 protein levels at 24 hours after cisplatin treatment in keeping with the reduction in [18F]FLT-PET NUV60 and LIPCNA at 24 hours after treatment. In contrast to the further decrease in [18F]FLT-PET NUV60 and LIPCNA at 48 hours posttreatment, there was an increase in TK1 protein. This recovery in protein levels was also seen previously after RIF-1 tumor–bearing mice were treated with 5-fluorouracil ( 21). The reason for this increase in TK1 protein levels at 48 hours is unclear. We speculate that this could be due to accumulation of a hyperphosphorylated form of TK1, as suggested by Chang et al. ( 42, 45), which is not readily targeted for degradation. In most cells including HeLa, TK1 protein levels increase at the G1-S border and reach a maximum during the M phase but the catalytic activity decreases at G2-M as replication is completed at this stage ( 42); in M-phase arrested HeLa cells, TK1 was heavily phosphorylated (8-fold over controls) but the phosphorylating pattern was different from that of asynchronized growing cells, and this may suppress degradation ( 42). Such an effect will be particularly pronounced if cells are arrested in G2-M of the cell cycle. A finding of a G2-M block in our study thus supports this speculation; the G1-S arrest (G0-G1 accumulation of cells) will potentially lead to a decrease in TK1 protein levels, suggesting the presence of mixed effects. The antibody that was used for determining TK1 protein was raised against the carboxy terminus of the protein; it could also recognize the differentially phosphorylated forms. Clearly, raising specific antibodies that will recognize phosphorylated epitopes of TK1 requires further investigation. Alternatively, DNA repair after a single dose of the drug ( 46– 48) may also account for the increased TK1. However, this is unlikely given the progressive decrease in BrdUrd incorporation. Whatever the mechanism, it is unlikely that the increased levels TK1 protein will correspond to an increased catalytic activity due to the requirement of TK1 for ATP ( 49, 50); the ATP levels in the tumors decreased in a stepwise fashion at 24 and 48 hours post-cisplatin. In the case of [18F]FDG-PET, hexokinase activity was found to decrease initially followed by a small but slight increase. Given the conditions of decreasing ATP levels within the tumor, it is possible that the overall enzyme activity will continue to decrease and, thus, account for the modest (nonsignificant) decrease seen in [18F]FDG-PET NUV60 at 48 hours compared with 24 hours. These findings therefore support the hypothesis of a coordinated interplay of TK1, hexokinase, and ATP levels to modulate the catalytic activity of the enzymes and, therefore, tumor [18F]FLT and [18F]FDG uptake.
In conclusion, we have shown in this mouse tumor model the suitability of [18F]FLT-PET to measure cytostatic and cytotoxic effects of cisplatin. [18F]FLT-PET was superior to [18F]FDG-PET for imaging changes in proliferation associated with early response. A combination of changes in TK1 protein levels, hexokinase activity, and ATP levels accounted for the radiotracer kinetics in control and treated tumors. This work strongly supports the use of [18F]FLT-PET for imaging early response to anti-cancer drug treatment.
Grant support: Cancer Research United Kingdom grant C2536/A3554 and the Medical Research Council (Core funding).
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Received November 18, 2004.
- Revision received February 11, 2005.
- Accepted March 4, 2005.
- ©2005 American Association for Cancer Research.