In vivo Biological Activity of the Histone Deacetylase Inhibitor LAQ824 Is detectable with 3′-Deoxy-3′-[¹⁸F]Fluorothymidine Positron Emission Tomography

Histone deacetylases (HDAC) are expressed in all organisms. They are evolutionarily conserved from prokaryotes to eukaryotes and together with acetyl transferases (HAT) in complexes with sequence-specific cofactors, control chromatin condensation, and gene expression (1–5). Eleven HDAC family members have been identified to date that are classified into the Zn-dependent class 1 and class 2 enzymes and the Zn-independent, NAD-dependent class 3 enzyme (6). The substrate for HDACs and HATs include ε-amino groups of lysine residues located in the NH₂-terminal regions of histones. In general, the positively charged hypoacetylated histones in the nucleosomes bind tightly to the phosphate backbone of DNA to maintain chromatin in a transcriptionally silent state, thereby preventing access of transcriptional regulatory complexes and RNA polymerases to DNA. Conversely, acetylation neutralizes the positive charge on the histones to form a more open configuration, thus allowing gene expression (4, 7, 8). Several reports have shown an association between altered HAT and HDAC activities and progression of colorectal cancers and promyelocytic leukemia (5, 9, 10).

Over the past few years, a number of HDAC inhibitors (HDACI) have been identified that inhibit the growth of cancer cells in vitro and in animal models (11–17). The HDACIs affect tumor growth by modulating gene expression through reversible inhibition of HDACs (13, 15, 18–20). For instance, LAQ824 is a structurally novel hydroxyamic acid derivative (Fig. 1A), which inhibits HDAC at nanomolar to low-micromolar concentrations in human colon cancer and myeloma cell lines (21). LAQ824 has also been shown to have efficacy in vivo in murine leukemia and myeloma models (22, 23). The biological effects of HDACIs on cancer cells include suppression of transformed cell morphology and inhibition of cell proliferation by induction of cell cycle arrest, differentiation, and/or apoptosis (24–28). The effects of HDACIs on gene expression are highly selective, leading to transcriptional activation of certain genes such as the cyclin-dependent kinase inhibitor CDKNIA that encodes p21WAF1/CIP1 (29, 30), but repression of others including the CCND1 gene encoding cyclin D1 (30, 31). Cyclin D, cyclin E, and p21 regulate the tumor suppressor retinoblastoma gene product (Rb), which when hyperphosphorylated (ppRB) releases E2F transcription factors (32–34) to activate other downstream genes and the protein products of such genes. E2F transcription factors are also targets for acetylation (35). Genes known to be regulated by the Rb/E2F system include thymidine kinase 1 (TK1; Fig. 1B; refs. 36, 37).

Lessons from the clinical development of HDAC inhibitors, including suberoylanilide hydroxamic acid, in patients with advanced cancer have indicated that such agents may predominantly induce tumor stabilization (34, 38, 39) with tumor shrinkage occurring in a few cases. In clinical studies, histones isolated from peripheral blood mononuclear cells and tumor biopsies were hyperacetylated (8, 38, 39). It is, however, most desirable to show such biological effects in tumors; also, the exact relationship between global acetylation and the biological effect of HDACIs on gene expression is not altogether clear. Given the difficulty in routinely obtaining biopsy material for these and other biochemical marker studies, we aimed to validate a noninvasive imaging method that measures a cognate biochemical effect of HDAC inhibition—the activity of the E2F-regulated TK1 protein product. As with the development of other molecularly targeted therapies, there is an increasing demand to incorporate such imaging methods in early clinical trials (40). Positron emission tomography (PET) imaging of TK1 activity is possible with [¹⁸F]fluorothymidine ([¹⁸F]FLT), a substrate for the enzyme (41–44). The level of TK1 protein is an important determinant of [¹⁸F]FLT uptake in tumors. We previously showed that this was the case in mouse lymphoma tumors containing the functional heterozygous TK1+/− allele versus the corresponding TK−/− variant derived from mutation of the TK1 locus (41). Barthel et al. (45), Leyton et al. (46), and Waldherr et al. (47), have shown the utility of [¹⁸F]FLT-PET for imaging the biological effects of 5-flurorouracil, cisplatin, and the pan-Erb B inhibitor PK1-166, respectively, in animal models of cancer. The advantage of [¹⁸F]FLT-PET over conventional imaging modalities, such as computed tomography and magnetic resonance imaging, lies in the ability of [¹⁸F]FLT-PET to provide information on tumor biology, effects that may occur much earlier than size changes (48).

The exact relationship between global acetylation, the current most widely used pharmacodynamic marker for HDAC inhibitor activity, and the biological effect of HDACIs on gene expression is not altogether clear. Thus, there is a need for newer pharmacodynamic end points. In our previous study (46), we focused on the use of [¹⁸F]FLT-PET for imaging chemotherapeutic agents that are largely cytotoxic, the time course of this effect, and comparison with the standard of PET imaging of drug response, [¹⁸F]FDG-PET. In this study, (a) we asked if [¹⁸F]FLT-PET could be used for imaging the biological effect of a class of molecularly targeted cell cycle inhibitors, HDAC inhibitors, that are largely cytostatic; and (b) because HDAC inhibitors act at the transcriptional level to modify gene expression, we asked if we could define the signaling mechanisms responsible for the changes in [¹⁸F]FLT kinetics by comparing the PET readout to biochemical measures of LAQ824 activity. We also assessed if early changes detected by FLT-PET correlate with changes in cell proliferation.

Radiopharmaceuticals. [¹⁸F]FLT was prepared on-site by Hammersmith Imanet Limited (MRC Cyclotron Building, Hammersmith Hospital, London, United Kingdom) by radiofluorination of the 2,3′-anhydro-5′-O-(4,4′-dimethoxytrityl)-thymidine precursor as previously described (49). All samples had >99% radiochemical purity as determined by high-performance liquid chromatography with radiochemical detection, and the specific radioactivity ranged between 24 and 465 GBq/μmol.

Cell cycle analysis. HCT116 cells were cultured in RPMI growth medium containing 10% (v/v) fetal bovine serum, 2 mmol/L l-glutamine, 100 units/mL penicillin, and 100 g/mL streptomycin and grown in a 5% CO₂ incubator at 37°C. Cells in exponential growth were used for subsequent studies. For cell cycle analysis, HCT116 cells (1 × 10⁶ each) were seeded in 10-cm-diameter Petri dishes and exposed to vehicle (0.1% DMSO) or increasing concentration of LAQ824 in vehicle (0.001, 0.01, 0.1, and 1.0 μmol/L). Studies were done in triplicate. After 48 hours of treatment, cells were harvested (adherent cells were scraped from the plates; floating cells were collected from the supernatant) and pelleted by centrifugation (500 × g for 5 minutes at 4°C). The cells were briefly suspended in ice-cold 90% ethanol and incubated in propidium iodide solution (70 μmol/L propidium iodide, 38 mmol/L sodium citrate, and 20 μg/mL RNase A) at 37°C for 30 minutes. Cells were analyzed by flow cytometry using a FACSCalibur flow cytometer (Becton Dickinson Immunocytometry System, San Jose, CA) and cell cycle profiles were subsequently determined using the CellQuest software (Becton Dickinson).

Tumor model and drug treatment. The experiments were done by licensed investigators in accordance with the United Kingdom Home Office's Guidance on the Operation of the Animal (Scientific Procedures) Act 1986. We used a human colon carcinoma xenograft model, HCT116, growing in BALB/c nu/nu mice to assess the biological activity of LAQ824. Six- to eight-week-old female BALB/c nu/nu mice were obtained from Harlan United Kingdom, Ltd. (Bicester, United Kingdom), and tumors were induced by inoculation of 5 × 10⁶ HCT116 cells s.c. on the back. Tumor dimensions were measured using a caliper and tumor volumes were calculated by the following equation: volume = (π/6) × a × b × c, where a, b, and c represent three orthogonal axes of the tumor. Mice were used when their tumors reached ∼100 mm³. LAQ824 in the form of the free base was kindly provided by Novartis Pharmaceuticals Corporation (Cambridge, MA). A stock solution of LAQ824 was prepared in DMSO and further diluted with PBS to give the final dose solution containing 10% DMSO. For imaging studies, size-matched tumor-bearing mice were randomized into groups of six and administered LAQ824 or vehicle (10% DMSO in PBS) i.p. once daily until imaging; the final dose was given 1 hour before [¹⁸F]FLT-PET scanning. We examined the dose-related effect of the drug by imaging at day 4 of treatment using 5 or 25 mg/kg/d schedules. In addition, we examined the time course of drug action at days 2, 4, and 10 following treatment with 5 mg/kg/d LAQ824. Mice did not tolerate daily i.p. dosing with the 25 mg/kg/d schedule beyond day 4 of treatment.

PET studies. [¹⁸F]FLT-PET studies were carried out on a dedicated small-animal PET scanner, quad-HIDAC (Oxford Positron Systems, Weston-on-the-Green, United Kingdom). The features of this instrument have been described previously (45). For scanning, the tail veins of vehicle- or LAQ824-treated mice were cannulated after induction of anesthesia 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 (80-100 μCi; 2.96-3.7 MBq) was administered i.v. via the tail vein cannula and scanning commenced. Dynamic 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 draw regions of interest. Regions of interest were defined on five tumor and five heart slices (each 0.5 mm thickness). Dynamic data from these slices were averaged for each tissue and at each of the 19 time points to obtain time versus radioactivity curves (TAC) for these tissues. Tumor radioactivity was corrected for physical decay and normalized to that of heart to obtain a normalized uptake value (NUV); LAQ824 had no effect on myocardial radiotracer uptake (data not shown). The NUV at 60 minutes postinjection (NUV60) was used for comparisons. The area under the NUV curve (AUC) was calculated as the integral of NUV from 0 to 60 minutes. The fractional retention of tracer (FRT) at 60 minutes relative to that at 1.5 minutes was also calculated.

Immunohistochemical examination of HCT116 tumors. For histologic evaluation of the degree of tumor proliferation, vehicle-treated tumors (n = 6) and tumors treated with LAQ824 at 5 mg/kg (n = 6) or 25 mg/kg (n = 6) were excised after imaging, fixed in formalin, embedded in paraffin, and cut into 5.0 μm sections. Each section was stained with both H&E and with mouse monoclonal antibody for Ki67 (Novocastra, Newcastle upon Tyne, United Kingdom). The primary Ki67 antibody was subsequently detected with biotinylated goat anti-mouse IgG antibody (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 analysis. The total number of cells and the number of Ki67-positive cells were counted in eight randomly selected fields of view per section using an Olympus BX51 microscope (Olympus UK, Ltd., London, United Kingdom) at ×400 magnification. The Ki67 labeling index (LIKi67) was calculated using the equation LIKi67 = (Ki67-positive cells / total number of cells) × 100%.

Western blot analysis. Western blot analyses were done in both cultured cells and tumor homogenates. In the cell culture study, HCT116 cells were exposed to vehicle or increasing concentration of LAQ824 (0.001, 0.01, 0.1, and 1.0 μmol/L) as per the flow cytometry studies above. Cells were harvested (adherent cells were scraped from the plates; floating cells were collected from the supernatant) and pelleted by centrifugation (500 × g for 5 minutes at 4°C). The cells were washed with PBS and stored at −80°C for before Western blot determination of TK1 protein levels. The levels of acetylated histone H4, pRb, ppRb, p21, and TK1 were determined in excised tumors by Western blot. The assays were done on vehicle- and LAQ824-treated HCT116 tumors that were excised after PET imaging and snap frozen. The tumors were pulverized in liquid nitrogen and homogenized in 100 μL ice-cold triple detergent [50 mmol/L Tris-HCl (pH 8.0), 150 nmol/L NaCl, 0.02% sodium azide, 0.1% SDS, 1% NP40, 0.5% sodium deoxycholate, and 10 μL/mL protease inhibitor mixture]. For TK1 determination, the tumors were homogenized in 100 μL of the above lysis buffer but without the addition of 0.1%SDS. Lysates were clarified by centrifugation (15,000 × g for 10 minutes 4°C) and the supernatants were analyzed for total protein content using a commercial BCA protein assay kit (Perbio Science, Ltd., Cheshire, United Kingdom). Samples were stored at −80°C before further analysis. For Western blot analysis, 30 μg of protein were resolved on precast 4% to 12% Bis-Tris gradient gel (Invitrogen, Ltd., Paisley, United Kingdom); for analysis of pRb and ppRb, separation was done using 12% SDS gel and transferred onto polyvinylidene difluoride membranes (Invitrogen, Ltd., Paisley, United Kingdom). TK1 protein levels were determined on 4% to 20% precast native Tris-glycine gels (Invitrogen Ltd., Paisley, United Kingdom) as described previously (45, 46, 50). Membranes were blocked overnight at 4°C in blocking buffer [5% (w/v) nonfat milk, 150 mmol/L NaCl, 10 mmol/L Tris (pH 8.0), and 0.05% (v/v) Tween 20]. Proteins were detected by incubation with primary antibodies diluted in blocking buffer at room temperature for 1 hour. The primary antibodies included rabbit polyclonal anti-pRb (C-15) antibody (Santa Cruz Biotechnology, Santa Cruz, CA; 1:2,000) and mouse monoclonal anti–phospho ppRb (Thr⁸²¹; Transduction Labs, 1:2,000 dilution; ref. 51), as well as rabbit polyclonal antiacetylated histone H4 antibody (Upstate, Lake Placid, NY; 1:1,000 dilution), mouse monoclonal anti-p21 (Autogen Bioclear, Caine, United Kingdom; 1:200 dilution), goat polyclonal anti-β-actin (Autogen Bioclear; 1:200 dilution), and mouse biotinylated monoclonal anti-TK1 (Svanova Biotech, Uppsala, Sweden; 1:20,000 dilution). The membranes were washed and incubated with IgG horseradish peroxidase (HRP) at 1:5,000 dilution (for anti-rabbit HRP; Amersham, Buckingham, United Kingdom), 1:1,000 (for anti-mouse HRP, Santa Cruz Biotechnology), 1:3,000 (for anti-goat HRP; Santa Cruz), and 1:20,000 (for avidin-HRP; Amersham). Bands were visualized by enhanced chemiluminescence (SuperSignal; Perbio Science). For quantification of band intensities, the films were scanned using a GS710 Calibrated Imaging Densitometer (Bio-Rad Laboratories, Hertfordshire, United Kingdom) and analyzed with the Quantity One Software (version 4.0.3, Bio-Rad Laboratories). Six tumor samples were analyzed from each of the treatment groups.

Determination of TK1 mRNA levels by quantitative real-time PCR. RNA from frozen tumor samples was extracted with TRIzol (Invitrogen Life Technologies, Paisley, United Kingdom) as per the directions of the manufacturer with a slight modification that the aqueous phase was cleaned-up with a chloroform/isoamyl alcohol mixture (24:1). Extracted RNA was dissolved in appropriate volumes of water and heated at 50°C for 2 minutes using a dry heating block to facilitate dissolution and to evaporate any traces of alcohol. The RNA was then treated with RNase-free DNase (RQ RNase-free DNase, Promega, Madison, WI) for 30 minutes to remove any genomic DNA contamination. DNase was inactivated by heating at 70°C for 10 minutes and the RNA was quantified spectrophotometrically. RNA (5 μg) was reverse transcribed using random primers (Amersham Pharmacia Biotech, Bucks, United Kingdom) and M-MLV RTase (Invitrogen Life Technologies) according to the instructions of the manufacturers with slight modifications. DTT was omitted from the reaction mix because its presence was found to reduce the efficiency of the real-time PCR.

Real-time PCR was done with a LightCycler (Roche Diagnostic Corp., Indianapolis, IN). The primers used for TK1 were as follows: forward primer 5′-AGAGTACTCGGGTTCGTGAACTT-3′, reverse primer 5′-CACTTGTACTGAGCAATC TGGAAG-3′; those used for the housekeeping gene, hypoxanthine phosphoribosyl transferase (HPRT), were as follows: forward primer 5′-TGTTGTAGGATATGCCCTTGACTA-3′; reverse primer 5′-GTCAATAGGACTCCAGATGTTTCC-3′. Sigma Jumpstart SYBR Green mix was used for amplifications that were done at an annealing temperature of 56°C, extension at 72°C, and fluorescence acquisition at 81°C. Transcripts of HPRT were used for normalization to account for any variation in reverse transcriptase efficiency or RNA loading. Copy numbers of the genes were calculated using a plasmid containing the amplified part of the HPRT gene as standard. TK1 mRNA levels were expressed as fold-change in copy number per microgram of RNA.

Determination of ATP levels. HCT116 tumor supernatants were prepared as described above for Western blot and used in this assay. 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.

Statistical analysis. Statistical analyses were done using the software GraphPad Prism, version 3.03 (GraphPad, San Diego, CA). Differences between vehicle- and LAQ824-treated groups with respect to radiotracer uptake, tumor volume, LIKi67, ppRb, pRb, p21, TK1 protein levels, and ATP were tested for significance using nonparametric Mann-Whitney test. Correlations between tumor radiotracer levels and LIKi67 were determined by linear regression analysis (Pearson test). Two-tailed P values ≤0.05 were considered significant.

Effect of LAQ824 on cell cycle profile and TK1 protein levels in HCT116 cells. LAQ824 induced a reduction in TK1 protein levels in HCT116. Figure 2A shows Western blots of TK1 and β-actin (control) proteins. A dose-related reduction in protein levels of TK1 is evident. A summary of the changes in TK1 protein levels normalized to β-actin from triplicate experiments is shown in Fig. 2B. This degree of inhibition was reflected in the cell cycle distribution profile of HCT116 cells as determined by flow cytometry analysis (Fig. 2C and D). The cells arrested in the G₁ phase at low concentration (0.001 and 0.01 μmol/L). At high concentrations (0.1 and 1.0 μmol/L), a high proportion of the cells were apoptotic (<2N DNA; sub-G₁; Fig. 2D). These findings supported the utility of a radiolabeled TK1 substrate for detecting biological effects of LAQ824 in vivo.

Effect of LAQ824 on tumor volume. LAQ824 inhibited the growth of HCT116 tumors in vivo. Figure 3A compares the growth of tumors in vehicle and LAQ824 (5 mg/kg)–treated groups over 10 days. In this time course study, no significant differences in tumor volume were seen at day 2 or 4 between the vehicle- and LAQ824 (5 mg/kg)–treated groups; there was a nonsignificant reduction in volume at day 4 following LAQ824. However, by day 10 of treatment, the LAQ824-treated group showed significantly lower tumor volumes compared with the vehicle group (395.4 ± 64.2 versus 249 ± 47.9 mm³, respectively; P = 0.03). Figure 3B shows the effect of dose on tumor size when animals were dosed with vehicle or LAQ824 for 4 days. At day 4, the tumor size for 25 mg/kg groups was significantly reduced compared with the vehicle-treated group (87.7 ± 13.9 versus 203.4 ± 21.9 mm³, respectively; P = 0.03). Taken together, these findings show that LAQ824 affected tumor growth in a dose- and time-dependent manner.

The intensity of [¹⁸F]FLT uptake is altered by LAQ824 treatment. Typical (0.5 mm) transverse [¹⁸F]FLT-PET images of HCT116 tumor-bearing mice are shown in Fig. 4A, to C. Within the field of view shown, tumor regions had the highest intensity of radioactivity. A qualitative reduction in [¹⁸F]FLT uptake is seen with LAQ824 treatment. The radioactivity in tumor regions normalized to that of the heart of the same animal (normalized uptake value, NUV) was used quantitatively to compare the effects of LAQ824 treatment. NUV increased over time in all tumors (Fig. 4D). LAQ824 induced a dose-dependent reduction in [¹⁸F]FLT kinetics. Kinetic variables are shown in Table 1 for both the time- and dose-dependent studies. All three kinetic variables that reflect the retention of [¹⁸F]FLT were found to discriminate between vehicle- and LAQ824-treated tumors in the order vehicle >5 mg/kg LAQ824 > 25 mg/kg LAQ824 in the dose-dependent study. In the time course study, there was a significant group difference in NUV60 between vehicle- and drug-treated groups at days 4 and 10 of treatment but not at day 2 (Table 1).

LAQ824 inhibits cell proliferation. The expression levels of the proliferation marker, Ki67, for the vehicle- and LAQ824-treated tumor sections are shown in Fig. 5A  to C.

Summary data are presented in Fig. 5D. LAQ824 induced a qualitative and quantitative dose-dependent reduction in LIKi67. LIKi67 decreased significantly to 38.0 ± 2.2% compared with that of vehicle-treated tumors after 5 mg/kg LAQ824 (P = 0.004). A more pronounced reduction in LIKi67 to 21.2 ± 4.2% compared with that of vehicle-treated mice was observed after 25 mg/kg LAQ824 (P = 0.002).

Relationship between tumor [¹⁸F]FLT kinetics and LIKi67. We sought to examine the degree of association between tumor radiotracer kinetics and cell proliferation as determined by Ki67 staining. This was done by comparing individual NUV60 values from the cohort of vehicle- and LAQ824-treated tumors with their corresponding LIKi67. Figure 5E shows that [¹⁸F]FLT NUV60 correlates positively with LIKi67 (r = 0.68; P = 0.0019). There were also positive correlations between [¹⁸F]FLT AUC and LIKi67 (r = 0.50; P = 0.013) and also between [¹⁸F]FLT FRT and LIKi67 (r = 0.64; P = 0.004).

In vivo biochemical targets of LAQ824. The acetylation status of histone H4, as well as the levels of p21, ppRb, pRb, and TK1 proteins, was examined in tumors excised after PET imaging (Fig. 6A). Figure 6A shows the acetylation status of histone H4 in HCT116 tumors. Our analysis indicated that histones were targets of LAQ824 in vivo. There was a significant difference in acetylation of histone H4 (both monoacetylated and hyperacetylated) between 5 mg/kg (P = 0.004) and 25 mg/kg (P = 0.002) LAQ824 compared with the vehicle-treated levels as determined by densitometer measurements (Fig. 6B and C). LAQ824 increased p21 levels (Fig. 6A). The densitometer measurements were significantly higher for p21 at both LAQ824 doses (P = 0.018 compared with vehicle; Fig. 6D). The drug decreased Rb hyperphosphorylation and conversely increased Rb hypophosphorylation (Fig. 7A). The densitometer measurements were significantly lower for cyclinE-CDK2-dependent phosphorylation of pRb on Thr⁸²¹ at both LAQ824 doses (P = 0.05 compared with vehicle; Fig. 7B), although this was not dose dependent. Under the conditions used, a band shift allowed resolution of total ppRb from that of pRb (Fig. 7A). There was a dose-dependent increase in pRb at both 5 and 25 mg/kg LAQ824 (P = 0.05; Fig. 7C), and a decrease in ppRb at both 5 mg/kg (2.0-fold compared with vehicle, P = 0.05; Fig. 7D) and 25 mg/kg LAQ824 (2.5-fold, P = 0.05). Thus, our analysis indicated that histone H4, p21, and Rb are targets of LAQ824 in vivo.

Of interest to the imaging of tumors with [¹⁸F]FLT, we investigated the effect of LAQ824 on tumor TK1 protein levels as well as the levels of ATP, which is a cofactor required for the catalytic activity of the TK1 enzyme (41). Figure 8A shows a typical Western blot of TK1 protein in vehicle- and LAQ824-treated HCT116 tumors. Analysis of the band intensities by densitometry (Fig. 8B) showed that compared with the vehicle group, TK1 protein levels decreased in a dose-dependent manner after treatment with LAQ824; this effect was statistically significant at the higher dose level only (P = 0.015). Figure 8C shows a time course modulation of TK1 levels over 10-day treatment with LAQ824. The TK1 levels were significantly inhibited by day 10 after daily treatment of 5 mg/kg LAQ824 compared with vehicle (P = 0.05) although not significantly at day 2 and day 4 compared with their respective vehicle-treated groups. ATP levels were similar in vehicle and 5 mg/kg LAQ824–treated groups at all time points (2, 4, and 10 days); at day 10, the ATP levels in both the vehicle and LAQ824 groups were ∼70% of day 4 levels, presumably due to the increased size of tumors and development of necrosis (data not shown). At day 4, there were no significant differences in ATP at both dose levels (5 versus 25 mg/kg; Fig. 8D). These results were in contrast to the effects of chemotherapeutic agents like 5-fluorouracil and cisplatin on tumor ATP levels (45, 46).

LAQ824 decreases TK1 mRNA expression in HCT116 tumors. The expression of TK1 mRNA in HCT116 tumors decreased in a dose-dependent manner in response to LAQ824 (Fig. 8E). Quantitative real-time PCR measurements showed a significant 1.3-fold decrease in TK1 mRNA copy number for the 5 mg/kg LAQ824–treated mice (62,207 ± 3,810 copy number/μg RNA) compared with vehicle-treated mice (80,384 ± 6,038 copy number/μg RNA; P = 0.05), and a 2.6-fold decrease in the 25 mg/kg LAQ824–treated tumors (32,006 ± 3,821 copy number/μg RNA) compared with vehicle (P = 0.05). These finding are consistent with the down-regulation of TK1 protein levels, as determined by Western blot.

We have shown that the biological activity of the HDACI LAQ824 can be quantified noninvasively in a mouse model of colon cancer by [¹⁸F]FLT-PET. The measurement of appropriate biomarkers is important in the preclinical and clinical development of anticancer drugs; however, due to the challenges associated with the development of such biomarkers, their use in early clinical trials is disappointingly low (40, 52). [¹⁸F]FLT-PET is being validated for preclinical and clinical noninvasive imaging of tumor cell proliferation. Our group has shown that TK1 protein and ATP are required for [¹⁸F]FLT uptake in vivo (41), in keeping with the in vitro studies of Rasey et al. (42) and Toyohara et al. (43). Furthermore, we have shown that treatment with cytotoxic agents, such as cisplatin and 5-fluorouracil, leads to a reduction in both [¹⁸F]FLT uptake and cellular proliferation as determined by immunostaining (41, 43, 46). These, together with baseline clinical studies demonstrating high correlations between [¹⁸F]FLT uptake and Ki67 immunostaining (53–58) suggest that [¹⁸F]FLT may be a useful generic marker for imaging cell proliferation in animal models of cancer and in cancer patients. We rationalized that, given the mechanism of HDAC inhibitor activity, we should see changes in thymidine kinase transcription and translation detectable with [¹⁸F]FLT-PET. We did not make comparisons with [¹⁸F]FDG in this present study because we cannot propose a similar hypothesis for the determinants of [¹⁸F]FDG uptake, Glut 1 and hexokinase. Also, in our previous studies (45, 46), we found a poorer correlation between cell proliferation and [¹⁸F]FDG uptake compared with [¹⁸F]FLT uptake.

Initial studies in this report showed that TK1 protein was modulated by LAQ824 in vitro at concentrations of the drug that caused both growth arrest and cell death. This finding supported the in vivo utility of [¹⁸F]FLT imaging for detecting the biological effects of LAQ824. LAQ824 was used in vivo at doses that were associated with hyperacetylation of histone H4, a unique biochemical signature of HDACIs. It should be noted, however, that although the changes in global histone acetylation occurred in a dose-dependent manner, the exact relationship between global histone acetylation by HDACIs and antitumor activity or gene expression is not altogether clear (34). On the other hand, because the uptake of [¹⁸F]FLT is dependent on TK1 (41–43), drugs that transcriptionally regulate TK1 levels, directly or indirectly, are good candidates for monitoring by [¹⁸F]FLT-PET. We hypothesized that this will be the case for HDACIs like LAQ824 (Fig. 1B). This hypothesis was strengthened by the observation of increased p21 levels and decreased ppRb levels. Both global changes in ppRb levels and cyclinE-CDK2–dependent phosphorylation of pRb on Thr⁸²¹ were seen; down-regulation of cyclin D but not cdk4 was also seen (data not shown). These properties of LAQ824, as well as other potential effects of HDACIs discussed earlier, may lead to transcriptional regulation of TK1. We showed that transcriptional mechanisms may play a role in the regulation of TK1 activity, as both TK1 mRNA and protein levels decreased after treatment with LAQ824 in a dose-dependent manner. In this model system, TK1 levels were modulated in a time- and dose-dependent manner following daily treatment compared with vehicle. We cannot, however, rule out contribution of posttranslational modifications of TK1 protein in regulating protein activity due to lack of appropriate phosphospecific antibodies. TK1 is subject to posttranslational modifications (59–61). These modifications and/or low ATP levels seen in vivo (62) with certain drugs (45, 46) may alter TK1 activity; however, no differences in ATP levels between vehicle- and drug-treated groups were observed in this case.

Given the observed changes in TK1 levels, it was perhaps not surprising to see a dose-dependent decrease in [¹⁸F]FLT uptake in tumors treated with LAQ824. In untreated tumors, [¹⁸F]FLT NUV60 increased over time. Treatment with 5 mg/kg LAQ824 had no effect at day 2 but caused a reduction in NUV60 at days 4 and 10. The effects of the drug on [¹⁸F]FLT uptake determined at 4 days after commencing treatment decreased in the order vehicle >5 mg/kg LAQ824 > 25 mg/kg LAQ824. Changes in [¹⁸F]FLT NUV60, therefore, mirrored the changes in TK1 protein levels. The kinetic variable AUC and FRT also showed this trend.

The time resolution versus noise characteristics of the PET data did not permit us to examine first passage of radiotracer from the tail vein to the heart and liver. The overall radiotracer time courses in heart and liver (continuous decrease) in this study, as well as in our previous studies with cisplatin and 5-fluorouracil, do not suggest any selective uptake in those tissues (45, 46). Normalizing the tumor TACs to heart TACs (mainly blood pool), given the low levels of myocardial uptake of [¹⁸F]FLT (45), reduces potential variability that could result from interindividual differences in radiotracer disposition, small errors in injected radioactivity, small changes in count rate during the 60-minute data acquisition period, as well as any potential changes in systemic radiotracer clearance. The quantitative normalized radioactivity variables obtained predicted the degree of growth delay observed after drug treatment. The finding that [¹⁸F]FLT-PET can be used to detect changes in growth delay is important as most HDACIs, given their proposed mechanisms of action, are likely to induce tumor stasis rather than shrinkage (34). An early marker of biological activity or efficacy like [¹⁸F]FLT-PET could be useful in the clinic soon after commencing therapy to determine if patients are responsive to treatment. This technology should be superior to determination of tumor size changes by cross-sectional imaging, which could take several weeks to allow for a change in patient management to be made. In keeping with the changes in [¹⁸F]FLT uptake, tumors excised after PET imaging also showed a reduction in cell proliferation as determined by Ki67 immunostaining. As previously shown for other model systems in vivo (45, 46), there was a statistically significant correlation between LIKi67 and [¹⁸F]FLT uptake (P = 0.003) when all data from vehicle- and LAQ824-treated tumors at day 4 were compared. This indicates that [¹⁸F]FLT uptake can be used to measure the reduction in proliferation following treatment with LAQ824. It should be noted, however, that the regulation of Ki67 and TK1 are different. Ki67 is a protein the expression of which is associated with all phases of the cell cycle in cycling cells (63). On the other hand, TK1 is dramatically stimulated at the G₁-S border (37, 64).

In summary, HDAC is now widely recognized as a suitable molecular target for anticancer drug discovery. However, many questions remain unanswered regarding the optimal evaluation of HDACIs. Ongoing clinical trials of HDACIs have used conventional methods of assessment such as the maximum tolerated dose, although, broadly speaking, most HDACIs are considered to induce tumor stasis at clinically acceptable doses (34). Most studies have also assessed global increase in histone acetylation, although the relationship between histone acetylation per se and biological activity is not very clear. In this study, we have shown that the biological activity of the HDACI LAQ824 is detectable by [¹⁸F]FLT-PET and relates to inhibition of cell proliferation in vivo. Thus, [¹⁸F]FLT-PET warrants further investigation as a noninvasive imaging method for quantifying the biological activity of HDACIs.

Grant support: Cancer Research United Kingdom grants CR-UK C2536/A3554 and CR-UK C2536/A5708.

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.