Original article - Springer Link

Original article The uptake of 3′-deoxy-3′-[18F]fluorothymidine into L5178Y tumours in vivo is dependent on thymidine kinase 1 protein levels Henryk Barthel1, 2, Meg Perumal1, John Latigo1, Qimin He3, Frank Brady4, Sajinder K. Luthra4, Pat M. Price5, Eric O. Aboagye1 1 Molecular

Therapy and PET Oncology Research Group, Faculty of Medicine, Imperial College London, London, UK of Nuclear Medicine, University of Leipzig, Leipzig, Germany 3 Clinical Research Laboratory, Department of Oncology, Huddinge University Hospital, Karolinska Institute, Stockholm, Sweden 4 Hammersmith Imanet, London, UK 5 Wolfson Molecular Imaging Centre, Christie Hospital NHS Trust, Manchester, UK 2 Department

Received: 31 March 2004 / Accepted: 18 May 2004 / Published online: 4 September 2004 © Springer-Verlag 2004

Abstract. Purpose: The aim of this study was to investigate the role of thymidine kinase 1 (TK1) protein in 3′-deoxy-3′-[18F]fluorothymidine ([18F]FLT) positron emission tomography (PET) studies. Methods: We investigated the in vivo kinetics of [18F]FLT in TK1+/− and TK1−/− L5178Y mouse lymphoma tumours that express different levels of TK1 protein. Results: [18F]FLT-derived radioactivity, measured by a dedicated small animal PET scanner, increased within the tumours over 60 min. The area under the normalised tumour time–activity curve were significantly higher for the TK1+/− compared with the −/− variant (0.89±0.02 vs 0.79±0.03 MBq ml−1 min, P=0.043; n=5 for each tumour type). Ex vivo gamma counting of tissues excised at 60 min p.i. (n=8) also revealed significantly higher tumour [18F]FLT uptake for the TK1+/− variant (6.2±0.6 vs 4.6±0.4%ID g−1, P=0.018). The observed differences between the cell lines with respect to [18F]FLT uptake were in keeping with a 48% higher TK1 protein in the TK1+/− tumours versus the −/− variant (P=0.043). On average, there were no differences in ATP levels between the two tumour variants (P=1.00). A positive correlation between [18F]FLT accumulation and TK1 protein levels (r=0.68, P=0.046) was seen. Normalisation of the data for ATP content further improved the correlation (r=0.86, P=0.003). Conclusion: This study shows that in vivo [18F]FLT kinetics depend on TK1 protein expression. ATP may be important in realising this effect. Thus, Eric O. Aboagye (✉) Molecular Therapy and PET Oncology Research Group, Faculty of Medicine, Imperial College London, Hammersmith Hospital Campus, Du Cane Road, London, W12 0NN, UK e-mail: [email protected] Tel.: +44-20-83833759, Fax: +44-20-83832029

[18F]FLT-PET has the potential to yield specific information on tumour proliferation in diagnostic imaging and therapy monitoring. Keywords: [18F]FLT – PET – Proliferation – Thymidine kinase 1 – ATP Eur J Nucl Med Mol Imaging (2005) 32:257–263 DOI 10.1007/s00259-004-1611-0

Introduction There is a growing need in oncology for more specific positron emission tomography (PET) radiotracers as alternatives to the current “gold standard”, [18F]fluorodeoxyglucose ([18F]FDG), for monitoring anti-cancer therapy. This is largely driven by the requirement to provide specific information on the mechanism of action and downstream biological effects of new cancer therapeutics targeting growth signal transduction, cell cycle control and differentiation [1–3]. For these new drugs, there is the optimism that PET imaging of proliferation could provide early proof of principle, thus expediting clinical testing [4, 5]. 3′-deoxy-3′-[18F]fluorothymidine ([18F]FLT) was first proposed as a PET marker of tumour proliferation in 1998 [6]. Since that first report, several investigators have demonstrated associations between [18F]FLT uptake and biochemical indices of proliferation (PCNA, Ki-67, S-phase fraction) in vitro [7–9], in animal models [10] and in human tumours [11–16]. While these studies have provided proof that [18F]FLT measures proliferation, there is still a need to clarify the mechanism by which this occurs.

European Journal of Nuclear Medicine and Molecular Imaging Vol. 32, No. 3, March 2005

258

The accumulation of [18F]FLT in tumours is thought to depend on the activity of cytosolic thymidine kinase 1 (TK1), the key enzyme of the exogenous (salvage) DNA pathway. The link between [18F]FLT uptake and proliferation is then probably due to the relationship between TK1 expression and cell cycle regulation. TK1 activity has been shown to dramatically increase during the late G1 and the S phase of the cell cycle [17–20]. The presumption that [18F]FLT uptake reflects TK1 activity is based on earlier work with unlabelled FLT [21], and more recently, in vitro studies showing a positive correlation between uptake of FLT labelled with 3H or 18F by different mouse and human tumour cell lines and TK1 activity as measured with [3H]thymidine [7–9]. Since in vitro TK1 activity is measured with [3H]thymidine and [3H]thymidine uptake correlates with [3H]FLT uptake [7], it is not clear from the literature whether differences in TK1 protein levels per se can lead to changes in [18F]FLT uptake. The present study was, therefore, undertaken to provide in vivo evidence for the hypothesis that [18F]FLT tumour uptake is a function of TK1 expression. For this purpose we assessed [18F]FLT uptake in a mouse lymphoma tumour containing the functional heterozygous TK1+/− allele and a corresponding TK1−/− variant derived from mutation of the TK1 locus.

where a, b and c represent three orthogonal axes of the tumour. The experiments were started when the tumours reached a volume of approximately 200 mm3.

Immunohistochemical examination of the L5178Y tumours For histological evaluation of the L5178Y tumours, sections of both the TK1+/− and −/− variants were stained with haematoxylineosin (H&E) and mouse monoclonal antibodies for proliferating cell nuclear antigen (PCNA; Novocastra, Newcastle-upon-Tyne, UK). The staining procedure has recently been described elsewhere [10]. The numbers of PCNA-positive and H&E-positive cells in adjacent sections were counted in five randomly selected fields of view per section using a BX51 Olympus microscope (Olympus Optical, Tokyo, Japan) at 600× magnification. The labelling index for PCNA (LIPCNA) was calculated using the equation: .

Western blot analysis of TK1 (protein) levels in L5178Y tumours

[18F]FLT was produced on-site by Hammersmith Imanet Limited (MRC Cyclotron Building, Hammersmith Hospital, London, UK) from the 2,3′-anhydro-5′-O-(4,4′-dimethoxytrityl)-thymidine precursor as previously reported [22]. Typical radiochemical yields of 30–35% and specific activities of 46.5–80.8 GBq mol−1 were achieved.

TK1 protein levels from excised tumours were determined by Western blotting as previously reported [10]. Briefly, identical (total cellular) protein amounts of tumour homogenates were loaded onto a pre-cast gel (pre-cast Tris-glycine (4–20%) gel; ICN Biomedicals, Aurora, USA). After standard electrophoresis of the samples and transfer of the separated proteins onto a nitrocellulose membrane (Hybond ECL; Amersham Pharmacia Biotech, Little Chalfont, UK), TK1 was detected using an anti-TK1 mouse monoclonal antibody (Svanova Biotech, Uppsala, Sweden) as primary antibody and anti-mouse immunoglobulin G horseradish peroxidase (Santa Cruz Biotechnology, Santa Cruz, USA) as secondary antibody. TK1 protein was visualised using the enhanced chemiluminescence method (SuperSignal West Pico Chemiluminescent Substrate; Pierce, Rockford, USA). Finally, band intensities were quantified using a GS-710 Calibrated Imaging Densitometer (BioRad Lab., Hercules, USA) and the Quantity One software (version 4.0.3; Bio-Rad Lab., Hercules, USA).

Animals and tumour models

Analysis of ATP levels in L5178Y tumours

Two variants of L5178Y-3.7.2C mouse lymphoma cells (TK1+/− and TK1−/−) that express different levels of TK1 protein [23, 24] were used. These cells were kindly provided by Prof. Ann Jackman and Dr. David Gibbs of the Institute of Cancer Research, Sutton, UK. The cells were expanded in tissue culture. Unlike in previous studies where the cells were injected into the gastrocnemius muscle [23, 24], the cells were suspended in a mixture of 25% Dulbecco’s phosphate-buffered saline and basement membrane matrix (BD Matrigel; BD Bioscience, San Jose, USA) and (5×106) inoculated subcutaneously into the right (TK1+/− cells) and left (TK1−/− cells) flanks of 10- to 12-week-old male DBA strain mice (Harlan United Kingdom Ltd, Bicester, UK). L5178Y cells are widely used in the Mouse Lymphoma Assay (MLA) to assess chemical mutagenesis caused by presumptive TK gene mutations or multiple loci mutations affecting the TK locus [25, 26]. In this assay, mutation of the TK1 locus (TK1+/− to TK1−/−) leads to the formation of small colonies resistant to the thymidine analogue trifluorothymidine [26, 27]. Tumour volumes were measured continuously using a calliper and calculated using the equation: ,

In addition to the TK1 (protein) levels, ATP levels in individual L5178Y tumours were determined with a standard bioluminescence assay kit (ENLITEN ATP assay system; Promega Corporation, Madison, USA) and a protocol which has been described in detail recently [10]. ATP levels were normalised to the total cellular protein content of the tumour samples that was determined by the BCA protein assay kit (Pierce, Rockford, USA).

Materials and methods Radiopharmaceutical preparation

[18F]FLT PET and biodistribution studies in L5178Y tumour-bearing mice The biodistribution of [18F]FLT in DBA mice bilaterally bearing TK1+/− and −/− variants of the L5178Y tumour was investigated in vivo by means of PET imaging, as well as ex vivo by gamma counting of excised tissues. For PET scanning, a second-generation dedicated small animal scanner was employed (quad-HIDAC, Oxford Positron Systems, Weston-on-the-Green, UK). The scanner characteristics have been described elsewhere [10]. Five mice


259 Statistical analysis

bearing both variants of the L5178Y tumour were scanned. For this purpose, 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 within the scanner. A bolus injection of ~3.0 MBq [18F]FLT was administered via the tail cannula and scanning commenced. Dynamic emission scans were acquired in list mode format over 60 min. The acquired data were sorted into 0.5-mm sinogram bins and 19 time frames (4×15 s, 4×60 s and 11×300 s) for image reconstruction, which was performed by filtered back-projection using a 2D Hamming filter (cut-off 0.6). The image data sets obtained were transferred to a SUN workstation (Ultra 10; SUN Microsystems, Santa Clara, USA) and visualised using the ANALYZE software (version 4.0; Biomedical Imaging Resource, Mayo Clinic, Rochester, USA). Regions of interest (ROIs) were defined on three to six coronal planes for all tumours. ROIs were defined to exclude central necrosis. Time versus activity curves (TACs) from the ROIs were averaged for each tumour and, in order to allow inter-individual comparison, normalised to the integral of the respective heart cavity TACs [28]. Biodistribution studies were performed in the five scanned mice together with three additional mice that had been similarly injected with [18F]FLT. All mice were sacrificed at 60 min post injection of [18F]FLT by exsanguination via cardiac puncture (under general isofluorane inhalation anaesthesia). Blood was collected and normal tissues (liver, kidneys, lung, brain, spinal cord, spleen, heart, leg muscle, leg bone and small intestine) and TK1−/− and TK1+/− tumours were rapidly excised. In addition, urine was collected and plasma was obtained from blood centrifugation. All samples were weighed, and the radioactivity was measured using a Cobra II Auto-Gamma counter (Packard Instrument, Meriden, USA) applying a decay correction. The results were expressed as percentage of injected dose per gram of tissue (%ID g−1).

The in vivo growth characteristics of L5178Y tumours are shown in Fig. 1a. The latency period before tumours were visualised was shorter in the TK1+/− than in the −/− variant by 5 days. Furthermore, tumour volume doubling time in the exponential growth phase was significantly shorter (1.20±0.13 vs 2.90±0.27 days, P=0.050) for the TK1+/− variant. Thus, in order to obtain sizematched bilateral tumours (ca. 200 mm3) at the time of the planned radiotracer experiment, the TK1−/− cells were implanted 7 days earlier than the TK1+/− cells. Figure 1b shows a representative PCNA-stained slice of a L5178Y TK1+/− tumour. As shown in Fig. 1c, the PCNA labelling indices (LIPCNA) obtained from histological examination of both L5178Y tumour types were not statistically different from each other at the time of the

Fig. 1a–f. Morphological, histological and biochemical characteristics of L5178Y tumours. a Growth curves (P was calculated for difference of tumour volume doubling times in exponential growth phase), b histologically determined degree of proliferation (typical

PCNA-stained tumour sample), c summary data for PCNA labelling index, d TK1 (protein) concentration (representative Western blot), e summary data for TK1 levels, and f ATP concentration in TK1−/− and TK1+/− variants of L5178Y tumours

Statistical analyses were performed using the software SPSS for Windows, version 10.0.7 (SPSS, Inc., Chicago, USA). Differences in histological, biochemical and radiotracer uptake parameters between the two variants of L5178Y tumours were tested for statistical significance using the Mann-Whitney U test, and in cases of TK1+/− and −/− tumours handled in parallel, using the Wilcoxon test. Correlations between these parameters were determined by linear regression analyses. Unless stated, data were mean ± 1 standard error of the mean (sem). P values of ≤0.05 were considered significant.

Results


260 Table 1. Uptake of [18F]FLT into TK1−/− and TK1+/− variants of L5178Y tumours, as determined in vivo by means of dynamic PET imaging using a dedicated small animal scanner

Mouse no.

TK1 variant

Uptake2 min p.i. [10−2 MBq ml−1]

Uptake60 min p.i. [10−2 MBq ml−1]

AUCTAC [10−2 MBq ml−1 min]

1

−/− +/− −/− +/− −/− +/− −/− +/− −/− +/− −/− +/−

1.30 1.53 0.88 1.70 0.74 1.37 0.98 1.25 1.09 1.22 1.00±0.08 1.42±0.08*

1.67 1.73 1.30 1.49 1.42 1.68 1.51 1.56 1.58 1.61 1.50±0.06 1.61±0.04*

91.3 95.3 71.9 88.1 70.0 90.6 81.2 87.0 80.7 86.1 79.0±3.4 89.4±1.5*

2 Dynamic PET imaging was performed 0–60 min after i.v. injection of ~3.0 MBq [18F]FLT on a quad HIDAC scanner (Oxford Positron Systems, Weston-onthe-Green, UK). The time versus activity curves (TACs) were normalised to the integral of the individual heart cavity TAC. AUC area under curve. *P