Preferential Tumor Targeting and Selective Tumor Cell Cytotoxicity of 5-[^131/125I]Iodo-4′-Thio-2′-Deoxyuridine Free

Radiolabeled nucleoside analogues are attractive candidates as imaging probes for noninvasive evaluation of tissue proliferation. For this purpose, appropriately radiolabeled nucleoside analogues should be preferentially transported into proliferating cells. Incorporation of proliferation probes into DNA is desirable when a radiotherapeutic effect is intended (1, 2). It has been shown that radionucleosides labeled with an Auger electron emitting halogen such as ¹²⁵I or ¹²³I and incorporated into DNA are highly cytotoxic (3–5). This affects preferentially proliferating tumor cells and may provide effective cytotoxic drugs for anticancer therapy.

Several thymidine and uridine analogues, labeled with γ or positron emitters, such as [¹¹C]dThd (6), [¹⁸F]FdUrd (7), [¹⁸F]FLT (8), [⁷⁶Br]BrdUrd (9), FMAU (10), FBAU (11) and [¹²⁴I]IdUrd (12), were synthesized and evaluated for noninvasive imaging to address tumor proliferation in patients. Moreover, the efficiency of radiolabeled 5-iodo-2′-deoxyuridine (IdUrd) for tumor imaging in vivo was successfully shown (4, 12–16). However, a potential therapeutic use of IdUrd is limited by its rapid enzymatic degradation by thymidylate phosphorylase (TP) leading to the low rate of [¹²⁵I]IdUrd DNA incorporation. To circumvent the rapid in vivo catabolism and to obtain preferential incorporation of radiohalogenated nucleoside analogues into DNA through the salvage pathway, Toyohara et al. synthesized a new thymidine analogue 5-iodo-4′-thio-2′-deoxyuridine (ITdU; refs. 17, 18). ITdU is phosphorylated by the cytosolic thymidine kinase and retained within the cell in a kinase-dependent manner. Moreover, ITdU was found to be resistant against enzymatic cleavage of the C-N glycosidic bond and more rapidly incorporated into nuclear DNA through the thymidine salvage pathway than IdUrd, making ITdU a highly promising DNA synthesis marker. Consequently, the new radiolabeled thymidine analogue ITdU was proposed as a proliferation marker (17). However, when used for conventional scintigraphy, imaging with ITdU was impaired by high background activity due to deiodination of ITdU and generation of [¹²⁵I] metabolites (17, 19, 20).

The thymidylate synthase (TS), an essential enzyme for the de novo synthesis of thymidylate, has been shown to deiodinate radiolabeled IdUrd (21–24). Several groups showed that 5-fluoro-2′-deoxyuridine (FdUrd)-mediated inhibition of TS-based deiodination resulted in increased incorporation of radiolabeled IdUrd into DNA (15, 21, 23, 24). Similarly, our own observations in a HL60 cell culture model showed a 20-fold increased cellular uptake and 18-fold increase of incorporation of radiolabeled ITdU into DNA as a result of inhibition of TS through FdUrd (25).

Based on these studies, we show here in a HL60 xenograft severe combined immunodeficient (SCID) mouse model that FdUrd-mediated TS inhibition results in a significantly increased uptake of ITdU in proliferating tumor cells and incorporation into tumor cell DNA. Already small doses of ITdU could selectively induce apoptosis in vivo in tumor cells, whereas apoptosis remained only marginally in normal proliferating cells of small intestine or hematopoietic cells in the spleen.

Reagents. Chemicals and solvents were purchased from Sigma-Aldrich and Merck or otherwise as indicated. All reagents and solvents were of highest commercially available grade and used without further purification. No carrier added sodium [¹³¹I] and no carrier added sodium [¹²⁵I] were obtained from GE Healthcare (formerly Amersham Biosciences); no carrier added sodium [¹²³I] was purchased from Zyklotron. The 5-trimethylstannyl precursor of ITdU and the standard compound ITdU were synthesized according to previously reported methods (17, 26–28). [¹³¹I]IdUrd was synthesized as reported previously (29). IdUrd was purchased from ABX.

Radiochemistry. Chloroamine T (80 μL; 2 mg/mL in 66% methanol) was added to a mixture containing 8.5 μL phosphate buffer [2 mol/L in 30% methanol (pH 2.0)], 1.4 μL of the 5-trimethylstannyl precursor solution of ITdU (50 mg/mL in 66% methanol), and sodium [^123/125/131I] solution in 0.05 mol/L NaOH. Reaction proceeded within 10 min at room temperature and was stopped by addition of 50 μL thiosulfate solution (0.1 mol/L). The product was purified by high-performance liquid chromatography using a Kromasil 100 5 μm C4 (250 × 4 mm) reverse-phase column (CS Chromatography) eluted with 10% ethanol at a flow rate of 1 mL/min. Quality control was done by analytic high-performance liquid chromatography [column: LiChrospher 100 RP-18 5μ-EC, 250 × 4 mm (CS Chromatography), eluent: 0.02 mol/L phosphate buffer (pH 3.6)/methanol = 34/66 (v/v), and flow: 1 mL/min]. Total radiochemical yields were 80% to 90% with radiochemical purities of >95%. Specific activities were 900 GBq/μmol for [¹²³I]ITdU, 45 GBq/μmol for [¹²⁵I]ITdU, and 200 GBq/μmol for [¹³¹I]ITdU.

Thymidine kinase 1 assay. Recombinant human thymidine kinase 1 (TK1) was kindly provided by Prof. S. Ericsson (Swedish University of Agricultural Sciences, Uppsala, Sweden) and prepared as described (30). The thymidine kinase assay was done at 37°C according to a previously reported method (31) with carrier added [¹³¹I]ITdU (40.5 μmol/L, 0.8 GBq/μmol) and carrier added [¹³¹I]IdUrd (40.3 μmol/L, 0.4 GBq/μmol) as substrates and 110 to 220 ng TK1. After 2, 5, 10, 30, and 60 min, 10 μL aliquots were removed and the reaction was terminated by addition of 10 μL TCA. The extent of nucleoside phosphorylation was determined by reverse-phase high-performance liquid chromatography using a LiChrospher 100 RP-18 5μ-EC (250 × 4 mm) reverse-phase column as described above.

Thymidine phosphorylase assay. Susceptibility of the substrates IdUrd and ITdU to glycosidic bond cleavage by thymidine phosphorylase was measured as reported previously (18). The assay mixture contained ∼30 nmol nucleoside, 168 mmol/L phosphate buffer (pH 7.6), and 6.5 × 10⁻³ to 1.3 × 10⁻⁶ units recombinant human thymidine phosphorylase (Sigma-Aldrich). The assay reactions were carried out at 37°C and terminated after 30 min by adding 30 μL TCA. The extent of nucleoside degradation was determined by reverse-phase high-performance liquid chromatography as described above using a Dionex UVD170S detector (254 nm).

Cell lines and culture conditions. Human myeloid leukemia (HL60) cells were grown in RPMI 1640 (Life Technologies) containing 10% FCS (Biochrom), 10 mmol/L HEPES (pH 7.3; Biochrom), 100 units/mL penicillin (Life Technologies), 100 μg/mL streptomycin (Life Technologies), and 2 mmol/L l-glutamine (Biochrom).

Tumor model. A suspension of 4 × 10⁶ HL60 cells was injected subcutaneously into the neck of 6-week-old male and female CB17ICR SCID mice (originally obtained from the Institute of Cancer Research). After 7 days, the mice developed subcutaneous tumors of ∼20 to 30 mm diameter. All animal experiments were done in accordance with national and local regulations for animal and radiation protection and good experimental practice.

Biodistribution. Thirty minutes before intravenous injection of 2 to 3 MBq [¹³¹I]ITdU in 150 to 200 μL saline, varying concentrations of FdUrd (8, 20, and 40 μmol/kg in 200 μL saline) were injected into the lateral tail vein of CB17ICR SCID mice bearing xenotransplanted HL60 tumors. Control mice had intravenous injections of 200 μL saline alone before [¹³¹I]ITdU injection. The aim of these studies was the assessment of in vivo biodistribution of ITdU without consideration of other effects like cytotoxicity. Thus, for this proposal, we used the [¹³¹I]-conjugated ITdU. Mice were sacrificed 0.5, 4, and 24 h postinjection by cervical dislocation. Tumors and organs were excised, weighed, and assayed for radioactivity in a γ counter (Cobra 2; GE Healthcare/Packard Instruments). Mean tumor and organ uptake was determined from decay-corrected tissue radioactivity normalized to injected dose and tissue sample weight (%ID/g tissue wet weight).

DNA extraction from tumor, spleen, and intestine. DNA isolation was accomplished by using DNeasy Blood and Tissue Kit (Qiagen) according to the manufacturer's instructions. Briefly, 24 h after intravenous injection of [¹³¹I]ITdU, tumor, spleen, and intestine tissues were dissected and incubated at 56°C for 3 h in tissue lysis buffer and proteinase K. After washing and elution steps, the amount of isolated DNA was determined. The radioactivity in DNA and in other tissue fractions was analyzed with a γ counter (Cobra 2; GE Healthcare/Packard Instruments) and results were given in cpm/μg isolated DNA.

Autoradiography. FdUrd (20 μmol/kg in 200 μL saline) was injected into the lateral tail vein of CB17ICR SCID mice bearing xenotransplanted HL60 tumors 30 min before intravenous injection of 25 MBq [¹²⁵I]ITdU in 150 to 200 μL saline. Control mice received intravenous injections of 200 μL [¹²⁵I]ITdU or FdUrd alone. Mice were sacrificed 30 min and 24 h postinjection and organs were excised. Histologic slices (2 μm) were prepared after fixation in formaldehyde and paraffin embedding. The slices were shortly dipped into a melted photographic emulsion (Kodak Emulsion NTB-2) and allowed to dry for ∼15 min in a darkroom. After an exposition period of 2 weeks at -20°C, the slices were developed (Kodak D-19 Film Developer) and fixed (Superfix Plus; Tetenal) followed by standard H&E staining. The cellular distribution of silver grains indicating radioactive sites was examined using a Zeiss Axiophot microscope equipped with a 3CCD digital camera (JVC).

Immunohistochemistry of tissue sections with Ki-67-specific antibody. Consecutive formalin-fixed, paraffin-embedded to autoradiography corresponding tissue sections (2 μm thick) were dewaxed in xylene and rehydrated through graded concentrations of ethanol to distilled water. Sections were then immersed in 10 mmol/L citrate buffer (pH 6.0) and processed in thermostatic water bath for 30 min at 98°C for antigen retrieval. The anti-Ki-67 antibodies (clone MIB-1 for human antigen and clone TEC-3 for murine antigen) were purchased from DAKO. After the antigen retrieval treatment, the tissue sections were incubated for 30 min with 1:150 and 1:25 diluted MIB-1 and TEC-3 antibodies, respectively. Subsequently, the sections were exposed for 30 min to peroxidase-linked anti-mouse immunoglobulin antibody (polymer from DAKO) or peroxidase-linked donkey anti-rat immunoglobulin antibody (dilution 1:250; Dianova). Color development used diaminobenzidine substrate and sections were counterstained with hemalaun.

In situ nick end labeling assay. Consecutive formalin-fixed, paraffin-embedded to autoradiography and Ki-67 staining corresponding tissue sections (2 μm thick) were dewaxed in xylene and rehydrated through graded concentrations of ethanol to distilled water. Terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assay was done according to the manufacturer's instructions (Roche). Color development used diaminobenzidine substrate and sections were counterstained with hemalaun. The cellular apoptotic effects were examined using a Zeiss Axiophot microscope equipped with a 3CCD digital camera (JVC).

Assessment of the apoptosis index. For the apoptosis index based on terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assay, five randomly selected sections were analyzed for the number of the positive cells in relation to the total number of cells using Image-Pro Express (Media Cybernetics). For each section, 500 ± 50 cells were analyzed.

Scintigraphy. Twenty-four hours before the experiment, mice were given drinking water containing potassium iodide (0.2 g/L) and potassium perchlorate (2 g/L) to avoid thyroid uptake of [¹²³I] and to block its secretion into the stomach, respectively (15). Scintigraphy of mice was done 0.5, 4, and 24 h after intravenous application of 4 MBq [¹²³I]ITdU with or without pretreatment with 20 μmol/kg FdUrd under xylacine (16 mg/kg) and ketamine (100 mg/kg) anesthesia with a gamma camera (Gamma Camera E.CAM, Siemens Medical Solutions) using HR collimator, matrix 1,024 × 1,024, zoom factor 2 (pixel size 0.3 mm), and energy window settings 159 keV ± 15%. The acquisition time was 30 min. A region of interest analysis was done to determine the cpm in the mouse and the tumor. Percent tumor uptake was determined as ratio of the obtained cpm.

Statistical analysis. Data are presented as mean ± SD. Mean values between groups were compared using the Mann-Whitney U test. Differences were considered to indicate statistical significance for a two-tailed P < 0.05. Statistical analysis was carried out using GraphPad Prism software version 4.00 (GraphPad Software).

Phosphorylation of [¹³¹I]ITdU by TK1. The phosphorylation rate of [¹³¹I]ITdU and [¹³¹I]IdUrd were determined as 125 ± 2 nmol monophosphate/(mg TK1 × min) and 270 ± 5 nmol monophosphate/(mg TK1 × min), respectively. [¹³¹I]ITdU showed ∼46% of the phosphorylation of [¹³¹I]IdUrd (Table 1).

Susceptibility of ITdU to glycosidic bond cleavage by thymidine phosphorylase. The recombinant thymidine phosphorylase assay showed that ITdU was highly resistant against enzymatic cleavage of C-N glycosidic bond. The 5-iodouracil (IU) formation caused by enzymatic degradation was 0.280 ± 0.002 μmol IU/(unit TP × min) for ITdU, whereas IdUrd was rapidly cleaved [7.8 ± 0.2 μmol IU/(unit TP × min); Table 1]. The relative enzymatic degradation of ITdU by TP was ∼3.6% compared with that of IdUrd.

Comparative biodistribution of [^123/131I]ITdU in tumor- bearing SCID mice in dependency on pretreatment with FdUrd. Biodistribution of [¹³¹I]ITdU injected alone or after pretreatment with FdUrd was measured after 0.5, 4, and 24 h in SCID mice bearing subcutaneous human leukemia HL60 xenografts. The results corrected for physical decay of the isotope showed that generally most of the activity was accumulated in proliferating tissues like tumor, spleen, and small intestine (Fig. 1). At 30 min postinjection, a high percentage of the initial activity was found in the kidney, indicating rapid urinary excretion of [¹³¹I]ITdU. Injection of [¹³¹I]ITdU without FdUrd pretreatment resulted in a considerably transient retention of ITdU in tumor tissue (3.2 ± 0.4, 0.7 ± 0.1, and 0.1 ± 0.1 %ID/g at 0.5, 4, and 24 h postinjection, respectively; Fig. 1). The highest residual radioactivity was detected in small intestine (5.9 ± 1.0, 5.2 ± 1.3, and 2.8 ± 0.3 %ID/g at 0.5, 4, and 24 h postinjection, respectively). To determine the most appropriate dose of FdUrd for favorable uptake and stable incorporation of [¹³¹I]ITdU in tumor tissue, we pretreated mice with different amounts of the TS inhibitor FdUrd. In contrast to animals receiving the [¹³¹I]ITdU alone, all tested FdUrd doses led to significantly increased radioactivity uptake and its retention in tumor tissue (1.8- to 3.1-fold at 0.5 h, 12.7- to 13.8-fold at 4 h, and 39.7- to 78.1-fold at 24 h postinjection; P < 0.05). Also, in rapidly proliferating normal tissue like the spleen, ITdU uptake was increased by FdUrd (6.3- to 8.8-fold at 0.5 h, 15.8- to 19.5 at 4 h, and 1.4- to 3.9-fold at 24 h postinjection). This transient, high concentration of ITdU in spleen detected at 4 h postinjection is probably due to incorporation of ITdU into cells undergoing apoptosis, because the red pulp of spleen in rodents is the site of extramedullary hematopoiesis and degradation of dead cells (32). Interestingly, whereas accumulated activity was extremely increased in tumor tissue (up to 78.1-fold), FdUrd had only a minor effect on uptake of [¹³¹I]ITdU in small intestine (1.2- to 2.1-fold at 0.5 h, 1.2- to 2.0-fold at 4 h, and 1.2- to 2.1-fold at 24 h postinjection, respectively). The concentration of radioactivity in normal tissues with low proliferation rate like liver or lung was minimally changed and diminished progressively. A dose of 20 μmol/kg FdUrd led to the highest relative retention of ITdU in tumor tissue and thus to the most favorable tumor to nontumor tissue radioactivity ratio (Fig. 1).

For comparative scintigraphy study, HL60-xenograft SCID mice were intravenously injected with [¹²³I]ITdU combined with FdUrd or alone. Analogous to biodistribution studies, gamma camera scintigraphy was done 0.5, 4, and 24 h postinjection to visualize the time-dependent retention of ITdU in tumor tissue. Pretreatment with FdUrd markedly enhanced uptake and retention of ITdU in malignant cells already 4 h postinjection causing improved tumor-to-nontumor tissue contrast due to low background activity in normal tissues (Fig. 2). The relative tumor uptake of the total activity in the mice was 7.1% and 6.8% at 0.5 h, 6.5% and 1.4% at 4 h, and 12% and 2% at 24 h postinjection with and without previous administration of FdUrd, respectively. Preadministration of KClO₄ and KI markedly reduced gastric secretion and intestine reabsorption as well as uptake in thyroid of free radioiodine. The comparative scintigraphy was carried out three times with similar results.

Effect of FdUrd pretreatment (20 μ mol/kg) on the DNA incorporation of [¹³¹I]ITdU. DNA incorporation of [¹³¹I]ITdU was measured using DNA extracted from tumor, spleen, and small intestine tissues samples 24 h postinjection of radiotracer in HL60 xenograft SCID mice allowing DNA incorporation of high percentage of the initial activity to occur (Table 2). In agreement with the biodistribution studies, the DNA-associated radioactivity was only moderately increased after pretreatment with FdUrd in small intestine (1.1-fold), whereas the incorporation into DNA isolated from spleen was ∼3-fold higher compared with untreated cells. The highest enhancement of ITdU incorporation into DNA as an effect of TS inhibition was observed in tumor cells (32-fold).

[¹²⁵I]ITdU-induced radiocytotoxicity. To determine a potential radiotoxic effect of [¹²⁵I]ITdU with or without pretreatment with FdUrd on tumor and normal proliferating tissues like spleen and small intestine, we measured apoptosis with terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assay 30 min and 24 h after ITdU application. Damage to tumor and normal tissue cells induced by [¹²⁵I]ITdU alone or in combination with FdUrd was compared (Fig. 3). No radiocytotoxicity was observed in all tissues examined 30 min after application of [¹²⁵I]ITdU (data not shown). In contrast, 24 h after [¹²⁵I]ITdU application, tumor tissue of mice treated with [¹²⁵I]ITdU and FdUrd showed highly increased apoptosis when compared with tumor tissue of mice injected with [¹²⁵I]ITdU alone (97% versus 4.2%; P < 0.05). No cytotoxicity was observed in tumor tissue from mice treated with FdUrd alone. DNA fragmentation in small intestine and spleen was not increased by combined FdUrd and ITdU application. Finally, all examined extratumoral tissue sections (heart, lung, liver, spleen, kidney, colon, small intestine, stomach, muscle, vertebral body, and bone) showed no histologic or morphologic changes after coadministration of [¹²⁵I]ITdU and FdUrd. Microautoradiography and Ki-67 staining analysis within corresponding tissue sections showed a clear correlation between cellular ITdU uptake and the proliferation state of cells (Fig. 4). As shown by microautoradiography, ITdU was virtually exclusively located within the cell nucleus in all tissues examined.

The thymidine analogue [^123/125I]ITdU is known to be incorporated into DNA of tumor cells and thus may provide a promising probe for delivering [^123/125I] Auger electrons into tumor DNA. In this study, we examined the efficiency of [^131/125I]ITdU for tumor cell and tumor DNA targeting in vivo with and without FdUrd-mediated modulation of ITdU biokinetics. In addition, potential cellular radiotoxicity in a HL60 xenograft SCID mouse model was investigated. We found that pretreatment with FdUrd (20 μmol/kg) preferentially increased ITdU uptake in tumor cells (up to 78-fold) and tumor cell DNA (32-fold). Thus, FdUrd pretreatment considerably increased tumor cell and tumor DNA selectivity of ITdU. These findings were confirmed by selective ITdU retention 24 h postinjection in tumor tissue as shown by whole-body scintigraphy and DNA analysis. Dupertuis et al. reported that FdUrd preadministration increased incorporation of the metabolically less stable radionucleoside IdUrd in human glioblastoma cell lines and in nude mice glioblastoma xenograft models (15). Incorporation of IdUrd into tumor cells and tumor DNA as well as tumor IdUrd retention were similarly affected in this study by FdUrd as FdUrd-mediated modulation of ITdU biokinetics in the present work. FdUrd-mediated TS inhibition and accumulation of proliferating cells in S phase have been shown to correlate to maximal cellular IdUrd accumulation in cell culture models (24) and might similarly affect ITdU uptake in the present study.

Interestingly, 24 h after ITdU injection, we observed high (>90%) apoptosis induction in tumor cells as opposed to negligible apoptosis induction in proliferating cells in the other tissues examined. In small intestine, apoptosis was strictly limited to proliferating crypt cells. Furthermore, we found virtually no apoptosis induction in the spleen. We did not observe any cytotoxicity of a single-agent treatment with FdUrd as measured with the terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling assay. These findings clearly link apoptosis induction in tumor cells to the combined treatment with ITdU and FdUrd. The study was not designed to measure gross therapeutic effects. Thus, apoptosis staining was used as surrogate marker for potential cellular cytotoxicity. Selective tumor apoptosis induction might be explained by higher tumor uptake, considerably longer tumor retention, and consequently increased radiation absorbed dose by tumor DNA mediated by stably incorporated ITdU emitting Auger electrons. We have shown previously in a HL60 cell culture model that [¹²³I]ITdU activates the mitochondrial, intrinsic apoptosis pathway, inducing efficiently apoptosis in HL60 cells (25). In this study, the time-dependent damage of DNA induced by metabolically incorporated [¹²³I]ITdU was shown. Furthermore, we showed stable tumor cell DNA incorporation and subsequently mediated tumor cell radiocytotoxicity, because apoptosis was affected by Auger electrons from [¹²³I] bound to DNA and not by the γ-radiation component of [¹²³I] or the pharmacodynamic effect of the tracer amounts of [¹²⁷I]ITdU (25).

Additionally, the role of p53 in the responses to therapeutics should be considered. The p53 tumor suppressor protein is involved in apoptosis and cell cycle checkpoints (33–35). Furthermore, p53 was described to be associated with several known DNA repair pathways, including nucleotide excision repair and base excision repair through transcriptional transactivation of DNA repair-associated genes and through an interaction with components of the repair and recombination machinery (36–38). In consideration of the fact that the high linear energy transfer deposition of Auger radiation in DNA produces a high proportion of double strand breaks, the loss of p53 in HL60 cells, increasing significantly genetic instability and reducing apoptotic threshold, might contribute to the exquisite increase of radiosensitivity of proliferating tumor cells compared with proliferating cells in normal tissues (39). The apoptotic effect detected in small intestine was attributed to high radiation susceptibility of proliferating stem cells located in the lower part of the crypts (40). Even small doses of radiation (0.05 Gy) raised the level of apoptosis by 5-fold. Importantly, the crypts were not reproductively sterilized by doses less than 8 or 9 Gy. The small intestine crypt contain a second category of stem cells, the clonogenic cells much more radioresistant, sustaining the crypts (40).

The time-dependent accumulation and retention of ITdU in tumor tissue in response to TS inhibition by FdUrd pretreatment was shown using biodistribution studies and whole-body scintigraphy 0.5, 4, and 24 h postinjection. During this time frame, the highest amount of residual radioactivity was detected in tumor tissue. Twenty-four hours post-application of ITdU and 8 μmol/kg FdUrd, we observed the highest accumulation of radiolabeled ITdU with 78.0-, 3.7- and 2.2-fold increased incorporation in tumor, spleen, and small intestine tissues, respectively. The most favorable tumor-to-normal tissue ratio we determined after pretreatment with 20 μmol/kg FdUrd was attributable to lower retention ratio of initially accumulated radioactivity (0.5 h) in spleen and small intestine. Similarly, the scintigraphy study on mice illustrated efficient tumor imaging, as a result of the selective and prolonged retention of ITdU in malignant cells and transient retention of radioactivity in normal tissue caused by FdUrd pretreatment. The detected residual background activity was concentrated in stomach content. Previously, it has been shown that more than 70% of ITdU was distributed in the DNA fraction 8 h postinjection (20). However, at 2 h postinjection, a significant amount of radioactivity was detected in the acid-soluble fraction of rapid proliferating tissues. Similarly, our biodistribution studies showed that 4 and 24 h post-application most of the initial detected activity (0.5 h) was accumulated in rapidly proliferating tissues like tumor and small intestine, indicating prolonged retention of ITdU mediated by its fast incorporation into DNA. Increased uptake and incorporation of ITdU in malignant cells, leading to high and selective damage of tumor tissue and minor radiocytotoxicity in normal tissues, indicates that the combined application of ITdU und FdUrd might provide a selective Auger radiation therapy approach in vivo. For application of ITdU in therapeutic settings, an evaluation of tolerable doses inducing only transient and acceptable cytotoxicity in normal tissues is of major concern. Owing to their physical properties, Auger emitters like ^123/125I develop high biological efficacy when closely associated with DNA. In this context, the most important requirement is an estimation of a preclinical Auger radiation dosimetry based on biodistribution data measured 24 h postinjection, when free iodine is eliminated and most radioactivity is stably DNA incorporated.

These in vivo studies showing significantly enhanced incorporation of [¹³¹I]ITdU into DNA of proliferating cells as consequence of preinjection with FdUrd are in the line with the observation reported for HL60 cells in vitro (25). In this study, we showed complete inhibition of TS-mediated dehalogenation of [¹²³I]ITdU and 18-fold increased DNA uptake of [¹²³I]ITdU after pretreatment with FdUrd. In the present study, we show the potential of FdUrd, which is a specific inhibitor of the de novo deoxythymidine synthesis (41), to improve incorporation of ITdU into DNA and consequently its preferential and prolonged retention in tumor tissue in vivo. The observed highest uptake occurring in tumor tissue might be attributed to some resistance to inhibition of TS by FdUrd in normal tissues, caused by different regulation of intracellular TS expression after exposure to TS inhibitor or lowered affinity of the target enzyme for FdUMP (42, 43). Therefore, a differential TS concentration-dependent blocking effect of FdUrd in tumor versus normal proliferating cells and hence a differential, tissue-specific modulation of ITdU metabolism by FdUrd cannot be excluded.

It has been shown that thymidine kinase in tumor cells exhibits a higher potential for conversion of nucleoside analogues into active nucleotides and FdUrd to inhibitory acting FdUMP (44, 45). This leads to a more efficient incorporation of radiolabeled nucleotides into DNA and inhibition of TS in tumor cells, respectively. Furthermore, the inhibition of the endogenous de novo pathway led to abnormally high intracellular dUTP/dTTP ratio and consequently to increased activity of TK1 and up-regulation of nucleoside transporter expression (46, 47). Additionally, evidence derived from the metabolic studies of [¹⁴C]thymidine and [¹⁴C]thymidine monophosphate in normal and leukemia cells indicates increased activity of catabolic enzymes thymidine phosphorylase and thymidine phosphatase in normal cells, leading to transient retention of intact thymidine analogues in normal tissue (48). Thus, up-regulation of thymidine salvage pathway in tumor cells in concert with up-regulation of thymidine phosphate catabolism in normal tissue might explain preferential tumor uptake and retention of ITdU in the present study.

The preferential uptake of ITdU into tumor cells was substantiated by DNA extraction analysis, showing the highest amount of radioactivity stably incorporated into DNA of tumor cells compared with other proliferating tissues. The radioactivity incorporated by the tumor cells was strictly confined to the cell nuclei as shown by microautoradiography done 24 h post-intravenous injection of ITdU.

The in vivo data of ITdU correspond closely to its biochemical characteristic. An affinity to cytosolic TK1 leading to production of active nucleotides is a critical requirement for incorporation of C-5 halogenated thymidine analogues into DNA. Using nucleoside kinase assay, we determined that ITdU was phosphorylated by TK1 with a phosphorylation rate of 125 ± 2 nmol monophosphate/(mg TK1 × min). The relative phosphorylation activity compared with the reference thymidine analogue IdUrd was ∼46%, which is consistent with previous results reported by Toyohara et al., established in whole-cell extracts, where the natural competing substrate was present (17). These results suggest that ITdU was effectively phosphorylated by cytosolic TK1 and that the phosphorylation level remains unaffected by naturally present nucleosides. This could be attributed to the relatively high affinity of ITdU to TK1. Another factor that affects incorporation of radiolabeled thymidine analogues into DNA is the C-N glycoside cleavage reaction catalyzed by thymidine phosphorylase and preventing their use in DNA synthesis (49, 50). The thymidine phosphorylase assay confirmed that the substitution of the 4′-oxygen by sulfur in the ITdU molecule increased the stability of the glycosidic bond (17). The relative enzymatic degradation of ITdU by TP was ∼3.6% of that of nonmodified substrate IdUrd. This is in accordance with previously reported results (17).

In conclusion, these results suggest that combined administration of ITdU and FdUrd as TS inhibitor improves significantly and selectively ITdU uptake in tumor cells. ITdU was incorporated into DNA resulting in prolonged intracellular retention of ITdU in malignant tissue. Our study provides distinct evidence for in situ induced apoptotic effects, mediated by Auger electron emitters stably incorporated into DNA and leading to extensive damage of tumor tissue and only moderate radiocytotoxicity in normal tissues. Therefore, ITdU is a promising candidate for selective targeting of malignant cells in vivo in a therapeutic setting.

No potential conflicts of interest were disclosed.