Bioconjugate Chem. 2007, 18, 754−764
754
Evaluation of Chemical, Physical, and Biologic Properties of Tumor-Targeting Radioiodinated Quinazolinone Derivative Ketai Wang, Agop M. Kirichian, Ayman F. Al Aowad, S. James Adelstein, and Amin I. Kassis* Harvard Medical School, Department of Radiology, 200 Longwood Avenue, Boston, Massachusetts 02115. Received September 18, 2006; Revised Manuscript Received November 28, 2006
Our group is developing a novel technology, enzyme-mediated cancer imaging and therapy (EMCIT), that aims to entrap radioiodinated compounds within solid tumors for noninvasive tumor detection and therapy. In this approach, a water-soluble, radioiodinated prodrug is hydrolyzed in vivo to a highly water-insoluble compound by an enzyme overexpressed extracellularly by tumor cells. We have synthesized and characterized the water-soluble prodrug, 2-(2′-phosphoryloxyphenyl)-6-[125I]iodo-4-(3H)-quinazolinone [125I]5, which is readily hydrolyzed by alkaline phosphatase, an enzyme expressed by many tumor cell lines, to a water-insoluble drug, 2-(2′hydroxyphenyl)-6-[125I]iodo-4-(3H)-quinazolinone [125I]1. In the course of our study, we discovered that ammonium 2-(2′-phosphoryloxyphenyl)-6-tributylstannyl-4-(3H)-quinazolinone, an intermediate in the radioiodination of the prodrug, exists as two isomers (3 and 4) whose radioiodination leads, respectively, to [125I]6 and [125I]5. These prodrugs have different in vitro and in vivo biologic activities. Compound 6 is not hydrolyzed by alkaline phosphatase (ALP), whereas 5 is highly soluble (mg/mL) in aqueous solution and is rapidly dephosphorylated in the presence of ALP to 1, a water-insoluble molecule (ng/mL). Mouse biodistribution studies indicate that [125I]6 has high uptake in kidney and liver and [125I]5 has very low uptake in all normal organs. Compounds 3 and 6 are converted, respectively, to 4 and 5 after incubation in DMSO. The stability of 5 in human serum is high. The minimum ALP concentration needed to hydrolyze 5 is much greater than the ALP level in the blood of patients with cancer, and the latter should not affect the pharmacokinetics of the compound. Incubation of 5 with viable human and mouse tumor-cell linessbut not with normal human cells and mouse tissuessleads to its hydrolysis and the formation of large crystals of 1. We expect that 5 will also be hydrolyzed in vivo by tumor cells that express phosphatase activity extracellularly and anticipate the specific precipitation of radioiodinated 1 within tumor cell clusters. This should lead to high tumor-to-normal-tissue ratios and enable imaging (SPECT/PET) and radionuclide therapy of solid tumors.
INTRODUCTION We are developing a novel strategy called enzyme-mediated cancer imaging and therapy (EMCIT), which is a method for enzyme-dependent, site-specific, in vivo precipitation of a watersoluble radioactive molecule (prodrug) within the extracellular space of solid tumors (1). The radiolabeled prodrug is hydrolyzed to its water-insoluble form by a hydrolase overexpressed on the extracellular surface of tumor cells, and the resulting water-insoluble drug is expected to be trapped within the extracellular spaces of the targeted solid tumor. Following tumor targeting and clearance from normal tissues and organs, the entrapped radioactive compound should lead to high tumor-tonormal-tissue ratios and enable the imaging (SPECT/PET) and radiotherapy of solid tumors. Alkaline phosphatase (ALP) was investigated to ascertain the validity of the proposed approach, since this hydrolase (i) is known to be minimally expressed on the plasma membrane of normal cells and overexpressed on the membrane of many human tumor cells (2-9), (ii) is post-translationally modified with two N-glycosylation sites and a glycosylphosphatidylinositol (GPI) anchor by which it is tethered to the cell membrane, (iii) has a reasonably high second-order rate constant (kcat/Km ) 5 × 107 M-1 s-1), and (iv) has already been used to hydrolyze several chemotherapeutic prodrugs in vitro and in * To whom correspondence should be addressed at Harvard Medical School, Armenise Building, Room D2-137, 200 Longwood Avenue, Boston, Massachusetts 02115. Telephone: (617) 432-7777. Fax: (617) 432-2419. E-mail:
[email protected].
tumor-bearing animals in vivo (10-14). Additionally, (i) a highresolution structure of the human placental isoform (PLAP) has been determined at 1.8 Å resolution and is available from the Protein Data Bank (code 1EW2) for analysis (15), and (ii) ALPexpressing tumors in cancer patients and tumor-bearing mice have been successfully imaged with radiolabeled antiALP antibodies (16-20). Consequently, we have been carrying out in silico molecular modeling to systematically investigate the structure-activity relationships of phosphorylated quinazolinone derivatives with PLAP. In 2002, we reported the synthesis and characterization of ammonium 2-(2′-phosphoryloxyphenyl)-6-iodo-4-(3H)-quinazolinone (5) and 2-(2′-hydroxyphenyl)-6-iodo-4-(3H)-quinazolinone (1) (1). In these studies, we demonstrated that nonfluorescent and water-soluble 5 is hydrolyzed by alkaline phosphatase to water-insoluble and highly fluorescent 1. We also synthesized ammonium 2-(2′-phosphoryloxyphenyl)-6-tributylstannyl-4(3H)-quinazolinone (4) (Scheme 1), a precursor for the synthesis of 5, and found that the intravenous injection of [125I]5 in normal mice leads to high activity (∼5%-7% ID/g) in normal liver and kidneys (1). Having observed that the pharmacokinetics of [125I]5 in mice are very different when the prodrug is radioiodinated using 4 that has been dissolved in DMSO, we have been reevaluating the findings and interpretation of some of our earlier studies. It has become evident that the tin precursor of 5 is a mixture of two compoundss3 (MW 590) and 4 (MW 608)swhose radioiodination results in the formation of a mixture of [125I]6 (MW 426) and [125I]5 (MW 444). We describe herein the
10.1021/bc0602937 CCC: $37.00 © 2007 American Chemical Society Published on Web 03/27/2007
Bioconjugate Chem., Vol. 18, No. 3, 2007 755
Enzyme-Mediated Cancer Imaging and Therapy Scheme 1. Pathway for Synthesis of Tributyltin and
125
I/127I Quinazolinone Derivatives
syntheses of the derivatives and their chemical and biologic characterization.
EXPERIMENTAL PROCEDURES General Remarks. Alkaline phosphatase (bovine intestinal mucosa, calf intestinal mucosa, human placental type XXIV, porcine kidney) and all other chemicals were obtained from Sigma-Aldrich Incorporated. Carrier-free sodium [125I]iodide was purchased from Amersham Corporation. HPLC separations (Waters) were performed on a reversed-phase Zorbax SB-C18 column, 9.4 × 250 mm (Agilent Technology), with UV absorption (Waters 486 detector) and γ-ray detection (gammaram, IN/US Systems) used to analyze the eluates. Syntheses (Scheme 1) and Chemical and Biophysical Characterization. Ammonium 2-(2′-Phosphoryloxyphenyl)-6tributylstannyl-4-(3H)-quinazolinone (3 and 4). 2-(2′-Hydroxyphenyl)-6-iodo-4-(3H)-quinazolinone (1) and 2-(2′-hydroxyphenyl)-6-tributylstannyl-4-(3H)-quinazolinone (2) were synthesized as previously reported (1). To phosphorylate the tin derivative, phosphorus oxychloride (5 µL, 53.6 µmol) in dry pyridine (0.5 mL) was added dropwise to a stirred solution of 2 (25 mg, 47.5 µmol) in dry pyridine (0.4 mL) cooled to 0 °C. The solution turned yellow. The reaction mixture was stirred for 1 h at 0 °C and then was neutralized by adding ammonium hydroxide (0.5 mL diluted from 15 µL 28% ammonium hydroxide solution). The solvent was evaporated under reduced pressure (rotary evaporator), and a light-yellow solid was produced. HPLC and 31P NMR showed two peaks, 3 and 4 (Figure 1A,C). When a mixture of 3 and 4 (0.5 mg) was dissolved in DMSO (100 µL) and left overnight, a solution containing only 4 was obtained (Figure 1B,D). ESI-HRMS: calcd for 3 [M - H]- 589.1296, found 589.1296; calcd for 4 [M + H]+ 609.1540, found 609.1539. 31P NMR for 3 and 4: -16.922 ppm and -2.958 ppm, respectively.
Ammonium 2-(2′-Phosphoryloxyphenyl)-6-[125I]iodo-4-(3H)quinazolinone ([125I]5 and [125I]6). Ammonium 2-(2′-phosphoryloxyphenyl)-6-tributylstannyl-4-(3H)-quinazolinone (mixture of 3 and 4, 0.5 mg) was dissolved in DMSO (100 µL) for 1 min, 3 h, or overnight; and 1 µL (5 µg/µL DMSO solution) was added to a reaction vial coated with 1,3,4,6-tetrachloro3R,6R-diphenylglycouril (Iodogen, 10 µg), along with phosphate buffer (12 µL, 0.1 M, pH 7.4) and the desired quantity of Na125I (0.2-2.0 mCi, 400 µCi/µL). After vortex mixing at ambient temperature for 5 min, the solution was analyzed by HPLC (linear gradient from 10% disodium hydrogen phosphate, 0.05 M, pH 2.5, to 100% methanol for 6 min; flow rate of 3 mL/ min) for the presence of [125I]5 and [125I]6. 2-(2′-Hydroxyphenyl)-6-[ 125I]iodo-4-(3H)-quinazolinone 125 ([ I]1). Compound [125I]5 (5 µL, 200 µCi) was incubated with ALP (10 U/2 µL) at 37 °C for 5 min. The product [125I]1 was identified by HPLC against a 2-(2′-hydroxyphenyl)-6-[125I]iodo4-(3H)-quinazolinone standard. In addition, 2, incubated for 5 min in DMSO, was radioiodinated using Iodogen and Na125I as described above, and the solution was analyzed by C18 HPLC. Ammonium 2-(2′-Phosphoryloxyphenyl)-6-iodo-4-(3H)quinazolinone (5 and 6). Compound 1 (25 mg, 47.4 µmol) was added to dried pyridine (0.5 mL) at 0 °C, followed by phosphorus oxychloride (5 µL, 53 µmol) dissolved in dried pyridine (0.5 mL) under argon gas. The reaction was complete within 1 h, as indicated by silica gel TLC. The solution was cleared and neutralized to pH 7 by adding ammonium hydroxide (0.5 mL diluted from 15 µL 28% ammonium hydroxide solution). At the end of neutralization, the solution turned light yellow. The solvent was evaporated, and a mixture of yellow solids including 5 and 6 was obtained. After an overnight incubation of the mixture in DMSO, 6 was converted to 5, as indicated by HPLC and 31P NMR with one peak present at -2.897. ESI-HRMS: calcd for 5 [M + H]+ 444.9451, found
756 Bioconjugate Chem., Vol. 18, No. 3, 2007
Wang et al.
Figure 1. HPLC (A and B) and 31P NMR (C and D) of stannylated intermediates 3 and 4, demonstrating conversion of 3 to 4 upon incubation in DMSO.
444.9443; calcd for 6 [M - H]- 424.9175, found 424.9188. 31P NMR for 5 and 6: -2.897 ppm and-16.568 ppm, respectively. Water Solubility of 2-(2′-Hydroxyphenyl)-6-[125I]iodo-4-(3H)quinazolinone ([125I]1) and 2-(2′-Phosphoryloxyphenyl)-6-[125I]iodo-4-(3H)-quinazolinone ([125I]5). Since 1 is highly waterinsoluble, a novel radiotracer method was developed to determine its solubility (21). In this approach, various volumes of a saturated solution of 127I-labeled 1, spiked with 125I-labeled 1, in DMSO are placed in centrifuge tubes, and sufficient deionized water is added to make the final volumes equivalent. After vortex mixing, the tubes are centrifuged. Since (i) the 127I- and 125I-labeled compounds have the same solubilities and are, therefore, equally distributed in the DMSO-water solution, and (ii) the amount of nonradioactive compound (added in the DMSO solution) is known, the concentration of the compound in solution can be calculated and plotted as a function of the DMSO-to-water ratio. The water solubility of the compound is then determined by extrapolation of the linear fit of the data points to zero DMSO, i.e., the y-axis intercept when x ) 0 (Origin 7.0, OriginLab Corporation, Northampton, MA). The water solubility of prodrug 5 was assessed by the conventional UV spectrophotometry method. Kinetics of 2-(2′-Phosphoryloxyphenyl)-6-[125I]iodo-4-(3H)quinazolinone ([125I]5) Hydrolysis by Alkaline Phosphatase. Compound 5 (4 µg) was incubated with bovine ALP (10 U/mL) at 37 °C for 8 min. The kinetics of hydrolysis of nonfluorescent, water-soluble 5 to fluorescent, water-insoluble 1 were followed with a luminescence spectrometer (Perkin-Elmer S50B, excitation at 340 nm, emission at 504 nm). To assess the kinetics of the hydrolysis, the Michaelis-Menten constant of the enzyme (Km) and the maximum velocity of the reaction (Vmax) were determined by a Lineweaver-Burke straight-line plot with a slope equal to Km/Vmax, where the y-intercept of the line is 1/Vmax and the x-intercept is -1/Km.
To determine the minimum concentration of ALP required to hydrolyze [125I]5 to [125I]1, [125I]5 was incubated with various concentrations of human, bovine, porcine, and calf ALP (between 0.0001 and 1 U/µL PBS) for 5 min, and the formation of [125I]1 was assessed by HPLC. After integrating the peak area of [125I]1, the percentage conversions of [125I]5 to [125I]1 were determined. The mean ((sem) values of [125I]1 (at the same ALP concentration for all four phosphatases) were calculated and plotted versus each ALP concentration. A fit of the data points furnished a sigmoidal line from which the IC50 value as well as the minimum concentration of ALP needed for hydrolysis was calculated. Compound [125I]5 (2 µL, 0.1 mCi/µL) was also incubated at 37 °C with various human ALP concentrations (0.000 01, 0.0001, 0.001, 0.01, and 0.1 U/µL PBS) for 5 min, 1 h, 12 h, 24 h, and 48 h. At each of these incubation times, the concentration of ALP needed to convert 50% [125I]5 to [125I]1 was determined (IC50) and plotted against time. Stability of Radiolabeled Prodrug ([125I]5) in Serum. Compound [125I]5 (2 µL, 0.2 mCi/µL) was incubated at 37 °C with human and other animal sera (18 µL) for 48 h. The percentage of [125I]5 remaining in the serum solution was measured at multiple times by HPLC, and the breakdown products of the prodrug were analyzed by HPLC. Alkaline phosphatase (10 U/20 µL) was also added to human serum-[125I]5 mixtures, and the ability of the hydrolase to dephosphorylate the quinazolinone was determined 5 min later. Biologic Studies. Mammalian-Cell-Mediated Hydrolysis of 2-(2′-Phosphoryloxyphenyl)-6-[125I/127I]iodo-4-(3H)-quinazolinone ([125I]5/5) to 2-(2′-Hydroxyphenyl)-6-[125I/127I]iodo-4-(3H)quinazolinone ([125I]1/1). The cell-mediated hydrolysis of the nonfluorescent, water-soluble prodrug (5) to the fluorescent, water-insoluble drug (1) was assessed in the presence of human and mouse tumor cells, including human breast cell lines (BT474 and MCF7), human ovarian carcinoma (SKOV-3, OVCAR-
Enzyme-Mediated Cancer Imaging and Therapy
Bioconjugate Chem., Vol. 18, No. 3, 2007 757
Figure 2. HPLC profiles of [125I]1, 1, [125I]5, 5, and [125I]6, demonstrating (i) time-dependent conversion of 3 to 4 upon preincubation of mixture in DMSO for 99%. After the mixture of 3 and 4 is incubated at room temperature in DMSO overnight and then radioiodinated, the radiolabeled yield is ∼100% 5 (Figure 2B). Similarly, the incubation of a mixture of 6 and 5 in DMSO for 1 h also results in the complete conversion of 6 to 5. Why the conversion of 6 to 5 occurs faster (within 2 h) than that of 3 to 4 (within 24 h) is unclear. Unlike the quantitative yield that occurs when 4 is radioiodinated (Figure 2B), the treatment of 2 with Iodogen and Na125I leads to the formation of a complex mixture of radioiodinated products that includes [125I]1 (∼16% yield) and several yet-to-
Enzyme-Mediated Cancer Imaging and Therapy
Bioconjugate Chem., Vol. 18, No. 3, 2007 759
Figure 5. Kinetics of hydrolysis (HPLC quantitative analysis) of [125I]5 to [125I]1 determined after 5 min incubation in various concentrations of human, bovine, porcine, and calf ALP (PBS, pH 7.4), demonstrating that >1000 U/L is necessary to initiate hydrolysis. Sample HPLC profiles (with human ALP) are shown on right.
Figure 6. IC50 of ALP (bovine intestinal mucosa)-mediated hydrolysis of [125I]5 as function of time (0.1 M PBS, 37 °C, pH 7.4).
be identified 125I-labeled products (Figure 3). Our current approach for the synthesis of pure [125I]1 is via the ALPmediated dephosphorylation of [125I]5. Water Solubility of 2-(2′-Hydroxyphenyl)-6-iodo-4-(3H)quinazolinone (1) and 2-(2′-Phosphoryloxyphenyl)-6-iodo-4(3H)-quinazolinone (5). The solubility of 1, as measured by the radiotracer method (21), is 1.9 ng/mL. The solubility of 5, determined by spectrophotometry, is 7.8 mg/mL. Therefore, 5 is 4.18 × 106 times more soluble than 1. Kinetics of 2-(2′-Phosphoryloxyphenyl)-6-iodo-4-(3H)-quinazolinone (5) Hydrolysis by Alkaline Phosphatase. The ALP conversion of water-soluble, nonfluorescent prodrug 5 to waterinsoluble, fluorescent drug 1 was determined spectroscopically. The fluorescence intensity at 504 nm increases with time (Figure 4). Further analysis of the data indicates that the enzymedependent appearance of 1 is biphasic with T1/2R of 81 s and T1/2β of 557 s.
Figure 7. Stability of [125I]5 in various types of mammalian serum (at 37 °C). T1/2R in mouse serum ) 2.0 h; T1/2R in rat serum ) 15.6 h; T1/2R in human serum ) several months. Complete hydrolysis of [125I]5 is seen upon 5 min incubation in ALP (0).
HPLC analysis has been utilized to determine the minimum concentration of ALP necessary for the hydrolysis of [125I]5. In these studies, [125I]5 was incubated (room temperature, 5 min) with various concentrations (U/µL) of ALP (human, bovine, porcine, and calf), and the appearance of dephosphorylated [125I]1 was quantified. The HPLC peak corresponding to that of substrate [125I]5 (tR ≈ 8 min) disappears as a function of increasing ALP concentration (Figure 5). Simultaneously, a new peak appears at tR ≈ 12 min that matches the retention-time value of nonradiolabeled 1. Complete conversion of [125I]5 to [125I]1 occurs when the ALP concentration is >106 U/L. A plot of formation of [125I]1 versus ALP concentration furnishes a sigmoidal curve (IC50 ) 23 000 U/L) indicating that the minimum ALP concentration needed to hydrolyze [125I]5 is g1000 U/L (Figure 5).
760 Bioconjugate Chem., Vol. 18, No. 3, 2007
Wang et al.
Figure 8. Fluorescence microscopy of viable human (A-G) and mouse (H) tumor cell lines incubated in vitro for 24 h with 5 showing its hydrolysis and crystallization of 1 (green). OVCAR-3 (A,B) and SKOV-3 (C) ovarian carcinoma cells; BT-474 (D) and MCF7 (E,F) breast cancer cells; HTB-173 small-cell lung carcinoma cells (G); and LLC1 Lewis lung carcinoma cells (H). Incubation of 5 with viable normal human mammary epithelial cells (HMEC) does not induce any observable hydrolysis (I). A and E: cells not washed, low power; B-D and F-I: cells washed, and counterstained with DAPI (blue nuclei), high power.
When [125I]5 is incubated with ALP (g1 U/µL) for 5 min at room temperature, the compound is rapidly hydrolyzed, and a new peak appears on HPLC whose retention time is identical to that of [125I]1 (Figure 2B,E). However, when a mixture of [125I]5 and [125I]6 is incubated with ALP, only [125I]5 is converted to [125I]1 (55% of mixture) (Figure 2A,D). HPLC has also been used to determine the IC50 when the incubation time of [125I]5 in varying ALP concentrations is changed. When the log IC50 is plotted as a function of log time, a linear relationship is obtained (Figure 6). Stability of Radiolabeled 2-(2′-Phosphoryloxyphenyl)-6-[125I]iodo-4-(3H)-quinazolinone ([125I]5) in Animal and Human Serum. Pure [125I]5 (20 mCi/mL) kept in buffer for a week is still 100% radiochemically unadulterated. However, this compound has very different stabilities in animal and human serum (Figure 7). In mouse serum, its half-life is 2.0 h, whereas in human serum, even after 48 h, 94% of [125I]5 is unchanged radiochemically. HPLC analysis indicates that the product of decomposition is the same in various types of serum and has the same retention-time peak as [125I]1. That the observed stability of [125I]5 in human serum is not due to the presence of alkaline phosphatase inhibitors is demonstrated by the quantitative and rapid hydrolysis of [125I]5 to [125I]1 upon the addition of alkaline phosphatase to the serum samples. Biologic Studies. Mammalian-Cell-Mediated Hydrolysis of 2-(2′-Phosphoryloxyphenyl)-6-iodo-4-(3H)-quinazolinone (5) to 2-(2′-Hydroxyphenyl)-6-iodo-4-(3H)-quinazolinone (1). We determined the ability of human ovarian carcinoma cells (OVCAR-3 and SKOV-3), human breast cancer cells (BT-474 and MCF7), human small-cell lung carcinoma cells (HTB-173), and mouse Lewis lung carcinoma (LLC1) to hydrolyze water-soluble 5 to
fluorescent, water-insoluble 1. The in vitro incubation (37 °C, pH 7.4) of all these tumor cell lines with 5 leads to the formation of very large (e20 µm) fluorescent crystals (Figure 8A-H), many of which are irreversibly bound to the cells (Figure 8B-D and 8F-H). Furthermore, autoradiography of viable human squamous-cell lung carcinoma cells (HTB-182) incubated with [125I]5 shows the presence of intense grains (Figure 9), consistent with the dephosphorylation of this radioiodinated analog and the formation of water-insoluble [125I]1. To ascertain that the hydrolysis of 5 by tumor cells is ALPmediated, human tumor cells (OVCAR-3) cells were incubated with the quinazolinone derivative in the presence and absence of levamisole. The presence of this specific inhibitor of alkaline phosphatase (22, 23) completely inhibits 5 dephosphorylation, as exemplified by the absence of any fluorescent crystal formation (Figure 10B). As expected, 1 is formed in the absence of levamisole (Figure 10A). We also assessed the ability of normal human cells and mouse tissues to hydrolyze 5. No hydrolysis (no fluorescent 1 crystal formation) is observed when human mammary epithelial cells (HMEC) are incubated with 5 (Figure 8I) or when this derivative is incubated in medium (in the absence of cells). Similar results (minimal fluorescence or its absence) have also been obtained following a 24 h incubation of normal mouse tissues (kidneys, liver, and spleen) with 5 (Figure 11A-C). Finally, the incubation of mammalian cells up to 48 h with 5 (1 mg/mL) does not lead to any microscopic alterations in their appearance and in their attachment to the incubation flasks. Ex ViVo ConVersion of 2-(2′-Phosphoryloxyphenyl)-6-iodo4-(3H)-quinazolinone (5) to 2-(2′-Hydroxyphenyl)-6-iodo-4(3H)-quinazolinone (1) by Solid Tumors Grown in ViVo. Human ovarian SKOV-3 tumors grown in the peritoneal cavity of nude
Bioconjugate Chem., Vol. 18, No. 3, 2007 761
Enzyme-Mediated Cancer Imaging and Therapy
Figure 9. Autoradiography of viable human squamous-cell lung carcinoma cells (HTB-182) after 24 h, in vitro incubation (37 °C, pH 7.4) with [125I]5, demonstrating its hydrolysis and formation of intense black grains where [125I]1 is formed. Cells counterstained with DAPI (blue nuclei).
mice and mouse B16F10 melanoma tumor metastases present within the spleen of mice were dissected and incubated ex vivo with 5. Under both circumstances, fluorescence microscopy indicates the presence of fluorescent crystals (1) that are associated with the SKOV-3 solid tumors (Figure 11D) and within the B16F10 melanoma-infiltrated spleens (Figure 11E). Biodistribution of 2-(2′-Phosphoryloxyphenyl)-6-iodo-4-(3H)quinazolinone [125I]5 and [125I]6 in Mice. Table 1 contains the biodistribution data obtained 1 and 24 h after the i.v. injection of [125I]5 and [125I]6 into mice. Compound [125I]5 has minimal uptake in all normal organs and tissues with less than 0.2% ID/g remaining by 24 h. On the other hand, the i.v. injection of [125I]6 (mixed with [125I]5, ∼50:50) demonstrates substantially higher uptake in the kidneys, liver, lungs, and blood (∼2%-12% ID/g). Since the % ID/g values for [125I]5 are all very low, the values obtained for the mixture of [125I]5 and [125I]6, in fact, reflect those of [125I]6. The blood clearance of [125I]5 is triphasic with T1/2R of 13 min, T1/2β of 1.3 h, and T1/2γ of 18.1 h, whereas the whole body clearance has a T1/2 of 5.3 h (Figure 12).
Figure 10. Fluorescence microscopy of human OVCAR-3 ovarian carcinoma cells after 24 h, in vitro incubation (37 °C, pH 7.4) with 5 (10 mM levamisole, showing absence of hydrolysis and lack of crystallization to 1 (green) in presence of ALP-specific inhibitor levamisole. Inserts: cells washed and counterstained with DAPI, high power. Table 1. Biodistribution of [125I]5 and [125I]6 Following Intravenous Injection into C57BL/6 Micea [125I]5-[125I]6 mixture (∼50:50)
[125I]5
DISCUSSION Ideally, radiolabeled agents for cancer imaging and therapy must (i) be stable in blood following their administration in an animal, (ii) be taken up rapidly by tumors (T1/2 in circulation much shorter than physical T1/2 of the radionuclide), (iii) be retained for long periods of time within tumors (T1/2 in tumor much longer than physical T1/2 of the radionuclide), (iv) be concentrated efficiently by tumors (i.e., high % ID/g), (v) be taken up minimally by normal tissue cells, (vi) have a short residence in normal tissues, (vii) achieve high tumor-to-normaltissue uptake ratios, and (viii) be labeled with an emitter whose decay characteristics are suitable for imaging (SPECT or PET) or therapy. We believe that the EMCIT strategy meets many of these requirements. First, being molecules of low molecular weight (MW 6-fold less soluble than 5) do not transit through the tissues and are therefore indefinitely entrapped (1). Our current results indicate the direct radioiodination of the stannylated quinazolinone derivative 2 produces a mixture of products (Figure 3) and the yield of [125I]1 among these products is only 16%. This mixture of radioiodinated products may be a consequence of the presence of the hydroxyl group on the benzene ring, an activating group that permits introduction at its neighboring 3-position of a second group by electrophilic substitution. When the hydroxyl group within stannylated 2 is phosphorylated to stannylated prodrugs 3 and 4 and this mixture is left in DMSO for >24 h, all of 3 converts to 4 (Figure 1). Since the reactivity of the various positions on the phosphorylated benzene ring to halogenation is now lost, the radioiodination of 4 leads to the production of a single radioiodinated quinazolinone derivative 5 in greater than 99% yield (Figure
2B). The hydrolase ALP is then used to quantitatively and rapidly (106 U/L, and (ii) the minimum ALP concentration needed to hydrolyze [125I]5 is g1000 U/L, a value that is 2-fold higher than the 500 U/L found in the blood of cancer patients (29, 42-48). Consequently, such radioiodinated quinazolinone analogues are not expected to be hydrolyzed en route following their intravenous injection into cancer patients.
CONCLUSIONS The prodrug ammonium 2-(2′-phosphoryloxyphenyl)-6-iodo4-(3H)-quinazolinone exists in two forms, 5 (MW 444) and 6 (MW 426). Compound 6 is not hydrolyzed by alkaline phosphatase and has a high uptake in key nontargeted organs; its presence needs to be avoided during preparation of the radiopharmaceutical, and this can be accomplished by leaving the mixture of 3 and 4 in DMSO for at least 24 h. Compound 6 is also readily converted to 5 in the presence of DMSO. Compound 5 is stable in human serum and is completely (>99%) and rapidly (
