Synthesis and Evaluation of Two Technetium-99m-Labeled Peptidic 2

Imaging Centre, 600 University Avenue, Toronto, Ontario M5G 1X5, Canada, and Resolution .... Pharmaceutical Sciences and University Medical Imaging...
1 downloads 0 Views 152KB Size
Download PDF
Bioconjugate Chem. 1999, 10, 897−904

897

Synthesis and Evaluation of Two Technetium-99m-Labeled Peptidic 2-Nitroimidazoles for Imaging Hypoxia Zi-Fen Su,†,‡ Xiuguo Zhang,†,‡ James R. Ballinger,‡,§ A. M. Rauth,*,†,‡ Alfred Pollak,| and John R. Thornback| Departments of Medical Biophysics and Pharmaceutical Sciences, University of Toronto, 610 University Avenue, Toronto, Ontario M5G 2M9, Canada, Division of Experimental Therapeutics, Ontario Cancer Institute, 610 University Avenue, Toronto, Ontario M5G 2M9, Canada, Division of Nuclear Medicine, University Medical Imaging Centre, 600 University Avenue, Toronto, Ontario M5G 1X5, Canada, and Resolution Pharmaceuticals Inc., 6850 Goreway Drive, Mississauga, Ontario L4V 1V7, Canada. Received May 4, 1999; Revised Manuscript Received June 29, 1999

The presence of hypoxic cells in solid tumors is a marker for therapy-resistant, aggressive disease. The noninvasive detection of hypoxic cells in tumors by radiolabeled 2-nitroimidazoles is a diagnostic technique under current evaluation. Two peptidic agents, dimethylglycyl-L-seryl-L-cysteinyl-lysyl{N[1-(2-nitro-1H-imidazolyl)acetamido]}glycine (RP435) and dimethylglycyl-tert-butylglycyl-L-cysteinylglycine-[2-(2-nitro-1H-imidazolyl)ethyl]amide (RP535) have been synthesized. Both agents contain an N3S class chelator for 99mTc and Re and a 2-nitroimidazole group which can be enzymatically reduced and selectively trapped in cells under hypoxic conditions. Two isomers of 99mTcO-RP435, which are assumed to be syn and anti conformations, were observed on HPLC analysis. The interconversion of the two isomers in aqueous solution was investigated. In contrast, RP535 chelated 99mTc to form a single isomer and no conversion to its counterpart has been observed on HPLC analysis. The tertbutyl group on the chelator may inhibit the formation and interconversion of the syn and anti isomers of 99mTcO-RP535. Both tracers showed a significant degree of hypoxia-specific accumulation in an in vitro assay, with 99mTcO-RP535 showing higher selectivity for hypoxic cells than 99mTcO-RP435. These results suggest that 99mTcO-RP535 represents a lead compound worthy of further investigation as an agent for imaging hypoxia in tumors.

INTRODUCTION

Detection of regions of hypoxia is important in several medical conditions. It is important to delineate hypoxic but viable tissue following interruption of blood flow in cerebral or myocardial infarction (1). Furthermore, it has recently been recognized that hypoxia in tumors plays a role not only in response to radiation, through the oxygen effect in fixation of DNA damage, but also in metastatic potential and response to other forms of therapy (2-4). However, hypoxia is heterogeneous, and tumors which appear identical by all other clinical measures can vary greatly in their proportion of hypoxic cells. At present, hypoxia is measured clinically with a polarographic oxygen electrode inserted into tumors; however, this is invasive, operator dependent, limited to accessible tumors, and not readily repeatable (2, 4). In recent years, several radiopharmaceuticals have been developed which are targeted to hypoxic tissues via a 2-nitroimidazole (2NI) moiety (1, 5). 2-NIs are electron-affinic compounds which undergo an enzyme-mediated one-electron reduction; in normoxic tissue, the resultant radical anion is immediately oxidized back to the starting compound, whereas under hypoxic conditions there is further reduction to products which are trapped by binding to macro* To whom correspondence should be addressed. Phone: (416) 946-2977. Fax: (416) 946-2984. E-mail: [email protected]. † Department of Medical Biophysics. ‡ Ontario Cancer Institute. § Pharmaceutical Sciences and University Medical Imaging Centre. | Resolution Pharmaceuticals Inc.

molecules (6). A radiolabeled 2-NI would therefore be selectively trapped in hypoxic tissues. Clinical studies have been performed with several radiolabeled 2-NIs, including [18F]fluoromisonidazole (FMISO) (7, 8) and [123I]iodoazomycin arabinoside (IAZA) (9, 10). However, the search continues for a 99mTc-labeled analogue that would be more widely applicable. BMS181321, a 2-NI linked to a PnAO chelator, has been evaluated in several model systems and shows promise, but its high lipophilicity and metabolic instability result in high background levels of radioactivity (11-14). BRU5921 (previously known as BMS194796) appears to possess improved biodistribution characteristics, but only limited information is available about it (15, 16). In addition, butyleneamine oxime (BnAO or HL91; Prognox, Nycomed-Amersham) is a nonnitro 99mTc complex that shows localization in hypoxic myocardium and tumors via an undetermined mechanism (17-20). Recent work involving radioiodinated sugar derivatives of 2-NI has suggested that an octanol/water partition coefficient (PC) of ∼1 confers optimal biodistribution properties for tumor imaging (5). BMS181321 and BRU5921 have PCs which are much higher (40 and 12, respectively) and cannot be readily modified to achieve a lower PC. As part of our radiopharmaceutical development program, we designed two compounds in which peptidic chelators for 99mTc were linked to 2-NI and evaluated the complexes in an in vitro model of tumor hypoxia. Dimethylglycyl-L-seryl-L-cysteinyl-glycinamide (RP294, Figure 1), which contains a sulfur atom (protected by an acetoamidomethyl group), one amine nitrogen atom, and two

10.1021/bc9900542 CCC: $18.00 © 1999 American Chemical Society Published on Web 08/27/1999

898 Bioconjugate Chem., Vol. 10, No. 5, 1999

Su et al.

the oxo moiety in the apical position. The metal complexes exist in syn and anti isomers with respect to the position of the metal-oxo bond and serine hydroxymethyl side chain. Interconversion of the two isomers in aqueous solution at room temperature has been observed (21). In the present work, N,N-dimethylglycyl-L-seryl-Lcysteinyl-lysyl{N-[1-(2-nitro-1H-imidazolyl)acetamido]}glycine (RP435, Figure 1) was prepared by attaching 2-NI-acetic acid to a lysine linker connected to the RP294 chelator. RP435 chelates 99mTc at room temperature, and two peaks, assumed to be the syn and anti isomers, were detected by HPLC. However, the PC of 99mTcO-RP435 is far lower than what is believed to be the optimum value. Accordingly, a new N3S class peptide, N,N-dimethylglycyl-L-tert-butylglycyl-L-cysteinyl-glycin[2-(2-nitro1H-imidazolyl)ethyl]amide (RP535, Figure 1) was designed and synthesized. The chelator (RP455, Figure 1) contains a tert-butyl group, which should increase the lipophilicity of the corresponding 99mTc complex and inhibit the interconversion of syn and anti isomers. The labeling, stability, and in vitro cellular accumulation of these compounds was studied. EXPERIMENTAL PROCEDURES

Figure 1. Chemical structures of two peptidic N3S chelators, their 2-nitroimidazole-linked derivatives, and the ReO complex of RP535. Numbers refer to proton assignments.

amide nitrogen atoms, is an N3S class chelator for 99mTcvO and RevO in a distorted square pyramidal geometry with

2-Nitroimidazole, 2-bromoethylphthalimide, N,N-dimethylformamide (DMF), 1-(3-dimethylaminopropyl)-3ethylcarbodiimide hydrochloride (DECI), 2,3,5,6-tetrafluorophenol (TFP), and trifluoroacetic acid (TFA) were purchased from Aldrich Chemical Co. (Milwaukee, WI). N,N-Dimethylglycyl-L-tert-butylglycyl-L-cysteinyl-glycine (RP455) was custom synthesized by Bachem Bioscience Inc. (Philadelphia, PA). N,N-Dimethylglycyl-Lseryl-L-cysteinyl-glycinamide (RP294) and ReO-RP455 precursor were provided by Resolution Pharmaceuticals Inc. (Mississauga, ON). Sodium 99mTcO4 in saline was obtained from a 99Mo/99mTc generator (DuPont Pharma, Billerica, MA). R-MEM and fetal bovine serum were purchased from Sigma Chemical Co. (St. Louis, MO). NMR data (1H, 13C, COSY, HSQC, HMQC, TOCSY) were recorded on a Varian UNITYplus-500 spectrometer (500 MHz) with TMS as external standard. Electrospray mass spectra were obtained on a Sciex API#3 spectrometer in the positive ion mode. Reversed-phase highpressure liquid chromatography (HPLC) analyses of peptide and 99mTc-peptide complexes were carried out on a Beckman model 125 System Gold (Fullerton, CA), with Zorbax SB or Beckman ODS 4.6 × 250 mm 5 µm C18 columns. Purification of peptides was accomplished by HPLC with a Waters C18 RCM 8 × 10 semipreparative column. UV and radiometric detectors were connected in series. Two mobile phases were used for HPLC analysis of 99mTc tracers. Mobile phase I was composed of H2O/ ACN containing 0.1% TFA, whereas mobile phase II was 0.02 M, pH 4.6, NH4OAc buffer/MeOH; both were run as gradients. The flow rate was 1.0 mL/min for analysis and 2.0 mL/min for purification. Synthesis of 2-Nitroimidazole Acetic Acid. To 1.0 g (8.8 mmol) 2-nitroimidazole and 2.0 g (14.5 mmol) potassium carbonate in 10 mL dry acetonitrile was added 1.20 mL (10.6 mmol) of bromoethyl acetate. The reaction mixture was stirred for 24 h at room temperature followed by filtration and washing of the solid product with acetone. Evaporation of the pooled filtrates under reduced pressure yielded 2.29 g of crude 2-nitroimidazole ethyl acetate as a yellow oil. To the crude oil was added a solution of NaOH (4 N, 2.5 mL), water (10 mL), and MeOH (10 mL). The solution was stirred and monitored

Tc-99m Peptidic 2-Nitroimidazoles

by thin-layer chromatography (silica gel; CH2Cl2:MeOH: NH3, 50:7:1) until none of the ester remained (∼30 min). The solution was acidified by addition of a cationexchange resin (Bio-Rad, 4 g), which had been converted to the protonated form by washing with H2SO4 (1 N, 60 mL). Following filtration and evaporation of MeOH, the concentrated solution was lyophilized for 4 days to yield 1.16 g (6.8 mmol, 77%) of 2-nitroimidazole acetic acid. 1 H NMR (275 MHz, DMSO): δ 7.5 (s, 1H), 7.1 (s, 1H), 4.8 (s, 2H). Synthesis of N,N-Dimethylglycyl-L-seryl-L-cysteinyl-lysyl{N-[1-(2-nitro-1H-imidazolyl)acetamido]}glycine (RP435). The peptide backbone was prepared on an automated peptide synthesizer (Applied Biosystems Inc., model 433A, Foster City, CA) using Sasrin resin and 9-fluorenylmethoxycarbonyl-protected amino acids. When the backbone was complete, the 1-(4,4dimethyl-2,6-dioxocyclohexylidine)-ethyl protecting group on the -amine of lysine was cleaved with 2% hydrazine and the free amine was coupled with 2-nitroimidazole acetic acid. The peptide was cleaved from the resin by stirring with 95% aqueous TFA at 0 °C for 30 min and then at room temperature for 60 min. The resin was filtered, and the filtrate was evaporated under vacuum. The residue was washed with a minimum amount of tertbutyl methyl ether, which caused the precipitation of the peptide. The ether was removed, and the residue was dissolved in a minimum amount of water and then lyophilized to yield 190 mg of crude product, of which 110 mg was further purified by HPLC, yielding 35 mg (31%) pure RP435. 1H NMR (500 MHz, d6-DMSO): δ 1.33 [m, 2H, H(18)], 1.43 [m, 2H, H(19)], 1.59 [m, 1H, H(17a)], 1.72 [m, 1H, H(17b)], 1.86 [s, 3H, H(30)], 2.58 [broad, 6H, H(1, 2)], 2.68 [dd, 1H, H(13a), JH13a-H11 ) 9.89 Hz, JH13a-H13b ) 14.04 Hz], 2.95 [dd, 1H, H(13b), JH13b-H11 ) 3.91 Hz, JH13b-H13a ) 13.92 Hz], 3.05 [m, 2H, H(20)], 3.61 [m, 2H, H(7)], 4.13 [m, 1H, H(15)], 4.20 [dd, 1H, H(27a), JH27a-H28 ) 5.86 Hz, JH27a-H27b ) 13.67 Hz], 4.32 [dd, 1H, H(27b), JH27b-H28 ) 5.13 Hz, JH27b-H27a ) 13.67 Hz], 4.44 [m, 1H, H(6)], 4.56 [m, 1H, H(11)], 5.07 [s, 2H, H(23)], 5.09 [broad, 1H, H(8)], 7.18 [d, 1H, H(25), JH25-H24 ) 0.97 Hz], 7.61 [d, 1H, H(24), JH24-H25 ) 0.98 Hz], 8.0 [d, 1H, H(14), JH14-15 ) 7.57 Hz], 8.31 [m, 2H, H(10, 21)], 8.56 [t, 1H, H(28), JH28-H27 ) 6.47 Hz]. Electrospray mass spectrum: m/z 646.15 ([M + H]+, C24H40N9O10S, calculated ) 645.69). Synthesis of 2-(2-nitro-1H-imidazolyl)ethylamine hydrochloride was accomplished by minor modification of a published method (22). A total of 0.56 g (5.0 mmol) of 2-nitroimidazole, 1.33 g (5.2 mmol) of N-(2-bromoethyl)phthalimide, and 0.72 g (5.2 mmol) of potassium carbonate was mixed in 15 mL of DMF and heated at 110 °C for 2 h. After removing the solvents, water was added to the residue to dissolve the salts. The precipitate was collected, washed by water, and dried under vacuum to yield 0.8 g (2.8 mmol, yield 56%) of raw product. The raw product was refluxed with 0.28 g (5.6 mmol) hydrazine monohydrate in 20 mL of EtOH for 2 h and then cooled to 4 °C, filtered, and evaporated to dryness. The residue was dissolved in 20 mL of 1 N HCl, filtered, and brought to dryness again. The residue was recrystallized from MeOH/EtOAc to yield 240 mg (1.25 mmol, 44%) of product. Synthesis of N,N-Dimethylglycyl-L-tert-butylglycyl-L-cysteinylglycin[2-(2-nitro-1H-imidazolyl)ethyl]amide (RP535). A total of 80 mg (0.18 mmol) of RP455, 55.4 mg (0.29 mmol) of DECI, and 48 mg (0.29 mmol) of TFP was dissolved in 1 mL of DMF. The mixture was stirred in an ice bath for 10 min, followed

Bioconjugate Chem., Vol. 10, No. 5, 1999 899

by 30 min at room temperature, and then in a 45 °C oil bath for 60 min. A total of 37.2 mg (0.19 mmol) of 2-(2nitro-1H-imidazolyl)ethylamine hydrochloride was added to the mixture which was set in an ice bath. To the mixture was added dropwise 74.7 mg (0.58 mmol) of diisopropyl ethylamine (DIEA) in 1 mL of DMF. The mixture was stirred at 45 °C for 1 h. The solvent was removed after the reaction was completed. About 2 mL of water was added and then removed. The crude product was purified by HPLC, yielding 50 mg (0.085 mmol, 47%) of pure RP535. 1H NMR (500 MHz, d6-DMSO): δ 0.93 [s, 9H, H(8, 9, 10)], 1.86 [s, 3H, H(28)], 2.71 [dd, 1H, H(14a), JH14a-H13 ) 9.3 Hz, JH14a-H14b ) 13.9 Hz], 2.79 [t, 6H, H(1, 2), J ) 4.0 Hz], 2.90 [dd, 1H, H(14b), JH14b-H13 ) 4.6 Hz, JH14b-H14a ) 13.9 Hz], 4.25 [d, 1H, H(25a), JH25a-H26 ) 6.59 Hz], 4.27 [d, 1H, H(25b), JH25b-H26 ) 6.59 Hz], 4.37 [d, 1H, H(6), JH6-H5 ) 9.27 Hz], 4.47 [m, 1H, H(13)], 7.16 [d, 1H, H(23), JH23-H22 ) 1.2], 7.51 [d, 1H, H(22), JH22-H23 ) 0.98 Hz], 7.89 [t, 1H, H(19), JH19-H20 ) 5.98 Hz], 8.14 [t, 1H, H(16), JH16-H17 ) 5.6 Hz], 8.39 [d, 1H, H(12), JH12-H13 ) 7.57 Hz], 8.45 [t, 1H, H(26), JH26-H25 ) 6.59 Hz], 8.61 [d, 1H, H(5), JH5-H6 ) 9.27 Hz]. Electrospray mass spectrum: m/z 586.11 ([M+H]+, C23H40N9O7S, calculated ) 585.69). Synthesis of 99mTcO-RP294, 99mTcO-RP435, 99mTcORP455, and 99mTcO-RP535. To a 3 mL tube were added RP294, RP435, RP455, or RP535 (100-200 µg), Na99mTcO4 solution (2-10 mCi in 200 µL of saline), and 100 µL of stannous gluconate solution (which contained 10-40 µg of SnCl2 and 0.25-1.0 mg of sodium gluconate). The mixture was incubated at 20-100 °C for 15-30 min. The 99mTc complexes were analyzed by HPLC (Zorbax SB column, H2O/ACN containing 0.1% TFA; gradient 100 to 90% H2O over 45 min for 99mTcO-RP294, 100 to 70% H2O for 99mTcO-RP435, and 100 to 50% H2O for 99mTcO-RP455 and 99mTcO-RP535). UV detection at 320 nm was used to monitor the disappearance of the 2-NI absorption peak of the ligand. The radiochemical yields were 99mTcORP294, >97%; 99mTcO-RP435, >68%; 99mTcO-RP455, >99%; and 99mTcO-RP535, >78%. Formation of 99mTccolloid was 98%. RP535 was prepared by coupling 2-(2-nitro-1H-imidazolyl)ethylamine to RP455 (Figure 1). The purity of RP535 was analyzed by two HPLC gradients (100 to 70% H2O and 100 to 50% H2O over 45 min) and shown to be >97% based on UV detection at 214 and 320 nm. The molecular weight of RP435 determined by mass spectrometry was consistent with the proposed structure. The 1H NMR (500 MHz) spectrum of RP435 revealed the two protons of the 2-nitroimidazole at 7.19 and 7.61 ppm with a weak mutual coupling (J ) 0.97 Hz). The proton of the hydroxyl H(8) appeared at 5.09 ppm as a broad peak. Four of the five amide protons of RP435, H(10), H(14), H(21), and H(28), were identified. H(10) and H(21) overlapped each other at 8.31 ppm. The six protons of H(1) and H(2) formed a broad peak at 2.57 ppm, while the two protons of H(3) were merged in the solvent peak ranging from 3.26 to 3.36 ppm. MS analysis of RP535 showed the molecular ion signal at the expected mass. The 1H NMR spectrum of RP535 exhibited signals of all five amide protons (Figure 1), with H(5) at 8.61 ppm, H(26) at 8.45 ppm, H(12) at 8.39 ppm, H(16) at 8.13 ppm, and H(19) at 7.89 ppm, respectively. Similar to RP435, the two protons of the 2-nitroimidazole of RP535 showed a weak mutual coupling (JH22-H23 ) 1.1 Hz). The nine protons of the tert-butyl group, H(8, 9, 10), appeared as a singlet peak at 0.91 ppm. It is interesting

Su et al.

Figure 2. HPLC analysis of 99mTc labeling of RP435 demonstrating formation of two main products, peak A and peak B, in a total yield of ∼70%. Zorbax C18 column eluted with H2O/ ACN containing 0.1% TFA; gradient 100% to 70% H2O over 45 min; flow rate 1 mL/min.

to note that the six protons, H(1, 2), of the two methyl groups of dimethylglycine produced triplet peaks, instead of a singlet peak, at 2.80 ppm. The signals of H(3), H(17), and H(20) were partly merged in the solvent peak ranging from 3.50 to 3.66 ppm, while H(13) and H(21) overlapped each other at 4.45 ppm (m, 3H). These data indicated that RP535 contained the N3S chelator, a 2-nitroimidazole group, and the acetamidomethyl protecting group for the sulfur atom of cysteine. 99m Tc Labeling of RP294, RP455, and RP435. 99mTc labeling of RP294 and RP435, which have the same N3S chelator, can be carried out at room temperature via transchelation from 99mTc-gluconate. However, 99mTc labeling of RP455, which has a tert-butyl group on the chelator, requires heating in a boiling water bath for 1015 min. 99mTcO-RP294 showed two peaks on HPLC in 0.1% aqueous TFA/ACN in a total yield >97%. In contrast, 99mTcO-RP455 showed one single peak (yield >99%). When analyzed in acetate/MeOH, 99mTc-RP294 showed two major peaks, at 33.69 min (39.6%) and 36.81 min (59.3%), with a gradient of 100% to 10% acetate over 45 min, while 99mTc-RP455 showed a major peak at 35.55 min (84.8%) with the same gradient (data not shown). Stannous gluconate was used as reducing agent for Na99mTcO4 in the presence of 100-200 µg of RP435. Two main peaks (peak A: 32.42 min, yield 26.4%; peak B: 35.59 min, yield 42.5%) were detected on HPLC (Figure 2) with aqueous TFA/ACN (gradient: 100% to 70% H2O over 45 min). The ratio of peaks A and B in the reaction mixture varied with time, while their retention times remained unchanged. The difference in retention times of peaks A and B of 99mTcO-RP435 was 3.1 min, while that of the two peaks from 99mTcO-RP294 was 0.7 min. Stability of 99mTcO-RP435. Peaks A and B of 99mTcORP435, after being separated, were reanalyzed by HPLC after different time intervals. A slow interconversion of the two isomers in aqueous solution containing 0.1% TFA was observed. For example, 4.2% of B was detected 103 min after the isolation of A from the labeling mixture, and 4.0% of A was detected 118 min after isolation of B (data not shown). The interconversion of A and B could be greatly accelerated when they were mixed with an equal volume of 0.25 M pH 7.4 phosphate buffer (PB). Figure 3 shows the interconversion of purified A to B and of purified B to A, and it is evident that an approximate 1:1 equilibrium was reached within 1 h. 99m Tc Labeling of RP535. RP535 (100-200 µg) was labeled with 2-10 mCi 99mTc with stannous gluconate

Tc-99m Peptidic 2-Nitroimidazoles

Figure 3. Time course of interconversion of syn and anti conformations of 99mTcO-RP435 in phosphate buffer (PB). The isomers (peak A and peak B) were isolated by HPLC, mixed with an equal volume of PB, and aliquots reanalyzed after different time intervals. (A) Conversion of peak A to peak B. (B) Conversion of peak B to peak A.

Figure 4. HPLC analysis of co-injection of 99mTcO-RP535 and ReO-RP535. 99mTcO-RP535 was monitored by radiometric detector (lower trace) and ReO-RP535 by UV detector at 320 nm (upper trace). The UV peak at 23.4 min is free RP535 while the peak at 33.0 min is ReO-RP535. 99mTcO-RP535 eluted at 33.4 min. Zorbax C18 column eluted with H2O/ACN containing 0.1% TFA; gradient 100 to 30% H2O over 45 min; flow rate 1 mL/ min.

as reducing agent. Higher temperatures (60-100 °C) were required to label the compound. One major labeled peak (yield >78%) was obtained when the reaction mixture was incubated in a boiling water bath for 30 min (Figure 4, lower trace). Varying the pH from 2 to 7 did not significantly affect the labeling yield. Although the nitroimidazole group of RP535 was stable under the labeling conditions, significant reduction was noted (analyzed by HPLC and monitored by the UV detector at 320 nm) when larger amounts of SnCl2 were added, resulting in a dramatic decrease in the yield of 99mTcO-

Bioconjugate Chem., Vol. 10, No. 5, 1999 901

RP535. For example, the yield of 99mTcO-RP535 dropped to less than 5% of the total detected radioactivity on HPLC when 100 µg of SnCl2 was used, and no intact RP535 remained when 600 µg of SnCl2 was added. Determination of Octanol/Water Partition Coefficients (PC) of 99mTcO-RP435 and 99mTcO-RP535. The PC values obtained were 99mTcO-RP435, peak A, 0.0009 ( 0.0001; 99mTcO-RP435, peak B, 0.0011 ( 0.0001; and 99mTcO-RP535, 2.8 ( 0.1, n ) 4. Thus, the difference in partition coefficients of 99mTcO-RP535 and 99mTcORP435 is over 2000-fold. Preparation of ReO-RP535 and Co-Injection with 99m TcO-RP535. It is difficult to determine directly the structure of a 99mTc complex because of the extremely small amount of the compound present. 99Tc is a longlived radioactive isotope (t1/2 ) 2.1 × 105 years) which restricts its application. However, the composition and structure of 99mTcO-RP535 could be indirectly determined by comparison with ReO-RP535, which can be prepared in weighable quantities and characterized. Re belongs to the same group (VIIb) of the periodic table as Tc and thus possesses similar chemical properties; furthermore, Re and Tc are similar in size because of the lanthanide contraction. ReO-RP535 was prepared by coupling 2-(2-nitro-1Himidazolyl)ethylamine with ReO-RP455. The compound was purified by HPLC. 1H NMR data showed that the two amide protons, H(5) and H(12), disappeared after the chelation of RP535 to ReO3+, while those of H(16) and H(19) of the uncoordinated RP535 (Figure 1) remained. The protons of the acetamidomethyl protecting group also disappeared, indicating the involvement of sulfur in the coordination with ReO. The two methyl groups and the methylene group of dimethylglycine could no longer freely rotate around the C(3)-Namine bond in ReO-RP535, which made the protons of these groups become chemically nonequivalent. For example, the protons of H(1) shifted from 2.80 to 2.44 ppm, and H(2) shifted from 2.80 to 3.55 ppm upon the formation of ReO-RP535. This indicated that the N atom of the dimethylglycine was coordinated with the ReO3+ core. The two protons of H(3) of ReORP535 were in a chemically nonequivalent environment as well. H(3a) appeared at 4.02 ppm and H(3b) at 4.82 ppm with a coupling constant JH3a-H3b at 14.9 Hz. H(14a) coupled H(13) with JH14a-H13 for 6.8 Hz, while H(14b) coupled H(13) with JH14a-H13 for 12.2 Hz. However, no signals of H(14a) coupling with H(14b) were found. Taken together, this evidence supported the assumption that the ReO3+ core was complexed with the N3S chelator. No evidence of the coexistence of the syn and anti isomers of ReO-RP535 was found in the 1H NMR spectrum. The prepared ReO-RP535 could be either a syn or anti isomer. The electrospray mass spectrum exhibited the characteristic two molecular ions at m/z 712.93 ([M + H]+), which corresponded to a 1:1 185ReO to RP535 complex, and m/z 714.94 ([M + H]+), which corresponded to another 1:1 187ReO to RP535 complex. Moreover, the relative abundances of the two peaks (36.5 and 63.5%) are consistent with the natural abundances of 185Re and 187Re (37.4 and 62.6%). An aqueous solution of ReO-RP535 was mixed with the 99mTcO-RP535 labeling mixture and co-injected into the HPLC for analysis. Two gradients of aqueous TFA/ACN were used: 100 to 30% and 100 to 50% water over 45 min. Two detectors, UV and radiometric, were connected in series with the eluent entering the UV detector first and then the radiometric detector. There was a time delay of 0.3-0.4 min for signal response between the UV and radiometric detectors. Figure 4 shows the chromato-

902 Bioconjugate Chem., Vol. 10, No. 5, 1999

Su et al.

Figure 6. Time course of accumulation of 99mTcO-RP435 and 99mTcO-RP535 in Chinese hamster ovary cells under aerobic and hypoxic conditions. Ordinate is ratio of radioactivity concentration inside cells (Cin) to that in supernatant (Cout). Each point is mean of eight or nine experiments and error bars are SEM.

Figure 5. Stability of 99mTcO-RP535 diluted in an equal volume of phosphate buffer (PB). (A) HPLC analysis of sample obtained 235 min after mixing demonstrates appearance of decomplexation products with retention times of 22-24 min but little release of free pertechnetate (retention time 3 min). Zorbax C18 column eluted with H2O/ACN containing 0.1% TFA; gradient 100 to 50% H2O over 45 min; flow rate 1 mL/min. (B) Time course of stability of 99mTcO-RP535 diluted in PB.

gram of the co-injection of ReO-RP535 with 99mTcORP535 with the 100 to 30% water gradient. The retention time of ReO-RP535 was 33.04 min (upper trace), while that of 99mTcO-RP535 was 33.35 min (lower trace). The UV peak at 23.37 min was free RP535 (upper trace). The retention times of ReO-RP535 and 99mTcO-RP535 became 39.91 and 40.51 min, respectively, when a 100 to 50% water gradient was used. The retention times of ReORP535 and 99mTcO-RP535 are the same when the time delay in signal response is taken into account. This implies that 99mTcO-RP535 has the same composition and structure as ReO-RP535, i.e., 99mTc is coordinated by the N3S chelator of RP535, there is a 99mTcdO core, and the ratio of 99mTc to RP535 is 1:1. Stability of 99mTcO-RP535. 99mTcO-RP535 was isolated by HPLC, and the collected 99mTcO-RP535 in aqueous TFA/ACN was reinjected into the HPLC for analysis after different time intervals at room temperature. It was evident that 99mTcO-RP535 was more stable than 99mTcO-RP435 in aqueous solution, which contained 0.1% TFA, because neither decomplexation nor change in the radioactive peak was seen over 26 h. Theoretically, the formation of syn and anti isomers of 99mTcO-RP535 is possible, when the tert-butyl side chain on the N3S chelator is in the syn and anti conformations with respect to the 99mTcdO core. When the isolated 99mTcO-RP535 was mixed with an equal volume of 0.25 M, pH 7.4, phosphate buffer (PB) and reinjected after different time intervals, a decomplexation of the complex, instead of interconversion of syn and anti isomers, was observed (Figure 5A). 99mTcO-RP535 (retention time 40.27 min) decomplexed in PB and formed three products, which appeared in the region of 22.1-23.7 min on the HPLC. These decomplexed compounds cannot be the isomers of 99mTcO-RP535 because they did not exist

in the labeling mixture (Figure 4, lower trace; note different gradient) and they did not reach any fixed equilibrium with 99mTcO-RP535 over time. The decomplexation might occur when one or more coordination bonds of the 99mTcO-RP535 complex are broken. No free pertechnetate (retention time 3 min) was detected from the decomplexation of 99mTcO-RP535. The half-time for decomplexation was about 200 min (Figure 5B). Interestingly, 99mTcO-RP535 decomposed immediately with release of pertechnetate when mixed with 0.01 N NaOH. To summarize, 99mTcO-RP535 with its tert-butyl group on the N3S chelator, is more difficult to label than 99m TcO-RP435 and does not interconvert to another isomer. This suggests that the tert-butyl group acts as an important steric hindrant that inhibits the coexistence and interconversion of syn and anti isomers of 99mTcORP535 and indeed forces the complex into a single conformation. However, the configuration of 99mTcORP535 remains to be determined. In Vitro Tests of 99mTcO-RP294, 99mTcO-RP455, 99m TcO-RP435, and 99mTcO-RP535. The ability of these 99mTc-labeled compounds to be taken up by mammalian cells under aerobic or hypoxic conditions was tested using CHO cells in suspension culture in equilibrium with air or nitrogen gas mixtures. Samples were removed as a function of time and ratios of cell-associated radioactivity (Cin) to that in the medium (Cout) were calculated. The control compounds, 99mTcO-RP294 and 99mTcO-RP455, which do not contain 2-NI groups, did not show selective accumulation in hypoxic CHO cells. In contrast, 99mTcORP435 showed differential accumulation between hypoxic and aerobic cells. The Cin/Cout for aerobic cells at 4 h was 0.42, while that of hypoxic cells was 0.90 (Figure 6A), for a hypoxic/aerobic differential of 2.1 ( 0.4 (mean ( SEM, n ) 8). In comparison, 99mTcO-RP535 revealed a greater differential accumulation between hypoxic and aerobic cells than 99mTcO-RP435. When 99mTcO-RP535 was added to a suspension of CHO cells in vitro, there was a modest accumulation of radioactivity in the cells under aerobic conditions, reaching a Cin/Cout value of 0.68 at 4 h (Figure 6B). In contrast, under hypoxic conditions, there was a further increase to a Cin/Cout value of 2.10, for a 3.1 ( 0.4 (n ) 9) fold hypoxic/aerobic differential. Previous work (24) with the 99mTc-labeled 2-NI BMS181321, using the same cell line and in vitro system, showed

Tc-99m Peptidic 2-Nitroimidazoles

that a large molar excess of the 2-NI compound misonidazole prevented selective accumulation in hypoxic cells while the 5-nitroimidazole metronidazole stimulated it. In the present work, addition of 5 mM misonidazole abolished the hypoxia-specific accumulation of 99mTcORP535, whereas 8 mM metronidazole selectively increased hypoxic accumulation to 130% of the control value. DISCUSSION

On the basis of promising results obtained with [18F]FMISO and [123I]IAZA (7-10), there has been great interest in the development of 99mTc-labeled markers of hypoxia which would be convenient to prepare and widely applicable. However, the standard 99mTc-chelation systems have not all proved to be useful. The first attempt using the BATO approach resulted in a complex which was not efficiently reduced and trapped, and which showed inadequate permeability to lipophilic membranes (25). A bis(amine-phenol) complex similarly failed to cross cell membranes (26). Greatest success thus far has been obtained with amineoxime-type chelators, specifically BMS181321 (11), BRU59-21 (27), and BnAO (17); however, each has its limitations. BMS181321 has a PC of 40, which results in slow clearance from the blood and background tissues and extensive elimination via the gastrointestinal tract (14). BRU59-21 has a lower PC of 12 with resultant improved clearance from the blood, but the extent of hepatobiliary excretion remains high (15, 16). In contrast, BnAO undergoes extensive renal elimination, but lacking a nitro group, its mechanism of hypoxia-specific localization is unclear (20). BnAO has a low PC of 0.1, although this does not appear to limit its penetration of tissue (20). Work in the radioiodinated sugar 2-NI series has suggested an optimum PC of ∼1 (5). The RP294 chelation system for 99mTc was developed by Resolution Pharmaceuticals Inc. (21) and is used in RP128, a peptide for imaging inflammation (28). Acetamidomethyl was selected as the thiol-protecting group because it can be cleaved at room temperature. The free carboxylic acid group of the molecule contributes to the hydrophilicity of the complex and RP435, the 2-NIcontaining derivative of RP294, has an extremely low PC of 0.001. Therefore, an analogue containing tert-butylglycine in place of serine was developed. In the present work, the utility of this chelator on a 2-NI compound targeted to hypoxic tumor cells was evaluated. The partition coefficient of 99mTcO-RP535 was 2.8, which is similar to that reported for IAZA (9). Although the presence of the tert-butyl group was desirable to increase the PC of the complex, it made the chelator more difficult to label. While 99mTc labeling of RP294 and RP435 by transchelation from gluconate proceeded efficiently at room temperature, the tert-butyl analogues RP455 and RP535 required heating. Radiochemical yields obtained were 26 and 43% for the two peaks of 99mTcO-RP435 (Figure 2) and 78% for 99mTcORP535 (Figure 4, lower trace). It was shown that the nitro group was not reduced by the typical quantities of stannous chloride used in 99mTc labeling, but that it could be reduced by excessive amounts. The structure of RP294 was determined by 13C and 1H NMR, except the assignments of the amide protons (21). In addition, ReO-RP294 and 99TcO-RP294 have been structurally confirmed by 1H NMR and X-ray crystallography, which indicated that the metal ions were coordinated by the N3S chelator. RP435, containing the

Bioconjugate Chem., Vol. 10, No. 5, 1999 903

same N3S chelator as RP294, can be presumed to chelate the 99mTcO3+ core in the same manner. As has been shown previously for 99mTcO-RP294 (21), two complexes of 99mTcO-RP435 were formed, presumably syn and anti isomers (Figure 2). Following separation by HPLC, these were quite stable in acidic aqueous solution (4% interconversion in 2 h) but equilibrated in 0.25 M, pH 7.4, phosphate buffer with a half-time of ∼1 h (Figure 3), suggesting that the rate of interconversion is a function of hydroxide concentration. In contrast, 99mTcO-RP535 existed as a single isomer (Figure 4, lower trace), confirmed by co-injection with its Re analogue (Figure 4, upper trace), which was extremely stable in acidic aqueous medium. When diluted in PB, 99mTcO-RP535 decomplexed with a half-time of ∼3 h, resulting in the formation of several hydrophilic species but no release of free pertechnetate (Figure 5). However, when 99mTcORP535 was diluted in 0.01 N NaOH it decomposed rapidly to pertechnetate. These results with 99mTcORP535 also implicate the concentration of hydroxide in the decomposition process. Following HPLC purification, 99mTcO-RP435 and 99m TcO-RP535 were evaluated in an in vitro model of cellular hypoxia that has been used extensively in radiation biology and investigation of bioreductive drugs, including other 99mTc-, 123I-, and 18F-labeled hypoxia tracers (14, 16, 20, 23). Both tracers showed a modest level of uptake in aerobic CHO cells, which increased slightly over the course of 4 h. In contrast, hypoxic cells accumulated 2-3 times as much radioactivity over that time period, a significant degree of hypoxia-specific accumulation (Figure 6). Moreover, the selective accumulation of 99mTcO-RP535 could be modulated by coincubation with millimolar concentrations of unlabeled nitro compounds (data not shown). The 2-NI misonidazole abolished the hypoxia-specific accumulation of 99mTcORP535 while metronidazole, a 5-NI of lower electron affinity, enhanced this accumulation. The same effects have been reported previously with the 2-NI 99mTcBMS181321 and suggest a bioreductive mechanism of localization of the tracer (24). CONCLUSIONS

RP435 and RP535 have a similar N3S chelator in structure, but their ability to form chelates with 99mTc is different. The 99mTcO-RP435 chelate could be prepared at room temperature, and showed interconversion of syn and anti isomers in aqueous solution. The interconversion of the two isomers was much faster in neutral than in acidic medium. In contrast, the tert-butyl group on the backbone of the N3S chelator of RP535 hindered the chelation with 99mTc, presumably by steric effects, and blocked the formation and interconversion of the theoretically existing syn and anti isomers of 99mTcO-RP535 in acidic and neutral aqueous solution at room temperature. However, the tert-butyl group of RP535 did have the desired effect of significantly increasing the lipophilicity of the corresponding 99mTc chelate. Both 99mTcORP435 and 99mTcO-RP535 showed selective uptake in hypoxic cells, suggesting that 99mTcO-peptidic complexes containing the 2-nitroimidazole group can be a new class of hypoxia imaging agents. ACKNOWLEDGMENT

We would like to thank Dr. Ernest Wong, Tam Nguyen, Linda Lu, and Dr. Susan Peers of Resolution Pharmaceuticals Inc. for their support and suggestions. This work was funded by the University-Industry program of the

904 Bioconjugate Chem., Vol. 10, No. 5, 1999

Medical Research Council of Canada in conjunction with Resolution Pharmaceuticals Inc. (Grant UI-13558). LITERATURE CITED (1) Nunn, A., Linder, K., and Strauss, H. W. (1995) Nitroimidazoles and hypoxia imaging. Eur. J. Nucl. Med. 22, 265-280. (2) Hockel, M., Knoop, C., Schlenger, K., Vorndran, B., Baussmann, E., Mitze, M., Knapstein, P. G., and Vaupel, P. (1993) Intratumoral pO2 predicts survival in advanced cancer of the uterine cervix. Radiother. Oncol. 26, 45-50. (3) Graeber, T. G., Osmanian, C., Jacks, T., Housman, D. E., Koch, C. J., Lowe, S. W., and Giaccia, A. J. (1996) Hypoxia mediated selection of cells with diminished apoptotic potential in solid tumors. Nature 379, 88-91. (4) Brizel, D. M., Scully, S. P., Harrelson, J. M., Layfield, J. M., Bean, J. M., Prosnitz, L. R., and Dewhirst, M. W. (1996) Tumor oxygenation predicts for likelihood of distant metastases in human soft tissue sacroma. Cancer Res. 56, 941943. (5) Chapman, J. D., Engelhardt, E. L., Stobbe, C. C., Schneider, R. F., and Hanks, G. E. (1998) Measuring hypoxia and predicting tumor radioresistance with nuclear medicine assays. Radiother. Oncol. 46, 229-237. (6) Rauth, A. M. (1984) Pharmacology and toxicology of sensitizers. Int. J. Radiat. Oncol. Biol. Phys. 10, 1293-1300. (7) Grierson, J. R., Link, J. M., Mathis, C. A., Rasey, J. S., and Krohn, K. A. (1989) A radiosynthesis of fluorine-18 fluoromisonidazole. J. Nucl. Med. 30, 343-350. (8) Koh, W. J., Rasey, J. S., Evans, M. L., Grierson, J. R., Lewellen, T. K., Graham, M. M., Krohn, K. A., and Griffin, T. W. (1992) Imaging of hypoxia in human tumors with [18F]fluoromisonidazole. Int. J. Radiat. Oncol. Biol. Phys. 22, 199212. (9) Mannan, R. H., Somayaji, V. V., Lee, J., Mercer, J. R., Chapman, J. D., and Wiebe, L. I. (1991) Radioiodinated 1-(5iodo-5-deoxy-beta-D-arabinofuranosyl)-2-nitroimidazole (Iodoazomycin arabinoside: IAZA): a novel marker of tissue hypoxia. J. Nucl. Med. 32, 1764-1770. (10) Parliament, M. B., Chapman, J. D., Urtasun, R. C., McEwan, A. J., Golberg, L., Mercer, J. R., Mannan, R. H., and Wiebe, L. I. (1992) Noninvasive assessment of human tumor hypoxia with 123I-iodoazomycin arabinoside: preliminary report of a clinical study. Br. J. Cancer 65, 90-95. (11) Linder, K. E., Chan, Y. W., Cyr, J. E., Malley, M. F., Nowotnik, D. P., and Nunn, A. D. (1994) TcO(PnAO-1-(2nitroimidazole)) [BMS-181321], a new technetium-containing nitroimidazole complex for imaging hypoxia: synthesis, characterization, and xanthine oxidase-catalyzed reduction. J. Med. Chem. 37, 9-17. (12) Di Rocco, R. J., Bauer, A. A., Pirro, J. P., Kuczynski, B. L., Belnavis, L., Chan, Y. W., Linder, K. E., Narra, R. K., Nowotnik, D. P., and Nunn, A. D. (1997) Delineation of the border zone of ischemic rabbit myocardium by a technetiumlabeled nitroimidazole. Nucl. Med. Biol. 24, 201-207. (13) Barron, B., Grotta, J., Lamki, L., Villar, C., Ephron, V., Patel, D., Linder, K. E., and Nunn, A. D. (1996) Preliminary experience with technetium-99m BMS-181321, a nitroimidazole, in the detection of cerebral ischemia associated with acute stroke. J. Nucl. Med. 37, 272P-273P. (14) Ballinger, J. R., Kee, J. W., and Rauth, A. M. (1996) In vitro and in vivo evaluation of a technetium-99m-labeled 2-nitroimidazole (BMS181321) as a marker of tumor hypoxia. J. Nucl. Med. 37, 1023-1031. (15) Johnson, L. L., Schofield, L., Mastrofrancesco, P., Donahay, T., and Nott, L. (1998) Technetium-99m-nitroimadazole uptake in a swine model of demand ischemia. J. Nucl. Med. 39, 1468-1475.

Su et al. (16) Melo, T., Duncan, J., Ballinger, J. R., and Rauth, A. M. (1998) BRU59-21, a second-generation technetium-99mlabeled 2-nitroimidazole for imaging hypoxia in tumors. J. Nucl. Med. 39, 219P. (17) Archer, C. M., Edwards, B., Kelly, J. D., King, A. C., Burke, J. F., and Riley, A. L. M. (1995) Technetium labeled agents for imaging tissue hypoxia in vivo. In Technetium and rhenium in chemistry and nuclear medicine (M. Nicolini, G. Bandoli, and U. Mazzi, Eds.) pp 535-539, S G Editorali, Padova. (18) Okada, R. D., Johnson, G., Nguyen, K. N., Edwards, B., Archer, C. M., and Kelly, J. D. (1997) 99mTc-HL91: Effects of low flow and hypoxia on a new ischemia-avid myocardial imaging agent. Circulation 95, 1892-1899. (19) Cook, G. J., Houston, S., Barrington, S. F., and Fogelman, I. (1998) Technetium-99m-labeled HL91 to identify tumor hypoxia: correlation with fluorine-18-FDG. J. Nucl. Med. 39, 99-103. (20) Zhang, X., Melo, T., Ballinger, J. R., and Rauth, A. M. (1998) Studies of 99mTc-BnAO (HL-91): a nonnitroaromatic compound for hypoxic cell detection. Int. J. Radiat. Oncol. Biol. Phys. 42, 737-740. (21) Wong, E., Fauconnier, T., Bennett, S., Valiant, J., Nguyen, T., Lau, F., Lu, L. F. L., Pollak, A., Bell, R. A., and Thornback, J. R. (1997) Rhenium(V) and technetium(V) oxo complexes of an N2N′S peptidic chelator: evidence of interconversion between syn and anti conformations. Inorg. Chem. 36, 57995808. (22) Hay, M. P., Wilson, W. R., Moselen, J. W., Palmer, B. D., and Denny, W. A. (1994) Hypoxia-selective antitumor agents. 8. Bis(nitroimidazolyl)alkanecarboxamides: A new class of hypoxia-selective cytotoxins and hypoxic cell radiosensitizers. J. Med. Chem. 37, 381-391. (23) Ballinger, J. R., Cowan, D. S., Boxen, I., Zhang, Z. M., and Rauth, A. M. (1993) Effect of hypoxia on the accumulation of technetium-99m-glucarate and technetium-99m-gluconate by Chinese hamster ovary cells in vitro. J. Nucl. Med. 34, 242245. (24) Melo, T., Hua, H. A., Ballinger, J. R., and Rauth, A. M. (1997) Modifying the in vitro accumulation of BMS181321, a technetium-99m-nitroimidazole, with unlabeled nitroaromatics. Biochem. Pharmacol. 54, 685-693. (25) Linder, K. E., Chan, Y. W., Cyr, J. E., Nowotnik, D. P., Eckelman, W. C., and Nunn, A. D. (1993) Synthesis, characterization, and in vitro evaluation of nitroimidazole-BATO complexes: new technetium compounds designed for imaging hypoxic tissue. Bioconjugate Chem. 4, 326-333. (26) Ramalingam, K., Raju, N., Nanjappan, P., Linder, K. E., Pirro, J., Zeng, W., Rumsey, W., Nowotnik, D. P., and Nunn, A. D. (1994) The synthesis and in vitro evaluation of a 99mtechnetium-nitroimidazole complex based on a bis(aminephenol) ligand: comparison to BMS-181321. J. Med. Chem. 37, 4155-4163. (27) Wedeking, P., Yost, F., Wen, M., Patel, B., Eaton, S., Romero, V., Linder, K. E., Rumsey, W., and Nunn, A. D. (1995) Comparison of the biologic activity of the isomers of Tc-99m-nitroimidazole complex BMS-194796. J. Nucl. Med. 36, 17P. (28) Caveliers, V., Goodbody, A. E., Tran, L. L., Bossuyt, A., and Thornback, J. R. (1996) Human dosimetry of Tc99mRP128, a potential inflammation imaging agent. Eur. J. Nucl. Med. 23, 1131.

BC9900542