Targeting Hypoxia in Tumors Using 2-Nitroimidazoles with Peptidic

Inc., 6850 Goreway Drive, Mississauga, Ontario L4V 1V7, Canada. Received ... Four of the 11 compounds showed selective accumulation in hypoxic cells...
0 downloads 0 Views 98KB Size
Download PDF
Bioconjugate Chem. 2000, 11, 401−407

401

Targeting Hypoxia in Tumors Using 2-Nitroimidazoles with Peptidic Chelators for Technetium-99m: Effect of Lipophilicity Xiuguo Zhang,†,‡ Zi-Fen Su,†,‡ James R. Ballinger,†,§,| A. M. Rauth,*,†,‡ Alfred Pollak,⊥ and John R. Thornback⊥ Division of Experimental Therapeutics, Ontario Cancer Institute, 610 University Avenue, Toronto, Ontario M5G 2M9, Canada, Departments of Medical Biophysics and Pharmaceutical Sciences, University of Toronto, 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 November 12, 1999; Revised Manuscript Received February 14, 2000

Tumor hypoxia is an important prognostic factor for response to therapy. Radiolabeled 2-nitroimidazoles have been used for imaging hypoxia, and the octanol/water partition coefficient (P) of these compounds appears to play a crucial role in their suitability for imaging. A series of 11 2-nitroimidazoles coupled to peptidic chelators for 99mTc with divergent P was developed and evaluated in an in vitro system. Two classes of N3S chelators were used: dialkyl-Gly-Ser-Cys-linker-2-nitroimidazole (Class I) and dialkyl-Gly-Lys(2-nitroimidazole)-Cys (Class II). The chelators were prepared by automated solidphase peptide synthesis. Xanthine oxidase was able to reduce the 2-nitroimidiazole moiety on the ligands, but the rate of reduction varied 5-fold among the different chelators. The chelators were labeled by transchelation from [99mTc]gluconate at temperatures between 22 and 100 °C. The reaction mixtures were analyzed by HPLC and their P values determined. The accumulation of each complex in suspension cultures of Chinese hamster ovary cells incubated under aerobic or extremely hypoxic conditions was determined. Radiochemical yields ranged from 5 to 80% for the 11 compounds. HPLC showed that some of the compounds formed two complexes with 99mTc, possibly syn and anti conformations with respect to the TcdO bond. In general, the Class I chelators labeled more readily than the class II chelators. The P values of the 99mTc complexes varied from 0.0002 to 5 and were generally in accordance with predictions based on structure. There were also differences in P as a function of pH; the free acids had a lower P at pH 7.4 than at pH 2.0 due to ionization, whereas the amides did not show this effect. Accumulation levels in aerobic cells were related to P but varied over a narrow range. Four of the 11 compounds showed selective accumulation in hypoxic cells. The peptidic class of 2-nitroimidazoles, with flexible design and convenient solid-phase synthesis, deserves further study as agents for imaging hypoxia in tumors.

INTRODUCTION

For a number of years it has been recognized that hypoxia in tumors is a factor in poor response to radiotherapy (1, 2), because the presence of oxygen is required for fixation of radiation-induced damage in DNA. More recently, it has become evident that hypoxia may also be a predictive factor for poor response to chemotherapy, metastatic spread of a tumor, and generally aggressive disease (3, 4). However, assessment of hypoxia in tumors is difficult. Tumors which seem to be identical in their type, size, stage, and radiographic appearance can vary widely in their extent of hypoxia. At present, the “gold standard” for measurement of hypoxia in tumors involves a polarographic electrode which is inserted directly into a tumor (2, 3). This procedure is invasive, expensive, operator-dependent, not readily repeated, and limited to accessible tumors. These * To whom correspondence should be addressed. Phone: (416) 946-2977. Fax: (416) 946-2984. E-Mail: [email protected]. † Division of Experimental Therapeutics. ‡ Department of Medical Biophysics. § Division of Nuclear Medicine. | Department of Pharmaceutical Sciences ⊥ Resolution Pharmaceuticals Inc.

limitations have led to the exploration of nuclear imaging techniques for noninvasive detection of tumor hypoxia (5, 6). The radiopharmaceutical approach generally involves labeled 2-nitroimidazoles (2-NI), a class of compounds which can undergo an enzymatic one-electron reduction to a radical anion. In the presence of normal physiological levels of oxygen, the radical anion is immediately backoxidized to the starting compound in a futile cycle. However, at low levels of oxygen, the radical anion undergoes further reduction to products which bind to macromolecules, and thus the associated radiolabel is selectively retained in hypoxic cells. Halogenated 2-NI, such as [18F]fluoromisonidazole (FMISO) and [123I]iodoazomycin arabinoside (IAZA), have been used clinically to detect hypoxia in tumors and other conditions (7-12). However, their use is limited by the choice of radiolabel. This has led to great interest in the development of 99m Tc-labeled analogues which would be less expensive and more convenient to use. Two such 99mTc-labeled compounds have been investigated by Bracco Research USA: BMS181321 and BRU5921, both of which are 2-NI coupled to PnAO-type chelators (Figure 1). The octanol/water partition coefficients (P) of the two agents are 40 and 12, respectively. These

10.1021/bc9901595 CCC: $19.00 © 2000 American Chemical Society Published on Web 04/15/2000

402 Bioconjugate Chem., Vol. 11, No. 3, 2000

Zhang et al.

Figure 1. Structures of three previously studied technetium99m labeled hypoxic cell markers.

complexes show selective accumulation in hypoxic cells in vitro and localization in animal tumor models, but their utility is limited by high blood levels and extensive hepatobiliary excretion related to their high P values (13-18). A drug development program at NycomedAmersham produced [99mTc]HL91 or butylene amine oxime (Figure 1), a nonnitro complex which shows hypoxic-specific localization via an undetermined mechanism (19-21). There is some uncertainty whether the 99mTc in [99mTc]HL91 is in the mono-oxo form, as drawn, or the di-oxo form (22). In the development of radiolabeled 2-nitroimidazoles as diagnostic agents, just as in the therapeutic use of 2-nitroimidazoles, P appears to play a crucial role. With therapeutic agents, it was necessary to move to compounds with lower P in order to minimize penetration of nerve tissue and reduce the incidence of peripheral neuropathy; however, at extremely low P, inadequate levels of drug reached the tumor due to inability to penetrate membranes (23). The PnAO-type chelator in BMS181321 and BRU59-21 offers only limited scope for modification of P, although Wiebe and colleagues made a glucuronide conjugate of a PnAO-2-NI to decrease P (24). The present work explores the use of the hypoxiatargeted 2-NI coupled to peptidic N3S chelators for 99mTc, which have been shown to form strong, stable complexes (25). These chelators can be conveniently prepared by automated solid-phase peptide synthesis, and P can be varied widely by the choice of amino acids and pendant groups. Eleven compounds, comprising two classes of structures (Figure 2, Table 1), have been designed and synthesized. Following labeling with 99mTc, each complex was purified by HPLC, its P was measured, and its behavior in an in vitro model of hypoxic tumor cells was investigated. With this approach, P was varied over a range of 4 orders of magnitude, and several complexes which show hypoxia-specific localization in vitro were identified. Studies with one of these compounds (RP435, compound 4 in Table 1) have been reported previously (26). EXPERIMENTAL PROCEDURES

General Synthesis. The peptide backbones of the two classes (Figure 2), consisting of a total of 11 compounds

Figure 2. Chemical structures of class I and class II 2-nitroimidazole-linked peptidic chelators. R1, R2, X, and Y are defined in Table 1. Nuclear magnetic resonance assignment numbers are included for the class II structure. Table 1. Chemical Structures and Mass Spectroscopic Analyses of Peptidic 2-Nitroimidazoles compound code class I 1 2 3 4 5 class II 6 7 8 9 10 11

r1,r2 dimethyl dibenzyl dibenzyl dimethyl dimethyl dimethyl dimethyl dimethyl -(CH2)5(i.e., piperidino) dibenzyl dibenzyl

X

Y

MW

M + H+

Gly Gly Val none none

Gly acid Gly acid Gly acid none (acid) dibenzylamide

759.4 911.7 653.6 645.3 824.7

760.0 912.4 654.4 646.0 825.4

acid amide dibenzylamide acid

558.0 557.3 737.4 598.2

558.8 558.2 738.0 599.0

acid amide

710.3 709.4

711.0 710.2

listed in Table 1, were synthesized on an automatic peptide synthesizer (model 433A, Applied Biosystems Inc., Foster City, CA) using standard 9-fluorenylmethoxycarbonyl (Fmoc) chemistry starting with the carboxyterminal amino acid preloaded on Sasrin resin (Bachem Biosciences Inc., Philadelphia, PA). The R-amino groups were protected with Fmoc, the -amine on Lys with 1-(4,4-dimethyl-2,6-dioxocyclohexidine)ethyl (Dde), and the thiol of Cys with acetamidomethyl (Acm). When the backbone was complete, the Dde on Lys was cleaved with 2% hydrazine and the free amine was coupled with 2-nitroimidazole acetic acid [synthesized from 2-nitroimidazole and bromoethyl acetate as described previously (26)]. Products were cleaved from the resin with trifluoroacetic acid (TFA), precipitated with tert-butyl methyl ether, dissolved in water, and lyophilized. Dibenzylamides were prepared by reacting the crude product with

2-Nitroimidazole Peptidic Chelators for

99mTc

1.1 equiv of dibenzylamine in dry dichloromethane containing 1-hydroxybenzotriazole (HOBT). Condensation was initiated by the addition of 1-[(3-dimethylamino)propyl]-3-ethylcarbodiimide hydrochloride. All products were purified using an HPLC system which consisted of a pump (System Gold model 125, Beckman Instruments Inc., Fullerton, CA), reversed-phase C-18 column (Zorbax Stable Bond, 9.4 mm × 25 cm, Agilent Technologies Inc., Newport, DE), and UV absorbance detector (model 166, Beckman). The mobile phase was a gradient of water and acetonitrile, each containing 0.1% TFA. The eluate was monitored at 214 nm for total UV absorbance and at 325 nm for 2-nitroimidazole absorbance. The desired peak was collected and lyophilized to produce a white powder. Each compound was characterized by electrospray mass spectroscopy on a Sciex API#3 spectrometer in positive ion mode (Table 1). One compound, 4, from the class I group and one compound, 9, from the class II group were further characterized by 500-MHz NMR spectroscopy on a Varian Unity plus-500 spectrometer with TMS as external standard. The results for 4 were presented previously (26). The results for 9 were as follows (see Figure 2 for assignment). δH (DMSO): 9.55 (s, 1H, OH, Y), 8.73 (d, J ) 8, 1H, N2), 8.53 (t, J ) 6.4, 1H, N4), 8.47 (d, J ) 8, 1H, N3), 8.30 (t, J ) 5.7, 1H, N1), 7.59 (d, J ) 1, 1H, A2), 7.28 (d, J ) 1, 1H, A1), 5.05 (s, 2H, A3), 4.36-4.42 (m, 2H, C1 and B1), 4.22 (d, J ) 5.4, 2H, E), 3.83-3.93 (broad q, 2H, D1), 3.30-3.43 (broad, 4H, H2 on piperidino group), 3.05 (q, J1 ) 7, J2 ) 6.8, 2H, B5), 2.99 (dd, J1)4.6, J2)13.9, 1H, C2 or C2′), 2.93 (broad, 4H, H3 on piperidino group), 2.82 (dd, J1 ) 9.3, J2 ) 13.7, 1H, C2 or C2′), 1.83 (s, 3H, F1), 1.641.74 (m, 2H, H4 on piperidino group), 1.48-1.58 (m, 2H, B2), 1.34-1.46 (m, 2H, B4), and 1.22-1.34 (m, 2H, B3). δC (DMSO): 181.81 (C3), 181.03 (B6), 179.53 (F2), 175.32 (A5), 173.85 (D2), 154.93 (A4), 138.77 (A2), 137.42 (A1), 66.48 (D1), 62.90 (C2 on piperidino group), 62.38 (C1 or B1), 62.32 (C1 or B1), 61.54 (A3), 50.53 (E), 48.71 (B5), 41.80 (B2), 41.61 (C2), 38.68 (B4), 32.56 (B3), 32.42 (C3 on piperidino group), 32.11 (F1), 31.03 (C4 on piperidino group). Reduction of Ligands by Xanthine Oxidase. The ability of the 2-nitroimidazole group of the peptide chelators to be reduced by xanthine oxidase was assessed for 10 of the 11 peptidic ligands. The procedures used were based on previous work concerning reduction of nitroimidazoles by xanthine oxidase (27, 28). Briefly, samples of each ligand (28-36 µM) dissolved in phosphate buffer (0.1 M) containing 160 µM xanthine (Sigma Chemical Co., St. Louis, MO) were placed in a quartz spectrophotometer cell in a 37 °C water-jacketed cell holder in a UV-vis spectrophotometer (Perkin-Elmer, Lambda 3B, Oak Bank, IL). After bubbling for 15 min with argon (Praxair Products Inc., Mississauga, ON, Canada) to remove oxygen, xanthine oxidase (EC 1.1.3.22, Sigma Chemical Co., St. Louis, MO, Type X4500 from buttermilk, specific activity 1-2 units/mg protein) was added to the vial at a final concentration of 0.45 units/ mL. Absorption of the sample was monitored at 325 nm as a function of time. The initial slope for the loss of absorbance was calculated and expressed as percent absorbance lost per minute per unit enzyme. Misonidazole, 1-(2-nitro-1-imidazolyl)-3-methoxy-2-propanol, a gift from the Division of Cancer Treatment, National Institutes of Health, Bethesda, MD, was used as a standard substrate. In addition, BRU59, a gift from Drs. K. Linder and A. Nunn, Bracco Research USA, Princeton, NJ, was also tested for comparison.

Bioconjugate Chem., Vol. 11, No. 3, 2000 403

Labeling and Purification. For labeling, each ligand was dissolved at 1-2 mg/mL in water, except for compounds 5, 8, 10, and 11, which were dissolved in acetonitrile:water 1:1 because of limited solubility in water. Labeling was performed by transchelation from [99mTc]gluconate under the following conditions: 100 µg of ligand, 100 µL of 99mTc-pertechnetate (∼200 MBq), and 100 µL of stannous gluconate solution, which contained 25 µg of stannous chloride and 2 mg of sodium gluconate. The pH of the mixture was adjusted to values between 3.5 and 6.0 by addition of 0.1 N HCl or 0.1 N NaOH, and the reaction was carried out at various temperatures between 22 and 100 °C using a water bath, with reaction times generally being 20 min. The reaction mixture was then analyzed with the HPLC system described above, except that an analytical C-18 column was used (Ultrasphere, 4.6 mm × 25 cm, Beckman) and a radiometric detector (model 171, Beckman) was connected in series after the UV detector. The standard linear gradient used was 90% water to 10% water (balance acetonitrile, both containing 0.1% TFA) over 40 min. The major radioactive peaks were collected for measurement of P and for in vitro studies. The purity and stability of the collected peaks was determined by repeated HPLC analysis. By varying pH, temperature, and time the optimal labeling conditions were determined for each compound. Measurement of Partition Coefficient. The P of each complex was measured at both pH 7.4 and pH 2. Ten microliters (∼1 MBq) of each HPLC-purified compound was added to a tube containing 2 mL of n-octanol and 2 mL of phosphate-buffered saline (pH 7.4) or 2 mL of 0.1% TFA (pH 2). Each phase had been presaturated with the opposite phase. The phases were mixed by vortexing for 2 min, then separated by centrifugation at 12000g for 5 min. Aliquots of both phases were transferred in triplicate to counting tubes and assayed in a γ well counter. Each measurement was performed at least twice. Cellular Accumulation in an in Vitro Model. Each complex was evaluated in an in vitro system which has been used in the study of other tracers of hypoxia (14, 15, 18, 21). Suspensions of Chinese hamster ovary (CHO) cells or human Hela cells were incubated under an atmosphere of air or nitrogen to generate aerobic or hypoxic conditions, respectively. The tracer was added, and aliquots were removed over the course of 4 h and centrifuged, and the radioactivity associated with the cell pellet was measured in a γ well counter. The results were expressed as the ratio of concentration of radioactivity in the cell pellet to that in an equivalent volume of supernatant medium (Cin/Cout) as a function of time, as described previously (14). RESULTS

The 11 compounds were conveniently synthesized by automated solid-phase techniques, purified by HPLC, and their structures were confirmed by mass spectroscopy and limited NMR data on one lead compound, compound 4 and compound 9, from each class (Experimental Procedures and Table 1). In the class I series, the 2-NIderivatized Lys was a pendant group on the chelator, while in class II, the Lys was integral to the chelation site. The nitro groups in all of the ligands were capable of reduction by xanthine oxidase, although the rates of reduction, expressed as initial slopes, varied among the compounds. Compound 11 was reduced at a similar rate to misonidazole, while the remainder were reduced more slowly, at 17-60% of the rate of misonidazole (Table 2).

404 Bioconjugate Chem., Vol. 11, No. 3, 2000

Zhang et al.

Table 2. Reduction of 2-Nitroimidazole Moiety of Peptidic Chelators by Xanthine Oxidase compound code class I 1 2 3 4 5 class II 6 7 8 9 10 11 BRU59 misonidazole a

initial slopea (% abs/min/unit)

% of misonidazole rate

-2.38 -1.46 -1.57 -0.80 -1.18

51 31 34 17 25

-1.14 -0.77 n/ab -1.02 -2.77 -4.71 -0.94 -4.64

25 17 n/a 22 60 102 20 100

Initial concentration: 28-36 µM. bNot available.

Table 3. Peak Number, HPLC Retention Time, Radiochemical Yield, and Partition Coefficient for 11 99mTc-Labeled Peptidic 2-Nitroimidazoles

compound code class I 1 2 3 4 5 class II 6 7 8 9 10 11 a

octanol/water partition coefficient pH 2 pH 7.4

peak no.

retention time (min)

yield (%)

SPa A B A B A B SP

10.8 17.2 18.1 17.5 19.6 11.0 12.1 29.4

68-80 8 20 8 30 15-17 26 5-10

0.002 n/ab 1.0 0.4 0.8 0.03 0.03 n/a

0.0002 0.004 0.001 0.06 0.07 0.001 0.002 3

A B A B SP A B SP SP

13.0 14.3 12.0 12.8 27.4 15.3 16.0 20.0 21.6

16-25 20-43 7-13 8-12 5 21 19 6 8

0.07 0.29 0.05 0.05 3.8 1.0 1.9 n/a n/a

0.006 0.01 0.02 0.03 4.4 0.01 0.005 0.2 5

Single peaks. bNot available.

Labeling with 99mTc was performed by transchelation from [99mTc]gluconate, but the optimal pH and temperature varied among compounds, and radiochemical yields ranged from 5 to 80% (Table 3). HPLC analysis showed that some compounds formed a single peak (SP) and others two complexes (A and B) with 99mTc. Representative labeling data is shown in Figure 3 for compound 6, demonstrating the formation of two complexes (peaks A and B) in a total of 82% yield, with 14% free pertechnetate remaining. In the class I series, compound 1, which contained the dimethylGly-Ser-Cys chelator, formed a single product in up to 80% yield at room temperature (Table 3). Compound 2, the dibenzylGly analogue, required heating at 50 °C and formed two peaks in a 1:2 ratio and a total yield of 30%. On the basis of studies with the chelator alone, and the chelator linked to 2-nitroimidazole, the two peaks are believed to be syn and anti conformations (25, 26). The remaining class I compounds (3-5) required at least 70 °C for the labeling reaction to take place. Substitution of Gly by Val (compound 3) or eliminating two Gly residues (compound 4) resulted in two peaks again in an approximate 1:2 ratio but at a slightly higher

Figure 3. Representative HPLC radioactivity trace for compound 6 labeled with technetium-99m by transchelation from [99mTc]gluconate. Peaks A and B are thought to be interconvertable syn and anti forms of the TcdO group.

total yield of 40%. Capping the free acid on the Cterminus of compound 4 with dibenzyl amide (compound 5) resulted in one peak with lower yield (5-10%). In general, the optimal pH for labeling was 3.5 for the free acids and 6.0 for the amides. In the class II series, the Lys linker to the 2-NI targeting group was incorporated into the chelator, replacing Ser, rather than being attached to the end of the chelator. The first compound in the series, 6, produced two peaks in an approximate 1:2 ratio and up to 70% yield. Capping the free acid on the C-terminus of compound 6 with an amide (compound 7) resulted in a much lower total yield (maximally 25%) of two peaks of equal size. The dibenzylamide of compound 6 (compound 8) produced an extremely low yield of 99mTc-labeled compound. The piperidino analogue of compound 6 (compound 9) resulted in two equal peaks in 40% yield. Substitution of the dimethyl group on Gly in compound 6 with dibenzyl (compound 10) resulted in a longer retention time with one peak at lower yield (Table 3). Capping the free acid on the C-terminus of compound 10 with an amide (compound 11) did not increase the yield of 99mTc-labeled compound. The P of the 11 complexes varied over a wide range (0.0002-5) but generally were in accordance with predictions based on structure (Table 3). There were also differences in P as a function of pH. In general, the free acids, which would be neutral at pH 2.0, had much lower P at pH 7.4 where they would be ionized, while the amides showed little effect of pH. The P values measured at pH 2.0 showed a significant correlation with HPLC retention times when run in 0.1% TFA (Table 3). Representative data for compound 6 is shown in Figure 4 for its accumulation in CHO cells under aerobic and hypoxic conditions. Under aerobic conditions, this compound showed a ratio of activity inside of the cells (Cin) to that outside the cells (Cout) of 0.3 in the first 5 min, which increased 2-fold over the next 4 h. In contrast, hypoxic cells, although showing the same initial Cin/Cout ratio at the first 5 min, steadily accumulated activity inside the cells, so at 4 h there was a 5-fold increase in activity over the 5 min value. Similar experiments were done for all 11 compounds and Cin/Cout values at 4 h for aerobic and hypoxic incubations are plotted in Figure 5 as a function of P. In the in vitro tracer accumulation studies, the tracers varied in their levels of partitioning into aerobic CHO

2-Nitroimidazole Peptidic Chelators for

99mTc

Figure 4. Accumulation of compound 6 in CHO cells measured at 37 °C under aerobic (closed squares) versus hypoxic (closed circles) incubation conditions. The Cin/Cout values for peaks A and B have been averaged together for each time point. Representative data from four experiments are shown. Error bars are standard error of the mean for n ) 4 experiments.

Bioconjugate Chem., Vol. 11, No. 3, 2000 405

Figure 6. Relationship between the molecular weight of nine of the 99mTc-labeled ligands from Table 1 and the hypoxic/aerobic differential in CHO cells after 4 h of incubation. Numbers indicate compounds as listed in Table 1. The dotted line indicates a hypoxic/aerobic differential of 1.5.

convertable syn and anti conformations with respect to TcdO (25, 26). Therefore, the data in Figure 5 is the average accumulation for both peaks; similarly, the P values at pH 7.4 for both peaks in Table 3 have been averaged. Also shown in Figure 5 are previous results obtained in the same in vitro system for BMS181321 (14), BRU59-21 (18) and [99mTc]HL91 (21) for comparison. It was noted that when a plot was made of the hypoxic differential at 4 h of (Cin/Cout)hypoxic/(Cin/Cout)aerobic versus molecular weight there was a tendency for the ratio to increase with decreasing molecular weight (Figure 6). DISCUSSION

Figure 5. Relationship between lipophilicity (partition coefficient) of nine compounds from Table 1 and 4-h accumulation (Cin/Cout) in aerobic (open squares) and hypoxic (closed circles) cells. Error bars are standard error of the mean for n ) 3 or more experiments. Included for comparison are previous results obtained in the same system for BMS181321 (14), BRU59-21 (18), and [99mTc]HL91 (21). Double-headed arrows indicate compounds with significant hypoxic-aerobic differentials (g1.5). Numbers indicate compounds as listed in Table 1.

cells (Figure 5). These levels showed some correlation with P, but, whereas P varied by 4 orders of magnitude, accumulation varied only 30-fold. Although a number of the 99mTc complexes did not show selective accumulation in hypoxic cells, others showed 1.8-3.8-fold greater accumulation under hypoxia after 4 h. Two compounds of class I showed this differential (compounds 4 and 5) and two compounds from class II were selectively taken up in hypoxic cells (compounds 6 and 7). Two of the compounds, 10 and 11, were not tested in the in vitro assay with CHO cells because of difficulties in labeling. Single experiments in Hela cells showed no significant hypoxic versus aerobic differential for these compounds (data not shown). For compounds which formed two 99m Tc-labeled complexes, both complexes tended to behave similarly, consistent with the suggestion they are inter-

In the search for a radiolabeled 2-NI with optimal properties for imaging hypoxia in tumors, P is felt to be an important factor (6). Nuclear imaging involves a compromise between two factors: signal and signal-tonoise [or target-to-background (T/B)] ratio. In other words, there needs to be an adequate signal but there also needs to be sufficient contrast in order for an agent to be useful. P plays a role in distribution of drugs within the body, penetration of drugs through membranes, and routes of metabolism and excretion. High P is often associated with retention in background tissues, including blood. This is certainly the case with BMS181321, 2-NI with a PnAO chelator, which exhibits a P of 40, extensive hepatobiliary excretion, prolonged clearance from blood and muscle, and achieves only modest T/B ratios (14). The second-generation compound BRU59-21 has a P of 12 which contributes to more rapid blood clearance and modestly improved T/B ratios but excretion is still primarily into the GI tract (18). Studies in the radioiodinated sugar 2-NI derivatives have shown that a P of ∼1 is optimal for achieving renal excretion and high T/B ratios (6). The objective of the present work was to prepare a series of 99mTc-labeled 2-NI with a wide range in P and to evaluate them in an in vitro model of tumor hypoxia. This range in P was achieved using the peptidic 99mTc chelation sequence dialkylGly-Ser-Cys, which contains one sulfur atom, one amine nitrogen, and two amide nitrogens as an N3S class chelator and was developed by Resolution Pharmaceuticals (25). By varying the amino acids, dialkyl groups, and pendant groups, P was varied over 4 orders of magnitude. The class I compounds use

406 Bioconjugate Chem., Vol. 11, No. 3, 2000

the dialkylGly-Ser-Cys chelator with the 2-NI attached through Lys at the end of the chelator. To reduce molecular weight, the class II compounds were designed with the 2-NI-derivatized Lys as part of the chelator in place of Ser. However, it became apparent that these modifications were not innocuous. While the dimethylGly-Ser-Cys chelator binds 99mTc in extremely high yield at room temperature (25), changing the dialkyl group or substituting Lys for Ser results in a compound which requires harsher labeling conditions and which does not achieve high yields. With most of the complexes, two peaks were evident on HPLC analysis. Ongoing work in this laboratory suggests that these are syn and anti conformations as described by Wong et al. for the dimethylGly-Ser-Cys chelator (25) and that these conformations interconvert after isolation (26, unpublished results). This is supported by the observations that the peaks do not differ greatly in their P values (Table 3) and that both peaks tend to behave similarly in the in vitro test system. The xanthine oxidase results show that attachment of a peptidic chelator to 2-nitroimidazole does not prevent the nitro group from being reduced enzymatically; however, it may alter the rate of reduction (Table 2). One of the ligands, 11, was reduced at a rate similar to misonidazole, while the remainder were reduced more slowly, at 17-60% of that rate. Linder et al. have shown that BMS181321 is reduced at 41% of the rate of misonidazole, which is in the same range as these ligands (13). The results for BRU59 showed that it was reduced somewhat more slowly, 20% of the misonidazole rate, than BMS181321. The accumulation of the tracers in CHO cells in vitro was related to P but did not vary over as wide a range as P. It has been shown recently by Oya et al. that P is not the sole determinant of brain uptake of 99mTc chelates but that the nature of the chelator plays a role; in particular, amides (such as the peptidic series in the present work) are not taken up as efficiently as amines (e.g., BMS181321 and BRU59-21) (29). Several of the compounds showed selective uptake in hypoxic cells, although the magnitude of this selective uptake was not as great as has been obtained with BMS181321, BRU5921, and [99mTc]HL91 (refs 14, 18, and 21 and Figure 5). The explanation of why only 4 of the present 11 compounds showed a significant hypoxic/aerobic differential ()1.5) is not known at this time. It is clearly not a strict function of P. However, as shown in Figure 6 for this limited series, it does appear that the lower molecular weight compounds tended to show a differential while the higher molecular weight compounds did not. In conclusion, the present studies indicated the usefulness of the automated solid-phase synthetic system in the design of peptidic chelators for 99mTc covering a broad range of P values. Four of the 11 compounds, all containing a 2-nitroimidazole targeting group, were selectively accumulated in hypoxic compared to aerobic cells. Though the degree of selectivity of the present compounds was not as great as that seen previously for several other 99mTc-labeled hypoxic markers, the present compounds represent a novel approach to the design of hypoxic markers that warrant further investigation. ACKNOWLEDGMENT

We thank Tam Nguyen and Dr Susan Peers of Resolution Pharmaceuticals Inc. for their assistance. Funding was provided by the Medical Research Council of Canada

Zhang et al.

and Resolution Pharmaceuticals Inc. through the University/Industry grant program (Grant UI-13558). LITERATURE CITED (1) Bush, R. S., Jenkin, R. D. T., Allt, W. E. C., Beale, F. A., Bean, H., Dembo, A. J., and Pringle, J. F. (1978) Definitive evidence for hypoxic cells influencing cure in cancer therapy. Br. J. Cancer 37 (Suppl.), S302-S306. (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) Brizel, D. M., Scully, S. P., Harrelson, J. M., Layfield, L. J., Bean, J. M., Prosnitz, L. R., and Dewhirst, M. W. (1996) Tumor oxygenation predicts for likelihood of distant metastases in human soft tissue sarcoma. Cancer Res. 56, 941943. (4) 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 tumours. Nature 379, 88-91. (5) Nunn, A., Linder, K., and Strauss, H. W. (1995) Nitroimidazoles and imaging hypoxia. Eur. J. Nucl. Med. 22, 265280. (6) 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. (7) 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. (8) Koh, W.-J., Bergman, K. S., Rasey, J. S., Peterson, L. M., Evans, M. L., Graham, M. M., Grierson, J. R., Lindsley, K. L., Lewellen, T. K., Krohn, K. A., and Griffin, T. W. (1995) Evaluation of oxygenation status during fractionated radiotherapy in human nonsmall cell lung cancers using [F-18]fluoromisonidazole positron emission tomography. Int. J. Radiat. Oncol. Biol. Phys. 33, 391-398. (9) Yeh, S.-H., Liu, R.-S., Wu, L.-C., Yang, D. J., Yen, S.-H., Chang, C.-W., Yu, T.-W., Chou, K.-L., and Chen K. Y. (1996) Fluorine-18 fluoromisonidazole tumour to muscle retention ratio for the detection of hypoxia in nasopharyngeal carcinoma. Eur. J. Nucl. Med. 23, 1378-1383. (10) 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. (11) 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) Non-invasive assessment of human tumor hypoxia with 123I-iodoazomycin arabinoside: preliminary report of a clinical study. Br. J. Cancer 65, 90-95. (12) Urtasun, R. C., Parliament, M. B., McEwan, A. J., Mercer, J. R., Mannan, R. H., Wiebe, L. I., Morin, C., and Chapman, J. D. (1996) Measurement of hypoxia in human tumours by non-invasive SPECT imaging of iodoazomycin arabinoside. Br. J. Cancer 74 (Suppl.), S209-S212. (13) 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. (14) Ballinger, J. R., Wan Min Kee, J., 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) 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 unlabelled nitroaromatics. Biochem. Pharmacol. 54, 685-693.

2-Nitroimidazole Peptidic Chelators for

99mTc

(16) 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. (17) Johnson, L. L., Schofield, L., Mastrofrancesco, P., Donahay, T., and Nott, L. (1998) Technetium-99m-nitroimidazole uptake in a swine model of demand ischemia. J. Nucl. Med. 39, 1468-1475. (18) Melo, T., Duncan, J., Ballinger, J. R., and Rauth, A. M. (2000) BRU59-21, A second-generation technetium-99mlabelled 2-nitroimidazole for imaging hypoxia in tumors. J. Nucl. Med. 41, 169-176. (19) Archer, C. M., Edwards, B., Kelly, J. D., King, A. C., Burke, J. F., and Riley, A. L. M. (1995) Technetium labelled 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 Editoriali, Padova. (20) 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. (21) Zhang, X., Melo, T., Ballinger, J. R., and Rauth, A. M. (1998) Studies of 99mTc-BnAO (HL91): A non-nitroaromatic compound for hypoxic cell detection. Int. J. Radiat. Oncol. Biol. Phys. 42, 737-740. (22) Brauers, G., Archer, C. M., and Burke, J. F. (1997) The chemical characterization of tumor imaging agent 99mTcHL91. Eur. J. Nucl. Med. 24, 943.

Bioconjugate Chem., Vol. 11, No. 3, 2000 407 (23) Brown, J. M., and Workman, P. (1980) Partition coefficient as a guide to the development of radiosensitizers which are less toxic than misonidazole. Radiat. Res. 82, 171-190. (24) Kumar, P., Mannan, R., Zhang, Z., McEwan, A. J. B., and Wiebe, L. I. (1995) 99mTc-PAO-Azomycin glucuronide (99mTcPAO-AZG): synthesis and biological evaluation as a marker of hypoxic tissues. Q. J. Nucl. Med. 39, 65-66. (25) Wong, E., Fauconnier, T., Bennett, S., Valliant, 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 the syn and anti conformations. Inorg. Chem. 36, 5799-5808. (26) Su, Z. F., Zhang, X., Ballinger, J. R., Rauth, A. M., Pollak, A., and Thornback, J. R. (1999) Synthesis and evaluation of two technetium-99m-labeled peptidic 2-nitroimidazoles for imaging hypoxia. Bioconjugate Chem. 10, 897-904. (27) Clarke, E. D., Wardman, P., and Goulding, K. H. (1980) Anaerobic reduction of nitroimidazoles by reduced flavin mononucletide and by xanthine oxidase. Biochem. Pharmacol. 29, 2684-2687. (28) Josephy, P. D., Palcic, B., and Skarsgard, L. D. (1981) Reduction of misonidazole and its derivatives by xanthine oxidase. Biochem. Pharmacol. 30, 849-853. (29) Oya, S., Plossl, K., Kung, M. P., Stevenson, D. A., and Kung, H. F. (1998) Small and neutral Tc(V)O BAT, bisaminoethanethiol (N2S2) complexes for developing new brain agents. Nucl. Med. Biol. 25, 135-140.

BC9901595