Toxicity, biodistribution and radioprotective capacity of L-homocysteine thiolactone in CNS tissues and tumors in rodents: comparison with prior results with phosphorothioates.

L-Homocysteine thiolactone (L-HCTL) was evaluated for its potential as an intravenously-administered central nervous system (CNS) radioprotector in C3H mice and F344 rats. Toxicity assessments in the mouse yielded a LD50 of 297 mg/kg and in the rat 389 mg/kg. Biodistribution studies in tumor-bearing mice showed that brain specimens contained more label at 10 min than the tumors but less at 30 or 60 min. Brain uptake relative to the tumors, the brain/tumor ratio, ranged between 0.5 and 3.3. The cervical spinal cord of non-tumor-bearing rats was irradiated with 32 Gy 137Cs with or without prior treatment with L-HCTL following which the time to forelimb or hindlimb paralysis was measured to determine the relative protective factors (RPFs) for this radiation dose. For forelimb paralysis the RPF was 1.9 (+/- 1.0, SD) and for hindlimb it was 2.0 (+/- 1.1, SD). 36B-10 glioma cells irradiated in vitro with or without L-HCTL and assayed for colony forming capacity demonstrated a dose modifying factor (DMF) of only 1.15 (+/- 0.16, SE). Rats bearing intracerebral 36B-10 glioma received 137Cs irradiation with or without L-HCTL after which the tumors were similarly assayed in vitro. From this the glioma DMF was 1.2 (+/- 0.30, SE). Compared to prior results with phosphorothioates our data show that the toxicity of L-HCTL is roughly the same as WR2721, WR77913 and WR3689 and that it distributes at higher levels in the CNS after systemic administration. L-HCTL may well equal these phosphorothioates at protecting normal CNS tissue without requiring administration directly into the cerebrospinal fluid-containing spaces and it does not protect the 36B-10 glioma.

Magnetic resonance spectroscopic measurement of cellular thiol reduction-oxidation state.

Chromatographic and magnetic resonance spectroscopic measurements of thiol reduction-oxidation state in chemically constructed samples show close analytical agreement. This result, coupled with the synthesis of new probe molecules allowing greater sensitivity and lower toxicity, supports the development of an NMR method for non-invasive thiol redox measurement, an important variable in the response of tumors to radiation and chemotherapy.

Reduction of fluoromisonidazole, a new imaging agent for hypoxia.

[18F]Fluoromisonidazole (1-(3-[18F]fluoro-2-hydroxypropyl)-2-nitroimidazole, [18F]FMISO) is a nitroimidazole compound that is being used as a new imaging agent for hypoxia. Because its uptake in hypoxic tissue is dependent on reduction of the nitro group on the imidazole ring, it is necessary to verify the availability of nitroreductase enzymes in a variety of tissues. FMISO reduction was studied using chemical and enzymatic reducing systems and mammalian cells. FMISO reduction by iron/HCl eliminated the absorbance peak at 325 nm caused by the nitro group. FMISO reduction by xanthine oxidase, as measured by a decrease in absorbance at 325 nm, occurred at a rate of 2.4 +/- 0.3 nmol/min/unit enzyme (mean +/- SEM, N = 15). This reaction was inhibited by allopurinol. Separation of the parent drug from its reduction product following chemical and enzymatic reductions indicated that iron/HCl reduced the majority of the FMISO molecules present, while xanthine oxidase did not. Reduction of FMISO by NADH dehydrogenase could not be demonstrated spectrophotometrically. Measurement of the reduction of FMISO in V79 cells based on the binding of [3H]FMISO to cellular macromolecules was performed using a cell suspension in a three-neck flask. Hypoxic V79 cells bound [3H]FMISO at the rate of 0.26 +/- 0.07 pmol/10(6) cells/min (N = 8). When specific inhibitors of two nitroreductase enzymes and a general inhibitor of electron transport were added to the cell suspension, no consistent, statistically significant inhibition of FMISO binding could be shown. We conclude that while inhibition of FMISO reduction by a purified nitroreductase can be shown, nitroreductase activity in cells is not inhibited so easily. This supports the hypothesis that nitroreductases are plentiful and will not limit the rate of FMISO reduction and uptake in hypoxic tumors or nonmalignant tissues.

Characteristics of the binding of labeled fluoromisonidazole in cells in vitro.

Radiolabeled fluoromisonidazole (FMISO) is being investigated as an imaging agent for hypoxia in tumors and nonmalignant tissues in myocardial infarct or stroke. In this study in vitro cell cultures were used to characterize the oxygen dependency of FMISO uptake and to examine other modifying factors. The uptake of [3H]FMISO was measured in four cell lines in vitro: V-79, EMT-6(UW), RIF-1, and CaOs-1. The modifying effects of different O2 levels as well as cell growth state and concentration of glucose and nonprotein sulfhydryls were examined. In these cell types an O2 level between 720 and 2300 ppm inhibited FMISO binding by 50%, relative to binding under anoxic conditions. These values bracket the O2 level which confers full radiobiologic hypoxia, about 1000 ppm. Some bound label was released from cells in the first 1 to 3 h after a 3-h anoxic labeling with [3H]FMISO, but this does not represent tritium loss from the parent molecule. Cells from unfed plateau-phase cultures took up less [3H]FMISO than did exponentially growing cells incubated at comparable O2 levels. Reducing glucose to 1/10 or 1/100 of the usual concentration in medium had little effect on binding of micromolar levels of FMISO, except in V-79 cells, where reduced glucose levels were associated with increased FMISO accumulation. Adding cysteamine to the culture medium moderately increased FMISO uptake. We conclude that cell growth state, glucose, and nonprotein sulfhydryl concentrations affect FMISO binding, albeit less than varying O2 levels: anoxic/oxic binding ratios vary from 12.6 to 28 for the four cell types examined. Nonetheless these factors must be considered in evaluating the oxygen-dependent binding of this nitroimidazole in tumors or tissues.

Binding of aminoalkylphosphorothioate radioprotective drugs to rodent tissue proteins.

The formation of protein mixed disulfides which influences the pharmacodynamics of the phosphorothioate radioprotective drugs WR2721 [S-2-(3-aminopropyl)aminoethylphosphorothioic acid, Ethiofos] and WR3689 [S-2-(3-methylaminopropylamino)ethylphosphorothioic acid] and their metabolites was investigated. WR3689-derived thiols and disulfides bound to rat serum protein to about 45 and 40% of the total drug in the incubation when present at a 400 microM concentration. Metabolites of WR2721 were nearly indistinguishable from the corresponding metabolites of WR3689 in their mixed disulfide binding propensity. Mixed disulfide formation was saturable; binding sites on bovine albumin or rat serum protein amounted to 0.15 and 2.4 mumol/mg protein respectively. The sum of all WR3689 metabolites (when measured by NMR spectroscopy) was reduced to the same degree as drug binding, suggesting that a portion of the bound drug was not NMR observable. Approximately 2-4 nmol WR3689-thiol/mg protein was bound to homogenates of mouse tissues (liver, kidney, lung, brain, and serum) when incubated in vitro, whereas after in vivo injection drug binding appeared to be limited more by drug distribution than by the capacity for mixed disulfide formation.

Synthesis and radiobiological applications of [35S]L-homocysteine thiolactone.

[35S]L-Homocysteine thiolactone ([35S]L-HCTL) was synthesized and its biodistribution evaluated as a potential brain radioprotective agent and as a tissue hypoxia marker. Drug uptake in mouse brain exceeded that in s.c. tumor 3 h post injection only. Multiple indicator dilution experiments in the rabbit heart indicate that membrane permeability of [35S]L-HCTL does not limit its usefulness as a hypoxia marker. In addition, a positive correlation was observed between regional coronary blood flow and myocardial content of [35S]adenosylhomocysteine formed from [35S]homocysteine and adenosine.

Radiolabelled fluoromisonidazole as an imaging agent for tumor hypoxia.

Fluoromisonidazole labeled with H-3 or F-18 has been tested as a quantitative probe for hypoxic cells in vitro and in rodent and spontaneous dog tumors in vivo. In V-79, EMT-6(UW), RIF-1, and canine osteosarcoma cells in vitro, the binding of 50 microM [H-3]Fluoromisonidazole was 50% inhibited by 1000-2000 ppm O2, relative to binding under anoxic conditions. After a 3 hr incubation with labeled drug, the anoxic/oxic binding ratios ranged from 12 to 27 for the four cell types. Retention of [H-3]fluoromisonidazole 4 hr after injection was greater in large KHT tumors (400-600 mm3) with an estimated hypoxic fraction greater than 30%, than in smaller tumors (50-200 mm3) with an estimated hypoxic fraction of 7-12%. RIF-1 tumors, with an estimated hypoxic fraction of 1.5%, retained the least label, with tumor: blood ratios ranging from 1.7 to 1.9. Spontaneous dog osteosarcomas were imaged with a time of flight positron emission tomograph for up to 5 hr following injection of [F-18] fluoromisonidazole. Analysis of regions of interest in images allowed creation of dynamic tissue time activity curves and calculation of tissue uptake in cpm/gram. These values were compared to radioactivity in plasma. In all cases, retention in some tumor regions exceeded that in plasma and in normal tissue, such as muscle or brain, by 3 to 5 hr post injection. Uptake of fluoromisonidazole in tumors was heterogeneous, with ratios of maximum to minimum uptake as high as 4 in different regions of interest in the same tumor. Tumor:plasma values ranged from 0.28 to 2.02. The oxygen dependency of fluoromisonidazole retention was similar in a variety of cell types and was 50% inhibited by O2 levels in the transition between full radiobiological hypoxia and partial sensitization. The quantitative regional imaging of [F-18] fluoromisonidazole in spontaneous canine tumors at varying times post-injection lays the basis for imaging and modeling of oxygen-dependent drug retention in different regions of human neoplasms.

Measurement of tissue oxidation-reduction state with carbon-13 nuclear magnetic resonance spectroscopy.

The oxidation state of tissues influences their response to cancer therapy. We have devised a novel approach to the measurement of thiol redox which is based on the relative nuclear magnetic resonance signal intensity from carbon-13 adjacent to sulfur in metabolites of the redox-sensitive phosphorothioate drug, S-2-(3-methylaminopropylamino)ethylphosphorothioic acid (WR3689). Incubation of WR3689 metabolites under oxidizing conditions results in quantifiable changes in the 13C nuclear magnetic resonance spectrum stoichiometrically related to the degree of oxidation in mouse liver homogenate in vitro. Drug oxidation is competitive with the oxidation of tissue-derived thiol groups under these conditions. Noninvasive measurement of redox state may assist in designing more effective strategies for altering normal and malignant tissue response to cancer therapy.

Enhanced binding of the hypoxic cell marker [3H]fluoromisonidazole in ischemic myocardium.

The 2-nitroimidazole fluoromisonidazole is metabolically trapped in viable hypoxic cells in inverse proportion to PO2. This attribute suggests that [18F]fluoromisonidazole may be useful for imaging hypoxic tissue using positron emission tomography. To examine this potential, we studied the pharmacokinetics and biodistribution of [3H]fluoromisonidazole in six open chest dogs. In two normal dogs, plasma and urine samples were collected over a 4-hr period following i.v. injection of the drug. In four animals, regional myocardial ischemia was produced 2 hr prior to drug injection by occlusion of the circumflex coronary artery and maintained during the 4-hr sampling period. In all animals, postmortem samples of myocardium and other organs were obtained and tissue, plasma, and urine tritium activity were determined by liquid scintillation counting. In areas of reduced flow, [3H]fluoromisonidazole accumulated in myocardium in inverse proportion to myocardial blood flow measured by microspheres, indicating enhanced binding in hypoxic tissues. Maximum tissue concentrations in ischemic myocardium were two- to three-fold greater than in normal myocardium and plasma. Plasma clearance data indicate the drug is rapidly distributed into the total-body water, clears from the body with a half-life of 275 +/- 50 min, and undergoes minimal metabolism by 4 hr. We conclude [18F]fluoromisonidazole may be a suitable agent for radionuclide imaging of hypoxic myocardium.

Characterization of radiolabeled fluoromisonidazole as a probe for hypoxic cells.

Radiolabeled fluoromisonidazole has been characterized as a probe for hypoxic cells in vitro and in vivo. The uptake and retention of [3H]fluoromisonidazole and [3H]misonidazole were compared in V-79 cell monolayers and spheroids by varying incubation time and O2 levels in contact with the medium. The two labeled drugs were retained similarly in cell populations isolated from different depths in spheroids, and the amount of each drug bound in cells at the spheroid periphery increased with decreasing O2 level. The labeling patterns in autoradiographs were similar for spheroids incubated with the two labeled drugs, with most silver grains located over a zone of viable and presumed hypoxic cells intermediate between the necrotic center and the periphery of the spheroid. Biodistribution of the two tritiated drugs was compared in C3H mice bearing KHT tumors with 15% radiobiologically hypoxic cells. Tumor:blood and tumor:muscle ratios greater than 5.0 were achieved in mice sacrificed 4 h after the last of three injections of 5 or 20 mumol/kg of [3H]fluoromisonidazole. These ratios are compatible with imaging and are higher than those obtained with 50 mumol/kg misonidazole in a similar administration protocol. TLC analysis of plasma from mice injected with [3H]fluoromisonidazole indicated that the drug was stable in vivo for up to 2 h and that the metabolites formed were too polar to be dehalogenation products. Fluoromisonidazole labeled with 18F at the end of the alkyl side chain would retain the label on metabolites that bind in hypoxic cells in vivo. Fluoromisonidazole binds stably in the same populations of hypoxic cells as does misonidazole, and we conclude that [18F]fluromisonidazole has potential use as a hypoxia imaging agent in vivo.

Synthesis and characterization of congeners of misonidazole for imaging hypoxia.

Misonidazole is a known hypoxic cell sensitizer that binds covalently in hypoxic cells. Its congeners labeled with 77Br, 75Br, or 18F, are likely candidates for imaging hypoxia. We have synthesized and tested [82Br]-4-bromomisonidazole, [3H]-4-bromomisonidazole, [3H]fluoromisonidazole and [3H]misonidazole as prototype radiopharmaceuticals and have compared their uptake in normal and malignant tissues. The higher lipophilicity of brominated misonidazole increased its concentration in the hypoxic portion of tumors at 2 hr, but high blood levels contributed to excessive background, incompatible with imaging. Hydrogen-3-fluoromisonidazole diffused into tumors at a slower rate than misonidazole but it also cleared from normal tissues so that after 2 hr tumor-to-blood ratios favorable for imaging were achieved. In the compounds that were studied, fluorine at the end of the alkyl chain is more stable in vivo than bromine on the imidazole ring. Our results indicate that [18F] fluoromisonidazole may be a useful tracer for imaging hypoxia at approximately 4 hr after injection.

Toxicity and biodistribution of the radioprotectors, WR2721, WR77913, and WR3689, in the CNS following intraventricular or intracisternal administration.

The radioprotective capacity of the phosphorothioate compounds, WR2721, WR77913, and WR3689, in the CNS is being evaluated following injection of the drugs into the lateral cerebral ventricle or the cisterna magna of F-344 rats. This approach circumvents the blood-brain barrier and permits an assessment of the CNS toxicity and regional distribution of these compounds. Following intraventricular injection in 150-200 gm female rats, the LD50 doses for WR2721, WR77913, and WR3689 were respectively 0.60 +/- 0.07 mg (S.E.), 2.36 +/- 0.13 mg, and 3.56 +/- 0.26 mg. Following intracisternal injection the LD50 doses were 0.71 +/- 0.18 mg, 4.12 +/- 1.09 mg and 3.03 +/- 0.68 mg, respectively. WR 2721 produced lethargy, unsteady gait, and dishevelment but these signs all resolved completely within 1-3 days in survivors. In addition to these signs, WR77913 and WR3689 produced severe convulsions. At high doses, following intraventricular administration, all three drugs were associated with cerebral and diencephalic periventricular necrosis and ipsilateral necrosis of the lateral hippocampus. Biodistribution studies were performed with [S-35]-labeled derivatives of the drugs and tissue sampling. The three drugs demonstrated similar patterns. Forty-five minutes following either the intraventricular or intracisternal route of drug delivery the highest drug concentrations were in the brainstem, cerebellum, and cervical cord. Additional studies with autoradiography revealed that intraventricular injection was associated with high drug uptake in the cerebral white matter, the periventricular diencephalon, and the periaqueductal mesencephalon. The biodistribution and toxicity data together suggest that the drugs can be ranked, WR3689 greater than WR77913 greater than WR2721, according to the level of drug thiol that can be achieved in the CNS tissues with intraventricular or intracisternal injection. Tissue levels achievable with WR2721 following these two routes of administration are as high as levels others have reported as radioprotective in rodent skin and gut.

Synthesis, biodistribution, and autoradiography of radiolabeled S-2-(3-methylaminopropylamino)ethylphosphorothioic acid (WR-3689).

35S- and 3H-labeled S-2-(3-methylaminopropylamino)ethylphosphorothioic acid (WR-3689) have been synthesized in our laboratory and used to study organ and cellular level distribution in C3H/Km mice bearing RIF-1 tumors. Tissue biodistributions obtained with 35S-WR-3689 showed that blood levels peak at 15 min postinjection and decline gradually over 60 min. At 30 min after drug injection the highest uptake is in kidney and submandibular salivary gland, with lowest levels in brain and moderate to low levels in the RIF-1 tumor, comparable to levels in skin and muscle. High resolution diffusible substance autoradiography with 3H-WR-3689 reveals a homogenous distribution of label over cells in liver and lung and nonuniform distribution of silver grains over the cytoplasm of cells in the kidney cortex, parotid and submandibular salivary glands, and small intestine. There are no indications of preferential nuclear location of label from protective drug in any tissue. Correlations of biodistribution and autoradiography data with measures of radioprotection in different tissues will be useful in interpreting mechanisms of radioprotection with this phosphorothioate.

Biodistribution of the radioprotective drug 35S-labeled 3-amino-2-hydroxypropyl phosphorothioate (WR77913).

3-Amino-2-hydroxypropyl phosphorothioate (WR77913), a less toxic phosphorothioate radioprotector than WR2721, has been labeled with 35S. The biodistribution of a radioprotective dose of 800 mg/kg was determined in C3H mice bearing RIF-1 tumors as a function of time after intraperitoneal injection and was expressed as percentage injected dose/gram (% ID/g). Levels of 35S in the blood peaked 10 min after injection, and radioactivity in most tissues was highest at 15 min. Label in most tissues declined markedly between 15 and 60 min, but in gut, salivary glands, tumor, and brain, the levels of radioactivity remained quite stable over 1 hr. At 30 min after injection the highest levels of labeled drug were found in submandibular salivary glands, gut, and kidney, with the lowest level in brain. Tumors had approximately the same amount of label as blood, muscle, skin, and esophagus. Two principal differences between the distribution of label from WR77913 and WR2721 were defined. Although blood levels of 35S-WR2721 also peaked 10 min after injection, the 10-min blood levels achieved for WR77913 were more than fourfold greater than those attained by WR2721. Maximum levels of WR2721 occurred in most tissues 30 to 60 min after administration of the drug, compared to 15 min for WR77913. The basis for these differences remains to be determined, but these results suggest that the optimum interval between administration of WR77913 and irradiation may be shorter than for WR2721.

Further characterization of 4-bromomisonidazole as a potential detector of hypoxic cells.

[14C]Bromomisonidazole was prepared by direct bromination of [ring-2] [14C]misonidazole in dioxane. The uptake and binding of the two labeled sensitizers were compared in vitro in 1-mm EMT-6 spheroids which contain a necrotic core. Using liquid scintillation counting it was shown that spheroids incubated with 50 microM [14C]bromomisonidazole concentrated drug above levels in the medium by 1 1/2 hr and achieved maximum concentration by 10 hr with no further increase at 23 hr. Spheroids incubated with 50 microM [14C]misonidazole may concentrate the sensitizer more slowly but ultimately reached the same fivefold increase over levels in the medium by 23 hr as was observed for bromomisonidazole. Autoradiographs prepared from spheroids after incubation with [14C]misonidazole or [14C]bromomisonidazole showed silver grains preferentially located over viable hypoxic cells in the inner half of the spheroid rim adjacent to the necrotic center, with lower grain density over nonviable necrotic areas and many fewer grains over oxic cells at the periphery of the spheroid. The results indicate that both severely and moderately hypoxic cells may preferentially bind [14C]bromomisondiazole. The data support the potential of radiolabeled bromomisonidazole for in vivo imaging pending additional studies of the metabolism of this agent.

Comparison of binding of [3H]misonidazole and [14C]misonidazole in multicell spheroids.

Uptake of [2-ring-14C]misonidazole and [3H]misonidazole with tritium in the side chain has been compared in 1-mm EMT-6/UW spheroids using liquid scintillation counting and autoradiography. The uptake of both labeled sensitizers as a function of incubation time was virtually identical. Uptake by the spheroids exceeded levels in the medium by 11/2 to 2 hr and was well modeled as a first-order binding process, with rate constants of 0.00324 hr-1 for 3H and 0.00388 hr-1 for 14C. The similar uptake of the two versions of this sensitizer labeled in different positions suggests that the metabolic actions which allow the drug to bind in hypoxic cells do not principally involve metabolites which separate the number 2 carbon of the imidazole ring from the side chain. The pattern of silver grains in autoradiographs was similar for both labeled sensitizers, with most labeled drug bound in an intermediate zone of cells between the necrotic center and the actively proliferating rim of the spheroids. The superior resolution possible with the tritiated compound showed that both nucleus and cytoplasm in viable looking cells were labeled while pycnotic cells were not labeled.

Analysis of S-35 labeled WR-2721 and its metabolites in biological fluids.

Studies with WR-2721 and related compounds have been hindered by the lack of a suitable assay for the drug and its major metabolites. We have developed a chromatographic method which requires no derivatization for the separation and detection of WR-2721, the free thiol, its symmetrical disulfide and other mixed disulfides. Our procedure involves ion-pairing for separation of ionizable compounds by causing polar molecules to become more lipophilic and hence separable using reverse phase HPLC. Detection is based upon liquid scintillation counting of S-35 incorporated during the synthesis of the parent compound. This method requires no pre-column preparation of samples and, by detecting the S-35 label, eliminates the chance that a coeluting species could interfere with detection, as might occur with post-column derivatization. Chromatography was done using a 10 micron C8RP column and 35% MeOH/65% 0.0113M NaH2PO4, 0.005 M hexanesulfonate, pH 5.9, flowing at 1 ml/min. Half-minute fractions were collected into scintillation vials for counting. Retention volumes for the various compounds were: column breakthrough (3.5 ml), WR-2721 (4.5 ml), WR-1065 (9 ml), and WR-33278 (24 ml). This analytical technique employing radiotracers can be used to study radioprotective mechanisms by time dependent measurements of the tissue distribution and chemical form of labeled drug. Such chemical information can then be correlated with biological measures of radiation protection.

Short-term thymidine uptake in normal and neoplastic tissues: studies for PET.

Uptake of H-3 thymidine was studied in mice, both normal and with spontaneous lymphoma, and in the organs and tumors of dogs with spontaneous tumors. Uptake was compared with relative blood flow as measured by the distribution of C-14 iodoantipyrine. Initial distribution of thymidine in normal mice measured 20 sec after injection, correlated with the relative perfusion measurements; however, all measurements of thymidine uptake made between 1 and 60 min after injection showed no correlation with perfusion. This indicates that the distribution more than 1 min after injection is primarily dependent on subsequent redistribution and/or metabolism of thymidine. A time-course study demonstrated that normal mouse organs with high rates of proliferation retained all the labeled thymidine initially taken up. Organs with low rates of proliferation lost their label in a nearly exponential washout. These studies provide further evidence of the feasibility of using C-11 thymidine for positron emission tomography (PET).

Positron imaging feasibility studies. I: Characteristics of [3H]thymidine uptake in rodent and canine neoplasms: concise communication.

Uptake [3H] thymidine was studied in BALB/c mice with EMT-6 sarcoma, in Buffalo rats with Morris 7777 hepatoma, and in nine dogs with spontaneous neoplasms: four lymphomas, two osteosarcomas, two soft-tissue sarcomas, and a thyroid carcinoma. High tumor-to-tissue ratios were observed for all tumor types assayed, and absolute uptakes, when computed as percent dose per gram tumor normalized for body weight, were similar for transplanted and spontaneous tumors. In the rodent tumors, radiothymidine was retained for at least 3 hr in the tumor without appreciable loss. In canine neoplasms, although the highest uptakes were observed in cellular tumors with many mitotic figures, tumor uptake showed significant variability that did not correlate with any obvious histologic change, and thus may reflect true biologic differences in metabolism among tumors at different sites in the same animal. These studies provide additional experimental evidence that the ratios of neoplastic to normal tissue and the kinetics of thymidine uptake by tumors are suitable for positron emission tomography of neoplasms in small and large animals, including both transplanted and spontaneous tumors.