Synthesis and evaluation of 4-bromo-1-(3-[18F]fluoropropyl)-2- nitroimidazole with a low energy LUMO orbital designed as brain hypoxia-targeting imaging agent

Article abstract of DOI:10.1248/bpb.25.616

In order to develop new imaging markers for brain hypoxia, 4-bromo-1-(3-fluoropropyl)-2-nitroimidazole (4-BrFPN) was designed based on molecular orbital calculations, synthesized and labeled with fluorine-18 as a lipophilic nitroimidazole analog with a lower energy LUMO orbital than those for fluoromisonidazole (FMISO) and 1-(3-fluoropropyl)-2-nitroimidazole (FPN). In an in vitro radiosensitization study, the sensitizer enhancement ratio for 4-BrFPN was found to be 1.65 at a 1 mM concentration, in comparison to 1.81 for FMISO. The preparation of ¹⁸F-labeled 4-BrFPN (4-Br¹⁸FPN) was achieved by [¹⁸F]fluoride ion displacement reaction of the tosylate precursor, in a reasonable radiochemical yield (33%, not corrected for decay). Metabolites in tumor and muscle extracts from methylcholanthrene-induced fibrosarcoma mice, as well as the tissue distribution of 4-Br¹⁸FPN in normal rat, were studied. The initial uptake into rat brain of 4-Br ¹⁸FPN was significantly higher relative to ¹⁸F-labeled FMISO (¹⁸FMISO), followed by a rapid washout from the brain. The tumor uptake of 4-Br¹⁸FPN was somewhat enhanced compared to those obtained with ¹⁸FMISO and ¹⁸F-labeled FPN ( ¹⁸FPN), but with lower tumor localization than ¹⁸FMISO. Analyses of tumor and muscle extracts showed metabolites remaining base line on the radio-TLC plates, and they were produced to a greater extent in tumor than muscle. The use of two drugs which increase hypoxic cell fraction in tumor, hydralazine or nitro-L-arginine, produced a significant increase in tumor levels of 4-Br¹⁸FPN, suggestive of a hypoxic mechanism of accumulation. The results imply that lowering of the LUMO energy of a molecule alone is not sufficient to improve its biodistribution properties for better imaging of regions of hypoxia.

Full text of DOI:10.1248/bpb.25.616

616
Biol. Pharm. Bull. 25(5) 616—621 (2002)
Vol. 25, No. 5
Synthesis and Evaluation of 4-Bromo-1-(3-[18F]fluoropropyl)-2-
nitroimidazole with a Low Energy LUMO Orbital Designed
as Brain Hypoxia-Targeting Imaging Agent
Fumihiko YAMAMOTO,a Mizuho AOKI,b Yoshiya FURUSAWA,b Koichi ANDO,b Yasuo KUWABARA,c
c
a
,a
*
Kouji MASUDA, Shigeki SASAKI, and Minoru MAEDA
a Faculty of Pharmaceutical Sciences, Kyushu University; c Faculty of Medicine, Kyushu University; 3–1–1 Maidashi,
b
Higashi-ku, Fukuoka 812–8582, Japan: and Clinical Radiation Biology and International Space Radiation Laboratory,
National Institute of Radiological Sciences; 4–9–1 Anagawa, Inage-ku, Chiba 263–8555, Japan.
Received December 3, 2001; accepted January 24, 2002
In order to develop new imaging markers for brain hypoxia, 4-bromo-1-(3-fluoropropyl)-2-nitroimidazole
(4-BrFPN) was designed based on molecular orbital calculations, synthesized and labeled with fluorine-18 as a
lipophilic nitroimidazole analog with a lower energy LUMO orbital than those for fluoromisonidazole (FMISO)
and 1-(3-fluoropropyl)-2-nitroimidazole (FPN). In an in vitro radiosensitization study, the sensitizer enhancement
ratio for 4-BrFPN was found to be 1.65 at a 1 mM concentration, in comparison to 1.81 for FMISO. The prepara-
tion of 18F-labeled 4-BrFPN (4-Br18FPN) was achieved by [18F]fluoride ion displacement reaction of the tosylate
precursor, in a reasonable radiochemical yield (33%, not corrected for decay). Metabolites in tumor and muscle
extracts from methylcholanthrene-induced fibrosarcoma mice, as well as the tissue distribution of 4-Br18FPN in
normal rats, were studied. The initial uptake into rat brain of 4-Br18FPN was significantly higher relative to 18F-
labeled FMISO (18FMISO), followed by a rapid washout from the brain. The tumor uptake of 4-Br18FPN was
somewhat enhanced compared to those obtained with 18FMISO and 18F-labeled FPN (18FPN), but with lower
tumor localization than 18FMISO. Analyses of tumor and muscle extracts showed metabolites remaining base
line on the radio-TLC plates, and they were produced to a greater extent in tumor than muscle. The use of two
drugs which increase hypoxic cell fraction in tumor, hydralazine or nitro-L-arginine, produced a significant in-
crease in tumor levels of 4-Br18FPN, suggestive of a hypoxic mechanism of accumulation. The results imply that
lowering of the LUMO energy of a molecule alone is not sufficient to improve its biodistribution properties for
better imaging of regions of hypoxia.
Key words nitroimidazole; fluorine-18; brain tissue; tumor; biodistribution
2-Nitroimidazole derivatives typified by misonidazole of nitroreduction products. In a previous paper, we reported
(MISO), originally developed as hypoxic radiosensitizers for the synthesis and evaluation of the two lipophilic 2-nitroimi-
radiation therapy, have been shown to be susceptible to re- dazoles, 1-(3-[18F]fluoropropyl)-2-nitroimidazole (18FPN) and
duction trapping in regions of low oxygen tension. Although 1-(8-[18F]fluorooctyl)-2-nitroimidazole (18FON), as brain-tar-
the mechanism of nitroimidazole binding to hypoxic tissue is geted hypoxic imaging agents.11) The high lipophilicity of
still not fully understood, the nitro group is believed to un- 18FPN and 18FON resulted in the increased initial uptake into
dergo a one-electron reduction in viable cells to produce a normal rat brain, relative to 18FMISO, but both compounds
radical anion, while in hypoxic cells this intermediate is fur- had significantly lower tumor uptake and lower tumor-to-
ther reduced to a species which reacts with cellular compo- blood ratios than 18FMISO, suggestive of a poor trapping
nents and is hereby trapped within the cell. In normoxic con- mechanism within the tumor tissue. One of the primary rea-
ditions reoxidation rapidly takes place, and the compound sons for these disappointing results appeared to be the lack
eventually diffuses out of the cell.1—3) This is the why most of a sufficiently high reduction potential (electron affinity)
efforts to develop selective markers for imaging hypoxic tis- undergoing more rapid bioreduction of the nitro group.
sue have used 2-nitroimidazole derivatives.4)
Therefore, we have now designed 4-bromo-1-(3-fluoro-
18F-Labeled fluoromisonidazole (18FMISO) has shown po- propyl)-2-nitroimidazle (4-BrFPN), based on the molecular
tential in humans as an agent to map hypoxic tissue with orbital calculations, to improve the electron affinic property
positron emission tomography (PET).5—8) However, this of the 2-nitroimidazole ring, with sufficient lipophilicity to
agent has drawbacks due to its relatively low concentration penetrate the BBB to image cerebral hypoxia. In this paper
within hypoxic tissue and the necessity to wait a long time we describe the design, synthesis and radiolabeling with fluo-
(2—4 h) to achieve acceptable target-background ratios rine-18 of 4-BrFPN. Additionally, the in vivo biodistribution
for imaging. The target tissues have mostly been the myo- of this designed compound is also characterized in normal
cardium or tumors, with less attention centered on the brain. rats and in tumor-bearing mice to evaluate its potential use as
The utility of 18FMISO as a hypoxic imaging agent for brain a marker of hypoxic tissue in comparison with 18FMISO and
studies appears to be limited because of its low blood–brain 18FPN.
barrier (BBB) permeability.9,10)
We have been interested in the development of new PET RESULTS AND DISCUSSION
markers exhibiting more rapid localization in the hypoxic re-
gion in the brain with a greater BBB permeability than
Design Reduction potential, as well as lipophilicity, is
18FMISO, based on the same principle of metabolic binding considered important in the rational design of hypoxia-selec-
To whom correspondence should be addressed. e-mail: maeda@phar.kyushu-u.ac.jp
© 2002 Pharmaceutical Society of Japan

May 2002
617
tive nitroimidazoles.11) To predict electron affinity as a mea-
sure of the ease of electron gain of a compound, we calcu-
lated the orbital energies of the lowest unoccupied molecular
orbital (LUMO), with MOPAC98-PM3, for nitroimidazole
analogs with additional substituents on the nitroimidazole
ring and containing a fluoropropyl group on imidazole nitro-
gen to maintain lipophilic character for brain targeting. We
hypothesized that the decrease of ELUMO of a molecule would
result in increased capacity to facilitate metabolic reduction
and trapping under hypoxia. The molecular orbital calcula-
tions show that the introduction of a bromine atom at the 4-
position in the nitro-bearing ring results in a significant low-
ering of ELUMO, when compared with FPN, FMISO and
MISO, as shown in Table 1. Moreover, 4-BrFPN shows the
distribution of LUMO on the 2-nitroimidazole ring. There-
fore, 4-BrFPN was chosen as a radiosynthetic target of this
work to investigate whether the increased electron affinity
would lead to improved biodistribution property and/or en-
hanced hypoxic selectivity.
Table 1. Partition Coefficients, Radiosensitizing Activity, and the Orbital
Energy of the LUMO of 2-Nitroimidazole Derivatives
e)
log P
(octanol/water)a) (1 mM)
SERd)
ELUMO
(eV)
Compound
MISO
Ϫ0.39b)
1.44b)
Ϫ0.79
FMISO
FPN
Ϫ0.40c)
1.60
Ϫ0.96
Ϫ1.02
0.28c)
1.81c)
4-BrFPN
1.09
1.65
Ϫ1.18
a) An octanol/water partition coefficient was determined by shaking solutions of
compounds in 2.5 ml 1-octanol with 2.5 ml potassium phosphate buffer (pH 7.4) for
20 min, and the data are the average of three separate experiments. b) Calculated
from reported p values, ref. 19. c) Values were from ref. 11. d) SER was calculated
from the ratio of radiation doses for 10% survival with and without compound, by read-
ing the curve in Fig. 1 for each sensitizer. e) The energy of the lowest unoccupied
molecular orbital (ELUMO).
Synthesis and Lipophilicity Chart 1 shows a synthetic
approach to the required 4-BrFPN. The tosylate (1) was pre-
pared by coupling di-tosyl propylate with 2-nitroimidazole in
the presence of triethylamine in N,N-dimethylformamide
(DMF), as reported previously.11) Direct bromination of the
tosylate (1) with an excess amount of Br2 in dioxane resulted
1
in the formation of the 4-bromo tosylate (2). H-NMR and
a) Br2, dioxane, r.t., 24 h; b) n-Bu4NF, THF, r.t., 6 h or K18F/Kryptofix 2.2.2.
mass spectra of the predicted isolated product indicated that
one bromine was substituted on the imidazole ring, but not a
mixture of the positional isomers. The preparation of 4-
BrFPN (3) was accomplished by fluoride substitution on the
tosylate precursor (2) using tetrabutylammonium fluoride at
room temperature in 71% yield. The regioisomeric identifica-
tion of the 4-bromo isomer was based on the comparison of
chemical shifts of 13C-NMR spectra of 4-BrFPN (3) and
FPN.
CH3CN, 80 °C, 15 min
Chart 1
The log P value for 4-BrFPN was measured as an index of
lipophilicity by octanol/water extraction at pH 7.4, and is
given in Table 1, in which comparison of the log P values for
FMISO, FPN and MISO are also included. 4-BrFPN has a
higher partition coefficient than FMISO, FPN and MISO.
Thus, the bromine atom on the imidazole ring, borne by 4-
BrFPN, has made the compound more lipophilic, but this
falls well within the range (log Pϭ2.0Ϯ1.0) of lipophilicity
thought to be compatible with good BBB penetration by dif-
fusion.
Radiosensitization Hypoxic cell radiosensitizers are
generally compounds with high electron affinity, and a corre-
lation between electron affinities (one-electron reduction po-
tentials) and radiosensitizing effectiveness with hypoxic cells
has been reported to exist among nitroimidazoles, including
MISO.12—15) The radiosensitizing effect of 4-BrFPN on hy-
poxic Chinese hamster V79 cells was thus investigated, com-
paring it with those of FMISO and FPN. Since 4-BrFPN has
low solubility in H2O, all in vitro radiosensitization experi-
ments were carried out in minimal essential medium (MEM)
containing 1% dimethylsulfoxide (DMSO). The profiles of
the dose–survival curves of V79 cells observed after X-ray
irradiation with 4-BrFPN are shown in Fig. 1. The SER (sen-
sitizer enhancement ratio) for 4-BrFPN was found to be 1.65
at a 1 mM concentration, in comparison to 1.81 for FPN and
1.60 for FMISO (Table 1), indicating that there was no sig-
Fig. 1. Radiation Survival Curves of Hypoxic V79 Cells
Cells were incubated with 1 mM concentration of 4-BrFPN (᭹) or FMISO (᭝) imme-
diately before X-ray irradiation under hypoxic conditions at room temperature, respec-
tively. Each point represents the meanϮS.E. for 3 experiments. Hypoxic control (᭛).
nificant difference in in vitro radiosensitizing potentials for
the three nitroimidazoles.
Radiosynthesis Radiochemical synthesis was achieved
by nucleophilic displacement of a tosylate precursor with a
[18F]fluoride ion. The no-carrier-added [18F]fluoride activity
obtained from the 18O(p,n)18F reaction was converted to
K18F/Kryptofix 2.2.2., as described earlier.16) We found that
optimal reaction conditions were a 15 min reaction time at
80 °C in acetonitrile, using the tosylate (2) and a Kryptofix
2.2.2.-potassium [18F]fluoride complex. The overall synthetic
time, including HPLC purification, was 40 min. The isolated
radiochemical yield of 4-Br18FPN at the end of synthesis was
33% with more than 96% radiochemical purity. The specific

618
Vol. 25, No. 5
Table 2. Biodistribution of Radioactivity after 4-Br18FPN Injection in Table 3. Biodistribution of 4-Br18FPN in Fibrosarcoma-Bearing C3H
Normal Wistar Rats (nϭ3)
Mice
Uptake (%dose/g)
15 min 30 min
Uptake (%dose/g)a)
Tissue
5 min
Tissue
60 min
30 min (nϭ4)
60 min (nϭ3)
Blood
Liver
0.353Ϯ0.108 0.403Ϯ0.060 0.308Ϯ0.021 0.228Ϯ0.023
0.811Ϯ0.578 1.705Ϯ0.213 1.263Ϯ0.302 1.471Ϯ0.158
Blood
Lund
Liver
S. Intestine
Adrenals
Kidneys
Heart
Tumor
Muscle
Bone
3.115Ϯ0.157
7.483Ϯ1.200
13.429Ϯ1.737
10.436Ϯ1.141
5.607Ϯ0.975
8.000Ϯ0.940
2.881Ϯ0.318
2.423Ϯ0.230*
1.728Ϯ0.237
6.131Ϯ1.642
1.585Ϯ0.255
0.777Ϯ0.053
1.410Ϯ0.121
1.822Ϯ0.195
6.722Ϯ0.432
11.464Ϯ1.168
4.211Ϯ1.835
3.531Ϯ1.024
4.154Ϯ0.226
1.527Ϯ0.025
2.152Ϯ0.258*
1.048Ϯ0.094
9.226Ϯ2.003
0.822Ϯ0.025
1.194Ϯ0.214
2.050Ϯ0.070
Adrenals 0.746Ϯ0.398 0.929Ϯ0.100 0.649Ϯ0.148 0.474Ϯ0.030
Kidneys
Heart
Bone
Brain
0.762Ϯ0.339 1.385Ϯ0.298 0.896Ϯ0.196 0.956Ϯ0.099
0.401Ϯ0.157 0.481Ϯ0.054 0.379Ϯ0.040 0.262Ϯ0.025
0.076Ϯ0.050 0.133Ϯ0.017 0.145Ϯ0.040 0.315Ϯ0.034
0.415Ϯ0.152 0.394Ϯ0.019 0.285Ϯ0.031 0.166Ϯ0.021
MeansϮS.D.
Brain
Tumor/blood
Tumor/muscle
a) MeanϮS.D. t-test: * pϽ0.01 compared to muscle data.
jection, in C3H mice bearing transplanted tumors (methyl-
cholanthrene induced fibrosarcoma) as rodent tumor models,
was further investigated to assess the accumulation of the
tracers in tumor (Table 3). It is generally known that tumors
provide a strong reducing environment when compared to
normal tissue, and many solid tumors contain hypoxic
areas.1,19) The tumor cells were implanted in the right leg
muscle of C3H mouse and allowed to grow for 9—10 d.
Tumor load did not significantly influence the uptake of 4-
Br18FPN in normal organs, and the tumors showed significant
retention of the radioactivity. The concentration of radioac-
tivity in the tumor was 2.4% dose/g at 30 min and 2.1%
dose/g at 60 min, which are actually higher than those ob-
tained with 18FPN (1.5% dose/g at 30 min and 1.6% dose/g at
Fig. 2. Rat Brain Uptake of Three 18F-Fluorinated Nitroimidazoles at Dif-
ferent Times after i.v. Injection
Data are the average from three animals. Values for 18FMISO and 18FPN were from
our previous report: ref. 11.
activity was estimated to be around 3700 GBq/mmol at the 60 min)11) in the same tumor model. The tumor-to-muscle
end of the synthesis. and tumor-to-blood ratios showed an increase with time: the
In Vivo Biodistribution As shown in Table 2, tissue dis- uptake ratios of the time-to-blood and tumor-to-muscle were
tribution of 4-Br18FPN was determined in normal Wistar rats 1.19 and 2.05 with 4-Br18FPN at 60 min postinjection, al-
at 5, 15, 30 and 60 min following intravenous injection. At though with insufficient selectivity, which are significantly
5 min after injection, the blood concentration was already at high ratios compared with those obtained with 18F-FPN (0.79
low levels, showing fast transport of the tracer to tissue. The and 1.38).11) Blood samples, obtained from the heart of mice
extent of defluorination for 4-Br18FPN, indicated by bone ra- at 60 min after injection of 4-Br18FPN were examined to de-
dioactivity, was kept at a low level until 30 min, then slightly termine the distribution of radioactivity in the plasma and red
increased at 60 min postinjection.
blood cells. We found 75% of the activity in the plasma, of
A biological characteristic of 4-Br18FPN is that both the which 74% was distributed in the protein-free plasma frac-
liver and kidneys showed higher uptake than most other nor- tion, with no radioactivity remaining as the intact parent
mal tissues, as observed with 18FMISO and 18FPN,17,18) in- compound, as shown by radio-TLC analysis (data not
dicative of extensive metabolism in the liver and urinary shown).
excretion. Brain uptake as a function of time after injection
We evaluated the selectivity of 4-Br18FPN for hypoxic tis-
is depicted graphically in Fig. 2 with data of 18FMISO and sues using two pharmacological modulators, hydralazine and
18FPN, which were previously reported. The initial brain ra- nitro-L-arginine, which have been shown to increase the hy-
dioactivity of 4-Br18FPN was almost similar to 18FPN, but poxic cell fraction in tumors.20,21) The modulators were in-
significantly higher than 18FMISO, indicating rapid brain jected 40 min before tracer administration, and the mice were
penetration, and its maximal brain uptake occurred at less killed at 30 and 60 min after injection of 4-Br18FPN. As
than 5 min post injection. It should be noted that rapid shown in Fig. 3, nitro-L-arginine-treated animals showed sig-
washout of radioactivity from the brain was observed with no nificantly higher uptake of 4-Br18FPN, by 40% at 30 min
indication of accumulation. Such an uptake and washout pat- postinjection, compared with the control animals, indicating
tern of radioactivity observed in the normal brain thus ap- that the accumulated radioactivity of 4-Br18FPN in the tumor
peared to be a desirable characteristic for a potential agent was at least related to tumor oxygenation. The difference in
for imaging hypoxic tissue in the brain.
the experiments with hydralazine did not reach statistical sig-
The biodistribution of 4-Br18FPN at 30 and 60 min post in- nificance regarding tumor uptake.

May 2002
619
mentation is necessary. Differential distribution of radioac-
tivity on the radio-TLC between tumor and muscle extracts,
closely similar to that found with 4-Br18FPN, was also ob-
served with 18FMISO (data not shown).
CONCLUSION
With the aim of developing a new hypoxia-imaging agent
in the brain, we have designed (from a simple frontier molec-
ular orbital standpoint), synthesized and evaluated a new 2-
nitroimidazole analog bearing a bromine atom on the imida-
zole ring, 4-Br18FPN. 4-Br18FPN offered enhanced tumor-to-
blood and tumor-to-muscle ratios compared with non-bromi-
Fig. 3. Uptake of Radioactivity in Mice Fibrosarcoma at 30 and 60 min nated analogs 18FPN, showing similar radioactivity clearance
Postinjection of 4-Br18FPN, after Injection of Hydralazine or Nitro-L-argi-
nine as a Pharmacological Modulator of Tumor Hypoxia (nϭ4)
from the blood and muscle for these two analogs, presumably
as a result of favorable changes in the LUMO energy level. A
hypoxic mechanism of accumulation is likely to play a role in
the tumor uptake of 4-Br18FPN, as revealed by metabolic
analysis and pharmacological modulator studies. However,
when compared with 18FMISO with a higher energy LUMO
orbital, 4-Br18FPN showed lower tumor localization than
18FMISO, although the tumor-to-background ratios for these
two analogs do not differ substantially at 30—60 min post in-
jection. The results provide an implication that lowering of
the LUMO energy of a molecule alone is not sufficient to im-
prove biodistribution properties for better imaging of regions
of hypoxia. Based on the findings of this study, we are now
focusing attention on designing analogs with more selective
Fig. 4. Radio-TLC Chromatograms of Mouse Tumor and Muscle Extracts
uptake in hypoxic tissue and less non-reductive metabolism.
at 60 min Postinjection of 4-Br18FPN
MATERIALS AND METHODS
In Vivo Metabolic Study Accumulated radioactivity in
Unless otherwise stated, chemical reagents were obtained
from commercial sources and were used directly. All melting
points are uncorrected. H-NMR spectra were obtained on a
the tumor and muscle of mice at 60 min post injection of 4-
Br18FPN were evaluated by radio-TLC for the presence of
metabolites. Both tumor and muscle tissues were homoge-
nized in MeOH and then centrifuged. The pellet was then re-
suspended in MeOH and again centrifuged. The combined
supernatant was evaporated, the residue was dissolved with
MeOH, and they were passed through Amberlite IRA-410
for the removal of free [18F]fluoride ion, because the [18F]flu-
oride ion and other polar metabolites were not differentiable
on radio-TLC. Both samples were analyzed for metabolites
by radio-TLC using CHCl3–MeOH. The extraction of ra-
dioactivity from the tumor and muscle was approximately
60%, respectively.
Radio-TLC analysis revealed a major radioactive peak re-
maining at the origin, indicative of the formation of polar
metabolites, in addition to a radioactive species (Rf 0.7) co-
migrated with authentic 4-BrFPN and a radioactive species
with an Rf value of 0.4—0.1 in both extract samples (Fig. 4).
Additional TLC analysis showed the absence of a debromi-
nated metabolite, i.e., 18FPN, which was presumed to be pro-
duced as one of the metabolites. Radioactivity remaining at
the origin of the TLC plate accounted for 69% in tumor ex-
tracts, whereas that in muscle extracts comprised 38%. Based
on the solvent extraction procedure used here, free [18F]fluo-
ride ion did not contribute significantly to the origin radioac-
tivity on TLC. Thus, metabolites remaining base line on the
TLC plates, which were found to a greater extent in tumor
than muscle, may reflect bioreductive metabolism and bind-
ing within the tumor hypoxic tissue, although further experi-
1
JEOL GX-270 spectrometer (270 MHz), and the chemical
shifts are reported in parts per million downfield from
tetramethylsilane. 13C-NMR spectra were obtained on a Var-
ian Unity 500 spectrometer (125 MHz), and the chemical
shifts are reported in parts per million downfield from
tetramethylsilane. Infrared (IR) spectra were recorded with a
JASCO IR Report-100 spectrometer. Mass spectra were ob-
tained with a JEOL JMS DX-610 or SX-102 mass spectrom-
eter. Column chromatography was performed on Kieselgel
60 (70—230 mesh, Merck), and analytical TLC and radio-
TLC was carried out on Silica gel 60F 254 (Merck). In the
synthetic procedures, organic extracts were routinely dried
over anhydrous Na2SO4 and evaporated with a rotary evapo-
rator under reduced pressure. HPLC was done using a
TOSOH CCP & 8020 series HPLC system fitted with a
Whatman Partisil 5PAC (10ϫ100 mm) with monitoring of
the radioactivity as well as UV absorption (at 254 nm). The
radioactivity was also quantified with a Capintec radioiso-
tope calibrator (CRC-30). The identity of radiolabeled com-
pounds was supported by HPLC co-injection studies.
Fluorine-18 was produced from 8 or 16% enriched
[18O]H2O by the 18O(p,n)18F reaction, as described previ-
ously.22) Aminopolyether (Kryptofix 2.2.2.)-supported potas-
sium [18F]fluoride (K18F/Kryptofix) was prepared by the ad-
dition of K2CO3·1.5H2O (1.0 mg) and Kryptofix 2.2.2.
(3.0 mg) to irradiated water in a TPX (polymethylpentene)

620
Vol. 25, No. 5
vessel, and by subsequent removal of the water under a
Partition Coefficients The log P value was measured
stream of argon at 110 °C by azeotropic distillation with dry using a standard shake flask method. The sample was shaken
CH3CN. Radiochemical yields were expressed at the end-of- well with a mixture of 1-octanol (2.5 ml) and 0.05 M phos-
synthesis (not corrected for decay), relative to the amount of phate buffer (2.5 ml, pH 7.4) for 20 min at 25 °C, after which
the [18F]fluorinating agent measured as total radioactivity aliquots of both phases were taken for analysis by HPLC
present in the reaction vessel. All animal experiments were quantitation. The reported log P value represents the mean of
carried out in accordance with the regulations on animal ex- three experiments.
periments of the Faculty of Pharmaceutical Sciences, Kyushu
University.
Radiosensitization Radiosensitization studies were car-
ried out by employing Chinese hamster V79 cells, and the ra-
4-Bromo-1-N-(3-p-toluenesulfonyloxypropyl)-2-ni- diation used was 200 kVp X-rays. All experiments were car-
troimidazole (2) Under argon, to a solution of 1-N-(3-p- ried out in glass culture dishes, gassed at room temperature
toluenesulfonyloxypropyl)-2-nitroimidazole11) (1) (50.7 mg, with humidified nitrogen or air for oxic controls. Just before
0.156 mmol) in dry 1,4-dioxane (2 ml) was added dropwise X-irradiation, the culture medium was exchanged to a fresh
bromine (100 ml) using a syringe. The mixture was stirred medium, including 1% DMSO and 1 mM concentration of 4-
at room temperature for 24 h and evaporated to dryness for BrFPN or FMISO. The cells were irradiated at room temper-
the removal of excess bromine. The residue was chromato- ature under hypoxic conditions. Cell survival was determined
graphed on silica gel (EtOAc : Hexaneϭ1 : 4) to give 4- using a conventional colony-forming assay. After post irradi-
bromo-1-N-(3-p-toluenesulfonyloxypropyl)-2-nitroimidazole ation treatment, the test compound was removed by two
(2) (49.7 mg, 79%) as a pale yellow powder, mp 70—75 °C. washings with PBS buffer solution. Subsequently, cells that
1H-NMR (CDCl3) d (ppm): 7.798 (d, 2H, Jϭ8.3 Hz, Ar-H), were trypsinized and suspended in fresh medium were
7.398 (d, 2H, Jϭ7.9 Hz, Ar-H), 6.984 (s, 1H, NCBrϭCHN), counted, diluted, and plated on culture dishes. After incuba-
4.500 (t, 2H, Jϭ6.6 Hz, NCH2CH2CH2OTs), 4.054 (t, 2H, tion for 6 d, the colonies were then fixed with 10% formalin
Jϭ5.6 Hz, NCH2CH2CH2OTs), 2.481 (s, 3H, Ar-CH3), 2.243 PBS buffer solution, stained with 1% methylene blue, and
(tt, 2H, Jϭ6.9, 5.6 Hz, NCH2CH2CH2OTs). IR (Nujol): 3120, then counted; survival curves were then constructed. All ex-
1740, 1585, 1530 cmϪ1. FAB-MS m/z: 404 (Mϩ), 406 periments were performed at least in triplicate. Radiosensiti-
(Mϩ2). High resolution (HR)-MS (FAB-MS). Calcd for zation was measured as the mean of the SER calculated at a
C13H15BrSN3O5: 403.9916 (MHϩ). Found: 403.9944.
10% survival level.
4-Bromo-1-N-(3-fluoropropyl)-2-nitroimidazole (4-
Biodistribution in Normal Rats Male Wistar rats (7
BrFPN) (3) Under argon, to a solution of 4-bromo-1-N-(3- weeks old, 215—240 g) given standard laboratory food and
p-toluenesulfonyloxypropyl)-2-nitroimidazole (2) (49.0 mg, water ad libitum were used in this investigation. Aliquots of
0.121 mmol) in dry tetrahydrofuran (THF) (1 ml) was added 4-Br18FPN in about 200 ml of saline solution, with activities
dropwise a 1 M-THF solution of n-Bu4NF (0.50 ml, 0.50 ranging from 1.75—2.98 MBq, were injected through the tail
mmol as FϪ). The mixture was stirred at room temperature vein of unanesthetized rats. After the given time intervals,
for 6 h and evaporated to dryness. The residue was chroma- the animals were killed by cervical dislocation while under
tographed on silica gel (EtOAc : hexaneϭ1 : 1) to give 4- ether anesthesia. All samples of either the blood or the organ
bromo-1-N-(3-fluoropropyl)-2-nitroimidazole (3) (21.9 mg, of interest that had been blotted free of blood, were taken and
71%) as a pale yellow oil. 1H-NMR (CDCl3) d (ppm): 7.171 weighed. The radioactivity of all samples was measured with
(s, 1H, NCBrϭCHN), 4.592 (t, 2H, Jϭ6.9 Hz, NCH2CH2- an Aloka AutoWell Gamma System ARC-400 (corrected for
CH2F), 4.517 (dt, 2H, Jϭ46.9, 5.5 Hz, NCH2CH2CH2F), decay), and the percent injected dose per gram of tissue
2.284 (dtt, 2H, Jϭ27.1, 6.9, 5.5 Hz, NCH2CH2CH2F). 13C- weight (%dose/g) was then calculated for each type of tissue.
NMR (CDCl3) d (ppm): 125.9 (s, 5-C), 115.3 (s, 4-C), 79.9
Biodistribution in Tumor-Bearing Mice 3-Methyl-
(d, Jϭ167.0 Hz, NCH2CH2CH2F), 47.03 (s, 2-C), 47.00 (s, cholanthrene-induced fibrosarcoma (NFSa) was inoculated
NCH2CH2CH2F), 30.8 (d, Jϭ19.4 Hz, NCH2CH2CH2F). IR s.c. into the right hind leg muscle of female C3H/He mice (5
(Nujol): 3125, 2960, 1530, 1500, 1465, 1365 cmϪ1. FAB-MS weeks old, 18.0—21.0 g). Tumors which developed with a di-
m/z: 252 (Mϩ), 254 (Mϩ2). HR-MS (FAB-MS). Calcd for ameter of about 1 cm at 9—10 d after inoculation were
C6H8BrFN3O2: 251.9784 (MHϩ). Found: 251.9769.
used.11) The animals were allowed free access to water and
4-Bromo-1-N-(3-[18F]fluoropropyl)-2-nitroimidazole (4- food at all times. Aliquots of 4-Br18FPN in about 200—
Br18FPN) (4) 4-Bromo-1-N-(3-p-toluenesulfonyloxypropyl)- 300 ml of saline solution, with activities ranging from 0.47—
2-nitroimidazole (2) (1 mg, 2.47 mmol) dissolved in dry 2.83 MBq, were injected through the tail vein of unanes-
CH3CN (300 ml) was added to a TPX vessel containing the thetized mice. At 30 and 60 min post injection, the animals
K18F/Kryptofix 222 (136.5 MBq). The vessel was closed and were killed by cervical dislocation while under ether anesthe-
heated in an oil bath at 80 °C for 15 min, then cooled to room sia. Samples of blood and the tissues of interest, blotted free
temperature. The reaction mixture was filtered through Sep- of blood, were taken, weighed, and assayed for radioactivity
Pak Silica with EtOAc (3 ml). The filtrate was evaporated in in an Aloka Auto Well Gamma System ARC-400 (corrected
vacuo. The residue was dissolved in hexane : EtOAcϭ3 : 1 for decay). The results are expressed as the percent injected
(1 ml) and injected into the HPLC (Whatman Partisil 5PAC, dose per gram (%dose/g) of tissue weight, and tumor-to-
10ϫ100 mm, hexane : EtOAcϭ3 : 1, 2 ml/min). The radioac- blood and tumor-to-muscle concentration ratios were calcu-
tive fraction eluted at tRϭ5.5 min, corresponding to authentic lated from the %dose/g tissue data.
4-BrFPN, was collected. Total synthetic time was 40 min.
The radiochemical yield (not corrected for decay) was 33%. jection of 4-Br18FPN were placed in heparinized vials, then
No radiochemical impurities were detected by HPLC. centrifuged (2000 rpm, 10 min), and the plasma was re-
Blood samples obtained from the heart at 60 min post in-

May 2002
621
moved. The red blood cells (RBC) were washed with a small Kyushu University Medical School Cyclotron Facility for
amount of ice-cold saline and the wash was added to the their assistance in producing 18F. The support of this work by
plasma. Aliquots of the plasma solution and the total RBC a Grant-in-Aid for Scientific Research from the Ministry of
pellet were counted for 18F activity. An equal volume of Education, Science, Sports and Culture of Japan is gratefully
freshly prepared 1 M HClO4 (PCA) was added to the plasma, acknowledged.
and the combined mixture was centrifuged (2000 rpm, 5 min)
to remove precipitated proteins. The supernatant and one REFERENCES
PCA washing were combined. The plasma protein and the
protein free-plasma were also counted for 18F activity.
1) Chapman J. D., N. Engl. J. Med., 301, 1429—1432 (1979).
2) Franko A. J., Koch C. J., Garrecht B. M., Sharplin J., Hughes D., Can-
cer Res., 47, 5367—5376 (1987).
3) Edwards D. I., J. Antimicrob. Chemother., 31, 9—20 (1993).
Aliquots of protein free-plasma were spotted on TLC plates,
then developed with hexane : EtOAcϭ1 : 1. The TLC plate
content was counted with a radio-TLC analyzer. The location
of 4-BrFPN on the plate was determined by co-spotting with
an authentic cold sample.
4) Nunn A., Linder K., Strauss H. W., Eur. J. Nucl. Med., 22, 265—280
(1995).
5) Casciari J. J., Rasey J. S., Radiat. Res., 141, 28—36 (1995).
6) Jerabek P. A., Patrick T. B., Kilbourn M. R., Dischino D. D., Welch M.
J., Appl. Radiat. Isot., 37, 599—605 (1986).
Pharmacological modulator studies were performed using
hydralazine and nitro-L-arginine. Hydralazine (5 mg/kg, i.p.)
or nitro-L-arginine (10 mg/kg, i.v.) was injected at 40 min be-
fore tracer administration. Thirty and sixty minutes after 4-
Br18FPN injection, the mice were killed and the tumor and
muscle were prepared and counted as described above.
In Vivo Metabolic Study Aliquots of 4-Br18FPN in
200 ml of saline solution, with activities ranging from 2.19—
13.3 MBq, were injected through the tail vein of unanes-
thetized fibrosarcoma-bearing mice (17.6—19.5 g). At
60 min postinjection, the animals were killed by cervical dis-
location while under ether anesthesia. Tumor and muscle tis-
sue was taken, homogenized and extracted with methanol
under an ice-cooled condition, respectively. The respective
extract solution was centrifuged (3000 rpmϫ10 min), and the
supernatant was concentrated in vacuo and filtered through
Amberlite IRA-410 (ClϪ form, 30 mg) to remove free
[18F]fluoride. A control experiment using the resin was per-
formed with authentic [18F]fluoride ion to determine the re-
moval capability of [18F]fluoride ion, and to validate the pro-
cedure. Aliquots of the filtered solution were spotted on TLC
plates, then developed with chloroform : methanolϭ4 : 1. The
TLC plates were cut into 0.5 mm wide strips, and their ra-
dioactivities were counted with an Aloka Auto Well Gamma
System ARC-400. The location of 4-Br18FPN and 18FPN on
the plates was determined by co-spotting with the authentic
sample, respectively.
7) Prekeges J. L., Rasey J. S., Grunbaum Z., Kenneth H., Biochem. Phar-
macol., 42, 2387—2395 (1991).
8) Martin G. V., Cerqueira M. D., Caldwell J. H., Rasey J. S., Embree L.,
Krohn K. A., Circ. Res., 67, 240—244 (1990).
9) Mathias C. J., Welch M. J., Kilbourn M. R., Jerabek P. A., Patrick T.
B., Raichle M. E., Krohn K. A., Rasey J. S., Shaw D. W., Life Sci., 41,
199—206 (1987).
10) Liu R. S., Yeh S. H., Guo W. Y., Pan D. H. C., Lee L. S., Chu L. S.,
Chang C. P., Wang J. K., Wu L. C., J. Nucl. Med., 36, 53P (1995).
11) Yamamoto F., Oka H., Antoku S., Ichiya Y., Masuda K., Maeda M.,
Biol. Pharm. Bull., 22, 590—597 (1999).
12) Adams C. E., Clarke E. D., Flockhart I. R., Jacobs R. S., Sehmi D. S.,
Stratford I. J., Wardman P., Watts M. E., Parrick J., Wallace R. G.,
Smithen C. E., Int. J. Radiat. Biol., 35, 133—150 (1979).
13) Adams C. E., Clarke E. D., Gray P., Jacobs R. S., Stratford I. J., Ward-
man P., Watts M. E., Parrick J., Wallace R. G., Smithen C. E., Int. J.
Radiat. Biol., 35, 151—160 (1979).
14) Adams C. E., Flockhart I. R., Smithen C. E., Stratford I. J., Wardman
P., Watts M. E., Radiat. Res., 67, 9—20 (1976).
15) Workmann P., Brown J. M., Cancer Chemother. Pharmacol., 6, 39—49
(1981).
16) Yamamoto F., Sasaki S., Maeda M., Appl. Radiat. Isot., 43, 633—639
(1992).
17) Hoffman J. M., Rasey J. S., Spance A. M., Shaw D. W., Lrohn K. A.,
Stroke, 18, 168—176 (1987).
18) Van Os-Corby D. J., Chapman J. D., Int. J. Radiat. Oncol. Biol. Phys.,
12, 1251—1254 (1986).
19) Grunbaum Z., Freauff S. J., Krohn K. A., Wilbur D. S., Magee S.,
Rasey J. S., J. Nucl. Med., 28, 68—75 (1987).
20) Wood P. J., Stratford I. J., Sanson J. M., Cattanach B. M., Quinney R.
M., Adams G. E., Int. J. Radiat. Oncol. Biol. Phys., 22, 473—476
(1992).
21) Wood P. J., Stratford I. J., Adams G. E., Szabo C., Thiemermann C.,
Vane J. R., Biochem. Biophys. Res. Commun., 192, 505—510 (1993).
22) Maeda M., Fukumura T., Kojima M., Appl. Radiat. Isot., 38, 307—310
(1987).
Statistical Analysis All values of radioactivity in ani-
mals were expressed as the meanϮS.D. Lack of significance
using Student’s t-test was considered to be pϾ0.05.
Acknowledgments We are indebted to the staff of the