Synthesis of 4-(5-[18F]fluoromethyl-3-phenylisoxazol-4-yl) benzenesulfonamide, a new [18F]fluorinated analogue of valdecoxib, as a potential radiotracer for imaging cyclooxygenase-2 with positron emission tomography

Article abstract of DOI:10.1016/j.bmcl.2005.07.065

Fluoroalkyl and fluoroaryl analogues of valdecoxib were found to possess potent inhibitory activities against cyclooxygenase-2 comparable to that of the parent valdecoxib. Among them, the fluoromethyl analogue was chosen for ¹⁸F-labeling. Thus,

Full text of DOI:10.1016/j.bmcl.2005.07.065

Bioorganic & Medicinal Chemistry Letters 15 (2005) 4699–4702
Synthesis of 4-(5-[18F]fluoromethyl-3-phenylisoxazol-4-yl)-
benzenesulfonamide, a new [18F]fluorinated analogue of
valdecoxib, as a potential radiotracer for imaging
cyclooxygenase-2 with positron emission tomography
Tatsushi Toyokuni,a,* J. S. Dileep Kumar,a Joseph C. Walsh,a
Alan Shapiro,a John J. Talley,b Michael E. Phelps,a Harvey R. Herschman,a
Jorge R. Barrioa and Nagichettiar Satyamurthya
aCrump Institute for Molecular Imaging and LA Tech Center, Department of Molecular and Medical Pharmacology,
David Geffen School of Medicine at University of California, Los Angeles, CA 90095, USA
bSearle Research and Development, St. Louis, MO 63198, USA
Received 18 June 2005; revised 27 July 2005; accepted 27 July 2005
Available online 8 September 2005
Abstract—Fluoroalkyl and fluoroaryl analogues of valdecoxib were found to possess potent inhibitory activities against cyclooxy-
genase-2 comparable to that of the parent valdecoxib. Among them, the fluoromethyl analogue was chosen for 18F-labeling. Thus,
4-(5-[18F]fluoromethyl-3-phenylisoxazol-4-yl)benzenesulfonamide (ꢀ2000 Ci/mmol at end of synthesis) was synthesized by [18F]fluo-
ride-ion displacement of the corresponding tosylate in ꢀ40% decay-corrected radiochemical yield within ꢀ120 min from end of
bombardment.
ꢀ 2005 Elsevier Ltd. All rights reserved.
Cyclooxygenase (COX) is the enzyme that catalyzes the
first step in the biotransformation of arachidonic acid
to prostanoids.1 COX exists as two distinct isoforms,
of which COX-1 is constitutively expressed in healthy
tissues mediating physiological responses, while COX-
2 is the inducible form responsible for the synthesis of
prostanoids involved in acute and chronic inflammatory
states. Inflammation is a common biological process
shared by many diseases. Indeed, elevated expression
of COX-2 has been implicated in many pathological
events, including rheumatoid arthritis, cancer, heart dis-
ease, stroke, and neurodegenerative disorders.1b,c,2
However, COX-2 is also found constitutively in some
tissues where it is considered to play a physiologically
important role.1b,c
from the disruption of COX-1. Since the discovery of
COX-2 in 1991,1a a rapid progress has been made in
the development of COX-2-selective inhibitors.3 In
1999, celecoxib (Celebrex) and rofecoxib (Vioxx) were
introduced to the market as non-ulcerogenic anti-
inflammatory drugs. Subsequently, in 2001, a second
generation inhibitor valdecoxib (Bextra) received the
US Food and Drug Administration (FDA) approval.
However, clinical trial studies with these drugs have
raised concerns about their potential cardiovascular
hazards.4 As a result, Vioxx and Bextra have recently
been withdrawn from the worldwide market and a
black-box warning is being required for Celebrex.5
To fully understand the role of COX-2 both in health
and disease, it would be of great benefit if we could mon-
itor Ôreal-timeÕ COX-2 expression in vivo, non-invasively
and repeatedly over time.6 COX-2-selective inhibitors
labeled with short-lived positron emitters are therefore
attractive molecules that could be used to image COX-
2 in living subjects with positron emission tomography
(PET), which is a powerful non-invasive, in vivo molec-
ular imaging technique currently available for biomedi-
The therapeutic effect of classical non-steroidal anti-in-
flammatory drugs (NSAIDs) is attributed to the inhibi-
tion of COX-2, whereas the undesired side effects arise
Keywords: PET; Radiotracer; Fluorine-18; COX-2; Inhibitor;
Valdecoxib.
*
Corresponding author. E-mail: ttoyokuni@mednet.ucla.edu
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.07.065

4700
T. Toyokuni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4699–4702
cal use.7 The [18F]fluorinated or [11C]methylated ana-
logues of celecoxib,8,9 rofecoxib,10 and their prototype
DuP-69711 have recently been synthesized. Herein, we
report the synthesis of a new [18F]fluorinated analogue
of valdecoxib 1-18F (Fig. 1) as a potential PET imaging
probe for COX-2.12
Table 1. Inhibition of COX-2 by fluorinated analogues of valdecoxib
Compound
IC50 (lM)a
Macrophage COX-2 assayd
Enzyme assayb
COX-2
S.I.c
1
0.002
0.008
>50,000
>12,500
0.002
0.002
0.002
0.002
2
e
e
3
—
e
—
—
Being bioisosteric with hydrogen, albeit having high
electronegativity,13 18F is often used to replace hydro-
gen of otherwise non-fluorinated molecules with
minimum effects on biomolecular interactions. Non-ra-
dioactive fluorinated analogues of valdecoxib were first
synthesized and evaluated for their COX-2 inhibitory
activities using human recombinant COX-1 and
COX-2 enzymes,14 and/or endotoxin-treated RAW
264.7 macrophages.15 The fluoroalkyl analogues 116
and 217 were synthesized by direct fluorination of the
corresponding hydroxy analogues 518 and 6,19 respec-
tively, with DAST (Scheme 1). The fluoroaryl ana-
e
4
Valdecoxib
—
0.005
e
>20,000
—
a Values are means of two experiments.
b Enzyme assays were performed against human recombinant COX-1
and COX-2 enzymes, as reported in Ref. 14.
c In vitro COX-2-selectivity index: COX-1 IC50/COX-2 IC50
.
d Cell-based assays were performed for prostaglandin E2 production as
a function of COX-2 inhibition using endotoxin-treated murine
RAW 264.7 macrophages, as reported in Ref. 15.
e Not determined.
by O-tosylation with Ts2O (!8) (Scheme 2). Radiosyn-
thesis of 1-18F was then effected by two consecutive reac-
tions in one pot, namely [18F]fluoride-for-tosylate
substitution of 8, followed by acidic deprotection.23
The radioactive product was purified by HPLC and
reconstituted in EtOH (1 mL).24 The specific activity
was ꢀ2000 Ci/mmol at end of synthesis (EOS). In a typ-
ical radiosynthesis, starting from ꢀ500 mCi of [18F]fluo-
ride, ꢀ80 mCi of chemically and radiochemically pure
1-18F was obtained in an injection-ready form within
120 min after end of bombardment (EOB). The decay-
corrected radiochemical yield was ꢀ40%. It was ob-
served that 1-18F experienced autoradiolysis in saline
containing 10% EtOH (ꢀ5% de[18F]fluorination after
6 h).25 The autoradiolysis was however effectively sup-
pressed in 100% EtOH (less than 0.5% de[18F]fluorina-
tion after 6 h).
logues
3 and 4 were synthesized, as previously
reported.20 As shown in Table 1, the fluoroalkyl and
fluoroaryl analogues were all potent COX-2 inhibitors.
Moreover, the fluoromethyl analogue 1 showed an
even higher COX-2-selectivity index than valdecoxib.
Therefore, the radiosynthesis of the [18F]fluoromethyl
analogue 1-18F was investigated.
The [18F]fluorination precursor 821 was synthesized from
5 via selective protection of the sulfonamide group with
a 4,40-dimethoxytrityl (DMTr) group (!722), followed
N
N
O
O
O
O
O
O
CH2R
CH2F
S
S
In vivo kinetics of 1-18F was preliminary evaluated using
a normal mouse with microPET.26 Rapid in vivo
de[18F]fluorination was evidenced by the conspicuous
emergence of the bony skeleton, which is characteristic
of [18F]fluoride peripheral formation with rapid uptake
in bone.27 A similar peripheral metabolism was also ob-
served in a vervet monkey, although it was somewhat
slower.28 It is likely that 1-18F is metabolized in vivo
NH2
NH2
Valdecoxib R = H
2
1
1-18
R = 19F
R = 18F
F
F
N
O
N
O
F
O
O
CH3
CH3
S
S
O
O
DMTr-Cl
Et3N, CH2Cl2-THF
overnight
NH2
NH2
Ts2O, Et3N, CH2Cl2
0 ˚C, 30 min
N
3
4
O
5
Figure 1. Valdecoxib and its [18F/19F]fluorinated analogues 1–4.
45%
85%
O
O
OH
S
NHDMTr
7
n = 1
1
N
DAST, THF 58%
-78˚C, 24 h
O
1. K18F, Kryptofix 222
MeCN, 90 ˚C, 5 min
N
O
O
1-18F
n = 2
33%
OH
n
O
S
2
2. 40% CCl3CO2H in MeCN
90 ˚C, 5 min
NH2
O
O
OTs
S
5 n = 1
6 n = 2
NHDMTr
~40% (decay corrected)
8
Scheme 1. Synthesis of non-radioactive fluoroalkylvaldecoxibs
and 2.
1
Scheme 2. Synthesis of [18F]fluoromethylvaldecoxib (1-18F).

T. Toyokuni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4699–4702
4701
similar to the parent valdecoxib, involving oxidative
Acknowledgments
hydrogen abstraction at the methyl group presumably
via the P450 enzyme-catalyzed oxidation process.29
The [18F]fluorohydroxymethyl group thus formed is
chemically labile leading to spontaneous de[18F]fluorina-
tion. It has been noted with drugs and other PET probes
that the rate of peripheral metabolism increases in the
order of mice > monkeys > humans.30 Indeed, the
metabolism of valdecoxib is reported to be rapid in
rodents but significantly slower in humans.29 Therefore,
it would be expected that 1-18F would undergo a much
slower metabolic de[18F]fluorination in humans. The
dosimetry data for a 70-kg adult were estimated from
the residence times determined with the monkey
whole-body biodistribution data using the MIRDOSE
program (Table 2).31 For a 10-mCi injection of 1-18F,
the highest absorbed dose found in the urinary bladder
wall was well below the dose limitation of 5 rad/mCi,
warranting the safe use of 1-18F in human PET imaging
studies.
We thank the UCLA Biomedical Cyclotron staff, the
UCLA microPET Imaging Facility staff, and the UCLA
LA Tech Center Chemistry staff for their support. This
research was supported in part by the Office of Science
(BER), US Department of Energy, Cooperative Agree-
ment No. DE-FC03-02ER63420, the National Institutes
of Health (CA86306), and University of California Bio-
technology STAR Grant 01-10184.
References and notes
1. (a) Herschman, H. R. Biochim. Biophys. Acta 1996, 1299,
125; (b) Turini, M. E.; DuBois, R. N. Annu. Rev. Med.
2002, 53, 35; (c) Simmons, D. L.; Botting, R. M.; Hla, T.
Pharmacol. Rev. 2004, 56, 387.
2. Critical roles of COX-2 in the pathogenesis of ParkinsonÕs
disease and tumor angiogenesis have recently been iden-
tified. See: (a) Teismann, P.; Tieu, K.; Choi, D.; Wu, D.;
Naini, A.; Hunot, S.; Vila, M.; Jackson-Lewis, V.;
Przedborski, S. Proc. Natl. Acad. Sci. U.S.A. 2003, 100,
5473; (b) Hunot, S.; Vila, M.; Teismann, P.; Davis, R. J.;
Hirsch, E. C.; Przedborski, S.; Rakic, P.; Flavell, R. A.
Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 665; (c) Chang, S.;
Liu, C. H.; Conway, R.; Han, D. K.; Nithipatikom, K.;
Trifan, O. C.; Lane, T. F.; Hla, T. Proc. Natl. Acad. Sci.
U.S.A. 2004, 101, 591.
In summary, the [18F]fluoromethyl analogue of valdec-
oxib 1-18F has been synthesized. The rapid metabolic
de[18F]fluorination in mice hinders the determination
of in vivo binding of 1-18F to COX-2 in experimental
rodent models, limiting our ability to investigate 1-18F
in detail. However, it should be mentioned that our pre-
liminary human imaging results have shown the poten-
tial usefulness of 1-18F in human studies, because only
minimum peripheral de[18F]fluorination occurred after
60 min. In addition to PET imaging determination in
humans, syntheses of radiofluorinated 2–4 are currently
underway.
3. Talley, J. J. Prog. Med. Chem. 1999, 36, 201.
4. (a) Finckh, A.; Aronson, M. D. Ann. Intern. Med. 2005,
´
142, 212; (b) Dogne, J.-M.; Supuran, C. T.; Pratico, D.
J. Med. Chem. 2005, 48, 2251.
5. A black-box warning is the most serious type of warning
in prescription drug labeling.
6. Herschman, H. R.; Talley, J. J.; DuBois, R. Mol. Imaging
Biol. 2003, 5, 286.
7. (a) Phelps, M. E. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
9226; (b) Fowler, J. S.; Wolf, A. P. Acc. Chem. Res. 1997,
30, 181.
Table 2. Dosimetry estimation for 1-18F for a 70-kg adult
Target organ
Total dose (rad/mCi)
8. (a) McCarthy, T. J.; Sheriff, A. U.; Graneto, M. J.;
Talley, J. J.; Welch, M. J. J. Nucl. Med. 2002, 43, 117;
(b) Kumar, J. S. D.; Prabhakaran, J.; Parsey, R. V.;
Underwood, M. D.; Simpson, N. R.; Majo, V. J.; van
Heertum, R.; Mann, J. J. Abstract of Papers, 50th
Annual Meeting of the Society of Nuclear Medicine,
New Orleans LA; Society of Nuclear medicine: VA,
2003; Abstract 1127; (c) Kumar, J. S. D.; Prabhakaran,
J.; Underwood, M. D.; Parsey, R. V.; Arango, V.; Majo,
V. J.; Simpson, N. R.; Arcement, J.; Cooper, A. R.; van
Heertum, R. L.; Mann, J. J. Abstract of Papers, 226th
National Meeting of the American Chemical Society,
New York, NY; American Chemical Society: Washing-
ton, DC, 2003; Abstract MEDI 375.
Adrenals
Brain
Breasts
4.2 · 10ꢁ3
3.1 · 10ꢁ5
1.1 · 10ꢁ2
8.5 · 10ꢁ3
5.0 · 10ꢁ3
6.4 · 10ꢁ3
2.6 · 10ꢁ3
5.0 · 10ꢁ2
1.8 · 10ꢁ2
1.9 · 10ꢁ2
3.4 · 10ꢁ2
2.3 · 10ꢁ3
1.8 · 10ꢁ3
6.4 · 10ꢁ3
3.8 · 10ꢁ3
2.2 · 10ꢁ3
1.2 · 10ꢁ3
8.9 · 10ꢁ4
1.6 · 10ꢁ3
2.8 · 10ꢁ3
2.0 · 10ꢁ3
2.0 · 10ꢁ4
1.4 · 10ꢁ1
1.0 · 10ꢁ2
3.2 · 10ꢁ3
1.3 · 10ꢁ2
Gallbladder wall
Lower large intestine wall
Small intestine
Stomach
Upper large intestine wall
Heart wall
Kidneys
Liver
Lungs
Muscle
Ovaries
Pancreas
9. The [123I]iodinated analogue of celecoxib has also been
synthesized for use with single photon emission computed
tomography (SPECT). See: Kabalka, G. W.; Mereddy, A.
R.; Schuller, H. M. J. Labelled Compd. Radiopharm. 2005,
48, 295.
Red marrow
Bone surfaces
Skin
Spleen
Testes
10. Wust, F. R.; Ho¨hne, A.; Metz, P. Org. Biomol. Chem.
¨
2005, 3, 503.
Thymus
Thyroid
11. de Vries, E. F. J.; van Waarde, A.; Buursma, A. R.;
Vaalburg, W. J. Nucl. Med. 2003, 44, 1700.
Urinary bladder wall
Uterus
12. The [11C]methylated analogue of the valdecoxib metabo-
lite 5 has been reported. See: Kumar, J. S. D.; Prabhak-
aran, J.; Underwood, M. D.; Simpson, N. R.; Parsey, R.
V.; Majo, V. J.; Arcement, J.; Arango, V.; van Heertum,
Total body
Effective dose

4702
T. Toyokuni et al. / Bioorg. Med. Chem. Lett. 15 (2005) 4699–4702
R.; Mann, J. J. Abstract of Papers, 50th Annual Meeting
(20:25:55 v/v/v) at 5 mL/min. The elution was monitored
using a UV detector (254 nm) and a c detector. The
chemically and radiochemically pure fraction (ꢀ10 mL; tR
ꢀ25 min) was collected, diluted with H2O (60 mL), and
passed through a Waters Sep-Pak C-18 cartridge (1 mL).
After washing the cartridge with H2O (20 mL), the
product was eluted with EtOH (1 mL) to a sterile vial
through a Millipore Millex-GS filter (0.22 lm) to give
1-18F as a 1-mL EtOH solution (ꢀ80 mCi). The chemical
and radiochemical purity, as well as chemical identity, was
confirmed using analytical HPLC (Phenomenex Aqua
C-18, 5 lm, 250 · 4.6 mm) with MeCN–H2O (1:1 v/v)
containing 0.01% TFA at 1 mL/min by reference to the
non-radioactive 1 (tR ꢀ 10 min). The radiochemical purity
was ascertained further by radio-TLC on silica gel (Rf
0.45; hexanes–EtOAc 1:1 v/v). The total synthesis time was
ꢀ120 min from EOB and the decay-corrected radiochem-
ical yield was ꢀ40%. The specific activity was determined
to be ꢀ2000 Ci/mmol at EOS based on the UV absorption
and concentration standard curve.
of the Society of Nuclear Medicine, New Orleans, LA;
Society of Nuclear Medicine: VA, 2003; Abstract 1110.
13. OÕHagan, D.; Rzepa, H. S. Chem. Commun. 1997, 645.
14. Gierse, J. K.; Hauser, S. D.; Creeley, D. P.; Koboldt, C.;
Rangwala, S. H.; Isakson, P. C.; Seibert, K. Biochem. J.
1995, 305, 479.
15. (a) Reddy, S. T.; Herschman, H. R. J. Biol. Chem. 1994,
269, 15473; (b) Wadleigh, D. J.; Reddy, S. T.; Kopp, E.;
Ghosh, S.; Herschman, H. R. J. Biol. Chem. 2000, 275,
6259.
16. 1H NMR (360 MHz; CD3OD, TMS) d 5.45 (2H, d,
J = 47 Hz, CH2F), 7.36–7.48 (7H, m) and 7.89–7.95 (2H,
m) (Ar). HRMS (EI): calcd for C16H13FN2O3S ([M]+)
332.0631, found 332.0624.
17. 1H NMR (360 MHz; CD3OD, TMS) d 3.23 (2H, dt,
J = 25, 5.8 Hz, CH2CH2F), 4.74 (2H, dt, J = 47, 5.8 Hz,
CH2CH2F), 7.32–7.44 (7H, m) and 7.89–7.93 (2H, m)
(Ar). HRMS (EI): calcd for C17H15FN2O3S ([M]+)
346.0780, found 346.0780.
18. Talley, J. J.; Brown, D. L.; Carter, J. S.; Graneto, M. J.;
Koboldt, C. M.; Masferrer, J. L.; Perkins, W. E.; Roger,
R. S.; Shaffer, A. F.; Zhang, Y. Y.; Zweifel, B. S.; Seibeert,
K. J. Med. Chem. 2000, 43, 775.
19. Talley, J. J. U.S. Patent 5,859,257, 1999; Chem. Abstr.
1999, 130, 110269.
24. The product was kept in EtOH until ready for use. For
in vivo applications, the solution was diluted with normal
saline such that the alcohol concentration was <5%.
25. De[18F]fluorination was monitored by radio-TLC on silica
gel, where [18F]fluoride stays at the origin. The non-
radioactive 1 was stable in PBS (pH 7.4).
20. Kumar, J. S. D.; Ho, M. M.; Leung, J. M.; Toyokuni, T.
Adv. Synth. Catal. 2002, 344, 1146.
26. A PET scanner for small animals. See: Chatziioannou,
A. F. Eur. J. Nucl. Med. Mol. Imaging 2002, 29, 98.
27. The mouse was anesthetized by xylazine (10 mg/kg) with
ketamine (200 mg/kg) and placed in the microPET scanner
(Concorde Microsystems Inc., Knoxville, TN). The mouse
was then injected via tail vein with 1-18F (200–300 lCi in
200 lL saline containing <5% EtOH). Dynamic data
acquisition commenced in a list mode at the time of
injection to obtain a time-dependent organ distribution of
radioactivity (frame times and sequence: 10 · 90 s and
21 · 300 s). The data were acquired in a 3-D mode with an
axial span of ꢀ8 cm. The images were reconstructed using
well-characterized 3-D filtered back-projection methods.
The bony skeleton was visible in 15 min post-injection.
28. The monkey, fasted after 9 PM on the evening prior to the
PET study, was anesthetized with ketamine (15–20 mg/kg)
and maintained anesthetically by inhalant isofluorane
techniques (1–2%). The animal was placed supine in the
scanner bed. The i.v. injection of 1-18F (ꢀ10 mCi) was
followed by the data collection, as described in Ref. 27,
except that a series of bed positions was used to cover the
entire animal with each position (300-s frames) repeated
3–6 times. The bone structure became visible in 45 min
post-injection.
21. 1H NMR (360 MHz; CDCl3, TMS) d 2.43 (3H, s, PhMe),
3.75 (6H, s, OMe), 5.11 (2H, s, CH2), 5.80 (1H, br s, NH),
6.70 (4H, d, J = 7.2 Hz), 6.94 (2H, d, J = 7.2 Hz), 7.16–
7.45 (18H, m) and 7.79 (2H, d, J = 7.2 Hz) (Ar). HRMS
(FAB): calcd for C44H39N2O8S2 ([M+H]+) 787.2148,
found 787.2147.
22. 1H NMR (360 MHz; CDCl3, TMS) d 3.75 (6H, s, OMe),
4.75 (2H, s, CH2), 5.85 (1H, s, NH), 6.68 (4H, d,
J = 7.2 Hz), 7.05 (2H, d, J = 7.2 Hz) and 7.15–7.45 (16H,
m) (Ar). HRMS (EI): calcd for C37H32N2O6S ([M]+)
632.1981, found 632.1977.
23. No-carrier-added [18F]fluoride (ꢀ500 mCi; specific activi-
ty: >10,000 Ci/mmol) was produced by 18O(p,n)18F nucle-
ar reaction of 95% [18O]H2O. The radioactivity
(ꢀ500 mCi) was transferred to a glass reaction vessel
containing azacrown ether Kryptofix 222 (10 mg) and
K2CO3 (1.0 mg) in MeCN–H2O (25:1 v/v, 1 mL). After
removal of H2O at 110 ꢁC with a stream of N2, the residue
was dried further by the azeotropic distillation with
MeCN (1.0, 0.5 and 0.5 mL). A solution of 8 (5 mg) in
MeCN (1 mL) was added and the mixture was heated at
90 ꢁC for 5 min. A 40% solution of trichloroacetic acid in
MeCN (0.5 mL) was added and the mixture was kept at
90 ꢁC for an additional 5 min. After partial neutralization
with saturated aq NaHCO3 (0.7 mL), the mixture was
diluted with H2O (7.3 mL) and passed through a Waters
Sep-Pak C-18 cartridge (1 mL). The cartridge was washed
successively with H2O (4 mL) and with 50 mM aq
NH4OAc (2 · 4 mL). The crude product was subsequently
eluted with MeOH (1.8 mL). The MeOH eluate was
purified by HPLC (Phenomenex Aqua C-18, 5 lm,
250 · 10 mm) with THF–MeOH–10 mM aq NH4OAc
29. (a) Yuan, J. J.; Yang, D.-C.; Zhang, J. Y.; Bible, R., Jr.;
Karim, A.; Findlay, J. W. A. Drug Metab. Dispos. 2002,
30, 1013; (b) Zhang, J. Y.; Yuan, J. J.; Wang, Y.-F.; Bible,
R. H., Jr.; Breau, A. P. Drug Metab. Dispos. 2003, 31,
491–501.
30. Barrio, J. R.; Satyamurthy, N.; Huang, S.-C.; Keen, R. E.;
Nissenson, C. H. K.; Hoffman, J. M.; Ackermann, R. F.;
Bahn, M. M.; Mazziotta, J. C.; Phelps, M. E. J. Cerebral
Blood Flow Metab. 1989, 9, 830.
31. Stabin, M. G. J. Nucl. Med. 1996, 37, 538.