Development of 2-Deoxy-2-[18F]fluororibose for Positron Emission Tomography Imaging Liver Function in Vivo

Article abstract of DOI:10.1021/acs.jmedchem.5b00569

Life-threatening acute liver failure can be triggered by a variety of factors, including common drugs such as acetaminophen. Positron emission tomography (PET) is rarely used to monitor liver function, in part because of a lack of specific imaging agents for liver function. Here we report a new PET probe, 2-deoxy-2-[¹⁸F]fluororibose ([¹⁸F]-2-DFR), for use in imaging liver function. [¹⁸F]-2-DFR was synthesized and validated as a competitive substrate for the ribose salvage pathway. [¹⁸F]-2-DFR was prepared through an efficient late stage radiofluorination. The desired selectivity of fluorination was achieved using an unorthodox protecting group on the precursor, which could withstand harsh S_N2 reaction conditions with no side reactions. [¹⁸F]-2-DFR accumulated preferentially in the liver and was metabolized by the same enzymes as ribose. [¹⁸F]-2-DFR could distinguish between healthy liver and liver damaged by acetaminophen. [¹⁸F]-2-DFR is expected to be a useful PET probe for imaging and quantifying liver functions in vivo, with likely significant clinical utility.

Full text of DOI:10.1021/acs.jmedchem.5b00569

Article
pubs.acs.org/jmc
18
Development of 2‑Deoxy-2‑[ F]fluororibose for Positron Emission
Tomography Imaging Liver Function in Vivo
,
†,§,●
‡,●
§,∥
‡
⊥,#
Nikolai M. Evdokimov,*
Peter M. Clark, Graciela Flores, Timothy Chai, Kym F. Faull,
Michael E. Phelps,§,∥ Owen N. Witte,‡,§,∞,○ and Michael E. Jung*
,†,§
†
Department of Chemistry and Biochemistry, University of California, Los Angeles, California 90095, United States
‡
§
∥
Department of Microbiology, Immunology, and Molecular Genetics, Department of Molecular and Medical Pharmacology, Crump
⊥
#
Institute for Molecular Imaging, Psychiatry and Biobehavioral Sciences, Pasarow Mass Spectrometry Laboratory, Semel Institute
∞
for Neuroscience and Human Behavior, Eli and Edythe Broad Center of Regenerative Medicine and Stem Cell Research, and
○
Howard Hughes Medical Institute, David Geffen School of Medicine, University of California, Los Angeles, California 90095,
United States
*
S Supporting Information
ABSTRACT: Life-threatening acute liver failure can be
triggered by a variety of factors, including common drugs
such as acetaminophen. Positron emission tomography (PET)
is rarely used to monitor liver function, in part because of a
lack of specific imaging agents for liver function. Here we
report a new PET probe, 2-deoxy-2-[ F]fluororibose ([ F]-2-DFR), for use in imaging liver function. [ F]-2-DFR was
synthesized and validated as a competitive substrate for the ribose salvage pathway. [ F]-2-DFR was prepared through an
1
8
18
18
18
efficient late stage radiofluorination. The desired selectivity of fluorination was achieved using an unorthodox protecting group on
1
8
the precursor, which could withstand harsh S 2 reaction conditions with no side reactions. [ F]-2-DFR accumulated
N
18
preferentially in the liver and was metabolized by the same enzymes as ribose. [ F]-2-DFR could distinguish between healthy
liver and liver damaged by acetaminophen. [ F]-2-DFR is expected to be a useful PET probe for imaging and quantifying liver
1
8
functions in vivo, with likely significant clinical utility.
10
INTRODUCTION
sampling. Noninvasive, quantitative methods for imaging
local changes in specific biological functions of the liver are
lacking. Yet in many cases, quantifying the extent of liver
damage and injury remains a significant clinical challenge and
can lead to unnecessary and expensive clinical interventions.
This led to our interest in developing PET probes to image
■
Positron emission tomography (PET) is a clinically relevant
imaging modality for quantitatively measuring biochemical
1
,2
pathways in vivo. PET images the selective accumulation of
radiolabeled imaging probes in specific cell populations due to
unique metabolic or protein expression profiles targeted by the
probes. PET is used both preclinically and clinically to study in
vivo metabolism, to provide a molecular imaging diagnostic of
the biology of disease, to aid in drug discovery and
development, and to stratify patient populations for drug trials
11
specific biochemical pathways in the liver. An early study
3
showed that [ H]-ribose, injected into rats intravenously,
accumulates in hepatocytes, and recently it was demonstrated
14
that [ C]-ribose, injected into mice intravenously, accumulates
1
−4
and treatments.
probes have played a central role in the use of PET. The
Historically, carbohydrate-based PET
preferentially in the liver through the activity of the ribose
1
,2
12,13
salvage pathway.
This suggested that molecules that
1
8
fluorinated glucose analogue 2-deoxy-2-[ F]fluoroglucose
measure the activity of the ribose salvage pathway in vivo
could be scaffolds upon which to build clinically useful hepatic
PET probes.
1
8
(
[ F]-FDG) has for decades been used as a tool clinically to
investigate alterations in glycolysis in normal and diseased
states of the brain and heart and to detect, stage, and assess
therapeutic efficacy in cancer patients.
The liver is a vital organ responsible for a remarkable
plethora of biological functions, including detoxifying blood,
producing albumin, and maintaining blood glucose levels. A
variety of blood tests exists that provide systemic assessments of
The synthesis of PET probes is often difficult because of the
1
−6
18
short half-life of the F radionuclide (109.8 min) and a limited
set of available efficient and selective methods with which to
18
14
introduce F into molecules. In most cases, this requires that
18
F be installed at or near the last step in the synthesis in a
7
−9
reaction that produces no fluorinated side products. Any
subsequent deprotection step must be fast, high yielding, and
preferably devoid of side products.
liver function.
Imaging modalities such as ultrasound and
magnetic resonance imaging have been used but suffer from
limited resolution and specificity in the case of ultrasound. The
PET imaging probe [ F]-FDG does not accumulate
significantly in the liver. Liver biopsies are invasive, associated
8
1
8
2
Received: April 11, 2015
with a risk of complications, and suffer from limited tissue
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 1. Ribose salvage and metabolism pathways.
1
8
Scheme 1. Proposed Synthesis of [ F]-2-DFR (1)
1
8
We recently developed and evaluated an F fluorinated
stereocenter. 2-Deoxy-2-fluororibose (2-DFR) differs from
ribose only in the replacement of the 2-hydroxyl with a
fluorine atom and is stereochemically identical to ribose. We
pursued the synthesis and evaluation of [ F]-2-DFR with the
goal of developing a PET imaging probe that would be an
improvement over DFA and 5-deoxy-5-fluororibose in
simultaneously being able to measure ribose salvage activity
while being resistant to defluorination in vivo.
1
8
18
ribose derivative, 2-deoxy-2-[ F]fluoroarabinose ([ F]-DFA),
as a PET imaging probe to measure the ribose salvage
pathway. When injected intravenous into mice, [ F]-DFA
accumulates strongly and selectively in the liver. However,
upon further analysis, we identified that mice injected with
F]-DFA and imaged with PET showed significant accumu-
lation of radioactivity in bone tissue that increased over the
1
2
18
18
12
1
8
[
course of an hour (Figure S1, Supporting Information). Mice
14
injected with [ C]-ribose and imaged by autoradiography after
an hour of accumulation failed to show any significant signal
enrichment in bone tissue (Figure S1, Supporting Information).
This strongly suggests that [ F]-DFA is defluorinated in vivo.
The undesired presence of radioactivity in bone tissue would
mask any specific accumulation of the probe in the bone
marrow, and hepatic probe defluorination could complicate
image quantification, particularly in the liver. There is a need
for an alternative PET imaging probe for measuring the ribose
salvage pathway. Recently the PET imaging probe 5-deoxy-5-
RESULTS AND DISCUSSION
■
Chemical Synthesis. Although modern synthetic chemistry
1
8
18
offers a range of methods to introduce F into mole-
1
4,18−20
cules,
undoubtedly the stereoselective labeling of an
aliphatic carbon in a carbohydrate is best performed using a
rapid S 2 reaction on a carefully designed precursor. We
N
18
envisioned achieving the synthesis of [ F]-2-DFR (1) by
performing an S 2 reaction on a protected arabinose bearing a
N
15,16
good leaving group, such as triflate, at C-2 (Scheme 1).
This seemingly trivial synthetic task was complicated by the
fact that the C-3 hydroxyl protecting group hinders the C-2
position from attack by the incoming fluoride. A more serious
concern was possible anchimeric assistance from ether or ester
protecting groups at C-3 that would stabilize a carbocation at
C-2 and possibly even lead to rearrangements. The same would
be true for protecting groups on the C-1 hydroxyl. After
exploring several synthetic avenues, we realized that an electron
deficient protecting group had to be installed on the hydroxyl
group at C-3.
1
8
[
F]fluororibose has been described.
5-Deoxy-5-fluoror-
ibose replaces the 5-hydroxyl of ribose with a fluorine, a
modification that blocks phosphorylation of this probe by the
ribose kinase ribokinase, the first enzyme in the ribose salvage
The inability of ribokinase to
phosphorylate 5-deoxy-5-fluororibose limits the utility of 5-
deoxy-5-[ F]fluororibose as a probe to measure the ribose
salvage pathway in vivo. Thus, we decided to pursue the
synthesis of a new ribose-based PET probe.
DFA differs from ribose both in the replacement of the 2-
hydroxyl with a fluorine atom and by epimerization at this
1
7
pathway (Figure 1).
18
B
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Scheme 2. Possible Mechanism for Side Product Formation
a
Scheme 3. Synthesis of Precursor 8 for Radiofluorination
a
(
a) (i) cat. I , acetone, rt, 4 h, (ii) NaOMe, MeOH, 50 °C, 30 min, 83% (2 steps); (b) 4, Ag O, CH Cl , 40 °C, 20 h, 87%; (c) (i) cat. TfOH, Ac O,
2 2 2 2 2
AcOH, rt, 3 h, (ii) cat. I , i-PrOH, THF, 67 °C, 8 h, (iii) NaOMe, MeOH, rt, 15 min, 64% (3 steps); (d) (COCl) , DMSO, Et N, CH Cl , −78 °C to
2
2
3
2
2
rt, 40 min, (ii) NaBH , H O, EtOH, 0 °C, 15 min, 87% (two steps); (e) Tf O, pyridine, CH Cl , 0 °C, 30 min, 96%.
4
2
2
2
2
1
8
Table 1. Optimization of the Synthesis of 1 and [ F]-1
1
8
number
reagent (equiv)
solvent
time (min)
temp (°C)
yield of 9 or [ F]-9 (%)
1
2
3
4
5
6
7
8
9
KF (1.0), K222 (1.0)
KF (1.0), K222 (1.0)
TBAF (1.0)
DMF
20
40
50
50
50
30
45
30
30
10
160
dec
5
DMF
120
t-BuOH
MeCN
MeCN
MeCN
MeCN
MeCN
MeCN
120
33
57
72
76
KF (1.0), K222 (1.0)
KF (1.0), K222 (1.0)
KF (1.0), K222 (1.0)
120
120, MW
a
154
1
8
a
b
K[ F], K222
154
84
1
8
a
,
bc
K[ F], K222
154
78 ± 5
1
8
d
K[ F], K222
154
34
1
8
a
b
10
K[ F], K222
MeCN
154
41
a
b
c
d
In house radiosynthesizer, sealed vessel, 10 mg of precursor. Decay corrected yield. n = 10. 5 mg of precursor.
2
2
Formation of C-1 fluoride (Scheme 2) was eliminated by
Woerpel group, while the NMR of compound C-1 fluorides
23
utilizing the pure β-anomeric precursor. Thus, the arabinofur-
anoside 8, with PNBn protecting groups at the C-3 and C-5
position, was synthesized (Scheme 3). This sequence involved
the two-step formation of the α-acetonide 3 from 2 followed by
shows a peak at −116 ppm.
Fluorination of the triflate 8 was examined under several
conditions (Table 1), and it was found that clean formation of
the protected fluororibose 9 was accomplished in 72% yield
after 50 min of heating in acetonitrile at 120 °C using
microwave irradiation and 76% yield after 30 min of heating in
acetonitrile under sealed conditions at 154 °C. The global
deprotection of 9 was achieved quantitatively by treatment with
Raney Ni at 130 °C in formic acid for 30 min, producing 2-
DFR (1) as the expected mixture of mostly the α,β-pyranose
21
protection of the diol as the bis(4-nitrobenzyl) ether 5. Next,
the 1,2-acetonide protection in 5 was converted to the 1,2-O-
diacetate, and selective formation of the 1-β-isopropoxy acetal,
with the help of the C-2 α-acetate group providing anchimeric
assistance, followed by saponification, finally gave the α-alcohol
6
. Inversion of the 2-alcohol via oxidation−reduction, and final
24
triflate formation afforded the substrate 8 in 40% overall yield.
The bulky β-isopropoxy group at the anomeric position was
chosen to increase the anti delivery of hydride during the
hydride reduction, allowing a more efficient overall inversion
anomers.
Radiofluorination was best performed at 154 °C in
acetonitrile for 30 min under sealed conditions using 10 mg
of triflate 8 and gave 78 ± 5% conversion with no detectable
side product, as determined by radio-TLC (Table 1, entry 8;
Figure 2a). Lower temperatures, shorter incubation times, and
lower amounts of precursor all decreased the fluorination
1
9
and higher yield of compound 7. The [ F] NMR spectrum of
the product 9 displayed the expected peak at −209 ppm for a
fluorine atom at C-2, which was previously observed by the
C
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
1
8
19
18
Figure 2. [ F]-2-DFR is an isotopologue of [ F]-2-DFR and is stable in vivo. (a) Radio-TLC of the radiolabeling reaction mixture. Red peak: [ F]-
. Green peak: 18F . (b) Radio-TLC of the deprotection of [ F]-9. Green peak: [ F]-1. (c) [ F]-2-DFR and (d) [ F]-2-DFR were analyzed by
high performance anion exchange chromatography connected to a pulsed amperometric detector and a γ detector. (e) [ F]-2-DFR was injected into
mice, and after 60 min, the mouse plasma was analyzed by an HPLC connected to a γ detector and compared with pure [ F]-2-DFR.
−
18
18
18
19
9
18
18
1
8
efficiency (Table 1). [ F]-9 was purified by a silica gel cartridge
amount of radioactivity in the bone was stable for at least 2 h
and subsequently deprotected with Raney Ni in formic acid in
(Figure 3a,c). Except for the difference in bone uptake, the
whole body biodistribution of [ F]-2-DFR is qualitatively and
18
7
0 ± 5% yield at 130 °C for 30 min, as determined by radio-
TLC (Figure 2b). The PET tracer 1 was produced in a total
synthetic time of 150 min (from [ F]-fluoride delivery to
formulation) with a final decay-corrected radiochemical yield of
quantitatively similar to the whole body biodistribution of
1
8
18
14
[ F]-DFA and [ C]-ribose (Figure S2, Supporting Informa-
12
tion). Whole body autoradiography of a mouse intravenously
injected with [ F]-2-DFR and imaged after 1 h of
18
18
12 ± 2%. The [ F]-2-DFR radiochemical purity was 95 ± 3%,
as measured by high performance anion exchange chromatog-
raphy in combination with a pulsed amperometric detector and
a γ detector, with the predominant impurity eluting at the same
accumulation showed uniform and strong labeling throughout
the liver with additional signal in the kidney and intestines,
consistent with the PET imaging (Figure 3d). No significant
accumulation of radioactivity was present in the bones.
Previously we identified that hepatocytes salvage ribose at a
1
8
12
retention time as [ F]-DFA (Figure 2c). To confirm the
1
8
18
19
fidelity of the [ F]-2-DFR synthesis, [ F]-2-DFR and [ F]-2-
1
8
12
DFR were injected into the same system. [ F]-2-DFR eluted at
the same time as the [ F]-2-DFR standard (Figure 2c,d),
much higher rate than Kupffer, stellate, or endothelial cells.
1
9
18
The strong, uniform [ F]-2-DFR accumulation in the liver is
1
8
18
suggesting that [ F]-2-DFR was correctly prepared and
consistent with [ F]-2-DFR accumulation in liver hepatocytes.
represents an 18F-fluorinated isotopologue of [ F]-2-DFR.
19
Additionally THLE-2 cells, an immortalized human hepatocyte
1
8
18
Finally the specific activity of [ F]-2-DFR was determined to
be 74 ± 30.9 TBq/mmol. These analyses collectively suggest
that [ F]-2-DFR is suitable for PET imaging experiments.
cell line, accumulated [ F]-2-DFR through a pathway that
could be blocked with excess ribose (Figure 3e). To determine
18
18
whether [ F]-2-DFR accumulated in the liver through the
Biological Analysis. The main motivation behind the
same pathway as ribose, a competition experiment was
1
8
18
development of [ F]-2-DFR was to identify a PET imaging
probe that measures the ribose salvage pathway but that is
resistant to defluorination in vivo. To assess the stability of
performed. Co-injection of [ F]-2-DFR with ribose signifi-
18
cantly decreased hepatic and intestinal [ F]-2-DFR accumu-
18
lation but failed to change renal [ F]-2-DFR accumulation
1
8
18
18
[
F]-2-DFR in vivo, [ F]-2-DFR was injected into mice and,
(Figure 3f), suggesting that hepatic and intestinal [ F]-2-DFR
after an hour, their serum was analyzed using an HPLC
connected to a γ detector. The HPLC chromatograms from the
plasma of the injected mice were nearly identical to the HPLC
chromatogram of pure [ F]-2-DFR with no additional peaks or
shifts in retention time (Figure 2e). This suggests that 2-DFR is
stable in vivo for up to 60 min and provides early evidence that
signal is due to specific accumulation. Collectively our data
18
suggest that [ F]-2-DFR is a PET imaging probe that measures
18
ribose salvage activity, that [ F]-2-DFR is not subject to
1
8
18
defluorination in vivo, and that [ F]-2-DFR accumulates
significantly in hepatocytes in vivo.
To better understand how 2-DFR is transformed inside cells
and to add further evidence that 2-DFR measures ribose salvage
activity, we studied its metabolism using cell culture and in vitro
enzymatic assays. During salvage, the glucose transporter Slc2a2
and likely other proteins transport extracellular ribose into cells
where ribokinase phosphorylates ribose to ribose-5-phosphate
2
-DFR is not subject to defluorination.
Mice intravenously injected with [ F]-2-DFR showed high
1
8
uptake in the liver and kidneys, with additional uptake in the
intestines, as measured and quantified by PET imaging, time
activity curves, and ex vivo biodistribution studies (Figure 3a−
c; ex vivo biodistribution values: liver, 11.38 ± 1.74% injected
dose/gram (% ID/g); kidney, 11.14 ± 1.59% ID/g; intestines,
1
2,17
(Figure 1).
In cell culture experiments, phosphorylated 2-
DFR levels were (9.1 ± 0.7)-fold higher in 293T cells that
overexpressed Slc2a2 and ribokinase than in 293T cells that
overexpressed ribokinase alone (Figure 4a), suggesting that 2-
DFR is a substrate for Slc2a2 and additionally that 2-DFR is
1
8
7
.37 ± 1.07% ID/g). [ F]-2-DFR accumulated in the liver
within 15 min and remained there for at least 2 h (Figure 3c).
Notably radioactivity was largely absent from the bone, and the
D
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Figure 4. 2-DFR is transported into cells by the glucose transporter
Slc2a2, phosphorylated by ribokinase, and further metabolized by the
enzyme transketolase. (a) 2-DFR-5-phosphate levels in cells that
overexpress ribokinase (control) or ribokinase and Slc2a2 (Slc2a2).
(b) 2-DFR-5-phosphate MRM trace of an in vitro reaction containing
ribokinase and 2-DFR or H O. (c) Phospho-2-deoxy-2-fluororibosyl
2
pyrophosphate MRM trace of an in vitro reaction containing PRPS1 or
PRPS2 and 2-DFR-5-phosphate or H O. (d) 4-Deoxy-4-fluorosedo-
2
heptulose-7-phosphate MRM trace of an in vitro reaction containing
transketolase and 2-DFR-5-phosphate or H O. (e) 4-Deoxy-4-
2
fluorosedoheptulose-7-phosphate levels in an in vitro reaction
containing 4-deoxy-4-fluorosedoheptulose-7-phosphate and transaldo-
lase or H O. (f) 2-DFR-5-phosphate levels in an in vitro reaction
2
containing 2-DFR-5-phosphate and RPIA or H O. (∗∗∗) P < 0.001.
2
phosphorylated in cells, potentially by ribokinase. To determine
whether 2-DFR is a substrate for ribokinase, bacterially
expressed human ribokinase was combined with ribose, 2-
DFR, or vehicle (H O) in vitro and the reaction products were
2
identified by LC/MS/MS-MRM. Ribokinase phosphorylated
ribose to ribose-5-phosphate and 2-DFR to 2-DFR-5-phosphate
(
Figure 4b; Figure S3, Supporting Information). Ribokinase has
−5
−5
a K of 3.5 × 10 M (standard error, 2.5 × 10 M) and a V
m
max
−11
−1
−1
of 2.6 × 10 molecules of 2-DFR-5-phosphate s μg for 2-
−5
DFR compared to a K of 6.9 × 10 M (standard error, 1.3 ×
m
−5
−11
−1
−1
18
10 M) and a Vmax of 8.2 × 10 ribose-5-phosphate s μg
Figure 3. [ F]-2-DFR strongly accumulates in the liver and kidney in
18
vivo. (a) PET/CT image of a mouse intravenously injected with [ F]-
-DFR and imaged for 10 min, 1 h later, after which a 10 min CT scan
for ribose. Thus, 2-DFR is an effective substrate for ribokinase
with greater affinity but slower kinetics than the endogenous
substrate ribose.
2
was performed. % ID/gram = percent injected dose per gram of tissue;
L = liver, K = kidney, I = intestines, B = bladder. (b) Ex vivo
In cells, ribose-5-phosphate is used in the de novo nucleotide
synthesis pathway or is metabolized through the nonoxidative
pentose phosphate pathway. The first enzymes that metabolize
ribose-5-phosphate for de novo nucleotide synthesis are
phosphoribosyl pyrophosphate synthetase 1 and 2 (PRPS1
and PRPS2). The first enzymes that metabolize ribose-5-
phosphate through the nonoxidative pentose phosphate
pathway are ribose-5-phosphate isomerase (RPIA) and the
sequential enzymes transketolase and transaldolase. RPIA
1
8
18
biodistribution of [ F]-2-DFR. (c) Time−activity curves of [ F]-2-
DFR accumulation in several organs from immediate postinjection to
2
h postinjection. (d) Whole body autoradiography of a mouse
18
intravenously injected with [ F]-2-DFR following 1 h of accumu-
lation. L = liver, K = kidney, I = intestines, B = bladder, Kn = Knee. (e)
1
8
[
F]-2-DFR accumulation in a hepatocyte-like cell line in the presence
of saline or excess ribose. (f) PET/CT images and quantification of
18
mice intravenously injected with [ F]-2-DFR and saline or ribose.
∗∗) P < 0.01; (∗∗∗) P < 0.001.
(
E
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
isomerizes ribose-5-phosphate to ribulose-5-phosphate while
transketolase and transaldolase convert ribose-5-phosphate and
the related pentose xylulose-5-phosphate first to sedoheptulose-
CONCLUSIONS
■
We have described the synthesis and validation of a novel PET
18
probe, [ F]-2-DFR, for quantifying ribose salvage activity in
7
-phosphate and glyceraldehyde-3-phosphate and then to
18
25
vivo. [ F]-2-DFR was synthesized in 12% radiochemical yield
fructose-6-phosphate and erythrose-4-phosphate (Figure 1).
and 74 ± 30.9 TBq/mmol specific activity from precursor 8.
Human PRPS1, PRPS2, transketolase, transaldolase, and
RPIA were expressed and isolated from bacteria and were active
against their natural substrates (Figure S3, Supporting
Information). PRPS1 and PRPS2 failed to metabolize 2-DFR-
18
18
[
F]-2-DFR contains an unusual cis-oriented F and hydroxyl
group and was challenging to synthesize because of the threat
of poor stereoselectivity during the fluorination step caused by
anchimeric assistance of nearby oxygen lone pair electrons. This
anchimeric assistance was completely suppressed by utilizing a
pure β-anomeric precursor, which allowed efficient stereo-
specific late stage fluorination in a short period of time, a
requirement for the synthesis of PET imaging tracers. 2-DFR is
a fluorinated ribose analogue that acts as a competitive
substrate with ribose for the Slc2a2 protein transporter and
ribokinase and transketolase enzymes. The accumulation of 2-
DFR and its metabolites is proportional to the activity of the
ribose salvage pathway, which can be examined in vivo with
5
-phosphate (Figure 4c). Transketolase metabolized 2-DFR-5-
phosphate to 4-deoxy-4-fluorosedoheptulose-7-phosphate, but
transaldolase failed to further metabolize 4-deoxy-4-fluorose-
doheptulose-7-phosphate (Figure 4d, e). Similarly RPIA failed
to metabolize 2-DFR-5-phosphate (Figure 4f). This suggests
that 2-DFR is metabolized by ribokinase and transketolase, at
which point it cannot be further metabolized.
Finally, we assessed whether [ F]-2-DFR could be used to
study and image changes in liver function. Acetaminophen
overdose is the leading cause of acute liver failure in the
developed world, and treatment often requires a liver
1
8
18
18
PET using [ F]-2-DFR. Unlike the previously described [ F]-
18
1
1,26,27
DFA PET imaging probe, [ F]-2-DFR is not subject to
defluorination in vivo. A model of liver dysfunction from a high
dose of acetaminophen in mice was used to demonstrate that
transplant.
Distinguishing between individuals who
require a liver transplant and those who will recover without
11
a transplant remains a significant clinical challenge.
18
PET imaging with [ F]-2-DFR provided an in vivo measure of
liver dysfunction in a clinically relevant model of hepatotoxicity.
Mice treated with a high dose of acetaminophen (300 mg/
kg) or saline control had significantly elevated blood levels of
the liver enzymes alanine aminotransferase and aspartate
aminotransferase 5 h after treatment (Figure 5a). At the same
18
Collectively, the data provide evidence that [ F]-2-DFR is a
novel PET probe that can be used to study and quantify ribose
salvage in vivo in experimental models and potentially patients.
EXPERIMENTAL SECTION
■
Chemical Procedures. All reagents and catalysts were purchased
from commercial sources (Acros Organics and Sigma-Aldrich) and
used without purification unless otherwise noted. Ethanol (200 proof)
was purchased from the UCLA Chemistry Department (Los Angeles,
CA, USA). Hydrochloric acid (1 N) and polypropylene filters (0.45
nm) were purchased from Fisher Scientific. tC18 (WAT036810), silica
(WAT020520 and WAT043400), alumina (WATO20505), and Accell
plus QMA (quarternary methylammonium; WATO20545) cartridges
1
8
were purchased from Waters. Cartridges for trapping and releasing
F
(MP1) were purchased from Ortg, Inc. (Oakdale TN). AG50 strong
cation exchange and AG11 ion retardation resins were purchased from
Biorad. The KT-500-2 column set was purchased from Isoflex. The
1
8
Figure 5. [ F]-2-DFR can be used to study acetaminophen toxicity, a
model of liver failure. (a) Alanine aminotransferase and aspartate
aminotransferase levels in mice treated with vehicle or acetaminophen.
QMA cartridges were preconditioned with KHCO (5 mL; 1 M) and
3
washed with water (10 mL) followed by K CO (0.4 mL; 0.605 M)
2
3
18
(
b) [ F]-2-DFR PET/CT image and quantification of mice treated
with vehicle or acetaminophen. All experiments were performed three
independent times. (∗∗) P < 0.01; (∗∗∗) P < 0.001.
and air (10 cc). The tC18 cartridges were preconditioned with EtOH
5 mL) followed by water (10 mL). The silica cartridges were
(
preconditioned with anhydrous hexane (6 mL). Alumina cartridges
were preconditioned with H O (10 mL). AG50 and AG11 resins were
2
manually packed into small cartridges and preconditioned with water.
All reactions were performed under inert atmosphere of dry argon and
monitored by thin layer chromatography (TLC) on precoated (250
μm) silica gel 60F254 glass-backed plates or on EMD silica gel 60
F254 TLC aluminum sheets and visualized with a UV-254 lamp or
time point, mice treated with acetaminophen showed a
18
decrease in [ F]-2-DFR accumulation from 14.5 ± 0.4% ID/
g to 11.3 ± 0.2% ID/g (Figure 5b), which could be caused by a
variety of mechanisms, including down-regulation of the ribose
salvage pathway, fewer hepatocytes, and decreased blood
perfusion. Future studies will be required to determine this
mechanism in detail. Additional studies will also have to be
performed to determine whether the change in signal during
acetaminophen toxicity is greater than the natural variation of
1
0% sulfuric acid in ethanol stain. NMR spectra were obtained on
Bruker AV300 and AV500 instruments at the UCLA MIC Magnetic
Resonance Laboratory. DART-MS spectra were collected on a
Thermo Exactive Plus MSD (Thermo Scientific) equipped with an
ID-CUBE ion source and a Vapur Interface (IonSense). Both the
source and MSD were controlled by Excalibur, version 3.0. The
analyte was spotted onto OpenSpot sampling cards (IonSense) using
methanol as the solvent. Ionization was accomplished using He plasma
with no additional ionization agents. Flash column chromatography
was performed on silica gel (32−63 μm, 60 Å pore size). Chemical
shifts (δ) are reported in ppm relative to the TMS internal standard.
Abbreviations are as follows: s (singlet), d (doublet), t (triplet), q
(quartet), m (multiplet).
18
hepatic [ F]-2-DFR signal in patients. However, these studies
1
8
provide initial data to suggest that [ F]-2-DFR may be a useful
tool for studying and diagnosing acetaminophen toxicity
preclinically and clinically.
F
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
1
,2-O-Isopropylidene-D-ribofuranose (3). To a solution of 3.05
(ddd, J = 5.7, 5.7, 5.7 Hz, 1H), 4.15 (ddd, J = 4.8, 4.8, 4.8 Hz, 1H),
4.12 (s, 1H), 3.94 (septet, J = 6.0 Hz, 1H), 3.66 (dd, J = 11.4, 4.9 Hz,
1H), 3.65 (dd, J = 11.6, 5.7 Hz, 1H), 2.54 (d, J = 3.6 Hz, 1H), 1.14 (d,
J = 6.3 Hz 3H), 1.12 (s, J = 6.3 Hz 3H). 13C NMR (125 MHz, CDCl3)
δ 148.0, 147.8, 145.9, 144.8, 128.1, 127.9, 124.1, 124.0, 105.9, 80.8,
g (9.6 mmol) of 1,2,3,5-tetra-O-acetyl-β-D-ribofuranose (2) in 30 mL
of dry (4 Å molecular sieves for 10 h) acetone was added 0.6 g (2.4
mmol) of elemental iodine, and the reaction was stirred under argon at
room temperature for 4 h. Dichloromethane (100 mL) was added, and
the reaction was washed with NaHCO (20 mL of sat. aq solution)
followed by Na S O (20 mL of 10% aq solution). The aqueous
+
80.2, 77.6, 74.0, 72.9, 72.4, 71.7, 69.9, 23.8, 21.7. MS (TOF ES ) m/z
3
+
calcd C H N O [M + NH ], 480.19820; found, 480.19355.
2
2
3
22 30
3
9
4
fractions were extracted with 2 × 30 mL of dichloromethane. The
combined organic fractions were washed with 50 mL of brine, dried
with Na SO , and the solvent was removed via rotary evaporation. The
Isopropyl 3,5-Di-O-(4-nitrobenzyl)-β-D-arabinofuranoside
(7). To a solution of oxalyl dichloride (1.07 mL, 12.5) in 20 mL of
dichloromethane was added dropwise a solution of DMSO (1.73 mL,
24.3 mmol) in 5 mL of dichloromethane at −78 °C. After the mixture
was stirred for 15 min, a solution of the ribofuranoside (6) (1.9 g, 4.1
mmol) in 25 mL of dichloromethane was added. After the mixture was
2
4
vacuum-dried crude material was dissolved in 30 mL of dry MeOH,
and 0.4 g of NaOMe was added to the solution. The reaction was
stirred for 30 min at 50 °C. Once TLC (4% MeOH in CH Cl )
2
2
showed a single spot with R = 0.34, Amberlyst-15 resin was added to
stirred at −78 °C for 40 min, Et
The solution was warmed to room temperature over 5 min, mixed
with H O (50 mL), and extracted with dichloromethane (50 mL
twice). TLC (EtOAc/hexanes = 1/2) showed one spot with R = 0.26.
The organic phase was dried with Na SO and concentrated. The
crude material was dissolved in 20 mL of EtOH−CH Cl (8:2)
mixture, and NaBH (0.31 g, 8.1 mmol) dissolved in 13 mL of EtOH−
H O (2:1) was added dropwise with stirring at 0 °C. The reaction was
2
3
N (3.4 mL, 24.7 mmol) was added.
f
neutralize the reaction. The resin beads were removed by filtration,
and mother liquor was concentrated via rotary evaporation. Upon
careful removal of solvent, the crude material was sufficiently pure for
use in the next step. Purification by column chromatography on silica
gel eluting with MeOH/CH Cl (0−4% gradient) afforded 1.51 g
2
f
2
4
2
2
2
2
(
83%) of 3 as a clear oil crystallizing with time. The spectroscopic data
4
28
were in agreement with that in the literature.
1
,2-O-Isopropylidene-3,5-di-O-(4-nitrobenzyl)-D-ribofura-
then warmed to ambient temperature over 1 h and quenched with 10%
nose (5). The diol 3 (1.5 g, 7.89 mmol) was coevaporated with 10 mL
AcOH. The mixture was extracted from NaHCO (aq sat. solution 100
3
of toluene and mixed with 10 g (43 mmol) of Ag O and 10 g of
mL) with 2 × 100 mL of dichloromethane. The combined organic
layer was washed with 2 × 100 mL of water and then dried over
Na SO . The solvent was removed under vacuum. TLC displayed
2
activated 4 Å molecular sieves (150 °C for 3 d). The solids were
suspended in dry dichloromethane (80 mL) and stirred for 1 h. Then
2
4
6.5 g (28 mmol) of 4-nitrobenzyl bromide was added in one portion
compound 7 with R = 0.21 (EtOAc/hexanes = 1/2) and a hardly
f
and the reaction was stirred at 40 °C under argon for 20 h. The
reaction was filtered through Celite, rinsed with 3 × 15 mL of
dichloromethane and the solvent dried in vacuo. Silica gel column
chromatography using EtOAc/hexanes (15−35% gradient) provided
visible spot of compound 6 with R = 0.18, HPLC t = 6.61. Column
f
R
chromatography of the residue on silica gel (30−40% of EtOAc in
1
hexanes) afforded 1.64 g (87%) of 7: H NMR (500 MHz, CDCl ) δ
3
8.21 (d, J = 8.8, 2H), 8.19 (d, J = 8.8 Hz, 2H), 7.51 (d, J = 8.8 Hz,
2H), 7.50 (d, J = 8.8 Hz, 2H), 5.11 (d, J = 4.9 Hz, 1H), 4.92 (d, J =
13.3 Hz, 1H), 4.76 (d, J = 13.3 Hz, 1H), 4.67 (s, 2H), 4.27 (ddd, J =
9.4, 6.1, 4.9 Hz, 1H), 4.17 (ddd, J = 5.8, 5.8, 5.7 Hz, 1H), 4.01 (septet,
J = 6.2 Hz, 1H), 3.87 (dd, J = 5.9, 5.9 Hz, 1H), 3.65 (d, J = 5.8 Hz,
2H), 2.64 (d, J = 9.4 Hz, 1H), 1.19 (d, J = 6.0 Hz, 3H), 1.18 (d, J = 6.0
3
.2 g (87%) of pure compound 5 (R = 0.28, 35% of EtOAc in hexanes,
f
1
HPLC t = 6.20 min). 5: H NMR (500 MHz, CDCl ) δ 8.19 (dd, J =
R
3
8
(
1
(
.5 Hz, 2H), 8.18 (d, J = 8.5 Hz, 2H), 7.53 (d, J = 8.5 Hz, 2H), 7.46
d, J = 8.5 Hz, 2H), 5.82 (d, J = 3.5 Hz, 1H), 4.88 (d, J = 13.0 Hz,
H), 4.70−4.63 (m, 4H), 4.25 (ddd, J = 9.0, 4.0, 2.0 Hz, 1H), 3.89
dd, J = 9.0, 4.5 Hz, 1H), 3.86 (dd, J = 11.5, 2.0 Hz, 1H), 3.71 (dd, J =
Hz, 3H). 13C NMR (125 MHz, CDCl ) δ 147.8 (2C), 145.9 (2C),
3
13
11.5, 4.5 Hz, 1H), 1.60 (s, 3H), 1.38 (s, 3H). C NMR (125 MHz,
128.1, 128.0, 124.0 (2C), 100.4, 85.8, 80.5, 77.9, 73.1, 72.4, 71.5, 70.8,
+
+
CDCl ) δ 147.9, 147.7, 145.9, 145.4, 128.3, 127.9, 124.0, 123.9, 113.5,
23.9, 22.2. MS (TOF ES ) m/z calcd C H N O [M + NH ],
3
22 30 3 9 4
+
1
04.6, 78.5, 78.2, 77.2, 72.6, 71.2, 69.2, 27.0, 26.9. MS (TOF ES ) m/z
calcd C H N O [M + NH ], 478.18255; found, 478.17804.
480.19820; found, 480.19401.
+
22
28
3
9
4
Isopropyl 2-O-(Trifluoromethylsulfonyl)-3,5-di-O-(4-nitro-
benzyl)-β-D-arabinofuranoside (8). To a solution of 0.025 g
(0.054 mmol) of the ribofuranoside (7) and 0.066 mL (0.81 mmol) of
pyridine in 1 mL of dry dichloromethane was added triflic anhydride
(0.11 mL of 1 M solution in dichloromethane) at 0 °C. The reaction
was stirred under argon in an ice bath for 30 min, then quenched with
Isopropyl 3,5-Di-O-(4-nitrobenzyl)-β-D-ribofuranoside (6).
To an ice-cold solution of the acetonide 5 (3.16 g, 6.85 mmol) in a
mixture of acetic acid (36 mL) and acetic anhydride (9 mL), 0.03 mL
(
0.34 mmol) of TfOH was added, and reaction was kept at room
temperature for 3 h. The reaction was carefully neutralized with
NaOH (1 M), and mixture of acetates (EtOAc/hexanes = 1/2: R =
5 mL of a 1:1 mixture of ice and saturated NaHCO solution. The
f
3
0
.27 β-anomer; R = 0.23 α-anomer) was extracted with dichloro-
mixture was extracted with 2 × 10 mL of dichloromethane, the
f
methane (3 × 75 mL). The organic phases were combined and dried
over Na SO , and the solvent was removed via rotary evaporation. The
crude material was dissolved in 15 mL of dry THF. To that solution
were added iodine (0.35 g, 1.39 mmol) and dry isopropanol (2 mL, 27
mmol), and the reaction was refluxed under argon for 8 h. The solvent
was removed by rotary evaporation and the residue dissolved in 100
mL of dichloromethane. The solution was washed with NaHCO (20
mL saturated aqueous solution) followed by Na S O (20 mL of 10%
aqueous solution), and the organic layer was dried over Na SO and
reduced in vacuo. TLC (1% MeOH in CH Cl ) indicated two major
combined organic layer was dried over Na SO and reduced under
2
4
2
4
high vacuum at room temperature. To this was added 5 mL of n-
heptane, and the resulting suspension was coevaporated in high
vacuum at room temperature. The crude triflate (10) was obtained (31
1
mg, 97%) (HPLC t = 5.30). 8: H NMR (300 MHz, CDCl ) δ 8.23
R
3
(d, J = 9.0, Hz, 2H), 8.19 (d, J = 9.0, Hz, 2H), 7.50 (d, J = 8.7 Hz, 2H),
7.43 (d, J = 9.0 Hz, 2H), 5.2 (d, J = 4.5 Hz, 1H), 5.11 (dd, J = 6.3, 4.5
Hz, 1H), 4.73 (s, 2H), 4.67 (s, 2H), 4.37 (dd, J = 6.0, 5.4 Hz, 1H),
4.20 (ddd, J = 6.0, 5.8, 5.4 Hz, 1H), 3.96 (septet, J = 6.3 Hz, 1H), 3.72
(dd, J = 9.9, 4.5 Hz, 1H), 3.67 (dd, J = 9.9, 6.0 Hz, 1H), 1.19 (d, J =
3
2
2
3
2
4
2
2
6.2 Hz, 3H), 1.17 (d, J = 6.2 Hz, 3H). 13C NMR (125 MHz, CDCl ) δ
spots with Rf = 0.54 (isopropyl 2-O-acetyl-3,5-di-O-PNBn-β-D-
ribofuranoside) and 6 with R = 0.12. The crude material was
dissolved in 30 mL of dry MeOH, and 0.2 g of NaOMe was added.
3
f
145.0, 144.3, 134.6, 130.1, 129.0, 128.3, 127.8, 123.8, 98.3, 87.8, 82.4,
+
78.9, 72.3, 72.1, 71.3, 23.4, 21.5. MS (TOF ES ) m/z calcd
+
The reaction was stirred for 30 min at 50 °C. TLC (EtOAc/hexanes =
C H F N O S [M + NH ], 612.14749; found, 612.14259.
23
29
3
3
11
4
1/2) showed one spot with R = 0.18. Amberlyst-15 resin was added to
Isopropyl 2-Deoxy-2-fluoro-3,5-di-O-(4-nitrobenzyl)-β-D-ri-
f
the reaction. The resin beads were removed by filtration, and the
solvent was removed using rotary evaporation. The crude material was
chromatographed on silica gel, eluting with a gradient of 20−40% of
EtOAc in hexanes to afford 2.02 g of 6 (HPLC t = 9.97). 6: H NMR
bofuranoside (9). KF (3 mg, 0.05 mmol) and Kryptofix (19 mg,
0.05 mmol) were azeotropically dried with acetonitrile (3 × 0.3 mL) in
a stream of dry nitrogen at 100 °C. To the obtained KF/K2.2.2
complex, 0.03 g (0.05 mmol) of triflate (8) was added in 0.2 mL of dry
acetonitrile. The reaction vessel was sealed and heated to 154 °C for
1
R
(
300 MHz, CDCl ) δ 8.21 (d, J = 8.7 Hz, 2H), 8.20 (d, J = 8.7 Hz,
3
2
4
H), 7.51 (dd, J = 9.0 Hz, 2H), 7.48 (d, J = 9.0 Hz, 2H), 5.11 (s, 1H),
.78 (d, J = 12.9 Hz, 1H), 4.71 (d, J = 12.9 Hz, 1H), 4.69 (s, 2H), 4.29
30 min. PTLC purification on silica gel with dichloromethane afforded
1
0.014 g (76%) of 9. 9: H NMR (300 MHz, CDCl ) δ 8.21 (d, J = 9.0
3
G
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
Hz, 2H), 8.20 (d, J = 9.0 Hz, 2H), 7.53 (d, J = 8.7 Hz, 2H), 7.52 (d, J
biodistribution studies. C57Bl/6 mice (male, 8−12 weeks old) were
used for the acetaminophen studies.
=
9.0 Hz, 2H), 5.27 (d, J = 11.1 Hz, 1H), 4.92 (dd, J = 53.5, 3.6 Hz,
1
4
1
3
6
1
H), 4.81 (d, J = 12.9 Hz, 1H), 4.70 (s, 2H), 4.69 (d, J = 12.9 Hz, 1H),
.36 (ddd, J = 7.6, 5.7, 3.6 Hz, 1H), 4.20 (ddd, J = 23.9, 7.8, 3.6 Hz,
H), 3.98 (septet, J = 6.3 Hz, 1H), 3.78 (dd, J = 10.7, 3.6 Hz, 1H),
.69 (dd, J = 10.7, 5.7 Hz, 1H), 1.15 (d, J = 6.5 Hz, 3H), 1.13 (d, J =
Micro-PET/Computed Tomography (CT) Imaging and Anal-
ysis. Mice were anesthetized with 1.5−2% (v/v) isofluorane and
immediately injected intravenous with approximately 740 kBq of
18
18
18
[ F]-2-DFR or [ F]-DFA. [ F]-DFA was synthesized as previously
1
2
.5 Hz, 3H). 13C NMR (125 MHz, CDCl ) δ 147.9, 147.8, 146.0,
described. Animals remained anesthetized throughout the imaging
experiment. PET scans were performed on an Inveon micro-PET
scanner (Siemens), Genisys 4 imager (Sofie Biosciences), and Genisys
3
45.1, 128.1, 127.9, 124.0, 123.96, 103.3 (d, J = 29.3 Hz, 1C), 92.5 (d,
J = 185.3 Hz, 1C), 79.8, 79.1 (d, J = 15.6 Hz, 1C), 72.4, 72.2, 71.6,
0.4, 23.7, 21.6. 19F NMR (282 MHz, CDCl ) δ −208.58 (ddd, J =
3.6, 23.9, 11.4 Hz). MS (TOF ES ) m/z calcd C H FN O [M +
8
imager (Sofie Biosciences). CT scans were performed on a micro-
7
5
3
+
CAT II CT scanner (Siemens). For dynamic scans, mice were
continuously imaged on a PET scanner from immediate to 2 h
postinjection after which the mice were subject to a 10 min CT scan.
The PET data were analyzed using the following imaging windows: 2
22
29
3
8
+
NH ], 482.19387; found, 482.18903.
4
2-DFR (1). The protected fluororibose 9 (0.014 g, 0.03 mmol) was
dissolved in 0.1 mL of 90% aqueous formic acid, and 0.03 g of Raney
Ni was added. The suspension was heated to 130 °C for 30 min and
passed through Amberlite resin (OH source) washing with 2 mL of
MeOH−CH Cl (1:9). The washings were purified by column
chromatography on silica gel using 5−10% of MeOH in dichloro-
methane gradient, yielding 4 mg of 1. The spectroscopic data were in
×
30 s, 4 × 1 min, 5 × 5 min, 9 × 10 min. For static scans, mice were
−
imaged on a CT scanner for 10 min, 50 min postinjection and then
immediately imaged on a PET scanner for 10 min. PET images were
reconstructed using a OSEM + MAP algorithm. PET/CT images were
manually co-registered. Data were imaged and quantified using the
2
2
1
2,30
24
AMIDE software as previously described.
Figure 3a represents a
agreement with that in the literature.
18
−
volume rendering of the PET/CT data. Figure 3f, Figure 5b, Figures
Radiosynthetic Procedures. K222/[ F]F Complex Formation.
S1a, and S2a represent coronal sections.
[ F]-2-DFR ex Vivo Biodistribution. [ F]-2-DFR ex vivo
All radiochemistry procedures were carried out in a high pressure
compact modular radiosynthesizer. No-carrier-added [ F]-fluoride
18
18
29
18
biodistribution experiments were performed as previously described
was produced in an RDS-111 cyclotron (UCLA, biomedical cyclotron)
1
8
18
12
by the nuclear reaction (p,n) of 1 O-enriched water ([ O]H O, 98%
8
18
except that [ F]-2-DFR was used in place of [ F]-DFA.
2
1
8
[
F]-2-DFR Whole Body Autoradiography. Mice were
isotopic purity, Rotem, Inc.) in a 1 mL silver target (11 MeV beam).
1
8
1
8
−
anesthetized and injected with 37 MBq of [ F]-2-DFR. After 1 h,
the mice were sacrificed, embedded in a 3% (wt/vol) solution of
carboxymethylcellulose, and frozen in a dry ice/methanol bath. Digital
images and 40 μm sections were cut with a Leica cryostat. Sections
were exposed to a Fujifilm imaging plate overnight, and the Fujifilm
cassettes were imaged on a Fuji BAS imager.
[
F]F (3700−9250 MBq) was captured and eluted from a QMA
cartridge in K CO (0.4 mL; 0.605 mM) into a reaction vessel
containing Kryptofix 222 (K222; 13 mg, 34.58 μmol) dissolved in
CH CN (0.7 mL). The mixture was heated to 120 °C in an oil bath to
complete dryness and anhydrous acetonitrile (0.5 mL) was added and
azeotropically evaporated 4 times to remove water and to form the
2
3
3
18
18
−
[ F]-2-DFR Cell Accumulation Experiments. THLE-2 cells
(ATCC; 0.3 × 10 cells/well) were plated in hepatocyte maintenance
K222/[ F]F complex.
6
-Isopropyl-2-fluoro-3,5-di-O-(4-nitrobenzyl)-β-D-[18F]-
1
1
8
18
−
media (Corning) onto poly-L-lysine-coated 24-well plates and allowed
ribofuranoside ([ F]-9). To the K /[ F]F complex, triflate (8)
222
1
8
18
to adhere overnight. 74 kBq of [ F]-2-DFR and saline or [ F]-2-DFR
and 50 mM ribose were added to each well. The cells were incubated
(
10 mg, 0.016 mmol), dissolved in acetonitrile (500 μL), was added
and heated in an oil bath (154 °C; 30 min). The reaction mixture was
loaded onto a silica Sep-pak cartridge and eluted with ethyl acetate (2
mL). The compound was heated to dryness in an oil bath (100 °C)
under nitrogen.
18
with the [ F]-2-DFR for 1 h, after which the cells exposed to trypsin,
collected, washed 3× with ice-cold hepatocyte maintenance media, and
the radioactive accumulation was measured on a γ counter.
[ F]-2-DFR and Ribose Co-Injection Studies. [ F]-2-DFR was
co-injected with ribose (25 μmoles in 50 μL of 1× PBS) or vehicle (50
μL of 1× PBS). After 50 min, the mice were CT and PET imaged as
described above.
18
18
18
18
18
2
-Deoxy-2-[ F]fluororibose ([ F]-2-DFR, [ F]-1). Raney Ni
18
(
30 mg) in formic acid (500 μL, 90% in water) was added to [ F]-9,
and the reaction mixture was heated (130 °C; 30 min) and filtered
through a polypropylene filter (0.45 μm). The filter was rinsed with
water (10 mL), and the water and filtrate were passed through the KT-
18
18
[
F]-2-DFR Transport and Metabolism Studies. [ F]-2-DFR
transport and metabolism studies were performed as previously
500-2 purification column and a filter (0.22 μm) to yield the final
12
described except that 2-DFR was used in place of DFA.
product (1). The final product was at a pH of ∼6.5 after the
purification column and filter and was further neutralized to pH ∼7.4
by the addition of 10× PBS to a final concentration of 1× PBS. Total
synthesis time from [ F]F delivery to the final product was 2.5 h.
Purified 1 and [ F]-1 were injected into an analytical HPLC coupled
to an electrochemical detector (ED50A, Dionex) and a γ detector
Acetaminophen Studies. Overnight fasted mice were injected
with acetaminophen (75 mg/mL in sterile saline; 300 mg/kg) or
vehicle (sterile saline). At 6 h postinjection, the mice were PET/CT
18
−
18
18
imaged with [ F]-2-DFR. Following PET/CT imaging, serum (100
μL) was collected via a retroorbital bleed in gel barrier blood collection
tubes (Capiject). Alanine aminotransferase and aspartate amino-
transferase measurements were performed by the UCLA Division of
Laboratory Animal Medicine Diagnostic Lab. Only those mice with
elevated alanine aminotransferase and aspartate aminotransferase
levels were analyzed as part of the acetaminophen treatment group.
PET images were quantified by an individual unaware of the treatment
groups.
(
Carroll and Ramsey Associates, model 105 S). A CarboPac MA1
column (4 mm × 250 mm) was used with a flow rate of 0.4 mL/min
of 150 nM NaOH. Both [ F]-1 and 1 eluted with the same retention
time. The radiochemical purity of [ F]-1 was 95 ± 3%.
18
18
18
18
Biological Procedures. [ F]-2-DFR Specific Activity. [ F]-2-
DFR specific activity was measured as previously described except that
1
1
8
8
18
12
[
[
F]-2-DFR was used in place of [ F]-DFA. The specific activity of
F]-2-DFR was 74 ± 30.9 TBq/mmol.
Statistics. Data are reported as mean ± standard error of the mean.
Treatment groups were compared using a Student t test in Microsoft
Excel. Graphs were created in Microsoft Excel and GraphPad Prism.
18
[
F]-2-DFR in Vivo Stability. Mice were injected intravenous
18
with ∼11.1 MBq of [ F]-2-DFR. After 1 h, the mice were sacrificed
and serum (400 μL) was collected via a cardiac bleed into gel barrier
18
blood collection tubes (Capiject). 25 μL of pure [ F]-2-DFR or the
mouse serum was injected into an HPLC instrument connected to a γ
detector.
Mice. Animal studies were performed in accordance with the
UCLA Animal Research Committee guidelines. C57Bl/6 mice
ASSOCIATED CONTENT
■
*
S
Supporting Information
Figures S1−S3, NMR, HPLC, and MS spectra. The Supporting
Information is available free of charge on the ACS Publications
website at DOI: 10.1021/acs.jmedchem.5b00569.
(
female, 8−12 weeks old) were used for the dynamic imaging and
H
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(
6) Tillisch, J.; Brunken, R.; Marshall, R.; Schwaiger, M.; Mandelkern,
AUTHOR INFORMATION
Corresponding Authors
N.M.E.: e-mail, nevdokim@chem.ucla.edu; phone, (310) 902-
552.
M.E.J.: e-mail, jung@chem.ucla.edu; phone, (310) 825-7954.
■
M.; Phelps, M.; Schelbert, H. Reversibility of cardiac wall-motion
abnormalities predicted by positron tomography. N. Engl. J. Med. 1986,
*
3
(
14, 884−888.
8
*
7) Dyson, J. K.; McPherson, S.; Anstee, Q. M. Non-alcoholic fatty
liver disease: non-invasive investigation and risk stratification. J. Clin.
Pathol. 2013, 66, 1033−1045.
Author Contributions
N.M.E. and P.M.C. contributed equally.
All authors contributed to the writing of this manuscript. All
authors have given approval to the final version of the
manuscript.
●
(8) Festi, D.; Schiumerini, R.; Marzi, L.; Di Biase, A. R.; Mandolesi,
D.; Montrone, L.; Scaioli, E.; Bonato, G.; Marchesini-Reggiani, G.;
Colecchia, A. Review article: the diagnosis of non-alcoholic fatty liver
diseaseavailability and accuracy of non-invasive methods. Aliment.
Pharmacol. Ther. 2013, 37, 392−400.
Notes
(9) Malnick, S. D.; Beergabel, M.; Knobler, H. Non-alcoholic fatty
liver: a common manifestation of a metabolic disorder. QJM 2003, 96,
699−709.
The authors declare no competing financial interest.
N.M.E., P.M.C., G.F., O.N.W., and M.E.J. are coauthors on a
patent application that includes the [ F]-2-DFR molecule
described in this manuscript.
1
8
(
10) Myers, R. P.; Fong, A.; Shaheen, A. A. Utilization rates,
complications and costs of percutaneous liver biopsy: a population-
based study including 4275 biopsies. Liver Int. 2008, 28, 705−712.
(11) Larson, A. M.; Polson, J.; Fontana, R. J.; Davern, T. J.; Lalani, E.;
ACKNOWLEDGMENTS
■
Hynan, L. S.; Reisch, J. S.; Schiodt, F. V.; Ostapowicz, G.; Shakil, A.
O.; Lee, W. M. Acetaminophen-induced acute liver failure: results of a
United States multicenter, prospective study. Hepatology 2005, 42,
1364−1372.
We thank Dr. David Stout, members of the Crump Institute for
Molecular Imaging, and members of the Pasarow Mass
Spectrometry Laboratory for their technical assistance and
advice. N.M.E. was supported by the Phelps Family
Foundation. P.M.C. was supported by the California Institute
of Regenerative Medicine Training Grant TG2-01169, the
UCLA Scholars in Oncologic Molecular Imaging Program
National Cancer Institute Grant R25T CA098010, and the
UCLA in Vivo Cellular and Molecular Imaging Center Career
Development Award P50 CA086306. T.C. was supported by
the University of California Amgen Scholars Program. O.N.W.
is an Investigator of the Howard Hughes Medical Institute and
is partially supported by the Eli and Edythe Broad Center of
Regenerative Medicine and Stem Cell Research. We thank
Ralph and Marjorie Crump for a donation made to the
University of California, Los Angeles (UCLA) Crump Institute
for Molecular Imaging. The National Science Foundation
Equipment Grant CHE-1048804 is acknowledged.
(12) Clark, P. M.; Flores, G.; Evdokimov, N. M.; McCracken, M. N.;
Chai, T.; Nair-Gill, E.; O’Mahony, F.; Beaven, S. W.; Faull, K. F.;
Phelps, M. E.; Jung, M. E.; Witte, O. N. Positron emission tomography
probe demonstrates a striking concentration of ribose salvage in the
liver. Proc. Natl. Acad. Sci. U.S.A. 2014, 111, E2866−E2874.
(
13) Goncalves, R. P.; Bennett, G. C.; Leblond, C. P. Fate of 3H-
ribose in the rat as detected by radioautography. Anat. Rec. 1969, 165,
43−557.
5
14) Miller, P. W.; Long, N. J.; Vilar, R.; Gee, A. D. Synthesis of 11C,
(
18
15
13
F, O, and N radiolabels for positron emission tomography.
Angew. Chem., Int. Ed. 2008, 47, 8998−9033.
15) Deng, H.; Cobb, S. L.; Gee, A. D.; Lockhart, A.; Martarello, L.;
(
McGlinchey, R. P.; O’Hagan, D.; Onega, M. Fluorinase mediated C-F-
18 bond formation, an enzymatic tool for PET labelling. Chem.
Commun. 2006, 652−654.
(16) Li, X. G.; Autio, A.; Ahtinen, H.; Helariutta, K.; Liljenback, H.;
Jalkanen, S.; Roivainen, A.; Airaksinen, A. J. Translating the concept of
peptide labeling with 5-deoxy-5-[F-18] fluororibose into preclinical
practice: F-18-labeling of Siglec-9 peptide for PET imaging of
inflammation. Chem. Commun. 2013, 49, 3682−3684.
ABBREVIATIONS USED
■
1
8
18
18
[
F]-2-DFR, 2-deoxy-2-[ F]fluororibose; [ F]-FDG, 2-
1
8
18
18
deoxy-2-[ F]fluoroglucose; [ F]-DFA, 2-deoxy-2-[ F]-
fluoroarabinose; PNBn, 4-nitrobenzyl; PRPS, phosphoribosyl
pyrophosphate synthetase; RPIA, ribose-5-phosphate isomer-
ase; % ID/g, percent injected dose per gram of tissue
(17) Park, J.; van Koeverden, P.; Singh, B.; Gupta, R. S. Identification
and characterization of human ribokinase and comparison of its
properties with E. coli ribokinase and human adenosine kinase. FEBS
Lett. 2007, 581, 3211−3216.
(18) Furuya, T.; Kuttruff, C. A.; Ritter, T. Carbon-fluorine bond
REFERENCES
formation. Curr. Opin. Drug Discovery Dev. 2008, 11, 803−819.
■
(
19) Huang, X. Y.; Liu, W.; Ren, H.; Neelamegam, R.; Hooker, J. M.;
(
1) Gambhir, S. S. Molecular imaging of cancer with positron
emission tomography. Nat. Rev. Cancer 2002, 2, 683−693.
Groves, J. T. Late stage benzylic C-H fluorination with [F-18]fluoride
for PET imaging. J. Am. Chem. Soc. 2014, 136, 6842−6845.
(
2) Phelps, M. E. Positron emission tomography provides molecular
imaging of biological processes. Proc. Natl. Acad. Sci. U.S.A. 2000, 97,
226−9233.
3) Phelps, M. E.; Mazziotta, J. C. Positron emission tomography:
human brain function and biochemistry. Science 1985, 228, 799−809.
4) van Tinteren, H.; Hoekstra, O. S.; Smit, E. F.; van den Bergh, J.
(20) Liu, W.; Huang, X. Y.; Cheng, M. J.; Nielsen, R. J.; Goddard, W.
A.; Groves, J. T. Oxidative aliphatic C-H fluorination with fluoride ion
9
(
catalyzed by a manganese porphyrin. Science 2012, 337, 1322−1325.
(21) Dahlman, O.; Garegg, P. J.; Mayer, H.; Schramek, S. Synthesis
of 3 3-C-hydroxymethylpentoses with the D-ribo-configurations, D-
xylo-configurations and L-lyxo-configurationsIdentification of the
latter with a monosaccharide isolated from phase-I Coxiella-burnetii
lipopolysaccharide. Acta Chem. Scand., Ser. B 1986, 40, 15−20.
(
H.; Schreurs, A. J.; Stallaert, R. A.; van Velthoven, P. C.; Comans, E.
F.; Diepenhorst, F. W.; Verboom, P.; van Mourik, J. C.; Postmus, P. E.;
Boers, M.; Teule, G. J. Effectiveness of positron emission tomography
in the preoperative assessment of patients with suspected non-small-
cell lung cancer: the PLUS multicentre randomised trial. Lancet 2002,
(22) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Smith, D. M.;
Woerpel, K. A. Stereoselective C-glycosylation reactions of ribose
derivatives: Electronic effects of five-membered ring oxocarbenium
ions. J. Am. Chem. Soc. 2005, 127, 10879−10884.
3
(
59, 1388−1393.
5) Pio, B. S.; Park, C. K.; Pietras, R.; Hsueh, W. A.; Satyamurthy, N.;
Pegram, M. D.; Czernin, J.; Phelps, M. E.; Silverman, D. H. Usefulness
of 3′-[F-18]fluoro-3′-deoxythymidine with positron emission tomog-
raphy in predicting breast cancer response to therapy. Mol. Imaging
Biol. 2006, 8, 36−42.
(23) Lopez, J. C.; Ventura, J.; Uriel, C.; Gomez, A. M.; Fraser-Reid,
B. Reaction of 1,2-orthoesters with HF-pyridine: A method for the
preparation of partly unprotected glycosyl fluorides and their use in
saccharide synthesis. Org. Lett. 2009, 11, 4128−4131.
I
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry
Article
(
24) Sanderson, P. N.; Sweatman, B. C.; Farrant, R. D.; Lindon, J. C.
Assignment of the H-1, F-19, and C-13 NMR spectra of 2-deoxy-2-
fluoro-D-ribose and characterisation of the isomeric equilibrium.
Carbohydr. Res. 1996, 284, 51−60.
(
25) Berg, J. M.; Tymoczko, J. L.; Stryer, L. Biochemistry; W. H.
Freeman: New York, 2002.
26) Larson, A. M. Acetaminophen hepatotoxicity. Clin. Liver Dis.
007, 11, 525−548.
27) Ward, J.; Kanchagar, C.; Veksler-Lublinsky, I.; Lee, R. C.;
(
2
(
McGill, M. R.; Jaeschke, H.; Curry, S. C.; Ambros, V. R. Circulating
microRNA profiles in human patients with acetaminophen hepatotox-
icity or ischemic hepatitis. Proc. Natl. Acad. Sci. U.S.A. 2014, 111,
1
(
2169−12174.
28) Lenagh-Snow, G. M. J.; Araujo, N.; Jenkinson, S. F.; Rutherford,
C.; Nakagawa, S.; Kato, A.; Yu, C. Y.; Weymouth-Wilson, A. C.; Fleet,
G. W. J. Inhibition of nonmammalian glycosidases by azetidine
iminosugars derived from stable 3,5-di-O-triflates of pentoses. Org.
Lett. 2011, 13, 5834−5837.
(
29) Amaraesekera, B.; Marchis, P. D.; Bobinski, K. P.; Radu, C. G.;
Czernin, J.; Barrio, J. R.; van Dam, R. M. High-pressure, compact,
modular radiosynthesizer for production of positron emitting
biomarkers. Appl. Radiat. Isot. 2013, 78, 88−101.
(
30) Loening, A. M.; Gambhir, S. S. AMIDE: A completely free
system for medical imaging data analysis. J. Nucl. Med. 2001, 42,
192P−192P.
J
DOI: 10.1021/acs.jmedchem.5b00569
J. Med. Chem. XXXX, XXX, XXX−XXX