A modular platform for the rapid site-specific radiolabeling of proteins with 18F exemplified by quantitative positron emission tomography of human epidermal growth factor receptor 2

Article abstract of DOI:10.1021/jm900420c

Receptor-specific proteins produced by genetic engineering are attractive as PET imaging agents, but labeling with conventional ¹⁸F-based prosthetic groups is problematic due to long synthesis times, poor radiochemical yields, and low specific activities. Therefore, we developed a modular platform for the rapid preparation of water-soluble prosthetic groups capable of efficiently introducing ¹⁸F into proteins. The utility of this platform is demonstrated by the thiol-specific prosthetic group, [ ¹⁸F]FPEGMA, which was used to produce site-specifically ¹⁸F-labeled protein (¹⁸F-trastuzumab-ThioFab) in 82 min with a total radiochemical yield of 13 ± 3% and a specific activity of 2.2 ( 0.2 Ci/μmol. ¹⁸F-trastuzumab-ThioFab retained the biological activity of native protein and was successfully validated in vivo with microPET imaging of Her2 expression in a xenograft tumor-bearing murine model modulated by the Hsp90 inhibitor, 17-(allylamino)-17-demethoxygeldanamycin. 2009 American Chemical Society.

Full text of DOI:10.1021/jm900420c

5816 J. Med. Chem. 2009, 52, 5816–5825
DOI: 10.1021/jm900420c
A Modular Platform for the Rapid Site-Specific Radiolabeling of Proteins with 18F Exemplified by
Quantitative Positron Emission Tomography of Human Epidermal Growth Factor Receptor 2
Herman S. Gill,† Jeff N. Tinianow,† Annie Ogasawara,† Judith E. Flores,† Alexander N. Vanderbilt,† Helga Raab,‡
Justin M. Scheer,‡ Richard Vandlen,‡ Simon-P. Williams,† and Jan Marik*,†
†Departments of Biomedical Imaging and ‡Protein Chemistry, Genentech Inc., 1 DNA Way, South San Francisco, California 94080
Received April 1, 2009
Receptor-specific proteins produced by genetic engineering are attractive as PET imaging agents, but
labeling with conventional 18F-based prosthetic groups is problematic due to long synthesis times, poor
radiochemical yields, and low specific activities. Therefore, we developed a modular platform for the
rapid preparation of water-soluble prosthetic groups capable of efficiently introducing 18F into proteins.
The utility of this platform is demonstrated by the thiol-specific prosthetic group, [18F]FPEGMA, which
was used to produce site-specifically 18F-labeled protein (18F-trastuzumab-ThioFab) in 82 min with a
total radiochemical yield of 13 ( 3% and a specific activity of 2.2 ( 0.2 Ci/μmol. 18F-trastuzumab-
ThioFab retained the biological activity of native protein and was successfully validated in vivo with
microPET imaging of Her2 expression in a xenograft tumor-bearing murine model modulated by the
Hsp90 inhibitor, 17-(allylamino)-17-demethoxygeldanamycin.
Introduction
18F-labeled prosthetic groups used for protein conjugations
are often limited by some combination of poor radiochemical
yield, long synthesis time, and low specific activity. Improved
methods for generating 18F-labeled proteins may facilitate
molecular imaging in humans and thereby address questions
in the clinical development process: What is the level of target
expression? How heterogeneous is it? And how does it change
over time? Therefore, we have developed a new platform for
the rapid production of 18F-labeled prosthetic groups capable
of efficiently introducing 18F into proteins.
Proteins and peptides make up a large part of the arma-
mentarium available for the molecular imaging of cell-surface
biomarkers. Although the development of “ideal” imaging
agents is an important goal, in practice many imaging agents
are developed from existing proteins. The development of
positron emission tomography (PET)a imaging agents from
monoclonal antibodies and their engineered fragments
(immuno-PET)1 holds promise as a tool for localizing and
quantifying molecular targets and may enhance the noninva-
sive clinical diagnosis of pathological conditions. Small im-
muno-PET imaging agents, such as Fab antibody fragments
(50 kDa) or diabodies (paired dimers of the covalently
associated VH-VL region of Mab, 55 kDa),2-4 may be
particularly useful since they exhibit a short circulation half-
life and high tissue permeability and reach a maximum tumor
to background ratio between 2-4 h after injection, facilitating
the use of short half-life isotopes such as the widely available
18F (T1/2 = 109.8 min). Unfortunately, methods for introdu-
cing 18F into proteins are far from satisfactory at present
as relatively mild aqueous reaction conditions are necessary
to preserve the function of most proteins. Hence, existing
Two conventional techniques for incorporating 18F into
proteins are the modification of ε-amino groups of lysine
residues and thiol groups of cysteine residues. Of these
techniques, the most common method reported is the amino-
specific prosthetic group, N-succinimidyl 4-[18F]fluoroben-
zoate ([18F]SFB).5-12 However, [18F]SFB is labile in the basic
aqueous conditions generally required for efficient amino
reactions,13 providing limited yield and specific activity of
the 18F-labeled conjugate. Furthermore, [18F]SFB often
impairs the biological activity of the resulting conjugate,
which may result from a combination of harsh reaction
conditions and the modification of amino groups near the
binding region. For example, 18F-labeled anticarcinoem-
bryonic agent (CEA) diabody was recently produced with
the compromised immunoreactivity (57%) characteristic of
[18F]SFB conjugations.2,9
Site-specific protein modification, where a selected amino
acid distant from the binding region is modified, is generally
preferred over a random modification since it potentially
provides a homogeneous product with unaltered biological
activity. Proteins engineered to contain cysteine, which is
relatively limited within the proteome and primarily present
as a nonreactive disulfide, are well suited for the development
of site-specific conjugates.14-16 A phage display-based assay
was recently developed to identify cysteine substitution sites
*To whom correspondence should be addressed: phone, 650-225-
8821; fax, 650-467-7991; e-mail, marik.jan@gene.com.
a Abbreviations: PET, positron emission tomography; MeCN, acet-
onitrile, DMF, N,N-dimethylformamide; DMSO, dimethyl sulfoxide;
K222, Kryptofix-222; % ID/g, percent ID per gram; CuAAC, copper-
(I)-catalyzed alkyne-azide cycloaddition; BPDS, bathophenanthroline-
disulfonate; TBTA, tris[(1-benzyl-1H-1,2,3-triazol-4-yl)methyl]amine;
17-AAG, 17-(allylamino)-17-demethoxygeldanamycin; PEG or PEGn,
monodisperse polyethylene glycol containing 2-10 ethylene glycol units;
Her2, human epidermal growth factor receptor 2; TOF LC/MS, time-of-
flight liquid chromatography/mass spectrometry; DIAD, diisopropyl
azodicarboxylate; NHS, N-hydroxysuccinimide; ThioMab, monoclonal
antibody containing engineered cysteine for site-specific modification;
ThioFab, antigen binding fragment of ThioMab.
r
pubs.acs.org/jmc
Published on Web 09/09/2009
2009 American Chemical Society

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19 5817
Scheme 1. (a) Convergent Methods for Introducing 18F into Proteins Using 18F-Labeled Prosthetic Groups Produced from Our
Platform and (b) Radiolabeling ThioFab Fragments Using [18F]FPEGMA
that provide superior thiol reactivity without compromising
the structure and function of the conjugate.17 The method was
used to produce a monoclonal antibody (Mab) containing
two reactive cysteine residues (ThioMab),18 from which an
antigen binding fragment (Fab) containing a single reactive
thiol (ThioFab) is conveniently generated. This prompted
the selection of trastuzumab-ThioFab, which is derived
from trastuzumab containing a V110C mutation (Kabat
numbering) in the light chain (trastuzumab-ThioMab), for
the development of site-specific 18F-labeled conjugates.
Prosthetic groups containing maleimide, such as
[18F]FBEM,19 [18F]FBAM,20 [18F]FBABM,15,21 [18F]FBOM,22
[18F]FDG-MHO,23 and [18F]FPyMe,24 may be used to site-
radiochemical yields of 18F-labeled protein. A similar platform
using the widely available radiotracer [18F]FDG instead of
[18F]FBAl was proposed for the production of the water-soluble
maleimide-bearing prosthetic group, [18F]FDG-MHO. How-
ever, [18F]FDG requires a longer synthesis time and provides a
lower radiochemical yield than [18F]FBAl.
An alternative platform using silicon-based fluoride accep-
tor (SiFA) synthons was recently investigated for the devel-
opment of 18F-labeled organosilicon prosthetic groups, such
as the thiol-bearing Si[18F]FA-SH25 and Si[18F]FA-isothio-
cyanate.26 The SiFA platform sets a high standard for total
synthesis time and ease of synthesis primarily by eliminating
the requirement for HPLC purification. However, the SiFA
synthon is highly lipophilic which necessitates the use of
organic cosolvent and may potentially alter the pharmacoki-
netic properties of the radiolabeled biomolecule.
These limitations inspired the development of a new
platform to robustly produce 18F-labeled prosthetic groups
with diverse physicochemical properties (Scheme 1a). The
foundation of this platform is the copper(I)-catalyzed
azide-alkyne cycloaddition (CuAAC).27,28 Since the
CuAAC is orthogonal to most functional groups, the
development of varied prosthetic groups is readily achieved
specifically introduce 18F into
a thiol-bearing protein.
[18F]FBEM, [18F]FBAM, [18F]FBABM, and [18F]FBOM were
developed from a common platform where the aromatic pre-
cursor, 18F-fluorobenzaldehyde ([18F]FBAl), was coupled to an
aminooxy-bearing maleimide precursor. However, the presence
of aromatic and (with the exception of [18F]FBOM) aliphatic
moieties enhances the lipophilicity of these prosthetic groups
which may limit the conjugation efficiency for thiol groups
located in a hydrophilic environment. Moreover, these prosthetic
groups require long synthesis times and typically provide limited

5818 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19
Gill et al.
by coupling modular “building blocks” containing azide or
alkyne functionalities (Scheme 2). Although previously
used for the preparation of 18F-radiolabeled peptides,29
our platform is the first reported set of methods to use the
CuAAC to introduce 18F into proteins.
variety of 18F-labeled prosthetic groups from our platform
was also investigated (Scheme 1a). This includes the thiol-
specific bromoacetamide compound [18F]FPEGBA, the
amino-specific N-hydroxysuccinimide (NHS) compound
[18F]FPEGNHS, the azide-specific compound [18F]FPEG-
propargyl, and the alkyne-specific compound [18F]FPEGN3.
Furthermore, important recommendations are provided for
the general development of 18F-labeled products from our
platform.
A compelling example of our platform is the maleimide-
bearing prosthetic group, [18F]FPEGMA, which was rapidly
produced in a single reaction vessel by a two-step synthesis
(Scheme 1b, approach A). Moreover, polyethylene glycol
(PEG) based building blocks were used to promote the water
solubility and conjugation efficiency of [18F]FPEGMA with
protein in aqueous conditions. [18F]FPEGMA was evaluated
as a prosthetic group with trastuzumab-ThioFab, and the
resulting conjugate (18F-trastuzumab-ThioFab) was vali-
dated in vivo as an imaging agent of the Her2 expression level
in a human tumor xenograft murine model modulated by the
Hsp90 inhibitor, 17-(allylamino)-17-demethoxygeldanamy-
cin (17-AAG).30,31 Her2, a receptor tyrosine kinase of the
epidermal growth factor family, is an important target for the
development of breast cancer therapeutics, and the inhibition
of the chaperone protein Hsp90, which is responsible for
maintaining the stability of a variety of oncoproteins includ-
ing Her2, is a promising therapeutic mode.
Results
Synthesis of Building Blocks. A set of heterobifunctional
polyethylene glycol based building blocks for constructing
18F-labeled prosthetic groups was prepared (Scheme 2). The
details for the synthesis of precursors needed to prepare
[18F]FPEGMA ([18F]5) are outlined in Scheme 3. The synth-
esis of 2 from hexaethylene glycol required two steps; the
Williamson ether synthesis using equimolar amounts of
NaH and propargyl bromide provided a 54% yield of
propargyl-PEG6-OH (1), which was reacted with excess
maleimide by the PPh3/DIAD-mediated Mitsunobu reac-
tion providing a 25% yield of 2. Compound 3 was obtained
from commercially available heterobifunctional HO-PEG8-
N3 and tosyl chloride in pyridine with a 25% yield. Azides 7
and 8 were synthesized in 30% and 19% yields, respectively,
by heating equimolar amounts of NaN3 and the appropriate
PEG-ditosylate in DMF to 110 °C. Bromoacetamide 9 was
prepared from commercially available NH2-PEG7-N3 and
bromoacetyl bromide. Compound 10 was prepared from 1,
and compound 12 was obtained in two steps from diethylene
glycol with propargyl-PEG2-OH (11) as a monopropargy-
lated intermediate. The synthetic procedures for compounds
1-6 are presented in the Experimental Section and the
preparation of intermediates 7-12 is described in detail in
the Supporting Information.
Although the development and biological validation of
[18F]FPEGMA is primarily considered, the production of a
Scheme 2. Alkynes and Azides Used as Building Blocks for
Preparation of 18F-Prosthetic Groups Using CuAAC
Microwave-Assisted Nucleophilic 18F-Fluorination. Tetra-
butylammonium hydrogen carbonate (TBAHCO3) was
selected over a Kryptofix-222/K2CO3 (K222/K2CO3)
system as a phase transfer catalyst for the nucleophilic
18F-fluorination, including [18F]FPEG2N3 (prepared
from 7), [18F]FPEG4N3 (from 8), and [18F]FPEG8N3
([18F]4, obtained from 3). Since TBAHCO3 exhibited
higher thermal stability relative to K222/K2CO3, micro-
wave heating at high temperatures accelerated the azeo-
tropic removal of water to less than 5 min. Additionally,
Scheme 3. Synthesis of [18F]5 and 18F-Labeled trastuzumab-ThioFab ([18F]6, [18F]FPEGMA-trastuzumab-ThioFab)a
a Reaction conditions: (i) NaH, propargyl bromide; (ii) maleimide, PPh3, DIAD; (iii) TsCl, pyridine; (iv) TBAHCO3, [18F]fluoride; (v) CuSO4 5H2O,
BPDS, sodium ascorbate; (vi) phosphate buffer, pH 8.
3

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19 5819
Table 1. Comparison of Three Methods for the CuAAC of [18F]4 and 2, Including Reaction Time, Conversion from [18F]4 to [18F]5, and Purity of Crude
18F]5 with Respect to 18F-Labeled Breakdown Products
[
method
catalysts
time (min)
% conversion
% purity
A
B
C
CuSO4/ascorbate
CuSO4/ascorbate/BPDS
Cu(MeCN)4PF6/TBTA/2,6-lutidine
20
1
79
96
66
89
90
5
100
Table 2. Conjugation Efficiency of [18F]5 (5 mCi, 1 nmol) to trastuzu-
mab-ThioFab (100 μg, 2 nmol) against Protein Concentration, pH, and
Timea
microwave heating provided a high 18F-fluorination effi-
ciency (89.0 ( 1.8%, n = 4) in 3 min for [18F]4 in acetoni-
trile, with similar yields observed for [18F]FPEG4N3.
Decreased yields of [18F]FPEG2N3, which is volatile in
the fluorination and evaporation conditions employed,
were prevented by using resealable reaction vessel caps
and gentle evaporation conditions.
yield (%)
pH
protein concn (mg/mL)
10 min
30 min
60 min
6.5
6.5
6.5
7.2
8.0
0.5
1.0
2.0
1.0
1.0
26
26
47
64
79
39
42
66
74
85
44
53
74
77
85
Optimization of the CuAAC and the Preparation of [18F]5
([18F]FPEGMA). Three catalytic systems were evaluated for
the reaction of [18F]4 and 2 using the CuAAC (Table 1).
a 10 mCi fractions of [18F]5 (in acetonitrile after SPE treatment) were
dispensed into ten reaction vials and evaporated, after which trastuzu-
mab-ThioFab was added.
Method A, which employs CuSO4 5H2O and sodium ascor-
3
bate as a source of CuI,29,32 provided the desired product in
20 min with a 79% conversion efficiency and a 66% purity.
The long reaction time and limited purity observed
prompted the investigation of ligands capable of accelerating
the CuAAC. Method B adds the CuI ligand, bathophenan-
throlinedisulfonate (BPDS), to method A which reduced the
reaction time to 1 min, increased the conversion efficiency to
96%, and improved the purity of crude [18F]5 to 89%. For
methods A and B, the formation of 18F-labeled degradation
products was decreased when the synthesis of [18F]4 was
catalyzed by TBAHCO3 compared to K222/K2CO3 and
when [18F]4 was purified with an Alumina N light cartridge
prior to the CuAAC (data not shown).
purified final product, 18F-trastuzumab-ThioFab, was
analyzed by HPLC system A (Supporting Information
Figure 2b), system C (Supporting Information Figure 2a),
and TOF LC/MS. The optimized conjugation proce-
dure provided 18F-trastuzumab-ThioFab in 30 ( 7%
conjugation efficiency, with greater than 90% radioche-
mical purity, and a specific activity of 2.2 ( 0.2 Ci/μmol
(n=5). The total synthesis time was 82 ( 4 min from the
start of synthesis, and the total decay corrected yield was
13 ( 3%.
Hsp90 Targeted Therapy and MicroPET Imaging. Mice
were distributed into two groups (n = 4) according to tumor
uptake and volume as determined by 18F-trastuzumab-
ThioFab 1 day prior to 17-AAG treatment. 17-AAG was
administered to the treated group at day 0, while the control
group was left untreated. PET imaging with 18F-trastuzu-
mab-ThioFab was performed at day 1 (14 h after the
administration of the last 17-AAG dose), day 5, and day 7.
Coronal slices through the tumor of control and 17-AAG-
treated groups 1 day before and 1 day after therapy are
shown in Figure 1a. The average tumor uptake in control and
treated groups over a 1 week period is shown in Figure 1b,
and the uptake in selected tissues is listed in Table 3. PET
imaging at day 1 revealed a significant reduction in tumor
uptake (0.65 ( 0.09% ID/g) compared to the control group
(1.33 ( 0.14% ID/g, P = 0.00046) and the uptake in the
treated group measured before therapy (1.31 ( 0.13% ID/g,
P = 0.00012) (Table 3). The correction for nonspecific
uptake using muscle as a reference tissue provided a 70%
decrease in 18F-trastuzumab-ThioFab tumor uptake after
therapy. The renal clearance of 18F-trastuzumab-ThioFab
was predominant with high uptake in the renal cortex and
accumulation of radioactivity in the bladder (Table 3).
Although tracer retention in the liver was relatively low,
the hepatobiliary route contributed to tracer excretion with
radioactive metabolites accumulating in the gall bladder and
large intestine.
Method C employs acetonitrile-compatible reagents such
as Cu(MeCN)4PF6 as a source of CuI, tris[(1-benzyl-1H-
1,2,3-triazol-4-yl)methyl]amine (TBTA) as a CuI ligand,
and 2,6-lutidine as a base.33,34 Using method C, the reaction
time (less than 5 min), purity of crude [18F]5 (90%), and
conversion efficiency (100%) were comparable to method
B. Furthermore, alumina purification of [18F]4 prior to the
CuAAC reaction did not impact the purity of crude [18F]5
and was omitted from method C.
After HPLC purification and SPE treatment, method B
provided [18F]5 with a yield of 61 ( 4%, a radiochemical
purity of 94% (Supporting Information Figure 1a), and a
specific activity of 5.6 ( 1.6 Ci/μmol (determined against an
HPLC standard curve of 19F-labeled 5) in 47 ( 2 min (n = 4).
The identity of [18F]5 was confirmed by HPLC coelution
with 19F-labeled standard (Supporting Information Figure
1b) and by LC/MS analysis of the decayed product. The
lipophilicity (log P) of [18F]5 was -2.41 ( 0.09.
Conjugation of [18F]5 to trastuzumab-ThioFab. The con-
jugation efficiency of [18F]5 and trastuzumab-ThioFab was
tested as a function of protein concentration, pH, and time
(Table 2). As the protein concentration and pH were in-
creased, reduced conjugation time and increased efficiency
were observed. The degradation of [18F]5 and aggregation of
trastuzumab-ThioFab were comparable at pH 6.5 and 8.
The purification of 18F-trastuzumab-ThioFab was per-
formed on a NAP-5 desalting column or a Bio-Sep S-2000
SEC-HPLC column (system D). The NAP-5 method pro-
vided coelution of unconjugated [18F]5 with trastuzumab-
ThioFab, particularly for reactions with a low conjuga-
tion efficiency. The Bio-Sep S-2000 SEC-HPLC separated
18F-trastuzumab-ThioFab from 18F-labeled aggregates
and [18F]5 (and 18F-labeled degradation products). The
Discussion
Important development considerations for [18F]FPEGMA
include the use of water-soluble building blocks derived
from PEG precursors, assembly of building blocks by the
CuAAC, and optimization of the site-specific conjugation

5820 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19
Gill et al.
Figure 1. (a) Representative microPET images (coronal slices through tumor) before (day -1) and after (day 1) 17-AAG therapy (top)
compared to a nontreated animal (bottom). (b) Tumor uptake of 18F-trastuzumab-ThioFab for treated and control groups over 7 days.
Table 3. Average Uptake of 18F-trastuzumab-ThioFab in Select Tissues before (day -1) and after (days 1, 5, and 7) Treatment with 17-AAGa
uptake (% ID/g)
17-AAG treated tissue
day -1
day 1
day 5
day 7
bladder
blood
brain
90.76 ( 23.40
0.80 ( 0.06
0.08 ( 0.01
3.59 ( 1.43
15.26 ( 3.02
3.96 ( 2.05
0.62 ( 0.06
0.29 ( 0.03
0.44 ( 0.09
1.31 ( 0.13
63.05 ( 23.08
1.31 ( 0.23
0.12 ( 0.02
4.05 ( 1.86
20.54 ( 4.97
4.05 ( 1.47
0.97 ( 0.08
0.33 ( 0.04
0.50 ( 0.05
0.65 ( 0.09
57.89 ( 9.11
0.97 ( 0.04
0.08 ( 0.01
3.58 ( 0.32
20.48 ( 6.47
3.54 ( 1.61
0.82 ( 0.10
0.25 ( 0.02
0.48 ( 0.02
1.36 ( 0.34
47.11 ( 32.10
0.67 ( 0.07
0.06 ( 0.01
1.97 ( 0.71
23.92 ( 6.98
4.19 ( 0.97
0.54 ( 0.04
0.21 ( 0.04
0.37 ( 0.04
1.66 ( 0.27
gall bladder
kidney
large intestine
liver
muscle
skin
tumor
a Kidney uptake was measured in the cortex, and blood uptake was measured in the left ventricle. Data are presented as mean ( SD.
to trastuzumab-ThioFab with subsequent biological vali-
dation.
[18F]5 was only 66%, which subsequently diminished due to
the instability of [18F]5 in the reaction conditions employed.
The low preparative yield (lessthan20%) provided bymethod
A prompted the investigation of CuI ligands, such as BPDS
and TBTA, which are known to increase the CuAAC reaction
rate and may further enhance the orthogonality of the
CuAAC.33,35,36 Method B, which adds BPDS to the reaction
system of method A, provided a dramatic improvement to the
reaction time, conversion, purity (Table 1), and preparative
yield (61%) of [18F]5. Using method B, the purity of crude
[18F]5 diminished after 1 min, which may be attributed to the
sensitivity of the maleimide moiety to hydrolysis in basic
aqueous conditions.
The development of a new synthetic platform to produce
18F-labeled prosthetic groups for protein conjugations
emerged from our unsuccessful application of the lipophilic
prosthetic group, [18F]FBAM, to trastuzumab-ThioFab.
From this observation, we hypothesized that hydrophilic
prosthetic groups would provide superior conjugation kinetics
and efficiency with trastuzumab-ThioFab and potentially
other hydrophilic proteins. Therefore, building blocks derived
from PEG precursors were investigated to improve water
solubility while maintaining solubility in organic solvents
conventionally employed for nucleophilic 18F-fluorination.
Prosthetic groups derived from our platform (Scheme 1a) were
more water-soluble than compounds derived from the
[18F]FBAl platform. For example, the measured log P of
[18F]5 (-2.41 ( 0.09) is three units lower (more hydrophilic)
than the value reported for [18F]FBOM and over five units
lower than [18F]FBAM.22 Our hypothesis was validated since
the conjugation efficiency of [18F]5 to trastuzumab-ThioFab
was dramatically superior to [18F]FBAM under comparable
conditions.
The limited stability of [18F]5 in the aqueous conditions of
method A or B prompted the development of method C,
which was performed in anhydrous acetonitrile. Method C
provided a higher conversion and more sustainable purity of
crude [18F]5 over time compared to method B. Moreover,
[18F]5 was stable in the presence of basic species, such as
TBAHCO3 and 2,6-lutidine, in anhydrous conditions obviat-
ing the alumina treatment between the 18F-fluorination and
CuAAC reaction steps. Method C also provided a higher yield
than method B for reactants with limited water solubility, such
as [18F]FPEG2N3, and for products which are labile in aqu-
eous conditions, such as [18F]FPEGNHS. However, methodB
was selected for the preparation of [18F]5 due to the extremely
short period of time needed to obtain nearly quantitative
conversion. Nevertheless, a consistent radiochemical process
Perhaps the most compelling design consideration of
[18F]FPEGMA and of our platform in general is the assembly
of azide and alkyne-bearing building blocks by the CuAAC.
Three CuAAC methods were developed to optimize the yield
of [18F]5 (Table 1). Method A provided a 79% conversion of
starting material [18F]4 in 20 min; however, the purity of crude

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19 5821
Scheme 4. Amino-Reactive Prosthetic Group 18FPEGNHS
increase, the formation of trastuzumab-ThioFab dimers or
aggregates did not increase across the range of pH (6.5-8)
and protein concentration (1-2 mg/mL) tested (Table 2).
Regarding the stability of [18F]5 during the conjugation reac-
tion, the formation of degradation products was also indepen-
dent of pH across the range tested (Table 2) or during scaled-up
production. Although [18F]5 was stable during small-scale
conjugations (using 10 mCi to generate Table 2), significant
product degradation was observed during scaled-up production
(more than 200 mCi) which limited the conjugation efficiency
(Supporting Information Figure 2c). The dependence of [18F]5
degradation on total radioactivity coupled with the formation
of a broad spectrum of byproducts may indicate that [18F]5 is
susceptible to radiolysis which is characteristic for maleimide
compounds.40,41 Irrespective of the decomposition mechanism,
the conjugation conditions were selected to maximize the
reaction rate while ensuring that limited aggregates are formed.
Considering the pH independence of the reactants’ stability,
basic conditions (pH 8) provided acceptable conjugation effi-
ciency (30%) for 18F-trastuzumab-ThioFab. However, the
complete degradation of [18F]5 ultimately reduced the conjuga-
tion efficiency and specific activity of 18F-trastuzumab-Thio-
Fab. Despite this limitation, a high specific activity of 18F-
trastuzumab-ThioFab was obtained, which may reduce the
potential for receptor saturation and improve target uptake,
particularly for cells with low receptor density. Notably, the
yield (13%), specific activity (2.2 ( 0.2 Ci/μmol), and total
synthesis time (82 ( 4 min) of 18F-trastuzumab-ThioFab are
favorable when compared against 18F-labeled anti-CEA dia-
body (1.4% yield and 0.05 Ci/μmol specific activity in 173 min
using [18F]SFB).9
and Thiol-Reactive Prosthetic Group 18FPEGBA
may be applied when producing varied 18F-labeled prosthetic
groups by either method B or method C. This has strong
implications for automation since a common reaction plat-
form eliminates the dedication of a synthesis module for a
specific prosthetic group, such as [18F]SFB.
The orthogonality of the CuAAC enabled the development
of multiple prosthetic groups capable of introducing 18F
into proteins (Scheme 1a). Method C was used to produce
[18F]FPEGNHS (Scheme 4) from a commercially available
precursor (N3-PEG4-NHS), and the product was subsequently
conjugated to the Fab fragment of trastuzumab (trastuzumab-
Fab, Supporting Information). [18F]FPEGNHS may limit the
reduction in immunoreactivity characteristic of amino-specific
modifications through a combination of high specific activity
and limited steric hindrance within the binding region (smaller
building blocks are under development). Therefore, [18F]-
FPEGNHS has strong potential as an alternative to
[18F]SFB for performing amino-specific modifications.
We have also developed the thiol-specific compound, [18F]-
FPEGBA (Scheme 4), using method B (Supporting Infor-
mation); however, the conjugation efficiency of [18F]-
FPEGBA with trastuzumab-ThioFab at pH 8 was insuffi-
cient for routine application.
After optimizing the conjugation of [18F]5 to trastuzumab-
ThioFab, the biological activity of 18F-trastuzumab-ThioFab
was validated by Scatchard analysis. The affinity of the
conjugate (KD = 11.9 ( 2.3 nM) was comparable to native
trastuzumab-ThioFab (9.6 ( 1.2 nM). Additionally, trastu-
zumab-ThioFab retained a comparable binding affinity to the
parent trastuzumab-Fab (8.3 ( 1.0 nM), indicating that the
modification of trastuzumab-ThioFab with [18F]5 did not
alter its Her2 bindingproperties. Further biologicalvalidation
of 18F-trastuzumab-ThioFab included a treatment response
study of a Her2-overexpressing tumor xenograft in a murine
model modulated by the Hsp90 inhibitor, 17-AAG.
The effect of the Hsp90 inhibitor, 17-AAG, on Her2-
positive tumors has been characterized.42 Immuno-PET was
recently employed to monitor the effect of 17-AAG on Her2
expression level in vivo using a 68Ga-labeled DOTA-conju-
gated Herceptin F(ab0)2 fragment (68Ga-DCHF).30,31 Smith-
Jones measured a 70% reduction in Her2 levels lasting 5 days
followed by a return to pretreatment levels after 12 days.
These data suggest that 68Ga-DCHF is capable of monitoring
tumor response to 17-AAG treatment earlier than 18FDG
PET. However, the short half-life of 68Ga (T1/2 = 68 min) is
not well suited for the plasma half-life of F(ab)02, resulting in a
signal loss during the 3 h period required prior to PET
acquisition. Therefore, improved imaging properties were
expected by coupling a shorter plasma half-life species, such
as monovalent Fab fragments, with an appropriate radio-
nuclide, such as 18F3.
The CuAAC was also used to directly label proteins
with [18F]FPEG-propargyl and [18F]FPEGN3 (Scheme 1b,
approach B), reducing the synthesis of 18F-labeled protein to
two radiochemical steps and providing an alternative method
for site-specific modification. For example, preliminary tests
reacting [18F]4 with propargyl-PEG6-trastuzumab-ThioFab
provided conjugation efficiencies up to 15% (Supporting
Information). A similar method was employed to couple
[18F]FPEG3-propargyl to an azide-PEG-modified nanoparti-
cle with a high conjugation efficiency due to the presence of
multiple azide sites.37 This highlights the high conjugation
efficiency and specific activity possible for 18F-labeled pro-
teins since multiple azide or alkyne groups can be positioned
on a protein without cross-reactivity. However, since the
combination of CuII and ascorbate tends to degrade pro-
teins,38 the CuAAC methods that we have developed are not
directly applicable to proteins. Instead, we continue to investi-
gate methods including CuI with the rigorous exclusion of
oxygen, CuII mixed with reducing agents, or CuI associated
with BPDS which may enhance the bioorthogonality of the
CuAAC with proteins.39 Nevertheless, [18F]FPEG-propargyl
and [18F]FPEGN3 have demonstrated promising results
as prosthetic groups for the development of 18F-labeled pep-
tides using method B with minor modifications (Supporting
Information).
A consistent response in Her2 levels to 17-AAG was observed
whether measured by 18F-trastuzumab-ThioFab (Figure 1)
or 68Ga-DCHF.30,31 The 70% decrease in tumor uptake of
18F-trastuzumab-ThioFab observed after therapy was identi-
cal to that measured by Smith-Jones using 68Ga-DCHF.30
The site-specific conjugation of [18F]5 to trastuzumab-Thio-
Fab was optimized with respect to the reactant stability and
reaction rate. Regarding the stability of thiol-bearing proteins,
which generally decreases as pH and protein concentration

5822 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19
Gill et al.
Additionally, since 18F-trastuzumab-ThioFab (Figure 1a) was
imaged only 1 half-life after injection, the image quality may
compare favorably to 68Ga-DCHF (imaged 3 half-lives after
injection).
TMS, and the 19F chemical shifts are reported using TFA as an
external reference standardized to -78.5 ppm. A Model 521
microwave heater (Resonance Instruments, Skokie, IL) was
used for radiochemical reactions. TBTA was synthesized ac-
cording the previously published procedure in 40% yield.33 The
lipophilicity (log P) was measured at pH 7.4 using a previously
described method suitable for radiopharmaceuticals.43
In summary, we have developed a modular platform for the
facile preparation of 18F-labeled prosthetic groups assembled
from building blocks using the CuAAC reaction. The applica-
tion of our platform was demonstrated by [18F]FPEGMA,
which was used to site-specifically modify trastuzumab-
ThioFab. Generally, the methods described herein for produ-
cing 18F-radiolabeled proteins provide high radiochemical
yields and specific activity with a short synthesis time. 18F-
trastuzumab-ThioFab is a viable alternative to 68Ga-DCHF
for the treatment response monitoring of Her2 levels. Ulti-
mately, our platform may be an important tool for the
discerning radiochemist and may simplify radiochemical pro-
cess development by providing a reliable, convenient, and
consistent method for producing diverse 18F-labeled prosthe-
tic groups.
3,6,9,12,15,18-Hexaoxahenicos-20-yn-1-ol (1). Sodium hydride,
60% dispersion in mineral oil (0.86 g, 21.4 mmol), was added in
portions to the solution of hexaethylene glycol (5.5 g, 19.4 mmol)
in anhydrous THF (40 mL) at 0 °C. The reaction mixture was
stirred at 0 °C for 15 min, and then propargyl bromide, 80% in
toluene (2.4 mL, 21.4 mmol), was added dropwise. The mixture
was allowed to warm to room temperature and stirred for 3 h. The
formed sodium bromide was removed by filtration, and the
solvent was evaporated. The crude product was purified on a
SiO2 column using a 0-100% methanol/dichloromethane gradi-
1
ent to yield 3.4 g (54%) of colorless oil. H NMR (500 MHz,
CDCl3) δ 2.45 (t, J = 2.4 Hz, 1H), 2.74 (br s, 1H), 3.61-3.71 (m,
24H), 4.21 (d, J = 2.4 Hz, 2H); 13C NMR (100.6 MHz, CDCl3)
δ 58.4, 61.7, 69.1, 70.3-70.6, 72.5, 74.5, 79.7; MS ESI (m/z)
[M þ H]þ calcd for C15H29O7, 321.38; found, 321.4.
1-(3,6,9,12,15,18-Hexaoxahenicos-20-yn-1-yl)-2,5-pyrroledione
(2). Diisopropyl azodicarboxylate (0.73 mL, 3.43 mmol) was
added dropwise into the ice-cooled solution of 1 (1000 mg,
3.12 mmol), triphenylphosphine (900 mg, 3.43 mmol), and mal-
eimide (456 mg, 4.70 mmol) in anhydrous THF (20 mL) under
nitrogen atmosphere. The resulting brown solution was allowed
to warm to room temperature and was stirred for 1 h at room
temperature. The solvent was evaporated at a reduced pressure,
and the crude product was purified on a SiO2 column using a
0-100% hexanes/ethyl acetate gradient to yield 650 mg of yellow
oil which was subsequently purified by semipreparative HPLC
(systemE) toprovide 310 mg (25%) ofcolorlessoilproduct free of
triphenylphosphine oxide. 1H NMR (400 MHz, CDCl3) δ 2.43 (t,
J = 2.36 Hz, 1H), 3.60-3.72 (m, 24H), 4.20 (d, J = 2.37 Hz, 2H),
6.70 (s, 2H); 13C NMR (100.6 MHz, CDCl3) δ 37.2, 58.4,
67.8, 69.1, 70.1, 70.4-70.6, 74.5, 79.7, 134.1, 170.6; MS ESI (m/z)
[M þ H]þ calcd for C19H30NO8, 400.19; found, 400.0.
Experimental Section
General. Solvents and chemicals were purchased from Al-
drich (Milwaukee, WI) unless stated otherwise. Heterobifunc-
tional polyethylene glycols were purchased from Iris Biotech
(Marktredwitz, Germany), and N3-PEG8-NHS was purchased
from Quanta Biodesign (Powell, OH). Egg phospholipid was
purchased from Avanti Polar lipids (Alabaster, AL), and
17-(allylamino)-17-demethoxygeldanamycin (17-AAG) was ob-
tained from InvivoGen (San Diego, CA). [18F]Fluoride was
purchased from PETNET Solutions (Palo Alto, CA), 18F trap
and release columns (8 mg) were purchased from ORTG, Inc.
(Oakdale, TN), and Strata-X 60 mg cartridges were purchased
from Phenomenex (Torrance, CA). The following reversed-
phase HPLC systems were used to determine purity and purify
˚
the products. System A: Phenomenex Jupiter C18 300 A (150 ꢀ
4.6 mm, 5 μm), 0.05% TFA þ 10-90% acetonitrile, 0-7 min,
90% acetonitrile, 7-10 min, 2 mL/min. System B: Phenomenex
Luna C18 (250 ꢀ 10 mm, 5 μm) 0.05% TFA þ 20% acetonitrile,
7 mL/min. System C: Phenomenex BioSep-SEC-S 2000 (300 ꢀ
4.60 mm, 5 μm) 50 mM PBS, pH 7.2, 0.5 mL/min. System D:
Phenomenex BioSep-SEC-S 2000 (300 ꢀ 7.80 mm, 5 μm) 20 mM
PBS, pH 7.2, 1.0 mL/min. System E: Altima C-18 (100 ꢀ 22.0
mm, 5 μm) 0.05% TFA þ 10-60% acetonitrile, 0-30 min,
24 mL/min. System F: Agilent Eclipse XDB-C18 (150 ꢀ 4.6 mm,
5 μm) 0.05% TFA þ 10-90% acetonitrile, 0-10 min, 1 mL/
min, 50 °C. The analytical and semipreparative HPLC systems
used for radiochemistry were equipped with an UV absorbance
and radioactivity detector (PMT); the purity of radioactive
products was determined by detecting UV absorbance at 214,
254, and 280 nm and radioactivity. Mass spectrometry and
purity analysis of low molecular weight products were per-
formed on a PE Sciex API 150EX LC-MS system equipped
with an Onyx Monolithic C18 column. The purity of low
molecular weight compounds determined at UV 214 and
254 nm was g95% if not stated otherwise. LC-MS analysis of
23-Azido-3,6,9,12,15,18,21-heptaoxatricos-1-yl p-Toluenesul-
fonate (3). The solution of 23-azido-3,6,9,12,15,18,21-heptaox-
atricosan-1-ol (800 mg, 2.02 mmol) in pyridine (4 mL) was
added dropwise to the solution of toluenesulfonyl chloride
(771 mg, 4.04 mmol) and 4-dimethylaminopyridine (100 mg)
in pyridine (10 mL) at room temperature. The mixture was
stirred for 24 h at room temperature. The solvent was evapo-
rated, and the crude product was purified on a SiO2 column
using a 0-100% hexanes/ethyl acetate gradient, followed by a
0-100% ethyl acetate/methanol gradient, and then purified on
a semipreparative HPLC (system E) to yield 280 mg (25%) of
colorless oil. 1H NMR (400 MHz, CDCl3) δ 2.45 (s, 3H), 3.38 (t,
J = 5.02 Hz, 2H), 3.58-3.70 (m, 28H), 4.16 (t, J = 5.00 Hz, 2H),
7.34 (d, J = 8.4 Hz, 2H), 7.80 (d, J = 8.4 Hz, 2H); 13C NMR
(100.6 MHz, CDCl3) δ 21.6, 50.7, 68.7, 69.2, 70.0, 70.5-70.8,
128.0, 129.7, 133.1, 144.7; MS ESI (m/z) [M þ H]þ calcd for
C23H39N3O10S, 549.24; found, 549.9.
23-Azido-1-fluoro-3,6,9,12,15,18,21-heptaoxatricosane (4).
The 23-azido-3,6,9,12,15,18,21-heptaoxatricosan-1-ol (500 mg,
1.26 mmol) was dissolved in DCM (2 mL) and cooled to -10 °C.
DAST (0.48 mL, 3.78 mmol) was added in portions to the cooled
solution, and the resulting mixture was stirred at -10 °C for
30 min, then allowed to warm to room temperature, and stirred
for an additional 20 h. The excess of DAST was quenched with
methanol (5 mL), and the solvents were subsequently evaporated
at a reduced pressure. The oily residue was dissolved in water
(3 mL) and adjusted to pH 5 using NaHCO3. The crude product
was purified bysemipreparative HPLC (system E) toyield122mg
(24%) ofyellowishoil. 1H NMR(500 MHz, CDCl3) δ3.39(t, J =
5.0 Hz, 2H), 3.65-3.68 (m, 26H), 3.75 (dt, J = 4.2 Hz, 29.6 Hz,
2H), 4.56 (dt, J = 4.0 Hz, 47.5 Hz, 2H); 13C NMR (125.7 MHz,
˚
proteins was performed on a PRLP-S, 1000 A, microbore
column (50 mm ꢀ 2.1 mm; Polymer Laboratories, Palo Alto,
CA) heated to 75 °C coupled to a TSQ Quantum Triple
quadrupole mass spectrometer with extended mass range
(Thermo Electron, Waltham, MA). A linear gradient from
30% to 40% B (solvent A, 0.05% TFA in water; solvent B,
0.04% TFA in acetonitrile) was used, and the eluant was directly
ionized using the electrospray source. Data were collected by the
Xcalibur data system and analyzed using ProMass (Novatia,
Monmouth Junction, NJ). NMR spectra were recorded on a
Bruker Avance II 500 or a Bruker Avance II 400 spectrometer at
298 K. The 1H and 13C chemical shifts are reported relative to

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19 5823
CDCl3) δ 50.8, 70.1, 70.4, 70.5, 70.7-70.8, 70.9, 83.2 (d, J =
169 Hz); 19F NMR (470.6 MHz, CDCl3) δ -225.9; MS ESI (m/z)
[M þ H]þ calcd for C16H33FO7N3, 398.2; found, 398.0.
Workup Procedure for Methods A-C. The crude product
[18F]5 was purified by semipreparative HPLC (system B) with a
retention time of 15-18 min. The collected fraction was
promptly diluted to 20 mL with H2O and loaded on a Phenom-
enex Strata-X 60 mg SPE cartridge, and the product trapped on
the cartridge was rinsed with H2O (10 mL), flushed with air,
eluted with acetonitrile (0.5 mL) into a 1 mL v-vial, and
evaporated at 30 °C to near dryness with 600 cm3 of argon.
MS ESI (m/z) [M þ H]þ calcd for C35H62FO15N7, 797.4; found,
797.4; HPLC (System A) retention time 2.98 min (Supporting
Information Figure 1a).
1-(1-Fluoro-3,6,9,12,15,18,21-heptaoxatricos-23-yl)-4-(1-(2,5-
pyrrolidone-1-yl)-3,6,9,12,15,18-hexaoxanonadeca-19-yl)-1,2,3-tria-
zole (5, FPEGMA). Cu(MeCN)4PF6 (7.5 mg, 0.02 mmol) dissolved
in acetonitrile (0.1 mL) was added at room temperature to the
solution of 2 (8.0 mg, 0.02 mmol), 4 (8.0 mg, 0.02 mmol), 2,6-
lutidine (2.5 mL, 0.02 mmol), and TBTA (2.0 mg, 0 004 mmol) in
acetonitrile (0.2 mL). The resulting mixture was stirred for 22 h at
room temperature and then diluted with water to 4 mL, and the
formed precipitate was removed by filtration. The filtrate contain-
ing the product was purified by semipreparative HPLC (system E)
[18F]FPEGMA-trastuzumab-ThioFab (6). Trastuzumab-
ThioFab (1.0 mg, 21 nmol) in 100 mM sodium phosphate, pH
8 (0.5 mL, 2 mg/mL), was added to [18F]5 and shaken on a test
tube rocker for 10 min. The crude product was purified by a
semipreparative SEC-HPLC column (system D, 10-12 min
retention time) and, if necessary, concentrated to desired volume
by centrifugation on an Amicon 10 kDa membrane. HPLC
(system A) retention time 3.3 min (Supporting Information
Figure 2b); SEC HPLC (system C) retention volume 3.0 mL
(Supporting Information Figure 2a).
1
to provide 7.5 mg (47%) of colorless oil. H NMR (500 MHz,
CDCl3) δ3.58-3.80 (m, 50H), 3.87 (t, J = 4.5 Hz, 2H), 4.53 (t, J =
5.0 Hz, 2H), 4.55 (dt, J = 47.3, 4.2 Hz, 2H), 4.68 (s, 2H), 6.70 (s,
2H), 7.73 (s, 1H); 13C NMR (125.7 MHz, CDCl3) δ 37.2, 50.3, 64.7,
67.9, 69.5, 69.7, 70.2, 70.6-70.7, 70.9, 83.2 (d, J = 169 Hz), 123.8,
134.2, 145.0, 170.7; 19F NMR (470.6 MHz, CDCl3) δ -225.8; MS
ESI (m/z) [M þ H]þ calcd for C35H62FO15N7,797.4; found, 797.4;
HPLC (system A) retention time 2.91 min.
1-[18F]Fluoro-3,6,9,12,15,18-hexaoxahenicos-20-azide ([18F]4).
[18F]Fluoride was eluted from a T&R cartridge into a 4 mL v-vial
using 10 μL of 1.3 M TBAHCO3 in 0.6 mL of acetonitrile/water
2:1 (v/v). Water was azeotropically removed by microwave heat-
ing at 60 W, 120°C, under a stream of argon (600 cm3) followedby
the addition of anhydrous acetonitrile (4 ꢀ 0.7 mL). Precursor 3
(2 mg, 3.6 μmol) dissolved in acetonitrile (0.6 mL) was added, and
the mixture was microwave heated in a septum-sealed vessel at
40 W, 120 °C, for 3 min. The reaction mixture was cooled to room
temperature, and the crude product was used with only partial or
without any further purification for CuAAC as described in the
next paragraph. HPLC-based radiochemical yield of [18F]4 was
86 ( 8% and HPLC retention time 7.5 min (system F).
17-AAG Formulation and Administration. Egg phospholipid
(0.5 g) in DCM (5 mL) was transferred into a 500 mL round-
bottom flask, the solvent was evaporated, and the residue was
stored at the reduced pressure for 16 h to remove remaining
solvent. The 5% (m/v) solution of dextrose (25 mL) was added
to the dried phospholipids, and the mixture was sonicated for
30 min at room temperature to obtain a milky emulsion.
17-AAG (25 mg) was dissolved in DMSO (1 mL) from which
an aliquot (0.25 mL) was mixed with the egg phospholipid
emulsion (4.8 mL) and sonicated at room temperature for
15 min. The resulting emulsion was administered within 30 min
of preparation by intraperitoneal injection in three 1 mL doses
(50 mg/kg each) over a 24 h period as described previously.30,31
Animal Mmodel. Beige nude XID mice of age 6-8 weeks were
obtained from Harlan Sprague-Dawley (Livermore, CA). Three
days prior to cell inoculation, the mice were implanted (sc, left
flank) with 0.36 mg, 60 day sustained release 17β-estradiol
pellets (Innovative Research of America) to maintain serum
estrogen level. Mice were inoculated in the mammary fat pad
with 5 ꢀ 106 BT474M1 cells in 50% phenol red-free matrigel.
BT474M1 is a subclone of breast cancer cell line BT474 that was
obtained from California Pacific Medical Center. Animal care
and treatment followed protocols approved by Genentech’s
Institutioned Animal Care and Use Committee which is accre-
dited by the Association for Assessment and Accreditation of
Laboratory Animal Care (AAALAC).
MicroPET Imaging. The images were acquired on a micro-
PET Focus 120 scanner (Siemens Preclinical Solutions). Mice
were anesthetized with 3% sevoflurane and inoculated iv via the
tail vein catheter with 0.3-0.4 mCi of 18F-trastuzumab-Thio-
Fab in isotonic solution (50-100 μL). The animals were allowed
to recover on a heated blanket until ambulatory and then
returned to the cage. After 2 h of conscious uptake the animals
were anesthetized with 3% sevoflurane and placed head-first,
prone position on the scanner bed. Body temperature was
measured by a rectal probe and maintained by warm air.
Dynamic 60 min scans were acquired. Full body image recon-
structions were obtained using maximum a posteriori algorithm
(MAP), and maximum intensity projections (MIPs) were cre-
ated with ASIPro software (CTI Molecular Imaging). MAP
reconstructed images were used to obtain quantitative activity
levels (presented as % ID/g) in each organ of interest using
ASIPro software (CTI Molecular Imaging).
1-(1-[18F]Fluoro-3,6,9,12,15,18,21-heptaoxatricos-23-yl)-4-(1-(2,5-
pyrrolidone-1-yl)-3,6,9,12,15,18-hexaoxanonadeca-19-yl)-1,2,3-tria-
zole ([18F]5, [18F]FPEGMA). Method A. The reaction mixture
containing crude [18F]4 was passed across a Sep-Pak Alumina N
light cartridge, the product [18F]4 was eluted with acetonitrile
(0.5 mL), the solvent was evaporated to near dryness, and vial
was cooled to 30 °C. Compound 2 (4 mg, 10.0 μmol) dissolved in
acetonitrile (0.2 mL) was added to [18F]4, followed by a solution of
freshly prepared sodium ascorbate (7.5 mg, 40 μmol) in 0.1 mL of
Tris buffer, pH 8, and CuSO4 5H2O (2.2 mg, 8 μmol) in Tris
3
buffer, pH 8 (0.2 mL). The resulting brown-orange solution was
stirred at room temperature for 20 min, followed by the addition of
H2O with 0.1% TFA (1 mL), and delivered to the HPLC loop for
purification.
Method B. The reaction mixture containing crude [18F]4 was
passed across a Sep-Pak Alumina N light cartridge, the product
[18F]4 was eluted with acetonitrile (0.5 mL), the solvent was
evaporated to near dryness, and the vial was cooled to 30 °C.
Compound 2 (4 mg, 10.0 μmol) dissolved in acetonitrile (0.2 mL)
was added to [18F]4, followed by freshly prepared sodium
ascorbate (7.5 mg, 40 μmol) in 0.1 mL of Tris buffer, pH 8,
and a mixture of CuSO4 5H2O (2.2 mg, 8 μmol) and bath-
3
ophenanthrolinedisulfonate (BPDS) (4.4 mg, 8 μmol) in Tris
buffer, pH 8 (0.2 mL). The resulting brown-orange solution was
stirred at room temperature for 1 min, followed by the addition
of H2O with 0.1% TFA (1 mL), and delivery to the HPLC loop
for purification.
Method C. The reaction mixture containing crude [18F]4 was
evaporated to near dryness (no alumina treatment required).
Compound 2 (4 mg, 10.0 μmol) dissolved in acetonitrile (0.1 mL)
was added, followed by a freshly prepared mixture of Cu-
(MeCN)4PF6 (3 mg, 8 μmol), TBTA (4.3 mg, 8 μmol), and
25 μL of 2,6-lutidine in acetonitrile (0.2 mL). The resulting light
yellow solution was stirred for 5 min at room temperature,
concentrated to 100 μL, diluted with H2O with 0.1% TFA
(1 mL), and delivered to the HPLC loop for purification.
Statistical Analysis. The graphs were constructed with R
software version 2.4.1 (R Foundation for Statistical Computing,
Vienna, Austria). Statistical significance was determined using a
two-tailed Student’s t-test, and P values of less than 0.05 were
considered significant; data are presented as mean ( SEM if not
stated otherwise.

5824 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19
Gill et al.
Acknowledgment. We thank Nina Aggarwal for acquiring
mass spectrometry data, Elizabeth Luis for performing Scatch-
ard affinity analysis, George Dutina’s and Phil Hass’ groups
for help with expression and purification of trastuzumab-
ThioMab, Jagath Reddy Junutula for providing the trastuzu-
mab-ThioMab construct, and Nick van Bruggen, Joan Greve,
Leslie Khawli for insightful discussions.
(17) Junutula, J. R.; Bhakta, S.; Raab, H.; Ervin, K. E.; Eigenbrot, C.;
Vandlen, R.; Scheller, R. H.; Lowman, H. B. Rapid identification
of reactive cysteine residues for site-specific labeling of antibody-
Fabs. J. Immunol. Methods 2008, 332, 41–52.
(18) Junutula, J. R.; Raab, H.; Clark, S.; Bhakta, S.; Leipold, D. D.;
Weir, S.; Chen, Y.; Simpson, M.; Tsai, S. P.; Dennis, M. S.; Lu, Y.;
Meng, Y. G.; Ng, C.; Yang, J.; Lee, C. C.; Duenas, E.; Gorrell, J.;
Katta, V.; Kim, A.; McDorman, K.; Flagella, K.; Venook, R.;
Ross, S.; Spencer, S. D.; Lee Wong, W.; Lowman, H. B.; Vandlen,
R.; Sliwkowski, M. X.; Scheller, R. H.; Polakis, P.; Mallet, W. Site-
specific conjugation of a cytotoxic drug to an antibody improves
the therapeutic index. Nat. Biotechnol. 2008, 26, 925–932.
(19) Cai, W.; Zhang, X.; Wu, Y.; Chen, X. A thiol-reactive 18F-labeling
agent, N-[2-(4-18F-fluorobenzamido)ethyl]maleimide, and synth-
esis of RGD peptide-based tracer for PET imaging of alpha v beta 3
integrin expression. J. Nucl. Med. 2006, 47, 1172–1180.
Supporting Information Available: Synthetic protocols for
additional building blocks 7-12, method for preparation of
18FPEGNHS and 18FPEGBA, procedure for preparation of
ThioFabs, optimized method for labeling peptides using
CuAAC, method for direct conjugation of 4 to proteins using
CuAAC, and supplementary Figures 1 and 2. This material is
available free of charge via the Internet at http://pubs.acs.org.
(20) Berndt, M.; Pietzsch, J.; Wuest, F. Labeling of low-density lipo-
proteins using the 18F-labeled thiol-reactive reagent N-[6-(4-[18F]-
fluorobenzylidene)aminooxyhexyl]maleimide. Nucl. Med. Biol.
2007, 34, 5–15.
(21) Toyokuni, T.; Walsh, J. C.; Dominguez, A.; Phelps, M. E.; Barrio,
J. R.; Gambhir, S. S.; Satyamurthy, N. Synthesis of a new hetero-
bifunctional linker, N-[4-(aminooxy)butyl]maleimide, for facile
access to a thiol-reactive 18F-labeling agent. Bioconjugate Chem.
2003, 14, 1253–1259.
References
(1) van Dongen, G. A.; Visser, G. W.; Lub-de Hooge, M. N.; de Vries,
E. G.; Perk, L. R. Immuno-PET: a navigator in monoclonal
antibody development and applications. Oncologist 2007, 12,
1379–1389.
(22) Wuest, F.; Kohler, L.; Berndt, M.; Pietzsch, J. Systematic compar-
ison of two novel, thiol-reactive prosthetic groups for 18F labeling
of peptides and proteins with the acylation agent succinimidyl-
4-[18F]fluorobenzoate ([18F]SFB). Amino Acids 2009, 36, 283–295.
(23) Wuest, F.; Berndt, M.; Bergmann, R.; van den Hoff, J.; Pietzsch, J.
(2) Shively, J. E. 18F labeling for immuno-PET: where speed and
contrast meet. J. Nucl. Med. 2007, 48, 170–172.
(3) Williams, L. E.; Wu, A. M.; Yazaki, P. J.; Liu, A.; Raubitschek,
A. A.; Shively, J. E.; Wong, J. Y. Numerical selection of optimal
tumor imaging agents with application to engineered antibodies.
Cancer Biother. Radiopharm. 2001, 16, 25–35.
(4) Holliger, P.; Hudson, P. J. Engineered antibody fragments and the
rise of single domains. Nat. Biotechnol. 2005, 23, 1126–1136.
(5) Vaidyanathan, G.; Bigner, D. D.; Zalutsky, M. R. Fluorine-18-
labeled monoclonal antibody fragments: a potential approach for
combining radioimmunoscintigraphy and positron emission tomo-
graphy. J. Nucl. Med. 1992, 33, 1535–1541.
(6) Vaidyanathan, G.; Zalutsky, M. R. Labeling proteins with fluor-
ine-18 using N-succinimidyl 4-[18F]fluorobenzoate. Int. J. Radiat.
Appl. Instrum. B 1992, 19, 275–281.
(7) Vaidyanathan, G.; Zalutsky, M. R. Improved synthesis of
N-succinimidyl 4-[18F]fluorobenzoate and its application to the
labeling of a monoclonal antibody fragment. Bioconjugate Chem.
1994, 5, 352–356.
(8) Wester, H. J.; Hamacher, K.; Stocklin, G. A comparative study of
N.C.A. fluorine-18 labeling of proteins via acylation and photo-
chemical conjugation. Nucl. Med. Biol. 1996, 23, 365–372.
(9) Cai, W.; Olafsen, T.; Zhang, X.; Cao, Q.; Gambhir, S. S.; Williams,
L. E.; Wu, A. M.; Chen, X. PET imaging of colorectal cancer in
xenograft-bearing mice by use of an 18F-labeled T84.66 anti-
carcinoembryonic antigen diabody. J. Nucl. Med. 2007, 48, 304–
310.
(10) Grierson, J. R.; Yagle, K. J.; Eary, J. F.; Tait, J. F.; Gibson, D. F.;
Lewellen, B.; Link, J. M.; Krohn, K. A. Production of [F-18]-
fluoroannexin for imaging apoptosis with PET. Bioconjugate
Chem. 2004, 15, 373–379.
Synthesis and application of [
18F]FDG-maleimidehexyloxime
([18F]FDG-MHO): a [18F]FDG-based prosthetic group for the
chemoselective 18F-labeling of peptides and proteins. Bioconjugate
Chem. 2008, 19, 1202–1210.
(24) de Bruin, B.; Kuhnast, B.; Hinnen, F.; Yaouancq, L.; Amessou,
M.; Johannes, L.; Samson, A.; Boisgard, R.; Tavitian, B.; Dolle, F.
1-[3-(2-[18F]fluoropyridin-3-yloxy)propyl]pyrrole-2,5-dione: design,
synthesis, and radiosynthesis of a new [18F]fluoropyridine-based
maleimide reagent for the labeling of peptides and proteins. Biocon-
jugate Chem. 2005, 16, 406–420.
(25) Wangler, B.;Quandt, G.;Iovkova, L.;Schirrmacher, E.;Wangler, C.;
Boening, G.; Hacker, M.; Schmoeckel, M.; Jurkschat, K.; Barten-
stein, P.; Schirrmacher, R. Kit-like 18F-labeling of proteins: synthesis
of 4-(di-tert-butyl[18F]fluorosilyl)benzenethiol (Si[18F]FA-SH)
labeled rat serum albumin for blood pool imaging with PET.
Bioconjugate Chem. 2009, 20, 317–321.
(26) Rosa-Neto, P.; Wangler, B.; Iovkova, L.; Boening, G.; Reader, A.;
Jurkschat, K.; Schirrmacher, E. [18F]SiFA-isothiocyanate: a new
highly effective radioactive labeling agent for lysine-containing
proteins. ChemBioChem 2009, 10, 1321–1324.
(27) Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A
stepwise huisgen cycloaddition process: copper(I)-catalyzed regio-
selective “ligation” of azides and terminal alkynes. Angew. Chem.,
Int. Ed. Engl. 2002, 41, 2596–2599.
(28) Tornoe, C. W.; Christensen, C.; Meldal, M. Peptidotriazoles on
solid phase: [1,2,3]-triazoles by regiospecific copper(I)-catalyzed
1,3-dipolar cycloadditions of terminal alkynes to azides. J. Org.
Chem. 2002, 67, 3057–3064.
(11) Page, R. L.; Garg, P. K.; Garg, S.; Archer, G. E.; Bruland, O. S.;
Zalutsky, M. R. PET imaging of osteosarcoma in dogs using a
fluorine-18-labeled monoclonal antibody Fab fragment. J. Nucl.
Med. 1994, 35, 1506–1513.
(29) Marik, J.; Sutcliffe, J. L. Click for PET: rapid preparation of
[
18F]fluoropeptides using Cu(I) catalyzed 1,3-dipolar cycloaddi-
tion. Tetrahedron Lett. 2006, 47, 6681–6684.
(12) Page, R. L.; Garg, P. K.; Vaidyanathan, G.; Zalutsky, M. R.
Preclinical evaluation and PET imaging of 18F-labeled Mel-14
F(ab0)2 fragment in normal dogs. Nucl. Med. Biol. 1994, 21, 911–
919.
(13) Li, J.; Trent, J. O.; Bates, P. J.; Ng, C. K. Labeling G-rich
oligonucleotides (GROs) with N-succinimidyl 4-[18F]fluoro-
benzoate (S18FB). J. Labelled Compd. Radiopharm. 2006, 49,
1213–1221.
(14) Olafsen, T.; Cheung, C. W.; Yazaki, P. J.; Li, L.; Sundaresan, G.;
Gambhir, S. S.; Sherman, M. A.; Williams, L. E.; Shively, J. E.;
Raubitschek, A. A.; Wu, A. M. Covalent disulfide-linked anti-
CEA diabody allows site-specific conjugation and radiolabeling
for tumor targeting applications. Protein Eng. Des. Sel. 2004, 17,
21–27.
(30) Smith-Jones, P. M.; Solit, D.; Afroze, F.; Rosen, N.; Larson, S. M.
Early tumor response to Hsp90 therapy using HER2 PET: com-
parison with 18F-FDG PET. J. Nucl. Med. 2006, 47, 793–796.
(31) Smith-Jones, P. M.; Solit, D. B.; Akhurst, T.; Afroze, F.; Rosen,
N.; Larson, S. M. Imaging the pharmacodynamics of HER2
degradation in response to Hsp90 inhibitors. Nat. Biotechnol.
2004, 22, 701–706.
(32) Glaser, M.; Arstad, E. “Click labeling” with 2-[18F]fluoroethyl-
azide for positron emission tomography. Bioconjugate Chem. 2007,
18, 989–993.
(33) Lewis, W. G.; Magallon, F. G.; Fokin, V. V.; Finn, M. G.
Discovery and characterization of catalysts for azide-alkyne
cycloaddition by fluorescence quenching. J. Am. Chem. Soc.
2004, 126, 9152–9153.
(15) Li, X.; Link, J. M.; Stekhova, S.; Yagle, K. J.; Smith, C.; Krohn,
K. A.; Tait, J. F. Site-specific labeling of annexin V with F-18 for
apoptosis imaging. Bioconjugate Chem. 2008, 19, 1684–1688.
(16) Tait, J. F.; Smith, C.; Levashova, Z.; Patel, B.; Blankenberg, F. G.;
Vanderheyden, J. L. Improved detection of cell death in vivo with
annexin V radiolabeled by site-specific methods. J. Nucl. Med.
2006, 47, 1546–1553.
(34) Chan, T. R.; Hilgraf, R.; Sharpless, K. B.; Fokin, V. V. Polytria-
zoles as copper(I)-stabilizing ligands in catalysis. Org. Lett. 2004, 6,
2853–2855.
(35) Rodionov, V. O.; Presolski, S. I.; Diaz, D. D.; Fokin, V. V.; Finn,
M. G. Ligand-accelerated Cu-catalyzed azide-alkyne cycloaddi-
tion: a mechanistic report. J. Am. Chem. Soc. 2007, 129, 12705–
12712.

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 19 5825
(36) Rodionov, V. O.; Presolski, S. I.; Gardinier, S.; Lim, Y. H.; Finn,
M. G. Benzimidazole and related ligands for Cu-catalyzed azide-
alkyne cycloaddition. J. Am. Chem. Soc. 2007, 129, 12696–12704.
(37) Devaraj, N. K.; Keliher, E. J.; Thurber, G. M.; Nahrendorf, M.;
Weissleder, R. 18F labeled nanoparticles for in vivo PET-CT
imaging. Bioconjugate Chem. 2009, 20, 397–401.
(38) Sereikaite, J.; Jachno, J.; Santockyte, R.; Chmielevski, P.; Bumelis,
V. A.; Dienys, G. Protein scission by metal ion-ascorbate system.
Protein J. 2006, 25, 369–378.
(39) Sen Gupta, S.; Kuzelka, J.; Singh, P.; Lewis, W. G.; Manchester,
M.; Finn, M. G. Accelerated bioorthogonal conjugation: a prac-
tical method for the ligation of diverse functional molecules to
a polyvalent virus scaffold. Bioconjugate Chem. 2005, 16, 1572–
1579.
(40) Ward, J. F.; Johansen, I.; Aasen, J. Radiosensitization by N-ethyl
maleimide;a model chemical system. Int. J. Radiat. Biol. Relat.
Stud. Phys. Chem. Med. 1969, 15, 163–170.
(41) Hayon, E.; Simic, M. Radiation sensitization reactions of N-ethyl-
maleimide with model compounds. Radiat. Res. 1972, 50, 464–478.
(42) Solit, D. B.; Zheng, F. F.; Drobnjak, M.; Munster, P. N.; Higgins,
B.; Verbel, D.; Heller, G.; Tong, W.; Cordon-Cardo, C.; Agus,
D. B.; Scher, H. I.; Rosen, N. 17-Allylamino-17-demethoxygelda-
namycin induces the degradation of androgen receptor and HER-
2/neu and inhibits the growth of prostate cancer xenografts. Clin.
Cancer Res. 2002, 8, 986–993.
(43) Wilson, A. A.; Jin, L.; Garcia, A.; DaSilva, J. N.; Houle, S. An
admonition when measuring the lipophilicity of radiotracers using
counting techniques. Appl. Radiat. Isot. 2001, 54, 203–208.