18F-Labeled wild-type annexin V: Comparison of random and site-selective radiolabeling methods

Article abstract of DOI:10.1007/s00726-015-2068-0

Early stage apoptosis is characterized by the externalization of phosphatidylserine (PS) from the inner leaflet of the plasma membrane to the outer periphery. Consequently, PS represents an excellent target for non-invasive imaging of apoptosis by positron emission tomography. Annexin V is a 36 kDa protein which binds with high affinity to PS. Radiolabeling of wild-type annexin V with fluorine-18 (¹⁸F) can be accomplished via random acylation of 23 amine groups (22 lysine residues and one N-terminal amine) with [¹⁸F]SFB or site-specific alkylation reaction on cysteine residue at position 315 with maleimide-containing prosthetic groups like [¹⁸F]FBEM. The effect upon random and site-directed ¹⁸F labeling of annexin V was studied with EL4 mouse lymphoma cells. Both, randomly and site-selectively radiolabeled annexin V demonstrated comparable binding to apoptotic EL4 cells. This finding suggests that the ¹⁸F radiolabeling method has no significant effect on the ability of ¹⁸F-labeled wild-type annexin V to bind PS in apoptotic cells.

Full text of DOI:10.1007/s00726-015-2068-0

Amino Acids
DOI 10.1007/s00726-015-2068-0
ORIGINAL ARTICLE
18F‑Labeled wild‑type annexin V: comparison of random
and site‑selective radiolabeling methods
Amanda Perreault1 · James C. Knight1 · Monica Wang1 · Jenilee Way1 · Frank Wuest1
Received: 30 April 2015 / Accepted: 3 August 2015
© Springer-Verlag Wien 2015
Abstract Early stage apoptosis is characterized by the
externalization of phosphatidylserine (PS) from the inner
leaflet of the plasma membrane to the outer periphery. Con-
sequently, PS represents an excellent target for non-invasive
imaging of apoptosis by positron emission tomography.
Annexin V is a 36 kDa protein which binds with high affin-
ity to PS. Radiolabeling of wild-type annexin V with fluo-
rine-18 (18F) can be accomplished via random acylation of 23
amine groups (22 lysine residues and one N-terminal amine)
with [18F]SFB or site-specific alkylation reaction on cysteine
residue at position 315 with maleimide-containing pros-
thetic groups like [18F]FBEM. The effect upon random and
site-directed 18F labeling of annexin V was studied with EL4
mouse lymphoma cells. Both, randomly and site-selectively
radiolabeled annexin V demonstrated comparable binding to
apoptotic EL4 cells. This finding suggests that the 18F radi-
olabeling method has no significant effect on the ability of
18F-labeled wild-type annexin V to bind PS in apoptotic cells.
Different cell death pathways have been studied exten-
sively, and molecular imaging of cell death after treatment
with radiation or chemotherapy has the potential to pro-
vide detailed information on cellular transformations and
dynamics of tumors before, during, and after therapeutic
interventions (Smith and Smith 2012).
Apoptosis (programmed cell death) is characterized by
a number of morphological and biochemical changes that
can be targeted for molecular imaging purposes (Neves
and Brindle 2014). In the early stages of apoptosis, phos-
phatidylserine (PS), a membrane phospholipid, is exter-
nalized from the inner leaflet to the outer leaflet of the cell
membrane. In a healthy cell, PS is completely confined
to the inner leaflet of the cell. Induction of apoptosis and
subsequent increase in cytosolic Ca2+ concentration leads
to redistribution of phospholipids and exposure of PS,
which is recognized by macrophages and other phagocytes
that engulf and remove the apoptotic bodies (Martin et al.
1995). PS is a well-characterized biomarker for early-stage
apoptosis, and a promising molecular target for apoptosis
imaging probes (Blankenberg 2008).
Keywords Wild-type annexin V · Fluorine-18 ·
Apoptosis · [18F]SFB · [18F]FBEM
The most commonly used PS-targeting agent is annexin
V, an endogenous, 36 kDa, intracellular protein that binds
to PS with nanomolar affinity (Kd ~2 nM) in the presence
of Ca2+ (Schaper and Reutelingsperger 2013; Tait et al.
1994). Annexin V modified with fluorescent dyes is used
as a standard staining technique to assess apoptosis in vitro.
Annexin V has also been labeled with various radionuclides
to give radiotracers for molecular imaging of apoptosis
in vivo using single-photon emission computed tomogra-
phy (SPECT) and positron emission tomography (PET).
Annexin V and annexin V mutants, generated for easier
bioconjugation, were radiolabeled with 125I, 123I, 111In and
99mTc for SPECT imaging of apoptosis (Vangestel et al.
2011). The most successful and extensively studied of these
Introduction
Non-invasive monitoring of tumor cell death in vivo is
an important clinical need to assess therapeutic response.
Handling Editor: D. Tsikas.
* Frank Wuest
wuest@ualberta.ca
1
Department of Oncology, University of Alberta, Edmonton,
AB T6G 1Z2, Canada
1 3

A. Perreault et al.
is 99mTc-HYNIC-annexin V, which was first reported by
Blankenberg et al. (1998). Usefulness of 99mTc-HYNIC-
annexin V for assessment of tumor response to therapy was
demonstrated in a wide array of preclinical (Blankenberg
et al. 1998; De Saint-Hubert et al. 2011) and clinical studies
(Belhocine et al. 2002). However, 99mTc-HYNIC-annexin
V shows high non-specific uptake in the kidneys, liver and
spleen, resulting in a poor tumor-to-background ratio and
limitations for imaging of apoptotic cells around these
areas. Moreover, SPECT is also less sensitive than PET.
Therefore, radiolabeling of annexin V with a positron-
emitting radionuclide for PET imaging of apoptosis is
more desirable as it would provide higher quality images
and allow for more sensitive imaging. Positron-emitting
2008). These concerns led to the generation of annexin V
mutants, or “second generation annexin V”, which contain
an additional short amino acid sequence on the N-terminus
bearing a cysteine residue for easy, site-specific labeling
by thiol-reactive prosthetic groups (Lahorte et al. 2004).
However, availability of cysteine-containing second gen-
eration annexin V is limited. On the other hand, commer-
cially available wild-type annexin V also contains a single
chemically accessible cysteine residue in the 315-position
of the protein. No loss of binding affinity was observed
for annexin V when Cys315 was deleted or replaced with
thiol-free amino acids (Tait et al. 2000). Thus, Cys315 is
not required for functionality of annexin V, as modifica-
tion of this residue does not alter its biological activity.
Maleimide-containing prosthetic groups, which react with
free thiol groups, can be used to target this single cysteine
residue and thus offer a controlled, site-selective approach
to radiolabel wild-type annexin V. However, to the best of
our knowledge, there is only one report exploiting this radi-
olabeling approach for wild-type annexin V (Wuest et al.
2008).
In the present study, we describe random and site-
selective labeling of wild-type annexin V with short-lived
positron emitter 18F, using [18F]SFB as amine-reactive
prosthetic group and N-[2-(4-[18F]fluorobenzamido)ethyl]-
maleimide ([18F]FBEM) as thiol group-reactive prosthetic
group. We examined the effect of both bioconjugation
methods on PS-binding ability of 18F-labeled annexin V
using a cell binding assay. In this assay, we compared the
binding of [18F]SFB- and [18F]FBEM-labeled annexin V to
EL4 mouse lymphoma cells that were treated with camp-
tothecin to induce apoptosis. We also examined binding
of 18F-labeled annexin V to apoptotic lymphoma cells as a
function of Ca2+ concentration, since annexin V binding to
PS is strongly Ca2+-dependent (Tait et al. 2004).
radionuclides for annexin V labeling include 124I, 64Cu, 18
F
and 68Ga (Lehner et al. 2014; Wängler et al. 2011; Keen
et al. 2005; Wuest et al. 2015). Among the reported PET
radiotracers based on annexin V, 18F-labeled annexin V has
received the most attention due to the favorable properties
of 18F for PET imaging.
18F has a physical half-life of 109.8 min, high positron
emission (97 %), low positron energy of 0.64 MeV result-
ing in high image resolution and lower radiation dose,
and 18F can be produced in large amounts with high spe-
cific activity by a small biomedical cyclotron (Smith et al.
2012).
Radiolabeling of annexin V with 18F is most commonly
accomplished via a random labeling approach involving the
NHS-containing prosthetic group N-succinimidyl-4-[18F]
fluorobenzoate ([18F]SFB), which can react with any of
the 22 lysine residues and N-terminus present in annexin
V. This labeling approach has been carried out by several
groups (Yagle et al. 2005; Zijlstra et al. 2003), and the
ability of [18F]SFB-labeled annexin V to image apoptosis
in vivo has been demonstrated in various preclinical mod-
els (Yagle et al. 2005; Murakami et al. 2004). However,
only one group has reported the utility of [18F]SFB-labeled
annexin V for imaging therapy-induced apoptosis as dem-
onstrated in head and neck squamous cell cancer UM-
SCC-22B tumor xenografts upon treatment with doxoru-
bicin (Hu et al. 2012. Tait et al. suggested that prosthetic
group [18F]SFB can conjugate to any of the 23 primary
amines present on annexin V, some in close proximity of
the active binding site of the protein. Such a non-specific,
amine-directed modification of annexin V not only results
in a poorly characterized radiotracer, but also might disrupt
crucial binding interactions as demonstrated in a recent
report describing that amine-directed chemical modifica-
tion of annexin V reduced its membrane-binding activity
even at low stoichiometries (Tait et al. 2006). In addition, if
annexin V is conjugated with low specific activity [18F]SFB
at multiple sites, this could lead to disproportional signal-
ing by low-affinity molecules in PET imaging (Li et al.
Materials and methods
General
Purified recombinant human annexin V (1 or 5 mg, lyo-
philized) was purchased from BioVision (Milpitas, USA),
reconstituted in phosphate-buffered saline (PBS), pH 7.4, to
1 mg/mL, and made into 100 μL aliquots which were kept
at −20 °C. Water was obtained from a Barnstead Nanopure
water filtration system (Barnstead Diamond Nanopure pack
organic free RO/DIS). All chemicals and reagents were
obtained from Sigma-Aldrich® unless otherwise stated.
18F was produced at the Edmonton PET Center using an
Advanced Cyclotron System TR 19/9 cyclotron. Radio-
TLC was performed using EMD Merck F254 silica gel 60
aluminum backed thin layer chromatography (TLC) plates
1 3

18F-Labeled wild-type annexin V: comparison of random and site-selective radiolabeling methods
(Bioscan AR-2000). Semipreparative high performance liq-
uid chromatography (HPLC) was performed on a Gilson sys-
tem consisting of a 321 pump and a 171 diode array detec-
tor, and a Berthold Technologies Herm LB 500 was used as
radio detector. UV absorbance was monitored at 254 nm.
Rotary evaporation of product solvents was carried out
using a Buchi HB 140 Rotavpor-M with a Fisher Maxima
C Plus Model M8C pump. Quantification of radioactive
samples during chemistry was achieved using a Biodex
ATOMLAB™ 400 dose calibrator. Reaction parameters
were carried out with an Eppendorf Thermomixer®. Cen-
trifugation of samples was carried out with a Hettich Zen-
trifugen Rotina 35R or a Fisher Scientific Mini Centrifuge.
Radio-sodium dodecyl sulfate polyacrylamide gel elec-
trophoresis (SDS-PAGE) was carried out on IDGel™-
Express IR121S-20 precast gels (12 % tris buffer), which
were run in a Bio-Rad Mini PROTEAN® Tetra Cell with
running buffer.
Gels were run with a constant current of 30 mÅ for
approximately 1 h, and Radio-SDS-page gels were scanned
on a Fujifilm BAS 5000 phosphor imager and analyzed by
AIDA Image Analyzer Software Version 4.50. The gels
were then stained with Coomassie® Brilliant Blue R-250
(Bio-Rad) for 30 min at 37 °C, and left in destaining solu-
tion overnight. Tryptan Blue stained cells were counted
using a Bio-Rad TC10™ Automated Cell Counter. A Per-
klinElmer 2480 Automatic Gamma Counter WIZARD2®
was used for the cell binding assays. Flow cytometry was
carried out using a Fluorescence Activated Cell Sorter (BD
FACSCaliber™) bench top analyzer. For fluorescence con-
focal microscopy, a Zeiss LSM 710 confocal microscope
was used. Protein content quantification was achieved
from absorbance measurements using a Molecular Devices
Spectramax 340PC microplate reader. In-house prepara-
tions of buffers were as follows: Phosphate-buffered saline
(PBS) (140 mM sodium chloride, 2.7 mM potassium chlo-
ride, 5.4 mM sodium phosphate dibasic, 570 μM potas-
sium phosphate monobasic, pH 7.4), borate buffer (50 mM
sodium borate, pH 8.4) 2.5 mM Ca2+ binding buffer
(10 mM HEPES (4-(2-hydroxyethyl)-1-piperazineethane-
sulfonic acid) (pH 7.4), 140 mM sodium chloride, 2.5 mM
calcium chloride), 1.25 mM Ca2+ binding buffer (10 mM
HEPES (pH 7.4), 140 mM sodium chloride l, 1.25 mM cal-
cium chloride), running buffer for SDS-PAGE (25 mM Tris,
192 mM glycine, 0.1 % SDS), Destaining solution (10 %
glacial acetic acid, 20 % methanol, 70 % deionized water).
achieved by radio-TLC. To increase the radiochemical purity
of [18F]SFB for radiolabelling, 1 mL of [18F]SFB in acetoni-
trile was purified by HPLC using a Phenomenex LUNA®
C18(2) (100Å, 250 × 10 mm, 10 μm) column and the fol-
lowing gradient: solvent A water; solvent B acetonitrile. 0 min
15 % B, 8 min 50 % B, 9–25 min 70 % B. The flow rate was
3 mL/min. At a retention time of about 15 min, the broad [18F]
SFB peak was collected into a 50 mL pear-shaped flask, and
solvent was removed by a rotary evaporator. The dry [18F]SFB
was then used for radiolabeling wild-type annexin V.
Radiosynthesis of [18F]FBEM
[18F]FBEM was synthesized manually from [18F]SFB in
a modified version of the procedure reported by Cai et al.
(2006). 3 mg of N-(2-aminoethyl)maleimidetrifluoroacetate
salt (the amine) was dissolved in 500 μL of borate buffer
(pH 8.4) and reacted with 500 μL of [18F]SFB (obtained
from the GE TRACERlab™ FX module) for 20 min at
50 °C.
[18F]FBEM was purified by HPLC using the same col-
umn and gradient that was used for [18F]SFB purification,
and it also eluted with a retention time of about 15 min.
The purified [18F]FBEM solution was placed on a rotary
evaporator until all solvent was removed.
Labeling of wild‑type annexin V with [18F]SFB
An annexin V aliquot (100 μg in 100 μL of either PBS
(pH 7.4) was warmed to room temperature and added to
the dry [18F]SFB (150–200 MBq) in a 50 mL pear-shaped
flask, vortexed vigorously for about 30 s, and then trans-
ferred to a microcentrifuge tube. Another 50 μL of PBS
was used to rinse the walls of the flask and added to the
reaction mixture. The reaction mixture was then placed
on the thermoshaker at 30 °C for 30 min to produce [18F]
SFB-annexin V. [18F]SFB-annexin V was purified by size
exclusion chromatography using either a PD-10 desalting
column (GE Healthcare) or a 10DG desalting column (Bio-
Rad Econo-Pac®) eluted with PBS into 250 μL fractions
(8 drops per fraction). The collected high molecular weight
fractions containing the highest activity were directly used
for cell binding studies. For quality control, a 5 μL sample
of the crude reaction mixture and a 5 μL sample of puri-
fied [18F]SFB-Annexin V were reduced with dithiothreitol
(DTT) at 95 °C for 5 min and subjected to SDS-PAGE. The
SDS-PAGE gels were then analyzed by phosphor imaging
and staining/destaining.
Radiosynthesis of [18F]SFB
The prosthetic group [18F]SFB was prepared via a three-step
procedure using a remotely controlled GE TRACERlab™
FX automated synthesis module according to Mäding et al.
(2005). Quality control of module-synthesized [18F]SFB was
Labeling of wild‑type annexin V with [18F]FBEM
Once [18F]FBEM (~200 MBq) was manually synthe-
sized from [18F]SFB and evaporated to dryness, the same
1 3

A. Perreault et al.
procedure that was used to label wild-type annexin V with
[18F]SFB was also used for [18F]FBEM. Like [18F]SFB-
annexin V, [18F]FBEM-annexin V was purified by size
exclusion chromatography and directly used for cell bind-
ing studies. SDS-PAGE was carried out on crude and puri-
fied samples for quality control.
The tubes were vortexed, covered with aluminum foil
and left to incubate in the dark for 15 min at room tempera-
ture. 400 µL of binding buffer was then added to each tube,
and apoptotic and necrotic cells in each tube were counted
by a fluorescence activated cell sorter.
Fluorescence confocal microscopy
Cell cultures
Fluorescence confocal microscopy was used to image the
apoptosis-inducing effect of camptothecin on EL4 lym-
phoma cells. Only two groups of cells were imaged by
fluorescence confocal microscopy: control (untreated)
cells, and cells treated with 1.5 µM camptothecin, the opti-
mal concentration to produce sufficient apoptosis levels
determined by flow cytometry. These cells were prepared
and treated using the same method that was used in the
flow cytometry studies, 1 day prior to fluorescence imag-
ing. Like the fluorescent staining done in the flow cytom-
etry studies, FITC-annexin V and PI were used to detect
apoptotic cells and necrotic cells, respectively. After 1 day
of treatment, cells were pelleted by centrifugation, washed
once in PBS and once in binding buffer, and then re-sus-
pended in binding buffer to a cell density of 2 × 105 cells/
mL. 30 µL FITC-annexin V and 30 µL PI was added to
500 µL of this cell solution, and the cells were incubated
in the dark at room temperature for 15 min. Cells were then
spun down and resuspended in 1 mL binding buffer, and
500 µL of each cell solution was added to its own well of a
Nunc™ Lab-Tek™ II Chambered Coverglass System. The
cells were then viewed and imaged using a confocal micro-
scope available in the Cell Imaging Facility of the Cross
Cancer Institute (Edmonton, Alberta).
EL4 murine T-cell lymphoma cells (TIB-39™), which
readily undergo apoptosis when introduced to cytotoxic
agents, were purchased from the American TypeCulture
Collection (ATCC®) (Manassas, VA, USA) and main-
tained in RPMI 1640 medium (prepared from Gibco® pow-
der) supplemented with 10 % fetal bovine serum (FBS)
(Gibco®), 2 mM l-glutamine (Invitrogen) and 1 % penicil-
lin/streptomycin (Invitrogen). Cell density was maintained
in the range of 2 × 105 cells/mL to 2 × 106 cells/mL, and
cell growth medium was added or changed every 2 days.
Cells were incubated at 37 °C in a humidified incubator
with a 5 % (v/v) CO2 atmosphere (ThermoForma Series II
Water Jacketed CO2 Incubator).
Induction of apoptosis in EL4 cells and flow cytometry
(s)-(+)-Camptothecin (Sigma®), a DNA-topoisomerase
I inhibitor, was used to induce apoptosis in EL4 cells. In
order to determine the optimal concentration of campto-
thecin to induce sufficient apoptosis, flow cytometry was
used to determine apoptosis levels in EL4 cells treated
with increasing doses of the drug. The FITC-Annexin V
Apoptosis Detection Kit 1 (BD Pharmingen™) was used
to detect apoptotic cells. From this kit, fluorescein iso-
thiocyanate conjugated annexin V (FITC-annexin V) was
used to stain apoptotic cells, while propidium iodide (PI)
was used to stain necrotic cells. The use of PI allowed for
discrimination between apoptotic and necrotic cells. A
6-well plate (Corning®) was prepared with 3 mL of RPMI
medium containing 3 million cells (cell density of 1 mil-
lion cells/mL) in each well. Each well had a different con-
centration of camptothecin: 0.5, 1.0, 1.5, 1.75 and 2.0 µM,
and the control well contained just NaOH (9 µL of 1.0 M
NaOH solution) without camptothecin. The cells were
left to incubate for 24 h at 37 °C. The next day, the cells
were pelleted by centrifugation, washed twice with PBS,
and resuspended in binding buffer (BD Pharmingen™) to
a cell density of 1 million cells/mL. A 100 µL aliquot of
each cell solution was transferred to its own flow cytom-
etry tube. To each tube, 3 µL of annexin V-FITC and 3 µL
of PI was added. 4 tubes of control cells were prepared:
one unstained, one stained with FITC-annexin V only, one
stained with PI only, and one stained with FITC-annexin
V + PI.
Cell binding assay of [18F]SFB‑annexin V and [18F]
FBEM‑annexin V in EL4 cells
24 h prior to the cell binding assay, two T75 flasks were
prepared containing 15 mL of EL4 cells at a cell density of
1 million cells/mL. To one flask, 9 µL of the 2.5 mM stock
solution of camptothecin was added to the 15 mL of media
to produce a camptothecin concentration of 1.5 µM. To the
control flask, 9 µL of 1.0 M NaOH solution was added. To
prevent direct contact of 1.0 M NaOH and 2.5 mM camp-
tothecin with the cells, NaOH/camptothecin was first added
to the media, which was then added to the cells. Con-
trol and treated cells were then left to incubate for 1 day
at 37 °C. On the day of experiment, control and treated
cells were pelleted by centrifugation, growth media was
removed, and cells were washed once with PBS, and once
with 2.5 mM Ca2+-containing binding buffer. Cells were
resuspended in binding buffer to 5 × 106 cells/mL. Using
two 12-well plates (one for control cells and one for treated
cells), 300 µL of cell solution was added to each well.
1 3

18F-Labeled wild-type annexin V: comparison of random and site-selective radiolabeling methods
Fig. 1 Random and site-spe-
cific labeling of annexin V with
[18F]SFB and [18F]FBEM
O
O
Annexin V
O
Annexin V
N
O
N
H
PBS, pH 7.4
30 min, 30 oC
O
18F
18F
[
18F]SFB
Randomly-labeled annexin V
O
O
Annexin V
O
Cys315 Annexin V
O
S
N
N
N
N
H
PBS, pH 7.4
30 min, 30 oC
H
O
O
18F
18F
[
18F]FBEM
Site-specific labeled annexin V
The 18F-labeled annexin V tracer being investigated was
diluted in binding buffer to 0.5 kBq/μL, and 200 μL of this
18F-annexin V solution was added to each well in order to
get approximately 0.1 MBq/well.
protein content samples. The microplate was left to incu-
bate for 25 min at 37 °C (Fisher Scientific Isotemp Incu-
bator Model 546) and then the protein content was deter-
mined using an absorbance microplate reader.
Tracer was not added to control and treated cells that
were to be used for protein quantification. Cells were left
to incubate at room temperature for 1, 15, 30 or 60 min.
At each time point, the cells of a well were transferred to
a microcentrifuge tube, rinsing the well with an additional
500 µL of binding buffer and adding this to the tube. Cells
were centrifuged for 2 min at 1500 rpm in order to form
a cell pellet. The supernatant was removed along with any
unbound tracer. The pellet was resuspended in 500 µL of
binding buffer, and this solution was transferred to a new
tube to remove unbound tracer that was non-specifically
bound to the tube. Cells were spun down to get a pellet, and
the supernatant removed. The pellet-containing tubes were
placed in a gamma counter to measure the amount of radi-
otracer bound to the cells. Calculated uptake values were
expressed as percentage of total activity normalized to mg
of the cellular protein content. In order to address the effect
of Ca2+ concentration, another set of cell binding assay was
carried out using a lower Ca2+ concentration, 1.25 mM,
which is suggested to be more representative of physiologi-
cal Ca2+ levels (Blankenberg 2008). To determine protein
content, cells (control and treated) were lysed with Cel-
Lytic™ M (300 μL) for 10 min at 4 °C, centrifuged at 4 °C
and 10,000 rpm for 8 min, and left in the freezer for at least
one night. Following the instructions of bicinchoninic acid
(BCA) BCA™ Protein Assay Kit (Pierce™), standards
containing diluted albumin (BSA) were prepared at concen-
trations of 0, 50, 100, 200, 300, 400, 600, and 800 μg/mL.
A 25 μL aliquot of each standard was pipetted into a
96-well microplate (Corning®) in descending concentra-
tion, along with 25 μL of the unknown protein content
sample, thawed to room temperature. 7 mL of BCA Rea-
gent A was thoroughly mixed with 140 μL of BCA reagent
B, and a multichannel pipette was used to dispense 200 μL
of this solution into a 96 well microplate that contained the
Data analysis
All data are expressed as mean
standard error of the
mean (SEM) from 3 or more experiments. All graphs were
constructed using GraphPad Prism 5.0 (GraphPad Soft-
ware, San Diego, CA, USA). Kd and Bmax values were
calculated using a nonlinear regression (curve fit) analysis
(GraphPad Prism 5.0). Statistical differences were con-
sidered significant if P < 0.05, and were tested using the
unpaired t test.
Results and discussion
Radiolabeling of wild‑type annexin V with [18F]SFB
and [18F]FBEM
Random labeling of the 23 primary amine groups acces-
sible in wild-type annexin V according to an acylation
reaction with prosthetic group [18F]SFB gave 18F-labeled
annexin V in decay-corrected radiochemical yields of
9
1 % (n = 7, based upon [18F]SFB).
Site-specific labeling of Cys315 in wild-type annexin V
was achieved by an alkylation reaction with [18F]FBEM to
afford 18F-labeled annexin V in decay-corrected radiochem-
ical yields of 4
2 % (n = 5, based upon [18F]FBEM).
Both prosthetic groups were purified by HPLC prior to the
reaction with wild-type annexin V. The synthesis route of
randomly labeled and site-specific labeled [18F]annexin V
is given in Fig. 1.
Both 18F-labeled annexin V derivatives were obtained
at high radiochemical purity >95 % after purification with
size exclusion chromatography. Figure 2 shows the results
of Coomassie stained SDS-PAGE and radio-SDS-PAGE
1 3

A. Perreault et al.
Fig. 2 SDS-PAGE analysis
of wild-type annexin V (lanes
1a and 1b), purified randomly
labeled annexin V (lanes 2a and
2b), and purified site-selectively
labelled annexin V (lanes 3a
and 3b)
O
HO
HO
analysis of wild-type annexin V as reference, and purified
18F-labeled annexin V derivatives.
OH
N
N
O
Conjugation of prosthetic groups [18F]SFB and [18F]
FBEM to wild-type annexin V proceeded under mild con-
ditions (PBS, pH 7.4, 30 min, 30 °C), and no degradation
was observed during the radiosynthesis of [18F]annexin V
as confirmed by SDS-PAGE analysis. Isolated radiolabeled
protein corresponds with the expected molecular weight of
wild-type annexin V derivatives of 36 kDa.
O
HO
18F
[
18F]FDG-MHO
Fig. 3 Structure of [18F]FDG-MHO
maleimide prosthetic group [18F]FDG-maleimidehexy-
loxime ([18F]FDG-MHO) gave 18F-labeled annexin V in
radiochemical yields of 43–58 % starting from 0.1 mg of
wild-type annexin V. This result makes [18F]FDG-MHO a
highly suitable prosthetic group for bioconjugation of wild-
type annexin V. However, [18F]FDG-MHO is a rather large
prosthetic group in comparison to [18F]SFB (Fig. 3).
In addition to the maleimide group it also contains a
hydrophilic sugar moiety prone to modify the pharmacoki-
netic properties of [18F]FDG-MHO-labeled annexin V dif-
ferently than [18F]SFB-labeled annexin V which lacks such
a polar substitution pattern. These differences in size and
pharmacokinetic profile make a direct comparison between
the random labeling method using [18F]SFB and a site-
directed labeling approach with [18F]FDG-MHO challeng-
ing. On the other hand, maleimide-containing compound
[18F]FBEM is an alternative thiol-selective prosthetic group
which is more comparable to the size of [18F]SFB. This
consideration also applies to [18F]FBEM-labeled annexin V
(Fig. 1).
Radiolabeling of wild-type annexin V with [18F]SFB has
been reported by several groups over the past decades, and
our results from this study further confirm the feasibility of
using prosthetic group [18F]SFB to radiolabel annexin V
with 18F according to this random labeling method (Blank-
enberg et al. 1998; Yagle et al. 2005; Zijlstra et al. 2003;
Lahorte et al. 2004). Although we obtained lower radio-
chemical yields (9 %) of 18F-labeled annexin V than other
groups, this can be explained by the significantly smaller
amount (0.1 mg) of wild-type annexin V used in our reac-
tions. Other groups reported radiochemical yields between
42 and 70 % using much higher annexin V amounts in the
range of 1 mg of protein (Yagle et al. 2005; Zijlstra et al.
2003; Toretsky et al. 2004). However, our obtained radio-
chemical yield of 9 % was comparable to the reported 10 %
radiochemical yield by Murakami and colleagues, who also
used small amounts (0.1 mg) of wild-type annexin V for the
radiolabeling with [18F]SFB (Murakami et al. 2004). Toret-
sky et al. (2004) described a linear relationship (r2 = 0.89)
between protein concentration and radiochemical yield of
[18F]annexin V, which provides a good explanation for the
obtained radiochemical yields.
Thus, [18F]FBEM is a suitable prosthetic group for a
comparative study with acylation agent [18F]SFB despite
the rather low radiochemical yield of 4 % for the bioconju-
gation reaction to wild-type annexin V.
The fairly high costs for wild-type annexin V make the
use of small amounts of the protein during the radiolabe-
ling reaction highly desirable. Moreover, smaller amounts
of protein provide radiolabeled annexin V at higher specific
radioactivity.
Successful bioconjugation of the single cysteine resi-
due at position 315 in wild-type annexin V was recently
reported by our group (Wuest et al. 2008). Glucose-based
Induction of apoptosis in EL4 cells with camptothecin
Characterization of randomly and site-specific 18F-labeled
annexin V derivatives was performed in murine lymphoma
EL4 cells, which were treated with camptothecin to induce
apoptosis. Flow cytometry was used to assess the effect of
1 3

18F-Labeled wild-type annexin V: comparison of random and site-selective radiolabeling methods
80
60
40
20
0
the camptothecin concentration on the level of apoptosis
induced in EL4 cells. The results are given in Fig. 4.
The data shown in Fig. 4 revealed a small amount of
basal apoptosis (17 %) in the untreated control cell popula-
tion due to the unstable nature of the cell line. This baseline
level of apoptosis in EL4 cells has been previously reported
(Guo et al. 2009). Treatment with 1.5 μM of camptoth-
ecin induced the highest increase in apoptosis levels rela-
tive to untreated cells, as 56 % of the cells were stained
with FITC-annexin V. Apoptosis levels reached a plateau
at higher camptothecin concentrations (1.75 and 2.0 μM)
resulting in 60 and 57 % apoptosis rates, respectively.
Based on these results, we selected a drug concentra-
tion of 1.5 μM to induce apoptosis in EL4 cells for the cell
binding assay. Drug-induced apoptosis in EL4 cells was
confirmed with confocal microscopy. Confocal microscopy
images of camptothecin-treated EL4 cells stained with
FITC-labeled annexin V and propidium iodide are shown
in Fig. 5.
0
0.5
1.0
1.75
2.0
[Camptothecin] (µM)
Fig. 4 Proportion of EL4 cells undergoing apoptosis when treated
with increasing concentrations of camptothecin, as determined by
flow cytometry with FITC-labeled annexin V (n = 3)
However, the lower Ca2+ concentration did not diminish the
preferential binding of [18F]SFB-annexin V to cells treated
with apoptosis-inducing agent compared to untreated cells.
After 60 min, binding of the radioligand was three-times
higher in treated cells versus control cells.
Fluorescence confocal microscopy images clearly dem-
onstrate the successful induction of apoptosis in EL4 cells
upon treatment with 1.5 μM of camptothecin as shown by
the green stain from FITC-annexin V binding to PS in the
treated cell population.
Our observed elevated binding of [18F]SFB-annexin V to
apoptotic EL4 cells confirms previous reports on the fea-
sibility of using randomly labeled [18F]SFB-annexin V for
the detection of apoptosis. The PS-binding ability of [18F]
SFB-annexin V was described in various other cell lines.
Grierson and colleagues demonstrated PS-binding of [18F]
SFB-annexin V through a cell binding assay using red
blood cells (RBCs) with exposed PS (Grierson et al. 2004),
while Zijlstra et al. (2003) demonstrated 60 % increased
binding of [18F]SFB-annexin V to UV-irradiated Jurkat
T-cells compared to non-irradiated cells. Another study by
Toretsky et al. (2004) demonstrated 88 % more binding of
[18F]SFB-annexin V to TC32 sarcoma cells treated with
etoposide compared to untreated cells . When using site-
specific labeled [18F]FBEM-annexin V, we observed a 2.6-
fold increase in binding of the radioligand to treated EL4
cells compared to untreated cells (Fig. 9).
Results of PS-binding shown in Figs. 6 and 9 confirm
that both randomly labeled [18F]SFB-annexin V and site-
specific labeled [18F]FBEM-annexin V retained the ability
to detect apoptotic cells in vitro. Both radiotracers dem-
onstrated significantly higher binding to EL4 cells treated
with 1.5 μM of camptothecin compared to untreated cells.
Binding to apoptotic cells in terms of extent and kinetics
was comparable for both 18F-labeled wild-type annexin V
derivatives.
Cell binding using EL4 cells with [18F]SFB‑annexin V
and [18F]FBEM‑annexin V
Cell binding studies with untreated (control) and drug-
treated EL4 cells were used to characterize PS-binding
properties of [18F]SFB- and [18F]FBEM-labeled annexin
V. Both radiotracers showed significantly higher binding of
18F-labeled annexin V derivatives to EL4 mouse lymphoma
cells treated with the apoptosis-inducing agent camptoth-
ecin (1.5 μM) compared to untreated cells over time. After
60 min, a fourfold higher uptake of [18F]SFB-annexin V to
treated EL4 cells was found compared to non-treated cells
(Fig. 6).
The Ca2+ concentration is known to be critical for PS-
binding of annexin V. It was shown that annexin V bind-
ing to PS is strongly Ca2+-dependent and declines rapidly
over a range of 2.5–1.25 mM (Tait et al. 2004). To address
this issue, we conducted another set of cell binding stud-
ies using Ca2+ concentrations of 1.25 and 2.5 mM. A Ca2+
concentration of 1.25 mM was recently suggested to be
representative of Ca2+ levels in vivo (Tait et al. 2006).
Binding of [18F]SFB-annexin V to both control
(untreated) EL4 cells (Fig. 7) and cells treated with 1.5 μM
camptothecin (Fig. 8) was significantly decreased by reduc-
ing the Ca2+ concentration from 2.5 to 1.25 mM.
This result suggests that random and site-selective radi-
olabeling methods with wild-type annexin V provide radi-
otracers with comparable PS-binding capabilities. In our
experiments, no significant differences upon PS binding in
EL4 cells for both 18F-labeled annexin V derivatives were
As expected, lowering the Ca2+ concentration from
2.5 to 1.25 mM significantly reduced binding of [18F]
SFB-annexin V to both treated and untreated EL4 cells.
1 3

A. Perreault et al.
Fig. 5 Fluorescence confocal
microscopy images of a Control
(untreated) EL4 cell population;
b EL4 cell population treated
with 1.5 μM of camptothecin.
Green stain—FITC-annexin
V (apoptosis); red stain—pro-
pidium iodide (necrosis) (color
figure online)
Binding of [18F]SFB-annexin V to untreated cells
600
[
[
18F]SFB-annexin V (Treated)
18F]SFB-annexin V (Control)
500
400
300
200
100
0
150
100
50
2.50 mM Ca2+
1.25 mM Ca2+
*
*
*
0
10
20
30
40
50
60
0
Time (min)
Time (minutes)
Fig. 6 Cell binding of randomly labeled [18F]SFB-annexin V to
untreated (control) EL4 cells and EL4 cells treated for 24 h with
1.5 μM of camptothecin (n = 3)
Fig. 8 Comparison of cell binding of [18F]SFB-annexin V to campto-
thecin-treated EL4 cells in the presence of higher (2.5 mM) and lower
(1.25 mM) concentrations of Ca2+ (n = 3)
Binding of [18F]SFB-annexin V to untreated cells
150
600
2.50 mM Ca2+
[
[
18F]FBEM-annexin V (Treated)
18F]FBEM-annexin V (Control)
1.25 mM Ca2+
500
400
300
200
100
0
100
*
50
*
*
0
0
10
20
30
40
50
60
Time (min)
Time (minutes)
Fig. 9 Cell binding of site-specific labeled [18F]FBEM-annexin V
to untreated (control) EL4 cells and EL4 cells treated for 24 h with
1.5 μM of camptothecin (n = 3)
Fig. 7 Comparison of cell binding of [18F]SFB-annexin V to
untreated EL4 cells in the presence of higher (2.5 mM) and lower
(1.25 mM) concentrations of Ca2+ (n = 3)
1 3

18F-Labeled wild-type annexin V: comparison of random and site-selective radiolabeling methods
noticeable. However, this finding is in contrast to a com-
parative study reported by Tait et al. (2006).
and more economical random labeling technique with pros-
thetic group [18F]SFB to radiolabel this protein with 18F.
In their study, the authors found that amine-directed ran-
dom modification of annexin V substantially reduced its
ability to bind to PS-expressing RBCs, though it did not
diminish it completely. They also reported that site-selec-
tive conjugation of an additional cysteine residue present
in the modified amino-terminal sequence of Ala–Gly–
Gly–Cys–Gly–His of annexin V-128 (a second generation
annexin V) resulted in an increased binding of the pro-
tein to RBCs. They also showed that the more conjugated
annexin V is, the lower its PS-binding ability becomes.
However, Grierson et al. (2004) determined that even high
[19F]SFB:annexin V molar ratios (32:1) during a 15 min
incubation time for the conjugation reaction resulted in a
rather low 2.1 average incorporation level of fluorobenzoic
acid into annexin V without loss of PS-binding proper-
ties. Only extension of the reaction time from 15 to 60 min
resulted in a conjugated product with compromised PS-
binding capabilities.
Our described conjugation method is using HPLC-
purified [18F]SFB and [18F]FBEM at high specific activity.
Moreover, mild reaction conditions (PBS, 30 min, 30 °C,
pH 7.4) during the conjugation reactions enable introduc-
tion of the radiolabel while preserving the structural and
functional integrity of wild-type annexin V as demonstrated
by the found comparable PS-binding ability for both, ran-
domly labeled and site-specific labeled annexin V deriva-
tives. However, given the rather low radiochemical yields
for [18F]FBEM-annexin V (4 %) combined with the length-
ier and more complex synthesis of maleimide-based pros-
thetic group [18F]FBEM, we propose to use the random
labeling technique with [18F]SFB to prepare 18F-labeled
wild-type annexin V. Radiolabeling of wild-type annexin
V with the readily available prosthetic group [18F]SFB pro-
vides 18F-labeled annexin V as a radiotracer for PS-binding
in apoptotic cells in reasonable radiochemical yields of
around 10 % while using only small amounts of wild-type
annexin V (0.1 mg) as starting material.
Acknowledgments This work was generously supported by Natu-
ral Sciences and Engineering Research Council of Canada (NSERC)-
CREATE Molecular Imaging Probes (cMIP), the Dianne and Irving
Kipnes Foundation, and Alberta Innovates—Health Solutions (AIHS).
The authors gratefully acknowledge the Edmonton PET Center and
Cyclotron Facility at the Cross Cancer Institute, as well as the Flow
Cytometry Lab and Cell Imaging Facility in the Department of Exper-
imental Oncology at the Cross Cancer Institute.
Compliance with ethical standards
Conflict of interest The authors declare that they have no conflict
of interest.
References
Belhocine T, Steinmetz N, Hustinx R, Bartsch P, Jerusalem G, Sei-
del L, Rigo P, Green A (2002) Increased uptake of the apopto-
sis-imaging agent (99 m)Tc recombinant human Annexin V in
human tumors after one course of chemotherapy as a predic-
tor of tumor response and patient prognosis. Clin Cancer Res
8:2766–2774
Blankenberg FG (2008) In vivo detection of apoptosis. J Nucl Med
49(Suppl 2):81S–95S
Blankenberg FG, Katsikis PD, Tait JF, Davis RE, Naumovski L, Oht-
suki K, Kopiwoda S, Abrams MJ, Darkes M, Robbins RC, Mae-
cker HT, Strauss HW (1998) In vivo detection and imaging of
phosphatidylserine expression during programmed cell death.
Proc Natl Acad Sci USA 95:6349–6354
Cai W, Zhang X, Wu Y, Chen X (2006) A thiol-reactive 18F-labeling
agent, N-[2-(4-18F-fluorobenzamido)ethyl]maleimide, and syn-
thesis of RGD peptide-based tracer for PET imaging of alpha v
beta 3 integrin expression. J Nucl Med 47:1172–1180
De Saint-Hubert M, Wang H, Devos E, Vunckx K, Zhou L, Reu-
telingsperger C, Verbruggen A, Mortelmans L, Ni Y, Mottaghy
FM (2011) Preclinical imaging of therapy response using met-
abolic and apoptosis molecular imaging. Mol Imaging Biol
13:995–1002
Grierson JR, Yagle KJ, Eary JF, Tait JF, Gibson DF, Lewellen B, Link
JM, Krohn KA (2004) Production of [F-18]fluoroannexin for
imaging apoptosis with PET. Bioconjug Chem 15:373–379
Guo MF, Zhao Y, Tian R, Li L, Guo L, Xu F, Liu YM, He YB, Bai S,
Wang J (2009) In vivo 99 mTc-HYNIC-annexin V imaging of
early tumor apoptosis in mice after single dose irradiation. J Exp
Clin Cancer Res 28:136
Hu S, Kiesewetter DO, Zhu L, Guo N, Gao H, Liu G, Hida N, Lang
L, Niu G, Chen X (2012) Longitudinal PET imaging of doxoru-
bicin-induced cell death with 18F-annexin V. Mol Imaging Biol
14:762–770
Keen HG, Dekker BA, Disley L, Hastings D, Lyons S, Reader AJ,
Ottewell P, Watson A, Zweit J (2005) Imaging apoptosis in vivo
using 124I-annexin V and PET. Nucl Med Biol 32:395–402
Lahorte CM, Vanderheyden JL, Steinmetz N, Van de Wiele C,
Dierckx RA, Slegers G (2004) Apoptosis-detecting radioligands:
current state of the art and future perspectives. Eur J Nucl Med
Mol Imaging 31:887–919
Conclusion
In conclusion, wild-type annexin V was successfully radi-
olabeled with 18F starting from rather low amounts of the
protein (0.1 mg) using both amino-targeting and thiol-tar-
geting bioconjugation methods, though with modest radio-
chemical yields. When compared with well-characterized,
randomly labeled [18F]SFB-annexin V, site-specific labeled
[18F]FBEM-annexin V showed no difference in the ability
to detect apoptotic cells. Since both radiolabeling methods
did not have any significant effect on PS-binding ability of
wild-type annexin V, it is more practical to use the simpler
Lehner S, Todica A, Vanchev Y, Uebleis C, Wang H, Herrler T, Wäng-
ler C, Cumming P, Böning G, Franz WM, Bartenstein P, Hacker
M, Brunner S (2014) In vivo monitoring of parathyroid hor-
mone treatment after myocardial infarction in mice with [68 Ga]
1 3

A. Perreault et al.
annexin A5 and [18F]fluorodeoxyglucose positron emission
tomography. Mol Imaging 13. doi:10.2310/7290.2014.00035
Li X, Link JM, Stekhova S, Yagle KJ, Smith C, Krohn KA, Tait JF
(2008) Site-specific labeling of annexin V with F-18 for apopto-
sis imaging. Bioconjug Chem 19:1684–1688
Mäding P, Füchtner F, Wüst F (2005) Module-assisted synthesis of the
bifunctional labelling agent N-succinimidyl 4-[(18)F]fluoroben-
zoate ([(18)F]SFB). Appl Radiat Isot 63:329–332
Tait JF, Smith C, Levashova Z, Patel B, Blankenberg FG, Vander-
heyden JL (2006) Improved detection of cell death in vivo with
annexin V radiolabeled by site-specific methods. J Nucl Med
47:1546–1553
Toretsky J, Levenson A, Weinberg IN, Tait JF, Uren A, Mease RC
(2004) Preparation of F-18 labeled annexin V: a potential PET
radiopharmaceutical for imaging cell death. Nucl Med Biol
31:747–752
Martin SJ, Reutelingsperger CP, McGahon AJ, Rader JA, van Schie
RC, LaFace DM, Green DR (1995) Early redistribution of
plasma membrane phosphatidylserine is a general feature of
apoptosis regardless of the initiating stimulus: inhibition by
overexpression of Bcl-2 and Abl. J Exp Med 182:1545–1556
Murakami Y, Takamatsu H, Taki J, Tatsumi M, Noda A, Ichise R, Tait
JF, Nishimura S (2004) 18F-labelled annexin V: a PET tracer for
apoptosis imaging. Eur J Nucl Med Mol Imaging 31:469–474
Neves AA, Brindle KM (2014) Imaging cell death. J Nucl Med
55:1–4
Schaper FL, Reutelingsperger CP (2013) 99 mTc-HYNIC-annexin A5
in oncology: evaluating efficacy of anti-cancer therapies. Can-
cers (Basel) 5:550–568
Smith BA, Smith BD (2012) Biomarkers and molecular probes for
cell death imaging and targeted therapeutics. Bioconjug Chem
23:1989–2006
Smith G, Carroll L, Aboagye EO (2012) New frontiers in the design
and synthesis of imaging probes for PET oncology: current chal-
lenges and future directions. Mol Imaging Biol 14:653–666
Tait JF, Cerqueira MD, Dewhurst TA, Fujikawa K, Ritchie JL, Strat-
ton JR (1994) Evaluation of annexin V as a platelet-directed
thrombus targeting agent. Thromb Res 75:491–501
Tait JF, Brown DS, Gibson DF, Blankenberg FG, Strauss HW (2000)
Development and characterization of annexin V mutants with endog-
enous chelation sites for (99 m) Tc. Bioconjug Chem 11:918–925
Tait JF, Gibson DF, Smith C (2004) Measurement of the affinity and
cooperativity of annexin V-membrane binding under conditions
of low membrane occupancy. Anal Biochem 329:112–119
Vangestel C, Peeters M, Mees G, Oltenfreiter R, Boersma HH, Els-
inga PH, Reutelingsperger C, Van Damme N, De Spiegeleer B,
Van de Wiele C (2011) In vivo imaging of apoptosis in oncology:
an update. Mol Imaging 10:340–358
Wängler C, Wängler B, Lehner S, Elsner A, Todica A, Bartenstein
P, Hacker M, Schirrmacher R (2011) A universally applicable
68 Ga-labeling technique for proteins. J Nucl Med 52:586–591
Wuest F, Berndt M, Bergmann R, van den Hoff J, Pietzsch J (2008)
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. Bioconjug
Chem 19:1202–1210
Wuest M, Perreault A, Kapty J, Richter S, Foerster C, Bergman C,
Way J, Mercer J, Wuest F (2015) Radiopharmacological evalu-
ation of 18F-labeled phosphatidylserine-binding peptides for
molecular imaging of apoptosis. Nucl Med Biol. doi:10.1016/j.
nucmedbio.2015.06.011
Yagle KJ, Eary JF, Tait JF, Grierson JR, Link JM, Lewellen B, Gib-
son DF, Krohn KA (2005) Evaluation of 18F-annexin V as a
PET imaging agent in an animal model of apoptosis. J Nucl Med
46:658–666
Zijlstra S, Gunawan J, Burchert W (2003) Synthesis and evaluation of
a 18F-labelled recombinant annexin-V derivative, for identifica-
tion and quantification of apoptotic cells with PET. Appl Radiat
Isot 58:201–207
1 3