High-Affinity “Click” RGD Peptidomimetics as Radiolabeled Probes for Imaging αvβ3 Integrin

Article abstract of DOI:10.1002/cmdc.201700328

Nonpeptidic Arg-Gly-Asp (RGD)-mimic ligands were designed and synthesized by click chemistry between an arginine–azide mimic and an aspartic acid–alkyne mimic. Some of these molecules combine excellent in vitro properties (high α_vβ₃ affinity, selectivity, drug-like logD, high metabolic stability) with a variety of radiolabeling options (e.g., tritium and fluorine-18, plus compatibility with radio-iodination), not requiring the use of chelators or prosthetic groups. The binding mode of the resulting triazole RGD mimics to α_vβ₃ or α_IIbβ₃ receptors was investigated by molecular modeling simulations. Lead compound 12 was successfully radiofluorinated and used for in vivo positron emission tomography/computed tomography (PET/CT) studies in U87 tumor models, which showed only modest tumor uptake and retention, owing to rapid excretion. These results demonstrate that the novel click RGD mimics are excellent radiolabeled probes for in vitro and cell-based studies on α_vβ₃ integrin, whereas further optimization of their pharmacokinetic and dynamic profiles is necessary for successful use in in vivo imaging.

Full text of DOI:10.1002/cmdc.201700328

Accepted Article
Title: High Affinity "Click" RGD Peptidomimetics as Radiolabeled
Probes for Imaging αvβ3 Integrin
Authors: Monica Piras, Andrea Testa, Ian N. Fleming, Sergio
Dall'Angelo, Alexandra Andriu, Sergio Menta, Mattia Mori,
Gavin D. Brown, Duncan Forster, Kaye J. Williams, and
Matteo Zanda
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemMedChem 10.1002/cmdc.201700328
Link to VoR: http://dx.doi.org/10.1002/cmdc.201700328
A Journal of
www.chemmedchem.org

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
High Affinity “Click” RGD Peptidomimetics as Radiolabeled
Probes for Imaging α Integrin
β
v 3
Monica Piras,* Andrea Testa, Ian N. Fleming, [a] Sergio Dall’Angelo, [a] Alexandra Andriu, [a] Sergio
[
a]
[a]
[b]
[c]
[d]
[d]
[e]
[a,f]
Menta, Mattia Mori, Gavin D. Brown, Duncan Forster, Kaye J. Williams, and Matteo Zanda*
Abstract: Non-peptidic Arg-Gly-Asp (RGD)-mimic ligands were
designed and synthesized by click chemistry between an arginine-
azide mimic and an aspartic acid-alkyne mimic. Some of these
Integrin α
v 3
β is a heterodimeric transmembrane protein that
[1,2]
belongs to the integrin superfamily receptors.
subtype attracted considerable attention because it plays a key
role in tumor angiogenesis, growth and metastasis, and in the last
This integrin
molecules combine excellent in vitro properties (high α
v 3
β affinity,
selectivity, drug-like logD, high metabolic stability) with a variety of
two decades it has become an established target for tumor
18
[3,4]
radiolabeling options (e.g. tritium and [ F]fluorine, plus compatibility
with radio-iodination), not requiring the use of chelators or prosthetic
treatment and diagnosis.
v 3
Integrin α β is a receptor for multiple
components of the extracellular matrix, including vitronectin,
fibrinogen and fibronectin, that expose the Arg-Gly-Asp (RGD)
tripeptide sequence.[5] The ligand-receptor recognition process is
mainly regulated by the aspartic acid carboxylic group and the
arginine guanidine group, both forming ionic interactions with a
charged region of the protein.[6] These two moieties need to be
groups. The binding mode of the resulting triazole RGD-mimics to α
v 3
β
or receptors was investigated by molecular modeling
IIb 3
α β
simulations. Lead compound 12 was successfully radio-fluorinated
and used for in vivo Positron Emission Tomography/Computed
Tomography (PET/CT) studies in U87-tumor models, which showed
only modest tumor uptake and retention, owing to rapid excretion.
These results demonstrate that the novel click-RGD mimics are
excellent radiolabeled probes for in vitro and cell-based studies on
separated by a 12–14 Å distance to allow the correct fitting into
7]
the α
v
β
3
binding pocket.[ Although a number of high affinity RGD
peptides have been described and some of them advanced to
clinical trials,[8] RGD peptide mimics may offer significant
advantages such as having better drug-like properties and
allowing more flexibility in the exploration of optimized structure-
activity-relationships with the receptor.[9] Structural modifications
made on the linear pharmacophore consist of variation of the
flanking amino acids and introduction of structural constraints,
which were achieved by cyclization or by insertion of a rigid moiety
v 3
α β integrin, whereas further optimization of their pharmaco-kinetic
and dynamic profile would be necessary for a successful use in in vivo
imaging.
Introduction
(
e.g. phenyl, isoxazole, 1,2,3-triazole rings and hydrazide bonds)
in the flexible peptide sequence.[
9–13]
In fact, the reduction of
conformational freedom towards a correct bioactive orientation
favorably affects both activity and selectivity. From several
studies focusing on the development of non-peptide RGD mimics,
L-2,3-diaminopropanoic acid emerged as an optimal replacement
of aspartic acid. Likewise, the substitution of the Arg guanidine
function with less basic moieties (2-aminoimidazoles, 2-
aminopyridines) has been proposed.[10,11] To improve the
pharmacological features of RGD peptides, a number of RGD
peptidomimetics have been developed as therapeutic agents.[
However, relatively few studies have investigated the use of these
molecules as radiotracers.[15–20] Kessler and co-workers assessed
the potential of non-peptidic RGD mimics as PET imaging agents
[
a]
Dr. Monica Piras, Dr. Andrea Testa, Dr. Ian N. Fleming, Dr. Sergio
Dall’Angelo, Alexandra Andriu and Prof. Dr. Matteo Zanda
Institute of Medical Sciences and Kosterlitz Centre for Therapeutics
School of Medicine, Medical Sciences and Nutrition
University of Aberdeen, Foresterhill, Aberdeen, AB25 2ZD
(Scotland, UK)
[
b]
c]
Dr. Sergio Menta*
Dipartimento di Chimica e Tecnologie del Farmaco
“Sapienza” Università di Roma, P.le A. Moro 5, 00185 Rome, Italy
14]
*
0
Current affiliation: IRBM Science Park SpA, Via Pontina km 30,600
0071 Pomezia (RM), Italy
[
Dr. Mattia Mori
Center for Life Nano Science@Sapienza
Istituto Italiano di Tecnologia
by equipping the integrin antagonists with a bifunctional chelator
Viale Regina Elena 291, 00161 Roma (RM), Italy
3+ [21]
(
NODAGA) and labeling with 68Ga .
More recently, the direct
[
d]
e]
Dr. Gavin D. Brown and Dr. Duncan Forster
Manchester Cancer Research Centre and Wolfson Molecular
Imaging Centre, The University of Manchester
Palatine Road, Manchester, M20 3JJ (UK)
radioiodination of non-peptide integrin ligands has been reported
and the resulting tracers assessed as SPECT imaging agents.[22]
In this study we designed small-molecule RGD-mimics that
combine excellent biological properties (high affinity, selectivity,
suitable logD value) with a variety of radiolabeling options not
requiring the use of chelators or prosthetic groups. To this end, a
small library of RGD peptidomimetics was synthesized and
biologically screened to identify the chemical modifications that
[
Prof. Dr. Kaye J. Williams
CRUK-EPSRC Cancer Imaging Centre in Cambridge and
Manchester, Manchester Cancer Research Centre
Division of Pharmacy and Optometry
The University of Manchester
Oxford Road, Manchester, M13 9PT (UK)
Prof. Dr. Matteo Zanda
[f]
C.N.R. – I.C.R.M., via Mancinelli 7, 20131 Milan (Italy)
confer the highest α
v 3 3
β affinity and selectivity against integrin αIIbβ ,
Supporting information for this article is given via a link at the end of
the document.
which mediates platelet aggregation. The main features of these
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
novel
RGD
mimics
are: (a)
α-N-(arylsulfonyl)-L-2,3-
d. N-methylation of the sulfonamide function of 5a afforded the
intermediate 5e. After acidic removal of the N-Boc protecting
group from 5a–e, an amide-forming coupling was performed with
propiolic acid, leading to the desired alkyne intermediates 6a–e.
The click partners 8a–c were synthesized by alkylation of N-Boc
protected 2-amino-3-methylpyridine 7. Coupling of alkynes 6a–e
with azides 8a–c via CuAAC reaction afforded 1,2,3-triazole
intermediates 17a–f and 17i. The iodinated 5-iodo-1,2,3-triazole
analogues 17g–h were prepared by coupling alkynes 6a and 6c
with the corresponding “click partners” 8c and 8b - respectively -
in the presence of CuI, N-iodosuccinimide (NIS) and
trimethylamine. Treatment of the methyl ester and N-Boc
protected intermediates (17a–i) with LiOH followed by TFA
afforded the desired products 1 and 9–16 (for structural details
see Table 1).
diaminopropanoic acid and 2-amino-3-methyl pyridine as Asp and
Arg substituent, respectively; (b) a distance of about 13 Å
between the acidic and basic groups in the predicted binding
conformation; (c) a triazole ring as Gly mimetic, which confers
favorable conformational restrictions to the linear scaffold and
[
23]
provides a versatile labeling site for the incorporation of tritium
and iodine[ radioisotopes; (d) an electron-poor aromatic ring
activated for 18F-labeling via nucleophilic aromatic substitution[25]
on the Asp-mimic portion (Figure 1). Notably, molecular modelling
studies suggest that this fluorine atom may significantly contribute
to the affinity and selectivity of these ligands through an attractive
24]
electrostatic interaction with the Tyr178 hydroxyl group of the α
subunit.
v
Figure 1. Structural scheme of triazole based RGD mimics with the triazole ring
as labeling site for tritium or iodine radioisotopes and phenyl ring as labeling site
1
8
for F.
Results and Discussion
Design and Synthesis. The copper-(I)-catalyzed 1,2,3-triazole
cycloaddition allowed a convenient assembly of all the target
compounds via “click chemistry” (Scheme 1). Two regioisomers
(
1 and 2) were initially prepared by swapping the orientation of the
triazole ring (Figure 2). The higher affinity ligand 1 was then
selected for the optimization of the biological properties and the
design of labeled derivatives 12 and 15.
Figure 2. Triazole based RGD mimic regioisomers 1 and 2 and evolution of 1
into the optimized α
β
v 3
ligands 12 and 15.
Scheme 1. Synthesis of compound 1 and its analogues (9–16). LG = Leaving
Group. Reagents and conditions: (i) TEA, DCM, 12h, 0°C to room temp (78–
9
2
9%); (ii) K
h; (iv) EEDQ, propiolic acid, DCM, 1h (50–90%); (v) NaH, K
(2%), Na ascorbate (10%), tBuOH/H
(vii) NIS, CuI, TEA, CH CN, 2h (45–47%) (viii) LiOH, H
2
2
CO
3
, TEBA, MeI, CH
3
CN, 2h, room temp (85%); (iii) TFA/DCM 1:2,
CO , DMF, 1.5 h
O 2:1, 12h (60–
O/THF 1:2, 20
2
3
The synthesis of compound 1 and its analogues 9–16
(90–95%); (vi) CuSO
4
2
(
substituents shown in Table 1) is outlined in Scheme 1. The
98%)
;
3
min then TFA/DCM 1:1, 2h (quantitative yields).
synthesis of the alkyne building blocks started from commercially
available (S)-methyl-2-amino-3-(tert-
butoxycarbonylamino)propanoate hydrochloride (3) that was
reacted with arylsulfonyl chlorides 4a–d to give sulfonamides 5a–
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
The regioisomer
2
was prepared by swapping the
Arg-mimetic fragments to allow a correct fitting of the ligand into
functionalization of the “click partners” relative to 1, namely
incorporating the azide moiety on the Asp-mimic fragment 18 and
the alkyne counterpart on the Arg-like side chain (19). Alkyne 19
was prepared from 7 as already described.[23] The synthesis of the
azide fragment 18 was accomplished by coupling of azidoacetic
acid with the amine 20, that was prepared according to the
literature.[26] “Click”-cycloaddition of the azide 18 with the alkyne
the binding pocket. Interestingly, the activity on α
selectivity against αIIb seem to be strongly dependent on the
substitution of the phenyl ring with fluorine and EWGs such as
nitro (NO ) and trifluoromethyl (CF ) moieties (entries 4–6). An
increased affinity (2 to 4-fold) and a remarkable improvement in
selectivity (up to 36-fold) were, in fact, observed in this series of
derivatives (compare entry 2 with 4–6).
v 3
β and the
β
3
2
3
1
9 gave the 1,2,3-triazole 21. Treatment of 21 with TFA afforded
The IC50 values of iodinated peptidomimetics 14 and 15 compared
with their non-iodinated analogues 1 and 12 revealed that
incorporation of iodine on the triazole ring did not affect affinity
and selectivity (entries 3 vs 7 and 5 vs 8). Conversely, the N-
the desired deprotected regioisomer 2 (Scheme 2).
Table 1. IC50 values of triazole-based RGD peptidomimetics and c(RGDfK)
pentapeptide concerning the inhibition of c[RGDfK(Biotin-PEG-PEG)] peptide
and biotinylated fibrinogen binding to the isolated α β and αIIbβ receptors,
v 3
respectively.
[
a]
[a]
IC50
IC50
[c]
entry
comp
structure
α
v
β
3
α
IIb
β
3
SI
(
nM)
(nM)
9
00 ±
8
6090 ±
135
1
2
3
4
10
9
6.8
12
6
70 ±
1
57 ± 1
Scheme 2. Synthesis of 2. Reagents and conditions: (i) azidoacetic acid, DCC,
DCM, 12 h (99%); (ii) CuSO
4 2
(2%), Na ascorbate (10%), tBuOH/H O 2:1, 12 h
(90%); (iii) TFA/DCM 1:1, 3 h (quantitative yield).
1
10 ±
500 ±
3
1
4.5
23
5
6
To investigate the possibility of radio-iodinating these RGD-
mimics, preliminary studies were performed using cold non-
radioactive conditions, applying a synthetic procedure reported by
Årstad et al for the synthesis of imaging agents containing a 5-
8
1 ±
1845 ±
63
11
51
1
±
1690
2840
[125
alkyne 6a and azide 8c was performed in the presence of a
stoichiometric amount of CuCl , TEA (1.5 equiv) and NaI (1 equiv).
I]Iodo-1,2,3-triazole group.[24] The click reaction between
5
12
25 ± 3
10 ± 3
467
2
2
4400
The reaction was monitored via LC-MS (Figure S1 top panel,
6
7
8
9
13
14
15
16
2
2444
22
±
450
Supporting Information) showing the non-iodinated triazole
1
2
analogue 17a (R, R , R , X = H, n=4) as the major by-product. In
1
10 ±
2450 ±
1390
order to optimize the labeling conditions, chloramine-T (1 equiv)
3
4
was
tetrakis(acetonitrile)copper(I)
[Cu(CH CN) ]PF ) (1 equiv) as a source of copper(I). Under
employed
as
an
oxidizing
agent
and
hexafluorophosphate
(
3
4
6
6670 ±
21
10 ± 1
3600
667
1
2
these conditions the iodinated compound 17g (R, R , R = H, X =
I, n=4) was detected as the major peak in the HPLC
chromatogram (see Figure S1 bottom panel, Supporting
Information) with no observed 17a formation.
-
-
-
±
18
Receptor binding assays. The design of triazole containing
RGD-mimics afforded a set of 10 molecules that were submitted
640 ±
12
1
0
1
-
to competition binding assays to determine their binding affinity
27]
toward integrin α
v
β [
3
and selectivity over the closely related
1
38 ±
20250
± 4630
integrin αIIbβ . Results are reported in IC50 values (Table 1). Cyclic
3
1
c[RGDfK]
146
5
1
c[RGDfK] was included in the tests as positive control in order to
compare the new RGD mimics with a well-established rigid ligand.
These tests showed that ligand length (n) has a significant impact
[
a] IC50 values were calculated as the concentration of ligand required for 50%
inhibition as estimated by GraphPad Prism software. Each data point is the
average from multiple independent experiments, performed in triplicate. [b]
on α
v 3
β affinity (entries 1–3), and a propyl chain (n = 3, entry 2)
was found to provide the optimal distance between the Asp- and
IIb 3 v 3
Selectivity Index (SI) = IC50 α β / IC50 α β .
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
methyl sulfonamide derivative 16 showed a considerable drop in
affinity (entry 9), confirming the crucial role that the free
sulfonamide function -NHSO
interactions. Comparison of the IC50 values of non-peptide ligands
2, 13 and 15 with that of c(RGDfK) reference peptide (entry 11)
indicates a considerable increase of binding affinity for α in this
series of analogues (to almost 14-fold) as well as an enhanced
selectivity over αIIb
2
-
plays in ligand-receptor
1
v 3
β
3
β .
Molecular Modelling. To provide structural support to biological
activity data shown in Table 1, the binding mode of triazole RGD-
mimics to α β or αIIbβ receptors was investigated by molecular
v 3 3
modeling simulations. Entries 1–10 were docked towards the
crystallographic structure of each receptor,[6,28] and lowest energy
poses were further relaxed by extensive energy minimization.
[
29,30]
Ligand theoretical affinity was computed by the MM-GBSA
and XSCORE[ methods providing a very good correlation with
experimental inhibitory activity data (R = 0.896 and R = 0.662,
respectively (see also Supporting Information). Notably, the
31]
2
2
Figure 3. Predicted binding mode of RGD-mimic 15 (yellow sticks) on the
crystallographic structure of α β receptor (PDB: 1L5G). α and β subunits are
v 3 v 3
predicted binding mode of RGD-mimics is highly comparable with
that of the crystallographic RGD ligand,[6] since they establish
canonical interactions with the α β receptor as observed for well-
v 3
colored cyan and green, respectively. Residues contacted by RGD-mimics are
shown as sticks and labeled. The divalent cation in the MIDAS region is shown
as a violet sphere.
known inhibitors. However, molecular docking suggested that
triazole RGD-mimics 11, 12, 13 and 15 bearing a nitro group or
fluorine atoms substituted at the phenyl ring are able to form an
Labeling. Based on the affinity data discussed above,
compounds 12 and 15, which are fluorinated and iodinated
versions of the initial lead 1, represent potential radiotracers for
PET and SPECT imaging. Furthermore, tritiated versions could
be useful tools for in vitro studies. We therefore investigated the
synthesis of 12 and 15 in different radiolabeled forms (tritiation
and radiofluorination).
additional H-bond interaction with the side chain of Tyr178 (α
v
subunit) (Figure 3). This interaction has never been exploited
before for improving the affinity of integrin inhibitors and, in
agreement with biological activity data of Table 1 and theoretical
binding affinity (see Supporting Information), it may be
responsible for the higher inhibitory potency against the α
receptor displayed by these molecules. Most notably, Try178 is
missing on the αIIb receptor, where it is replaced by Phe160 that
v 3
β
Tritiation. Compound 12 was successfully labeled with tritium
(
Scheme 4). To the best of our knowledge this is the first example
β
3
3
of H-labeled non peptidic molecule able to bind integrin α
v
β
3
,
is however unable to establish H-bond interactions with these
RGD-mimics, thus providing a structural explanation to the high
selectivity for α β observed for 11, 12, 13 and 15.
v 3
Iodine substitution is another unique feature of these triazole
RGD-mimics, which was found to increase the binding affinity of
3
therefore [ H]12 may represent a very useful tool for in vitro or ex-
vivo studies on angiogenesis. The palladium-mediated
dehalogenation was accomplished on the 5-iodotriazole 15
employing tritium gas in a tritium handling manifold in the
presence of Pd/C and TEA (excess) in ethanol. Complete
conversion of the iodinated starting material was obtained after
15 by experimental and molecular modelling studies (compare
entry 5 with 8). Indeed, the iodine atom in 15, the most potent
RGD-mimic of this series, is able to establish hydrophobic
interactions with Tyr178 and Trp179 (α ) and Ala218 (β subunit).
V 3
In contrast, iodine substitution in RGD-mimic 14, which bears a
four-carbon linker, did not provide affinity gain, as the iodine is
exposed to the solvent due to steric restrictions imposed by the
length of the linker (compare entry 3 with 7, see also Supporting
Information).
2
.5 hours and the desired product was isolated by analytical
HPLC with a radiochemical purity > 99% and a specific activity of
69 GBq/mmol.
4
The sulfonamide group of all molecules but 16 contacts the
backbone of Asn215 and the side chain of Arg214 of the β
3
subunit by H-bonds, thus providing a rational explanation for the
loss of inhibitory activity observed upon methylation of the
sulfonamide NH function (compare entry 2 with 9).
For additional details on binding modes and theoretical affinities
see the Supporting Information.
3
2
Scheme 4. Synthesis of [ H]12. Reagents and conditions: Pd/C cat, TEA, T ,
EtOH, 2.5 h, room temp.
3
A logD (pH 7.4) = –0.55 was experimentally measured for [ H]12
(
see Supporting Information for details). The compound was
evaluated in cells that are known to express no, moderate and
integrin.[32,33] The magnitude of [ H]12
3
very high levels of α
v
β
3
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
binding to each cell line was consistent with α
v
β
3
expression levels
The radiosynthesis was performed as a one-pot, two-step method
using the Tracerlab FX-FN platform (GE Healthcare). For the
initial step, a [K/Kryptofix222]+ 18F- complex was prepared and
dried azeotropically under a stream of helium and vacuum at 110
oC. Second, [18F]12 was prepared by Kryptofix-mediated
(
Figure 4A). Non-radiolabeled c(RGDfK) (10µM) was included to
3
v 3
competitively block specific binding to α β . This decreased [ H]12
binding to PC3 and U87MG cells by approximately 60%, but had
no effect in MCF7 cells (Figure 4B).
[18
F]nucleophilic fluorination at the nitro position of 13. Isolation of
1
8
pure [ F]12 was achieved by semi-preparative reverse phase
HPLC (Figure 5). The isolated product was then eluted through a
C18 SEP-PAK to remove the HPLC eluent and formulated in a
mixture of 0.9% saline/ethanol (92:8) for preclinical application.
1
9
After radioactive decay, the cold [ F]12 - which had formed in the
18
radiofluorination concomitantly with [ F]12 - was analyzed by
proton-decoupled 19F-NMR (376 MHz) to verify the fluorination
position (for a copy of the spectrum see Figure S5, Supporting
Information). A single peak at –110.1 ppm was observed, having
the same chemical shift as the fluoro group of authentic 12. The
18
radiochemical yield (decay corrected) of [ F]12 from nucleophilic
1
8
Figure 4. (A) Characterization of α
β
v 3
expression on MCF7, PC3 and U87MG
[ F]fluoride was 36–41% giving 2.54–2.90 GBq (n = 8) of product
at the end of the radiosynthesis (EOS). The radiochemical purity
was ≥99%, and the specific radioactivity was 248–281 GBq/mol
cells by flow cytometry. Results are expressed as Median Fluorescence
Intensity (MFI) of stained/unstained cells. (B) Cells were incubated with [ H]12
in the presence (+) or absence (–) of 10 µM c(RGDfK) peptide for 1 h. [ H]12
binding is expressed as counts per minutes (cpm), as a function of the
radioactivity added/plate and normalized for protein content. Each data point is
the average ± standard deviation of at least three independent data points. T-
test: Binding in presence vs absence of c(RGDfK) peptide; MCF7 p>0.05; PC3
p<0.001, U87MG p<0.0001.
3
3
(
EOS). The total radiosynthesis time was approximately 80 min.
In vivo Imaging. Promising in vitro data and successful 18F-
radiolabeling prompted pilot in vivo analyses. Four mice
implanted with U87 tumor xenografts on the back were subjected
1
8
to PET/CT imaging with [ F]12. Overall tumor retention of the
tracer was low throughout the dynamic scan, whereas clearance
was rapid, exhibiting both hepatobiliary and renal excretion.
Uptake was initially high in the kidneys and liver but dropped off
over the course of the 90 mins scan (Figure 6). By 10 mins most
of the injected dose could be found in the urinary bladder, where
by 90 mins virtually all of the activity could be detected (Figure 6
A). The low in vivo tumor uptake (0.9 ± 0.1% ID / g) and retention
1
8
F-radiolabeling. Radiolabeling of compound 12 with 18F-fluorine-
1
8
was achieved via [ F]-fluoride nucleophilic substitution using the
nitro derivative 13 as precursor (Scheme 5).
1
8
of [ F]12 appears to be partly due to low bioavailability resulting
in rapid excretion. The rapidity of this process and lack of kidney
retention suggest that the molecule is excreted unchanged
without any prior metabolism. This hypothesis is supported by
human and rat liver microsomal (HLM / RLM) stability tests
performed on 12, whose half-life upon incubation with HLM and
RLM at 37 ºC in the presence of the co-factor NADPH was very
high, namely 770 and 470 min, respectively, as detected by LC-
MS/MS analysis. In comparison, in similar U87-MG mouse
Scheme 5.
1
8
xenograft models [ F]fluciclatide was reported to have a tumor-
retention comprised between 1.28 ± 0.34 % ID/g and 1.88 ±
0
.73 % ID/g (13 days later) 90 mins after injection.[34]
Autoradiography. Sections of the U87-cells-derived tumors were
mounted on slides and, after standard preparations, treated
1
8
18
either directly with [ F]12 or [ F]12 incubation combined with
cold 12 for determining the presence of non-specific binding. After
18
18
overnight exposure to [ F]12 the sections treated with [ F]12
alone showed greater overall binding than the sections treated
with an excess of cold 12 as well (Figure 6B), indicating specific
binding to α
studies. Expression of integrin α
v
β
3
integrin comparable to that observed in in vitro
in the U87 tumor xenografts
v 3
β
Figure 5. Radio-RP-HPLC analysis of [18F]12 (above) and its UV profile (below).
used was confirmed by immunohistochemical staining (see
Figure S3, Supporting Information).
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
Chemicals were purchased from Sigma Aldrich and Bachem. NMR spectra
were recorded using a Bruker Avance-400 spectrometer equipped with a
5
mm single-axis Z-gradient quattro nucleus probe, operating at 376 MHz
for 19F NMR. The spectrometer was operated with TOPSPIN NMR
[18F]12 +
block
A.2 . Tumour
B.
[18
F]12
software (Version 2.0). Chemical shifts () are reported in parts per million
1
(
ppm), relative to trichloro-fluoro-methane (0.00 ppm).QMA SEP-PAK light
1
Tumour 1
Tumour 2
cartridges and C18 SEP-PAK plus cartridges were obtained from Waters
(Milford, MA). No-carrier-added [ 8F]fluoride was produced via the 18O(p,
n) F nuclear reaction by irradiation of a 95-97% [18O]enriched water target
with a 16.4 MeV proton beam on the GE PETtrace cyclotron at the Wolfson
Molecular Imaging Centre, Manchester, UK. Automated synthesis of
0
0
0
0
.8
.6
.4
.2
0
1
18
Tumour 3
Tumour 4
[18
F]12 was carried out using the Tracerlab FX-FN radiochemistry system.
0
20
40
60
80
100
-
0.2
Purification of [18F]12 by HPLC was carried out on the same system
equipped with a preparative reverse phase C18 column (Prodigy ODS-
Prep, 10µm particle size; 250 × 10 mm i.d.; Phenomenex, UK). For the
purpose of quality control, HPLC analysis was carried out on a Shimadzu
prominence system operated using LabLogic software (Laura 3.0) and
configured with a CBM-20A controller, LC-20AB solvent delivery system
and SPD-20A absorbance detector. The system was equipped with a C18
Prodigy ODS(3) column, 5 µm particle size, 250 × 4.6 mm. Radioactivity
was monitored with a radio-HPLC Bioscan Flowcount B-FC 3100 detector.
Time / Min
C. Kidney
D. Liver
6
5
4
3
2
1
0
0
0
0
0
0
0
6
5
0
0
40
30
20
10
0
0
20
40
60
80
100
0
20
40
60
80
100
Time / min
Time / min
E. Bladder
2
2
1
1
50
00
50
00
Chemistry.
Synthesis of 17e. Alkyne 6c (67.5 mg, 0.17 mmol, 1 equiv) and azide 8b
50
(50 mg, 0.17 mmol, 1 equiv) were suspended in a mixture of tBuOH/H
2
O
0
0
20
40
60
80
100
2:1 (3 mL/mmol). CuSO (2%) and sodium ascorbate (10%) were added
4
Time / min
and the mixture was stirred at room temperature for 12 h. The reaction
mixture was diluted with water, extracted with ethyl acetate three times and
Figure 6. (A) Time-activity curve of [18F]12 tumor uptake in U87 xenograft
tumors. (B) Autoradiography of tumor sections. Pre-treatment of U87 tumor
sections with cold 12 (“block”) reduces [18F]12 binding. (C–E) Time-activity
curves of kidney, liver and bladder. Data in A, C–E are mean values obtained
from 4 mice. Error bars indicate SD.
2 4
washed with brine. The collected organic phases were dried over Na SO ,
filtered and the solvent removed under vacuum. The crude mixture was
purified by preparative thin-layer chromatography (PLC) (silica gel, n-
Hex/EtOAc 4:6 as eluent) affording 96.5 mg (83% yield) of a white solid:
1
R
f
= 0.5 (n-Hex/EtOAc 3:7); H NMR (400 MHz, CDCl
3
) δ 8.24 (s, 1H), 8.20
(
1
(
1
d, J = 5.1 Hz, 1H), 8.12 – 7.89 (m, 3H), 7.43 (s, 1H), 7.20 (t, J = 9.2 Hz,
H), 6.89 – 6.84 (m, 2H), 4.49 (t, J = 7.1 Hz, 2H), 4.34 – 4.26 (m, 1H), 3.99
t, J = 6.7 Hz, 2H), 3.91 – 3.72 (m, 2H), 3.65 (s, 3H), 2.43 – 2.21 (m, 5H),
.50 (s, 9H); 19F NMR (376 MHz, CDCl
) δ –61.74 (d, J = 12.5 Hz), –107.17
q, J = 12.5 Hz); 13C NMR (101 MHz, CDCl
) δ 170.1, 161.7 (d, J = 262.3
Hz), 161.0, 154.2, 153.9, 148.5, 147.2, 142.1, 137.2 (d, J = 3.8 Hz), 133.3
Conclusions
3
(
3
Novel small-molecule RGD mimics were designed and
synthesized by click chemistry between an arginine-azide mimic
and an aspartic acid-alkyne mimic. Some of these compounds
(
(
d, J = 9.8 Hz), 126.9, 126.2, 121.6 (q, J = 273.3 Hz), 121.1, 119.9, 118.9
dd, J = 34.0, 13.5 Hz), 117.9 (d, J = 21.8 Hz), 81.6, 56.2, 52.9, 48.7, 43.9,
displayed excellent in vitro properties (high α
β
v 3
affinity and
41.2, 29.3, 28.3, 21.1; MS (ESI, m/z): C H F N O S [M+H] calc. 688.21
found 688.2. [α]
+
28
34
4
7
7
25
selectivity, drug-like logD, high HLM/RLM stability), are amenable
D
3
= –3 (c = 0.7, CHCl ).
18
to radiolabeling with iodine, or tritium, or [ F]fluorine, and do not
require the use of chelators or prosthetic groups. Molecular
modeling simulations provided an insight into the binding mode of
Synthesis of 12. The methyl ester and N-Boc diprotected intermediate
7e (91 mg, 0.13 mmol) was dissolved in THF (0.2 M) and a solution of
1
LiOH 4 M in water (500 μL/mmol) was added. The reaction mixture was
vigorously stirred at r.t. After cleavage of the methyl ester function (20 min)
the THF was removed in vacuo and the reaction mixture was partitioned
v 3 3
these compounds to α β or αIIbβ receptors. Compound 12 was
successfully radiofluorinated and used for in vivo PET/CT studies
in U87-tumor models, which showed only modest tumor uptake
and retention, partly owing to rapid excretion through the kidneys
into urine. The results above demonstrate that the novel click-
RGD mimics are excellent radiolabeled probes for in vitro and cell-
2 4
between DCM and HCl (1 M aq.). The organic layer was dried over Na SO
and concentrated under low pressure. The resulting colourless oil was
dissolved in a solution of TFA/DCM 1:1 and stirred for 2 h at r.t. for removal
of the Boc protecting group. The solution was concentrated bubbling
nitrogen into the flask and then lyophilized. After lyophilisation the product
v 3
based studies on α β integrin, whereas further optimization of
appeared as a white solid (90 mg, quantitative yield, MW = 12 · 1 TFA =
their pharmaco-kinetic and dynamic profile will be required for a
successful use in in vivo imaging via SPECT or PET/CT.
1
5
8
6
4
1
73.5 + 114 = 687.5): H NMR (400 MHz, CD
.12 (m, 2H), 7.73 (d, J = 6.6 Hz, 1H), 7.44 (t, J = 9.9 Hz, 1H), 6.86 (s, 1H),
.80 (dd, J = 6.7, 1.4 Hz, 1H), 4.63 (t, J = 6.8 Hz, 2H), 4.30 (dd, J = 8.8,
.9 Hz, 1H), 3.81 (dd, J = 13.8, 4.9 Hz, 1H), 3.55 (dd, J = 13.8, 8.9 Hz,
H), 3.42 (t, J = 6.8 Hz, 2H), 2.43 (s, 3H), 2.40 – 2.31 (m, 2H); 19F NMR
3
OD) δ 8.37 (s, 1H), 8.16 –
Experimental Section
(
376 MHz, CD
Hz); 13C NMR (101 MHz, DMSO) 171.4, 160.2, 161.0 (d, J = 260.0 Hz),
58.61, 153.2, 142.7, 138.63 (d, J = 3.6 Hz), 136.5, 134.26 (d, J = 10.5
3
OD) δ –63.20 (d, J = 12.7 Hz), –76.97, –110.65 (q, J = 12.7
Materials and instrumentation
1
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
Hz), 127.0, 126.5, 122.2 (q, J = 273.4 Hz), 118.78 (d, J = 21.8 Hz), 117.51
end of PET scanning, the mice will be removed from the scanner and
sacrificed. Tumors were excised and sectioned for autoradiography and
immunohistochemistry. Inveon acquired data: before reconstruction, the
list-mode data were histogrammed with a span of 3 and maximum ring
differences of 79 into 3D sinograms with 19 time frames (5 × 60 seconds
(dd, J = 33.4, 13.4 Hz), 114.6, 111.8, 55.7, 47.7, 40.5 (signal detected by
HSQC as it overlaps with solvent peak), 39.5 (partially overlaps with
+
solvent peak), 28.9, 21.8; HRMS (ESI, m/z): C22
H
24
F
4
N
7
O
5
S [M+H] calc.
25
574.14 found 574.1475. [α]
D
= +5 (c = 1.3, DMSO).
(s), 5 × 120s, 5×300s and 4 × 300s) for image reconstruction. Images were
reconstructed using the 3D-OSEM/MAP algorithm (4 OSEM3D iterations
and no MAP iterations, with a requested resolution of 1.5mm). Regions of
interest (ROIs) were drawn manually over tumors using Inveon Research
Workplace software (Siemens, Germany) and further normalization was
performed using the injected dose (from the dose calibrator) and animal
weight to give a standardized uptake value (SUV). SUVmax was
calculated from the maximum voxel value within the ROI and SUVmean
as the average over all voxels.
v 3
Flow cytometry analysis of α β integrin expression
Cells were seeded at 0.35 × 106 cells/60 mm plate and grown for 72 h
before the experiment. Cells were harvested from plates using cell
dissociation reagent. α
cytometry as previously described,
binds selectively to α integrin and detected with an Alexa Fluor 488
v 3
β cell surface expression was determined by flow
[33]
using the LM609 antibody which
v 3
β
secondary antibody. Results are expressed as Median Fluorescence
Intensity (MFI) of stained/ unstained cells (i.e. fluorescence in the
Autoradiography
v 3
presence/absence of α β antibody).
Sections of flash-frozen U87 tumors cut at 20m on a cryotome and
mounted on slides were pre-incubated in Tris-NaCl at 4 C for 5 mins. The
Radiotracer binding assays
o
Binding assays were performed as described in the literature.[
27,33]
slides were then incubated for 1 hour in Tris-NaCl plus ligands at room
temp (50MBq of [18F]12 in 50mL Tris-NaCl in both, plus 2 mL of 100ug/mL
cold [18F]12 for non-specific binding test). The slices were rinsed with 2 ×
Automated synthesis of [18F]12
o
o
5
2
mins in Tris-NaCl at 4 C, then quickly rinsed in H O at 4 C. Slides were
prepared for phosphor-imager (FujiFilm BAS-1800 II) and exposed
overnight.
The automated synthesis of [18F]12 was carried out on the TRACERlab
FX-FN radiochemistry system as follows. No-carrier-added [18F]fluoride
was delivered to the system and passed through a QMA SEP-PAK light
cartridge where the [18F]fluoride was trapped and [18O]water collected for
recycling. Trapped [18F]fluoride was eluted off the cartridge with a solution
of Kryptofix 2.2.2 (K222) / 0.1 M K CO / ACN (1 mL) into a reaction vessel.
2 3
Acknowledgements
The mixture was azeotropically dried under a stream of helium and
vacuum at 110 C for 4 min. The drying step was repeated with acetonitrile
We thank The Development Trust, University of Aberdeen, for
financial support and a fellowship to M.P. The work was also
supported by the CRUK-EPSRC Cancer Imaging Centre in
Cambridge and Manchester (KJW Co-I; reference 16465). We
thank Dr Massimiliano Baldassarre (University of Aberdeen) for
helpful discussions.
o
(
1 mL) then compound 13 (2.0 mg) in anhydrous DMSO (1.0 mL) was
o
added and the reaction mixture heated at 120 C for 20 min. At the end of
the reaction a 1 mL mixture of ACN / water (35/65, v/v; 0.01% TFA) was
added and the entire sample loaded onto a C18 Prodigy ODS-Prep column
(10µm particle size; 250 × 10 mm) eluted with ACN/water (35/65, v/v;
0.01% TFA) at a flow rate of 3 mL/min. Eluate from the column was
monitored for radioactivity and UV absorbance at 320 nm. The fraction
Keywords:
containing [18F]12 was collected in a round bottom flask containing water
[
1]
M. Shimaoka, T. A. Springer, Nat. Rev. Drug Discov. 2003, 2, 703–
16.
(40 mL) and the resultant mixture passed through a C18 SEP-PAK plus
7
cartridge. The cartridge was rinsed with water (10 mL) and the product
eluted with sterile ethanol (1 mL) and collected in sterile 0.9 % saline (10
mL). Finally, the product was dispensed into a sterile vial via a 0.2 micron
sterile filter for preclinical application. The radiochemical purity of the
product was determined by HPLC using a C18 Prodigy ODS (3) column (5
µm particle size; 250 × 4.6 mm) eluted with ACN/water (35/65, v/v; 0.01%
TFA) at a flow rate of 2 mL/min. The retention time for [18F]12 was 8.6 min.
[
[
2]
3]
H. Hamidi, M. Pietilä, J. Ivaska, Br. J. Cancer 2016, 115, 1017–1023.
P. C. Brooks, R. A. Clark, D. A. Cheresh, Science 1994, 264, 569–
571.
[
[
4]
5]
D. Arosio, C. Casagrande, Adv. Drug Deliv. Rev. 2016, 97, 111–143.
J. D. Humphries, A. Byron, M. J. Humphries, J. Cell Sci. 2006, 119,
3901–3903.
[
[
[
6]
7]
8]
J.-P. Xiong, T. Stehle, R. Zhang, A. Joachimiak, M. Frech, S. L.
Goodman, M. A. Arnaout, Science 2002, 296, 151–155.
L. Marinelli, A. Lavecchia, K.-E. Gottschalk, E. Novellino, H. Kessler,
J. Med. Chem. 2003, 46, 4393–4404.
C. Mas-Moruno, F. Rechenmacher, H. Kessler, Anticancer Agents
Med. Chem. 2010, 10, 753–768.
PET/CT imaging.
7
[9]
R. Haubner, D. Finsinger, H. Kessler, Angew. Chem. Int. Ed. Engl.
Four female cba nu/nu mice were implanted with 0.1 mL of a 5×10 /mL
1997, 36, 1374–1389.
suspension of U87 cells i/d on the back. Tumors were measured daily once
since a palpable tumor became apparent. Once tumors reached approx.
00mm , PET scans were performed to assess [18F]12 tracer uptake and
[
10]
W. J. Pitts, J. Wityak, J. M. Smallheer, A. E. Tobin, J. W. Jetter, J. S.
Buynitsky, P. P. Harlow, K. A. Solomon, M. H. Corjay, S. A. Mousa, et
al., J. Med. Chem. 2000, 43, 27–40.
D. Heckmann, A. Meyer, L. Marinelli, G. Zahn, R. Stragies, H. Kessler,
Angew. Chem. Int. Ed Engl. 2007, 46, 3571–3574.
A. Trabocchi, G. Menchi, N. Cini, F. Bianchini, S. Raspanti, A.
Bottoncetti, A. Pupi, L. Calorini, A. Guarna, J. Med. Chem. 2010, 53,
7119–7128.
G. A. Sulyok, C. Gibson, S. L. Goodman, G. Hölzemann, M. Wiesner,
H. Kessler, J. Med. Chem. 2001, 44, 1938–1950.
W. H. Miller, R. M. Keenan, R. N. Willette, M. W. Lark, Drug Discov.
Today 2000, 5, 397–408.
L. Peng, R. Liu, J. Marik, X. Wang, Y. Takada, K. S. Lam, Nat. Chem.
Biol. 2006, 2, 381–389.
3
5
[
[
11]
12]
biodistribution. Mice were anaesthetized by placing in an induction
chamber and turning the isofluorane to 2% (oxygen at 2 L/min). Mice were
transferred to an anesthetic facemask on a heated bed and settled for 5
minutes and the tail vein dilated and catheterized. Mice were then
transferred to the temperature-controlled imaging bed and attached to the
anesthetic facemask. After CT scanning, 10-15MBq [18F]12 in a total
volume < 200µL was injected i/v. Imaging commenced immediately after
tracer injection, during which time body temperature and depth of
anesthesia were monitored using readouts from the imaging bed. At the
[
[
13]
14]
[15]
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
[
16]
T. D. Harris, S. Kalogeropoulos, T. Nguyen, S. Liu, J. Bartis, C. Ellars,
S. Edwards, D. Onthank, P. Silva, P. Yalamanchili, et al., Cancer
Biother. Radiopharm. 2003, 18, 627–641.
B.-S. Jang, E. Lim, S. Hee Park, I. S. Shin, S. N. Danthi, I. S. Hwang,
N. Le, S. Yu, J. Xie, K. C. P. Li, et al., Nucl. Med. Biol. 2007, 34, 363–
[26]
[27]
[28]
[29]
M. E. Solomon, C. L. Lynch, D. H. Rich, Tetrahedron Lett. 1995, 36,
4955–4958.
M. Piras, I. Fleming, W. Harrison, M. Zanda, Synlett 2012, 23, 2899–
[
17]
2902.
J. Zhu, B.-H. Luo, T. Xiao, C. Zhang, N. Nishida, T. A. Springer, Mol.
Cell 2008, 32, 849–861.
P. A. Greenidge, C. Kramer, J.-C. Mozziconacci, R. M. Wolf, J. Chem.
Inf. Model. 2013, 53, 201–209.
M. Mori, F. Manetti, M. Botta, J. Chem. Inf. Model. 2011, 51, 446–454.
R. Wang, L. Lai, S. Wang, J. Comput. Aided Mol. Des. 2002, 16, 11–
26.
W. Cai, Y. Wu, K. Chen, Q. Cao, D. A. Tice, X. Chen, Cancer Res.
2006, 66, 9673–9681.
S. Dall’Angelo, Q. Zhang, I. N. Fleming, M. Piras, L. F. Schweiger, D.
O’Hagan, M. Zanda, Org. Biomol. Chem. 2013, 11, 4551–4558.
M. R. Battle, J. L. Goggi, L. Allen, J. Barnett, M. S. Morrison, J. Nucl.
Med. 2011, 52, 424–430.
370.
[
[
[
[
18]
19]
20]
21]
H. Chen, G. Niu, H. Wu, X. Chen, Theranostics 2016, 6, 78–92.
N. Withofs, R. Hustinx, Médecine Nucl. 2016, 40, 41–54.
J. Shi, F. Wang, S. Liu, Biophys. Rep. 2016, 2, 1–20.
S. Neubauer, F. Rechenmacher, A. J. Beer, F. Curnis, K. Pohle, C.
D’Alessandria, H.-J. Wester, U. Reuning, A. Corti, M. Schwaiger, et
al., Angew. Chem. Int. Ed. 2013, 52, 11656–11659.
F. Bianchini, P. Fabbrizzi, G. Menchi, S. Raspanti, A. Bottoncetti, A.
Passeri, E. Andreucci, A. Guarna, L. Calorini, A. Pupi, et al., Bioorg.
Med. Chem. 2015, 23, 1112–1122.
A. Testa, M. Piras, M. Hickey, I. Fleming, N. Bushby, E. Lenz, C.
Elmore, M. Zanda, Synlett 2014, 25, 1019–1025.
R. Yan, E. El-Emir, V. Rajkumar, M. Robson, A. P. Jathoul, R. B.
Pedley, E. Årstad, Angew. Chem. Int. Ed. 2011, 50, 6793–6795.
J. Becaud, L. Mu, M. Karramkam, P. A. Schubiger, S. M. Ametamey,
K. Graham, T. Stellfeld, L. Lehmann, S. Borkowski, D. Berndorff, et
al., Bioconjug. Chem. 2009, 20, 2254–2261.
[30]
[31]
[32]
[33]
[34]
[
22]
[
[
[
23]
24]
25]
This article is protected by copyright. All rights reserved.

ChemMedChem
10.1002/cmdc.201700328
FULL PAPER
Entry for the Table of Contents
FULL PAPER
Small-molecule ‘click’ RGD-ligands
displayed excellent in vitro properties
and offer a variety of radiolabeling
M. Piras,* A. Testa, I. N. Fleming, S.
Dall’Angelo, A. Andriu, S. Menta, M.
Mori, G. D. Brown, D. Forster, K. J.
Williams, and M. Zanda*
1
8
options (including [ F]). Binding mode
to α receptors was studied
v 3 3
β or αIIbβ
by molecular modeling.
Page No. – Page No.
Radiofluorination and PET/CT studies
on 12 in tumor models showed that
these mimics are excellent
High Affinity “Click” RGD
Peptidomimetics as Radiolabeled
Probes for Imaging α β Integrin
v 3
radiolabeled probes for in vitro and
v 3
cell-based studies on α β integrin, but
further pharmaco-kinetic and dynamic
profiling is required for in vivo imaging.
This article is protected by copyright. All rights reserved.