Polythiazole linkers as functional rigid connectors: A new RGD cyclopeptide with enhanced integrin selectivity

Article abstract of DOI:10.1039/c4sc00572d

Polythiazole amino acids clasp linear peptides to generate cyclic derivatives, however, the resulting species are not merely stapled peptides but bear a complex heterocyclic moiety displaying its intrinsic set of interactions. As a proof of concept, a bisthiazole moiety has been grafted onto an RGD sequence to deliver a new cilengitide analogue with improved integrin selectivity and remarkable in vivo antiangiogenic activity.

Full text of DOI:10.1039/c4sc00572d

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Polythiazole linkers as functional rigid connectors:
a new RGD cyclopeptide with enhanced integrin
selectivity†
Cite this: Chem. Sci., 2014, 5, 3929
J. Ruiz-Rodrıguez,a M. Miguel,b S. Preciado,b G. A. Acosta,af J. Adan,c A. Bidon-
´
d
d
c
be
afgh
*
*
*
*
Chanal, F. J. Luque, F. Mitjans, R. Lavilla
and F. Albericio
Polythiazole amino acids clasp linear peptides to generate cyclic derivatives, however, the resulting species
are not merely stapled peptides but bear a complex heterocyclic moiety displaying its intrinsic set of
interactions. As a proof of concept, a bisthiazole moiety has been grafted onto an RGD sequence to
Received 22nd February 2014
Accepted 22nd May 2014
DOI: 10.1039/c4sc00572d
deliver
a new cilengitide analogue with improved integrin selectivity and remarkable in vivo
www.rsc.org/chemicalscience
antiangiogenic activity.
by cyclization3 or by the introduction of non-natural amino acid-
like moieties, such as N-alkyl-,4 rigidied or chemically modi-
ed amino acids5 and b-amino acids.6 Cyclization was rst
implemented via a homodetic linkage through a “head-to-
tail”—as illustrated by cilengitide—7 or “side chain-to-side
chain” (Lys-Asp/Glu) topology. Later, Gilon and coworkers
introduced the backbone-to-backbone cyclization concept to
preserve the identity of the side chains.8 Although disulde
bridges frequently appear in many biologically active peptides
and efficiently contribute to their stabilization, their redox
properties jeopardize their introduction in the de novo design of
cyclic peptides. In this regard, several authors have proposed
the replacement of the S–S bonds with dicarba analogs through
metal-catalyzed ring-closing olen metathesis.9 This approach
has led to the “stapled” peptide concept introduced by Ver-
dine,10 where a exible hydrocarbon motif is introduced. A
further advance has been the development of cyclic peptides
through Sharpless click chemistry,11 which incorporate a tri-
azole ring on the cyclization branch.
Additionally, heterocycles are a recurring motif in the
medicinal chemistry toolbox,12 and their presence can overcome
some of the problems associated with the use of peptides as
drugs. In this context, the “heterocycle-connected” cyclic
peptide concept is described herein. In a broad sense, these
rigid connectors can be regarded as “functional staples”:
modular and synthetically amenable structural units based on
the suitable intercalation of (poly)thiazole moieties (or other
groups) by amide formation, which enables tting of the
appropriate length, introduces rigidity, and improves drug
likeness for the development of a new class of cyclic peptide.
Furthermore, they retain almost all of the features of classical
staples (conformational restriction and metabolic stability),
leading to new structural features that improve the peptide–
target interaction. On the other hand, this approach requires
more specic synthetic work (short sequences to prepare the
Introduction
In recent years, the pharmaceutical industry has undergone a
profound transformation. In the period 2009–2012, the US FDA
approved the release of only 116 new therapeutics; of these, 91
were new chemical entities and 25 were new biologics.1 An
analysis of the chemical entities by chemical species reveals that
the so-called “small molecules” have been losing ground over
the past few decades, since they represent only 48% of total FDA
approvals, as compared with the aforementioned biologics
(21%). Interestingly, peptides account for 14 new therapeutic
agents (12%).2 This change in scenario reects the general
acceptance that peptides may be used as drugs. The main
drawbacks of peptides as drugs are their relatively high exi-
bility, lack of stability, and poor pharmacokinetic properties
when compared with small molecules. Over the years, peptide
chemists have addressed these issues in two main ways, namely
aInstitute for Research in Biomedicine, Baldiri Reixac 10-12, 0802 Barcelona, Spain.
E-mail: albericio@irbbarcelona.org
bBarcelona Science Park, Baldiri Reixac 10-12, 08028 Barcelona, Spain. E-mail:
rlavilla@pcb.ub.es
cLeitat Technological Center, Barcelona Science Park, Baldiri Reixac, 15, 08028, Spain.
E-mail: fmitjans@leitat.org
dDepartment of Physical Chemistry and Institute of Biomedicine, Faculty of Pharmacy,
University of Barcelona, Avda. Prat de la Riva 171, 08921 Santa Coloma de Gramenet,
Spain. E-mail: luque@ub.edu
eLaboratory of Organic Chemistry, Faculty of Pharmacy, University of Barcelona, Avda.
Joan XXII s.n., 08028 Barcelona, Spain
fCIBER-BBN, Networking Centre on Bioengineering Biomaterials and Nanomedicine,
Barcelona Science Park, Baldiri Reixac 10, 08028 Barcelona, Spain
gDepartment of Organic Chemistry, Faculty of Chemistry, University of Barcelona,
´
´
Martı i Franques 1-11, 08028 Barcelona, Spain
hSchool of Chemistry & Physics, University of KwaZulu-Natal, 4001 Durban, South
Africa
† Electronic supplementary information (ESI) available. See DOI:
10.1039/c4sc00572d
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3929–3935 | 3929

View Article Online
Chemical Science
Edge Article
thioazole amino acids) and, although not as direct as the initial and transmembrane signaling. In this context, integrins
stapling techniques, is quite straightforward.
comprise a family of transmembrane heterodimeric receptors
and have emerged as a promising target because of their
prominent role in many biological functions.14 A subset of
integrins specically recognizes the tripeptide sequence Arg-
Gly-Asp (RGD) in their ligands. In particular, the avb3, avb5 and
a5b1 subfamilies participate in tumor metastasis and tumor-
induced angiogenesis. Furthermore, endothelial cell survival
and proliferation are strongly dependent on the adhesive
activity of avb3 and avb5. Thus, several integrin antagonists have
been designed for therapeutic purposes, including those with
an RGD scaffold as well as others that are structurally unre-
lated.4,15 Aer the pioneering studies of Kessler and coworkers,
which culminated with the development of cilengitide,7 the
rational design of antagonists has pursued the control of the
three-dimensional features of the RGD sequence in order to
improve selectivities and pharmacokinetic proles and to allow
for the development of strategies for drug delivery and bio-
imaging.16 There are a considerable number of distinct integ-
rins and their different cellular roles determine several events,
including among them the blockade of tumor progression in
some cancer cell lines. Furthermore, the oen complex binding
proles of most integrin antagonists severely restricts the
therapeutic and bioimaging applications of such compounds.
Therefore, although potent antagonists have been developed,
more selective agents are still needed, both as probes to study
the fundamental biology of integrins and as theragnostic
agents. Given the relevant presence of thiazole nuclei in several
bioactive natural products,17 we included these structures in an
RGD cyclopeptide in order to determine their antiangiogenic
activity. In particular, we synthesized and characterized
compounds 1–3 (Fig. 1), which bear 1, 2 or 3 thiazole units in
the tether, and were likely to t the specicity determining loop
(SDL)18 and display a range of affinities for the integrin
receptors.19
Results and discussion
Natural biosynthetic thiazoles (2,4-disubstituted) can be
considered derivatives of Cys-containing peptides. Thiazole
peptides are formed through a cyclodehydration step followed
by an oxidation step, thus rendering their synthesis extremely
challenging. A synthetic alternative involves the use of pre-
synthesized (poly)thiazole amino acids in a standard solid
phase peptide synthesis. The thiazole precursors, featuring a
2,4-disubstitution pattern, can be prepared through a Hantzsch
thiazole synthesis by the condensation reaction between thio-
amides and a-haloketones. (Scheme 1). Incidentally, oxazoles
are structurally isosteric to thiazoles, and they would be very
likely to behave in a similar way. However, although they are
biosynthesized through related cyclodehydration mechanisms,
they require harsher reaction conditions for chemical synthesis
and were not considered in this work.
The corresponding synthetic (poly)thiazole-containing
amino acids can be coupled to the N- and C-terminal sites of an
active peptide motif to prepare a head-to-tail cyclic peptide or to
side chains of Lys and Asp/Glu to give side chain-to-side chain
derivatives (or a combination of both). As a proof of concept,
analogs of the integrin antagonist cilengitide (cyclo[R-G-D-f-
(NMe)V]) of different sizes containing thiazole motifs have been
synthesized. Cilengitide is actually in advanced clinical phases
for use against glioblastoma and other cancers as an anti-
angiogenic agent, either alone or in combination with radio/
chemotherapy. Although so far no signicant efficacy has been
demonstrated in the concluded trials, preliminary data suggest
that this compound could become a useful antitumoral drug.7
Angiogenesis involves a complex sequence of biological events
that lead to the formation of blood vessels. Tumor growth
occurs by cell division and is limited by the availability of
nutrients. To go beyond this point, tumors promote angiogen-
esis in order to irrigate the new cell mass and, in this way,
maintain expansion. Therefore, antiangiogenesis strategies are
considered to be highly attractive ways to treat cancer.13
To this end, we envisioned the synthesis of the parent thia-
zole amino acids, which were prepared through a unied
approach involving an iterative protocol based on the Hantzsch
thiazole reaction (Scheme 2).20 In this way, treatment of
N-benzyloxycarbonylglycine with isobutyl chloroformate (IBCF),
in the presence of N-methylmorpholine (NMM) as a base, fol-
lowed by the reaction with aqueous NH3 afforded amide 4. This
compound was transformed into the corresponding thioamide
5 by interaction with the Lawesson reagent (L.R.), and its
The interaction between the extracellular matrix and
membrane proteins controls cell adhesion, tissue formation
Scheme 1 Biosynthetic and chemical access to thiazole peptides.
3930 | Chem. Sci., 2014, 5, 3929–3935
Fig. 1 Cilengitide, RGD peptides and thiazole analogs 1–3.
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
Scheme 3 Synthesis of the thiazole-based peptides 1–3.
Scheme 2 Synthesis of the thiazole-based amino acids 9–11.
reported in previous studies. On the other hand, these tests
(Fig. 2) provide a robust demonstration of the real activity of
these compounds under physiological conditions, which is
more reliable for prospective therapeutic applications. The
biological assays were performed with the following cell lines:
DAOY (human medulloblastoma cells, expressing both avb3 and
avb5 members of the av-integrin family), HT29 (human colon
adeno-carcinoma, expressing both avb5 and avb6 integrins), and
HUVEC (human umbilical vein endothelial cells as the primary
cell culture, expressing both avb3 and avb5 integrins). As integ-
rin ligands, in these experiments we used vitronectin (VN),
brinogen (FB), bronectin (FN) and collagen (COL). The
experimental protocol was set up via ligand coating followed by
analysis of the compound’s effect on cell adhesion to the ligand.
Briey, matrix proteins were coated onto 96-well plates, the
wells were blocked with bovine serum albumin (BSA) and seri-
ally diluted solutions of compounds were added followed by the
cells in the cell culture medium. Aer 1 hour at 37 ꢀC, the non-
reaction with ethyl bromopyruvate led to thiazole 6. Subsequent
hydrolysis of the ester group afforded carboxylic acid 7.
Deprotection of the N-Cbz group and introduction of the Fmoc
substituent yielded amino acid 9, which was suitable for
coupling by a subsequent solid-phase methodology using
Fmoc/tBu chemistry. The bis- and tris-thiazole amino acids 10
and 11 were prepared analogously (Scheme 2).
The linear precursors of the cyclopeptides were synthesized
by standard solid-phase methodology using 2-chlorotrityl resin
and Fmoc/tBu chemistry followed by a nal solution phase
cyclization. The protecting groups of the side chains of Asp and
Arg, tBu and Pbf, respectively, were stable to the triuoroacetic
acid (TFA) concentrations used for the peptide release from the
2-chlorotrityl chloride resin (Scheme 3). The cyclization reac-
tions between the amino group of Asp and the carbonyl group of
Gly were performed in solution using PyAOP as a coupling
agent, furnishing the protected cyclic peptides. The nal
deprotections to remove the tBu and Pbf groups were carried
out without triisopropylsilane (TIS) as a scavenger to prevent the
reduction of the thiazole rings. The corresponding cyclo-
peptides with thiazoline residues were detected in experiments
with TIS (M + 2 peaks were found in the HPLC-MS analysis),
thus suggesting an incompatibility between the thiazoles (or
related imine-containing structures) in acidic media and silicon
hydrides. Thus, the optimized conditions involved treatment
with TFA/H2O (95 : 5) (Scheme 3). In this way, a scalable
synthesis of thiazolopeptides 1,21 2 and 3 was ensured.
Although direct binding assays with puried integrins are
the most direct and specic method to determine the antago-
nists’ selectivities,4,7c,15e we have determined IC50 values from
cell-based competition experiments in order to facilitate the
direct comparison of the present data with the activities
Fig. 2 Inhibition of angiogenesis by peptide 2 in the in vivo experi-
ments by the Matrigel plug assay. Hemoglobin contents of control
vehicle, VEGF and VEGF + peptide 2 (2 mg mLÀ1, left graph) and plug
weight of the same samples (right graph). Mean and statistical signif-
icances are shown.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3929–3935 | 3931

View Article Online
Chemical Science
Edge Article
adherent cells were washed away and the amount of attached
To further conrm the activity of cyclopeptide 2, the capacity
cells remaining was measured using the hexominidase enzy- of this compound to inhibit angiogenesis was explored using
matic assay. The inhibition of attachment was calculated using in vivo assays in conjunction with both positive (VEGF) and
wells with no added compounds as a reference, and the results negative controls. The results obtained in the matrigel plug
are summarized in the table as IC50 values in mM concentrations assay22 revealed an antiangiogenic effect of 2 (p < 0.05) in vivo,
for compounds 1–3 and cilengitide (A; used as the reference showing a lower level of haemoglobin in the treated plug than in
control), tested in HUVEC, DAOY and HT-29 cells (Table 1).
the VEGF control (Fig. 2). Moreover, the average weight of the
In the experiments using VN as the integrin ligand, the plug was also signicantly lower than in the positive (VEGF) and
HUVEC and DAOY cells used avb3 and avb5 to bind to this negative controls. These data unequivocally show that this
protein, whereas HT-29 cells used only avb5. Complementarily, compound efficiently represses angiogenesis in living animals,
in the tests in which FB was the integrin ligand, HUVEC and thereby demonstrating its capacity as a potential therapeutic
DAOY cells used avb3 to bind to it. Experiments involving FN agent.
and COL were run to check that the new peptides did not block
To rationalize these ndings, compound 2 was docked in the
the cell adhesion mediated by other integrins (the b1 binding site of avb3 and avb5, and the structural integrity of the
subfamily). The results indicated that compounds 2 and 3 binding mode was further checked using molecular dynamics
showed satisfactory inhibitory activity in the HUVEC/DAOY cells (MD) simulations. Similar studies were performed for cilengi-
for VN, but they were less potent than cilengitide. However, tide (A), which was used as a reference system. To this end, the
these peptides displayed potencies similar to cilengitide with structure of 2 was determined by combining NMR-based
respect to blocking avb3-mediated cell adhesion with FB. This restraints with MD simulations of the cyclic compound, and a
observation suggests that 2 and 3 show different affinities to 3D model of avb5 was developed by homology modeling
avb3 and avb5. Furthermore, the results obtained for the HT-29– (see details in the ESI†).
VN system showed that these cyclopeptides do not exert
The docking of A to the binding site of avb3 reproduced the
signicant inhibitory activity against avb5, thereby indicating binding mode found in the X-ray structure of the complex
that, unlike cilengitide, they are extraordinarily specic for avb3. between A and avb3 (PDB ID: 1L5G), which was stable
Compound 2 displayed interesting potency and selectivity, even throughout the 50 ns MD simulation (the positional root mean
compared with cilengitide, against several cell lines (Table 1). square deviation (rmsd) of the modeled backbone of A versus
˚
These data reveal that the bisthiazole cyclopeptide has the the X-ray structure was 0.8 A). This can be primarily attributed
correct structure to accommodate an extended RGD sequence to the stabilizing contributions from the Asp and Arg residues
in the cle, similar to that of cilengitide.
in the RGD stretch: the former interacts with the metal-ion-
dependent adhesion site (MIDAS; in b3), and the latter forms a
Table 1 Activities (cell adhesion inhibition) of compounds 1–3 in the cell systems, expressed as IC50 in mM
VN or FB coating
A
[mg mLÀ1
]
Cell lines (integrin prole)
1
2
3
(cilengitide)
VN [0.75]
VN [0.5]
VN [4.0]
FB [4.0]
FB [5.0]
HUVEC (avb3/avb5)
DAOY (avb3/avb5)
HT-29 (avb5)
HUVEC (avb3)
DAOY (avb3)
31.13
ND
ND
368.4
ND
7.46
5.71
53.16
4.95
63.80
6.80
59.05
33.15
1.86
0.56
0.068
2.24
4.03
0.059
0.082
3932 | Chem. Sci., 2014, 5, 3929–3935
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
salt bridge with D248 and a cation–p interaction with Tyr208 Phe, mainly with Tyr148 in b3 and Leu148 in b5, and the
(both in av), respectively (Fig. 3A). The net effect is to keep cil- hydrogen bond between the NH group of Asp with the carbonyl
engitide tightly bound in the binding pocket. Importantly, the group of Arg242 in b3 and Arg243 in b5.
results obtained for the binding to avb5 showed that A assumed
Compound 2 bound to avb3 and avb5 retained the interac-
a binding mode that matched the binding to avb3 (Fig. 3B). tions with both the metal cation at MIDAS and Asp248 (av).
Analysis of the pairwise interactions with the residues in the Nevertheless, while 2 superposes well with the structure of cil-
binding site also points out that the binding of A to both avb3 engitide upon binding to avb3, a signicant relocation of the
and avb5 is assisted by the van der Waals contacts formed by cyclopeptide backbone is observed upon binding to avb5
Fig. 3 (A) Representation of cilengitide bound to avb3 in the X-ray structure (PDB entry: 1L5G; magenta sticks) and in the last snapshot of the MD
trajectory (yellow sticks). The av and b3 units are shown as cyan and green surfaces, and selected residues are shown as sticks. (B) Representation
of cilengitide in the last snapshots of the MD trajectories run for avb3 (yellow sticks) and avb5 (orange sticks). Selected residues in avb5 are shown
as orange sticks (labelled in italics). (C) Superposition of cilengitide (yellow sticks) and compound 2 (blue sticks) bound to the last snapshots of the
MD trajectories run for avb3. (D) Superposition of cilengitide (yellow sticks) and compound 2 (blue sticks) bound to the last snapshots of the MD
trajectories run for avb3 and avb5, respectively. In all cases the metal cations are shown as magenta spheres. Residues are numbered according to
the sequences in Uniprot ID P06756 (av), P05106 (b3) and P18084 (b5).
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3929–3935 | 3933

View Article Online
Chemical Science
Edge Article
Table 2 Energetic contributions (kcal molÀ1) to the interaction of
cilengitide and compound 2 and their predicted binding affinities
the design of and synthetic access to new analogues with
enhanced integrin selectivity, a key issue for the development of
theragnostic antiangiogenic agents.23 We envisage that the
“functional rigid connector” concept, exemplied here with the
bisthiazole motif, will also have a broad range of applications
for the synthesis of more rigid and drug like cyclic peptides with
improved stability and selectivity.
Integrin
van der Waals
Electrostatic
Binding affinity
A
2
avb3
avb5
avb3
avb5
À29.1
À31.4
À24.5
À24.2
À29.3
À24.0
À26.4
À22.2
À10.8
À11.5
À8.1
À6.1
Acknowledgements
(Fig. 3C and D). This structural rearrangement can be attributed
to the net effect arising from the electrostatic inuence of
Asp279, which points toward the NH groups proximal to the
thiazole ring and the Asp residue, and the lack of stabilizing van
der Waals contacts with Leu148, leading in turn to the loss of
the cation–p interaction formed by the Arg residue with Tyr208,
and of the hydrogen bond formed by the Asp NH unit with
Arg243.
The authors acknowledge nancial support from DGICYT
(CTQ2009-07758, SAF2011-27642, CTQ2012-30930), the Gen-
eralitat de Catalunya (2009SGR249) and Icrea Academia (FJL).
Dr Carles Mas-Moruno (Polytechnic University of Catalonia) is
thanked for helpful discussions. The Barcelona Super-
computing Center is acknowledged for providing computa-
tional facilities.
The similar structural arrangement of A in avb3 and avb5 was
reected in the similar van der Waals component of the ligand–
receptor interaction (À29.1 vs. À31.4, respectively; Table 2),
whereas the electrostatic component was less favourable in
avb5, which reects the electrostatic repulsion between the Asp
residue in A and Asp279 (replaced by Ala278 in b3). Overall, MM/
PBSA calculations predict similar affinities for the binding of
cilengitide to both avb3 and avb5 (Table 2), which is in agree-
ment with the similar experimental activities obtained from the
DAOY assays (Table 1). As expected from the loss of contact
formed by the Phe residue in cilengitide (replaced by a thiazole
ring in 2), the van der Waals interaction energy was less favor-
able for the binding of 2. However, the electrostatic term
favored the interaction with avb3 by ꢁ4 kcal molÀ1. As a result,
binding of 2 to avb3 was predicted to be more favorable than to
avb5, which agrees qualitatively with the ꢁ70-fold increase in
selectivity determined experimentally for compound 2 (Table 1).
Notes and references
1 (a) B. Hughes, Nat. Rev. Drug Discovery, 2010, 9, 89; (b)
A. Mullard, Nat. Rev. Drug Discovery, 2011, 10, 82; (c)
A. Mullard, Nat. Rev. Drug Discovery, 2012, 11, 91; (d)
A. Mullard, Nat. Rev. Drug Discovery, 2013, 11, 87.
2 F. Albericio and H. G. Kruger, Future Med. Chem., 2012, 4,
1527.
3 (a) S. A. Kates, N. A. Sole, F. Albericio and G. Barany, in
Peptides: Design, Synthesis and Biological Activity, ed. C.
Basava and G. M. Anantharamaiah, Birkhauser, Boston,
MA, 1994; (b) S. Royo-Gracia, K. Gaus and N. Sewald,
Future Med. Chem., 2009, 1, 1289.
4 C. Mas-Moruno, J. G. Beck, L. Doedens, A. O. Frank,
L. Marinelli, S. Cosconati, E. Novellino and H. Kessler,
Angew. Chem., Int. Ed., 2011, 50, 9496 and references cited
therein.
¨
5 For precedents, see inter alia: (a) J. Spiegel, C. Mas-Moruno,
H. Kessler and W. D. Lubell, J. Org. Chem., 2012, 77, 5271; (b)
M. Marchini, M. Mingozzi, R. Colombo, I. Guzzetti,
L. Belvisi, F. Vasile, D. Potenza, U. Piarulli, D. Arosio and
C. Gennari, Chem.–Eur. J., 2012, 18, 6195; (c) A. S. M. da
Ressurreiçao, A. Vidu, M. Civera, L. Belvisi, D. Potenza,
L. Manzoni, S. Ongeri, C. Gennari and U. Piarulli, Chem.–
Eur. J., 2009, 15, 12184; (d) A. O. Frank, E. Otto, C. Mas-
Moruno, H. B. Schiller, L. Marinelli, S. Cosconati,
A. Bochen, D. Vossmeyer, G. Zahn, R. Stragies, E. Novellino
and H. Kessler, Angew. Chem., Int. Ed., 2010, 49, 9278.
6 For precedent, see: S. Urman, K. Gaus, Y. Yang, U. Strijowski,
N. Sewald, S. De Pol and O. Reiser, Angew. Chem., Int. Ed.,
2007, 46, 3976.
Conclusions
In summary, a modular strategy for the preparation of a new
(poly)thiazole moiety containing cyclic peptides has been
described. As the key derivative is an amino acid that is
prepared by an iterative protocol based on the Hantzsch thia-
zole reaction, analogues can be prepared, and therefore size,
length and structural issues can be suitably modulated. This
approach has allowed the preparation of cyclopeptide 2 with a
similar potency and considerably increased integrin selectivity
compared to the parent cilengitide. From a molecular point of
view, it is likely that cyclopeptide 1 was overstrained or distorted
and thus failed to properly display the recognition sequence for
the RGD receptor cle.
On the other hand, compounds 2 and 3 showed a satisfactory
level of receptor occupancy, the former probably being better
suited for binding with the SDL in integrin avb3 than with the
avb5 protein (discriminating more efficiently than cilengitide).
This fact is particularly relevant, since avb3 integrin is highly
expressed in several tumor tissues, whereas avb5 is commonly
found in healthy cells. These results open up new avenues for
7 (a) C. Mas-Moruno, F. Rechenmacher and H. Kessler, Anti-
Cancer Agents Med. Chem., 2010, 10, 753; (b) C. Mas-
Moruno and H. Kessler, Methods Princ. Med. Chem., 2013,
56, 257; (c) C. Scaringi, G. Minniti, P. Caporello and
R. M. Enrici, Anticancer Res., 2012, 32, 4213; (d) For an
updated survey of the ongoing clinical trials, search this
drug in https://clinicaltrials.gov/.
3934 | Chem. Sci., 2014, 5, 3929–3935
This journal is © The Royal Society of Chemistry 2014

View Article Online
Edge Article
Chemical Science
8 Z. Hayouka, A. Levin, M. Hurevich, D. E. Shalev, A. Loyter, 16 For representative examples, see: (a) P.-P. Lv, Y.-F. Ma, R. Yu,
G. Gilon and A. Friedler, Bioorg. Med. Chem., 2012, 20, 3317
and references cited therein.
H. Yue, D.-Z. Ni, W. Wei and G.-H. Ma, Mol. Pharmaceutics,
2012, 9, 1736; (b) F. Bianchini, N. Cini, A. Trabocchi,
A. Bottoncetti, S. Raspanti, E. Vanzi, G. Menchi, A. Guarna,
A. Pupi and L. Calorini, J. Med. Chem., 2012, 55, 5024; (c)
K.-T. Yong, I. Roy, W.-C. Lawb and R. Hub, Chem.
Commun., 2010, 46, 7136; (d) J. Simecek, J. Notni,
T. G. Kapp, H. Kessler and H.-J. Wester, Mol. Pharm., 2014,
11, 1687; (e) S. Neubauer, F. Rechenmacher, A. J. Beer,
F. Curnis, K. Pohle, C. D'Alessandria, H.-J. Wester,
U. Reuning, A. Corti, M. Schwaiger and H. Kessler, Angew.
Chem., Int. Ed., 2013, 52, 11656.
9 As relevant examples, see: (a) V. R. Pattabiraman,
J. L. Stymiest, D. J. Derksen, N. I. Martin and J. C. Vederas,
Org. Lett., 2007, 9, 699; (b) R. M. J. Liskamp, M. J. Rob,
D. T. S. Rijkers, J. A. W. Kruijtezr and J. Kemmink,
ChemBioChem, 2011, 12, 1626; (c) D. Wang, M. Lu and
P. S. Arora, Angew. Chem., Int. Ed., 2008, 47, 1879.
10 (a) S. Y. Shim, Y.-W. Kim and G. L. Verdine, Chem. Biol. Drug
Des., 2013, 82, 635; (b) G. L. Verdine and G. J. Hilinski,
Methods Enzymol., 2012, 503, 3 and references cited therein.
11 As relevant examples, see: (a) S. Cantel, I. A. Le Chevalier, 17 For instance, see: (a) R. A. Hughes and C. J. Moody, Angew.
M. Scrima, J. J. Levy, R. D. DiMarchi, P. Rovero,
J. A. Halperin, A. M. D'Ursi, A. M. Papini and M. Chorev, J.
Org. Chem., 2008, 73, 5663; (b) C. W. Tornoee and
Chem., Int. Ed., 2007, 46, 7930; (b) Z. Jin, Nat. Prod. Rep.,
2009, 26, 382; (c) Also see: B. J. MacLean and P. G. Pickup,
J. Mater. Chem., 2001, 11, 1357.
M. Meldal, Organic Azides, ed. S. Braese and K. J. Banert, 18 L. Marinelli, K.-E. Gottschalk, A. Meyer, E. Novellino and
Wiley & Sons Ltd., Chichester, UK, 2010, pp. 285–310; (c) H. Kessler, J. Med. Chem., 2004, 47, 4166.
J. Zhang, J. Kemmink, D. T. S. Rijkers and 19 A couple of RGD peptides including 2-amino-1,3-thiazole-4
R. M. J. Liskamp, Org. Lett., 2011, 13, 3438; (d) B. Thomas,
M. Fiore, I. Bossu, P. Dumy and O. Renaudet, Beilstein J.
Org. Chem., 2012, 8, 421–427.
carboxylic acid and 4-amino-1,3-thiazole-2-carboxylic acid
were prepared but showed low antiangiogenic potency (see
ESI†).
12 C. T. Walsh and E. M. Nolan, Proc. Natl. Acad. Sci. U. S. A., 20 For related amino acids, see: (a) J. D. Butler, K. C. Coffman,
2008, 105, 5655.
K. T. Ziebart, M. D. Toney and M. J. Kurth, Chem.–Eur. J.,
2010, 16, 9002–9005; (b) Y. Singh, N. Sokolenko,
M. J. Kelso, L. R. Gahan, G. Abbenante and D. P. Fairlie, J.
Am. Chem. Soc., 2001, 123, 333; (c) I. G. Stankova,
G. I. Videnov, E. V. Golovinsky and G. Jung, J. Pept. Sci.,
1999, 5, 392.
13 (a) R. O. Hynes, Nat. Med., 2002, 8, 918; (b) J. S. Desgrosellier
and D. A. Cheresh, Nature Rev. Cancer, 2010, 10, 9 and Nature
Rev. Cancer, 2010, 10, 890 (corrigendum).
14 (a) M. Hoeing, H. Kessler and K.-E. Gottschalk, Angew.
Chem., Int. Ed., 2009, 48, 6590; (b) M. Schottelius,
B. Laufer, H. Kessler and H.-J. Wester, Acc. Chem. Res., 21 This compound has been preliminarily disclosed: Finsinger
2009, 42, 969; (c) T. Weide, A. Modlinger and H. Kessler,
Top. Curr. Chem., 2007, 272, 1 and references cited therein.
15 (a) L. Auzzas, F. Zanardi, L. Battistini, P. Burreddu, P. Carta,
G. Rassu, C. Curti and G. Casiraghi, Curr. Med. Chem., 2010,
17, 1255; (b) A. Bochen, U. K. Marelli, E. Otto, D. Pallarola,
D (1997) Peptidische und peptidanaloge Antagonisten des
¨
Universitat
avb3-integrins.
Dissertation,
Technische
¨
Munchen. Also, a related monothiazole RGD-derivative has
been recently reported: A. Nefzi and J. E. Fenwick,
Tetrahedron Lett., 2011, 52, 817.
C. Mas-Moruno, F. Saverio di Leva, H. Boehm, J. P. Spatz, 22 J. O. Nam, J. E. Kim, H. W. Jeong, S. J. Lee, B. H. Lee,
E. Novellino, H. Kessler and L. Marinelli, J. Med. Chem.,
2013, 56, 1509; (c) M. Aufort, M. Gonera, J. Le Gal,
J. Y. Choi, R. W. Park, J. Y. Park and I. S. Kim, J. Biol.
Chem., 2003, 278, 25902.
B. Czarny, L. Le Clainche, R. Thai and C. Dugave, 23 Part of these results are covered by a patent: F. Mitjans,
ChemBioChem, 2011, 12, 583; (d) S. Neubauer,
F. Rechenmacher, R. Brimioulle, F. Saverio Di Leva,
A. Bochen, T. R. Sobahi, M. Schottelius, E. Novellino,
C. Mas-Moruno, L. Marinelli and H. Kessler, J. Med. Chem.,
2014, 57, 3410.
J. Adan, C. Calvis, F. Albericio, R. Lavilla and J. Ruiz
Rodriguez, PCT Int. Appl., WO 2012062777 A1 20120518,
2012.
This journal is © The Royal Society of Chemistry 2014
Chem. Sci., 2014, 5, 3929–3935 | 3935