The Design of Potent, Selective and Drug-Like RGD αvβ1 Small-Molecule Inhibitors Derived from non-RGD α4β1 Antagonists

Article abstract of DOI:10.1002/cmdc.201900359

Up to 45 % of deaths in developed nations can be attributed to chronic fibroproliferative diseases, highlighting the need for effective therapies. The RGD (Arg-Gly-Asp) integrin αvβ1 was recently investigated for its role in fibrotic disease, and thus war

Full text of DOI:10.1002/cmdc.201900359

Accepted Article
Title: The Design of Potent, Selective and Drug-Like RGD αvβ1 Small
Molecule Inhibitors Derived from non-RGD α4β1 Antagonists
Authors: Richard Hatley
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.201900359
Link to VoR: http://dx.doi.org/10.1002/cmdc.201900359
A Journal of
www.chemmedchem.org

10.1002/cmdc.201900359
ChemMedChem
COMMUNICATION
The Design of Potent, Selective and Drug-Like RGD v1 Small Molecule Inhibitors
Derived from non-RGD 41 Antagonists
Richard J. D. Hatley*[a], Tim N. Barrett[a], Robert J. Slack[a], Morag E. Watson[a], Daniel J. Baillache[a],
Anna Gruszka[a], Yoshiaki Washio[a], James E. Rowedder[a], Peter Pogány[a], Sandeep Pal[a], Simon J. F.
Macdonald[a]
[a]
Dr. R. J. D. Hatley, Dr. T. N. Barrett, Dr. R. J. Slack, Dr. M. E. Watson, D. J. Baillache, A. Gruszka, Y. Washio, J. E. Rowedder, Dr. P. Pogány, Dr. S. Pal,
Dr. S. J. F. Macdonald
GlaxoSmithKline (GSK), Medicines Research Centre, Gunnels Wood Road, Stevenage, SG1 2NY (UK).
E-mail richard.j.hatley@gsk.com.
Supporting information for this article is given via a link at the end of the document.
Abstract: Up to 45% of deaths in developed nations can be attributed to
chronic fibroproliferative diseases, resulting in a requirement for effective
therapies. The RGD (Arg-Gly-Asp) integrin v1 has recently been
investigated for its role in fibrosis disease and warrants therapeutic
targeting. Herein, we describe (i) the identification of non-RGD hit small
molecule v1 inhibitors; (ii) that v1 activity is embedded in a range of
literature 41 (VLA-4) ligands; (iii) how, for the first time, a non-RGD
integrin inhibitor (of 41 in this case) was converted into a potent non-
zwitterionic RGD integrin inhibitor (of v1 in this case); (iv) the design of
urea ligands with excellent selectivity over 41 and the other v integrins
(v3, v5, v6, v8); (v) how in silico docking models and density
functional theory (DFT) calculations aided the discovery of the lead urea
series.
fibrosis in lung, liver and kidney in vivo models, leading the
authors (Reed et al) to state that “compounds based on this
lead compound could be broadly useful for treatment of
diseases characterized by excessive tissue fibrosis“.[7]
1
Figure 1. Reported v1 selective ligand
However, compound 1 has poor PK properties which
makes in vivo investigation difficult. In addition, while 1
delivers exquisite v selectivity, it has recently been shown
to have additional 1 activities, which may complicate
further TV studies.[8] Historically, identifying low dose, orally
bioavailable and selective molecules suitable for clinical use
has been highly challenging.[5c,9] Some of the reasons behind
this are the polar zwitterionic pharmacophore which usually
limits permeability and that many of the integrins have
similar binding domains. Approaches have been investigated
to improve oral PK properties by attenuating the high
polarity in the design of non-zwitterionic − non-RGD
mimetics. However, very few genuine non-RGD integrin
inhibitors exist, most are either pan-assay interference
compounds (PAINS), not easily optimised, or non-drug like,
thus make unsuitable leads for progression to clinical
candidates. [10, 11]
Fibrosis is characterized by the build-up of collagen rich
components of the extra cellular matrix (ECM) and the
layering of excessive scar tissue in response to chronic injury.
Eventually, this leads to organ dysfunction and failure, as
depicted by fibrotic diseases such as idiopathic pulmonary
fibrosis (IPF), liver cirrhosis and kidney fibrosis − all with poor
patient survival rates. Indeed, it is estimated that nearly 45%
of all deaths in the developed world are the result of various
fibrotic disorders, which highlights the requirement for
further treatments to meet this growing burden.[1] A large
number of biological mechanisms have been targeted for the
treatment of fibrotic conditions[2] but the v-integrins (RGD
family) remain an attractive proposition because of their role
in activation of TGF, a key cytokine regulator of fibrosis.[3]
The 24 known integrins, classified into various sub-
families, function as -heterodimeric signalling proteins
with structures spanning cellular membranes.[4] In addition
to fibrotic diseases, they have already been widely targeted
for the treatment of cancers, thrombosis and osteoporosis −
investigated with synthetic peptides, antibodies and small
molecules.[5] Integrins bind endogenous ligands containing
short peptidic recognition sequences, as observed in the
therapeutically important family, RGD (Arg-Gly-Asp). It
consists of v1 and homologous members v3, v5,
v6, v8, along with three other RGD integrins (81,
IIb3 and 51), not discussed here.
During this work to identify new ligands of the v1
integrin, we describe how we have started to address some
of frequently encountered drug design problems, especially
the selectivity and the zwitterionic nature. Specifically, we
detail; (i) the identification of non-RGD hit small molecule
v1 inhibitors; (ii) that v1 activity is embedded in a range
of literature 41 (VLA-4) ligands; (iii) how, for the first time,
a non-RGD integrin inhibitor (of 41 in this case) was
converted into a potent non-zwitterionic RGD integrin
inhibitor (of v1 in this case); (iv) the design of urea ligands
with excellent selectivity over 41 and the other v
integrins (v3, v5, v6, v8); (v) how in silico docking
models and density functional theory (DFT) calculations
aided the discovery of the lead urea series. With many
inhibitors of 41 known,[12] we show that this innovative
approach, of converting a non-RGD ligand to an RGD ligand,
could be of particular value in providing a rich source of new
chemotypes for inhibiting v1, but may also catalyse the
While extensive target validation for v-integrins in
fibrosis exists for v6,[6] some attention has recently
focussed on v1, after reports of compelling pre-clinical
target validation in fibrosis disease with the design and use
of the first v1 selective RGD small molecule mimetic 1
(Figure 1). It has been shown that 1 impressively reduced
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900359
ChemMedChem
COMMUNICATION
discovery of new ligands for the other closely related v- found that compound 1 also unexpectedly contained high
integrins.
1 activity, pKd 9.11.[8] These data collectively show that
There is a paucity of literature v1 SAR data and v1 activity is likely to be present in a number of distinct
structural information to guide the rational design of new 1 chemotypes and should be taken into account for
inhibitors. We therefore took the approach to identify future target validation studies. In light of uncovering this
potential hits from the GSK data bank of diverse compounds, dual activity, optimization of any v1 inhibitors would
already with a pedigree for inhibiting various integrins. benefit by having selectivity over 1 as well as the other
Compounds were selected and tested at
a fixed v integrins.
concentration (1 M) in a fluorescence polarization assay
against three v-integrins; v1, v5 and v6, allowing
rapid identification of potential inhibitors with simultaneous
generation of v-selectivity data. Assessment of the output
revealed unanticipated but striking results where one set of
structurally related molecules, containing the biphenyl -
amino acid core template, appeared to robustly show
exquisite selectivity for v1 over both v5 and v6, as
exemplified by 2 and 3 (Table 1). The high potency and
selectivity of the hit compounds was confirmed after wider
screening in all five vx cell adhesion assays (x=1,3,5,6,8)[13]
and in v1/1 radioligand binding assays (See Table 1
and SI for methodology). First indications of integrin cross
reactivity came with RGD peptides some time ago but has
not seen further exploration.[14]
Table 2. Selected literature molecules demonstrating dual v1-41 activity
Structure
Potency and
Selectivity#
v1 pKd 9.57
41 pKd 9.11
(x2.9 sel. for
v1)
Physchem
properties
MW 559
ChromLogD7.4 0.68
1[a]
4[b]
5[c]
CAS no. 1689540-62-2
MW 573
ChromLogD7.4 3.7
v1 pKi 8.51
41 pKi 9.70
(x15.5 sel. for
41)
CAS no. 217325-56-9
CAS no. 232271-19-1
MW 474
ChromLogD7.4 2.7
v1 pKi 6.50
41 pKi 9.13
(x427 sel. for
41)
Table 1. Exemplar v1 hit molecules obtained from 41 templates
Structure
Potency and
Physchem
properties
MW 499
ChromLogD7.4 2.8
Selectivity#
[a] Reported data: v1 IC50 0.089-0.63 nM.[7,18] [b] Reported data: 41 IC50 0.46
nM.[19] [c] Reported data: TR-14035 41 IC50 87 nM.[20] #GSK Radioligand data (pKi
or pKd) – mean of 2 assay test occasions.
2
3
v1 pIC50 5.8
v3, 5, 6, 8 pIC50 <5
v1 pKi 7.41
BEI* 11.6 (14.8)
The hit compounds containing the biphenyl -amino acid
backbone showed the expected binding mode to v1 as
depicted by the in silico docking and served as the basis for
future molecule design (See Figure 2 and SI for methodology).
Elevated potencies were readily achieved by the
incorporation of simple functional groups to the
unsubstituted bis-methoxy biphenyl 5, to give for example,
nitrile 2 and prolinol 3. The general finding was that the hit
compounds exhibited good potency at v1 with inherent
selectivity over the v integrin family members, although
with concomitant high activity at 1, as expected. It is
therefore highly notable that in contrast to almost all other
small molecule v-RGD inhibitor structures, hit molecules
here lacking the Arg mimetic (see 2, 3, 4, and 5) can be
considered as non-RGD inhibitors of the RGD integrin v1.
This is presumably driven in part by maximizing other
interactions within the 1-subunit, such as the acid-metal
interaction in the metal-ion dependent adhesion site
(MIDAS), the π-π interaction with Tyr178 and/or with
interactions to the benzamide moiety (See Figure 2).
41 pKi 9.85
(x275 sel. for 41)
MW 632
ChromLogD7.4 0.13
BEI* 11.7 (14.3)
v1 pIC50 7.4
v3, 5, 6, 8 pIC50 <5
v1 pKi 9.01
41 pKi 9.71
(x5 sel. for 41)
# GSK cell adhesion data (pIC50) – mean of at least 2 assay test occasions; radioligand
data (pKi) – mean of 2 assay test occasions. *BEI[15] calculated from v1 pIC50 and
(pKi).
This result was very surprising given that the close
homology in v integrin binding sites normally precludes
high selectivity for one v-isoform.[5c] Moreover, it was
unexpected to find that the hit compounds originated from
legacy series of 1 inhibitor molecules, thus
demonstrating unusual cross reactivity across the integrin
families (RGD and non-RGD) ─ the 1 integrin belongs to
the leukocyte integrin family (not RGD) with different ligand
recognition sequences to RGD.[16]
Before optimizing the dual active hit molecules (1-
v1) to improve potency and selectivity for v1, we
explored the generality of this finding and only found one
example from the literature − a series of aryl sulfonamide
pyrimidines with activity at both v1 (0.5M) and
1 (9nM).[17] However, with no other indication of this
dual activity being reported, we screened additional
structurally related 1 small molecules from the literature
The hit compounds are also considerably less polar
than 1 as determined by the higher ChromLogD values[21]
and may therefore help to increase permeability. Hit
molecule 2 was a reasonable start point for further
optimization because of its structural similarity to the oral
at v1 (eg. 4, Table 2) and found that the connection of clinical compound firategrast[5b] − a dosing route advantage
embedded dual activity began to strengthen, although the
activity at 1 was always higher. During this work, we also
over many other v integrin inhibitors. During the
optimization of the hits, the ligand efficiency metric (BEI
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900359
ChemMedChem
COMMUNICATION
efficiency measure)[15] was used to establish which structural
changes were most beneficial. At the outset, this revealed
that the hit compounds (eg. 2) generally exhibited low
efficiency at v1, mainly because of the moderate potency,
but we show how the molecules were efficiently optimized
to give better profiles (vide infra).
space held by these v1 GD type hits. Visual inspection of
v1 and 1 integrin in silico homology models (See SI for
methodology) indicated that incorporation of a polar motif
to the 1 molecules in the v subunit could potentially
reduce binding to 1 (due to lack of interaction with
lipophilic residues) but concomitantly elevate v1 activity
(due to an interaction with Asp218). We reasoned that a
non-basic binding motif to Asp218 in the form of a urea,
which has precedence in v3,[22] might serve to meet the
desired potency, selectivity and physicochemical
requirements. To test the hypothesis, novel phenyl urea 6
(Table 3) was rapidly synthesized in
3 steps from
commercially available materials (See Figure 3 and SI) then
tested for activity in cell adhesion and radioligand assays.
Figure 2. Docking of non-RGD hit (2) in an v1 homology model showing good π-π
stacking of the bismethoxy benzonitrile with Tyr 178 (Note the absence of interactions
with Asp218, the normal observed Arg binding amino acid in RGD inhibitors).
The wider v1 SAR around molecule 2 was quickly
established by the testing of available close analogues not
previously screened, but also through the synthesis of
additional analogues, still lacking an Arg binder (data not
shown). The first intent was to determine the minimum
structural features responsible for good potency (ie.
pharmacophore) and high ligand efficiency. However, it was
found that most of the functional groups already present in
these GD hits were required for binding. This was not a
surprise, given the curtailed binding opportunity ie. without
the Arg binder. Nevertheless, potential efficiency
improvements were envisaged in the -region by replacing
the bulky dichlorobenzamide group or bismethoxy benzene
appendages with lower MW alternatives, but this proved to
be challenging. A potency advantage came with installation
of a para-carboxylic acid, present in 3, which forms a
putative salt-bridge interaction with Lys182, giving a potency
increase of 1 log unit (see SI-homology models for docking
schematic). Yet, evaluation of this hit series early on
indicated that any potency advantages would likely be
outweighed by the sub-optimal oral drug-like properties,
therefore we took the decision to pursue an alternative lead.
Given that the non-RGD hit molecules were not of
suitably high v1 potency, we hypothesized that
improvements may be achieved by introducing a motif with
modulated binding to Asp218, in the v-subunit akin to the
traditional full RGD mimetics. This would also give a
potential property advantage, overcoming the requirement
for a strongly basic group, such as guanidine, which could
otherwise limit permeability − a common design problem
encountered in the quest for oral RGD mimetics and hence
our keen interest to maintain as much of the oral property
(a) Na2CO3, MeCN/H2O (b) RNCO, CHCl3, 110 oC (c) 4-NO2PhOCOCl, pyr, CH2Cl2 then
ArNH2 (d) (A), Pd(dppf)Cl2, K3PO4, THF/H2O, 110 oC.
Figure 3. General synthesis of v1 urea inhibitors
Phenyl urea 6 showed a dramatic selectivity switch now
in favour of v1 over 1 (x21 fold) and represented the
breakthrough lead that delivered on potency and selectivity
(See Table 3). It gave the expected docking pose in the v1
homology model, showing a bidentate interaction between
the urea and the Asp218 residue located in the -subunit,
Table 3. Urea hit
Structure
Potency and
Physchem
properties
Selectivity#
6
MW 500
ChromLogD7.4 2.1
BEI* 11.8 (16.4)
v1 pIC50 5.9
v3, 5, 6, 8 pIC50 <5
v1 pKi 8.22
41 pKi6.89
(x21 sel. for v1)
# GSK Radioligand data (pKi) – mean of 2 assay test occasions; cell adhesion data
(pIC50) – mean of at least 2 assay test occasions. *based on v1 cell adhesion and
(radioligand) data.
which is absent in 1 (Figures 4a & 4b) and thus provides
an explanation for the observed selectivity switch, favouring
v1. While the potency for 6 at v1 was within the range
of original hits 2 and 3, improvements in ligand efficiency
were clearly evident (BEI 16.4 vs. 14.8 & 14.3), thus allowing
greater scope for incorporation of additional structural
modifications that may be of benefit. The absence of the
methoxy groups aided an increase in BEI and is thought to
provide increased structural flexibility for binding to Asp218
while maintaining the acid-metal interaction. It was pleasing
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900359
ChemMedChem
COMMUNICATION
with the electron rich residue Tyr178 (See Figures 2 and 4).
DFT calculations (Table 6 and SI for methodology) showed
that (i) predominates over (ii) and additionally, from dihedral
scanning, the geometry of the urea is affected by the
electronics of the attached aryl ring, impacting the binding
conformation (vide infra).
to observe that activity at other v integrin members was
still absent (pIC50 <5).
a
Table 4. SAR of selected urea molecules
Structure
Potency and
Selectivity^
Physchem
properties
7[d]
MW 510
ChromLogD7.4 3.0
BEI* << 9.8
v1 pIC50 <5
v3 pIC50 6.2
v5 pIC50 6.1
v6 pIC50 <5
8
9
MW 514
ChromLogD7.4 2.2
BEI* 11.1
v1 pIC50 5.7
v3,5,6,8 pIC50 <5
b
MW 502
v1 pIC50 6.4
ChromLogD7.4 2.7
BEI* 15.7
v3,5,6,8 pIC50 <5
[d] close analogue to that already reported[22b] ^GSK cell adhesion data (pIC50) – mean
of at least 2 assay test occasions. *based on v1 cell adhesion data.
To test this hypothesis, the electronic properties were
initially probed by substituting fluorine atoms at each of the
four sites on the benzene ring (adjacent to the urea as in 10,
Table 5) – synthesized according to Figure 3. Increases in
potency at v1 were immediately obtained for three of the
four F analogues (v1 pIC50 6.4-6.8) vs. pIC50 5.9 for the
unsubstituted urea 6.
Figure 4. a) Docking of lead urea (6) in the v1 homology model showing a robust
urea bidentate interaction with Asp218. b) Docking of lead urea (6) in the 1
homology model, unable to effectively bind the urea motif.
Table 5. Advanced v1 leads
The biaryl urea amide series now offered an v
pharmacological profile clearly differentiated from the
reported v3 urea sulfonamide series,[22b] which is devoid
of v1 activity (see closely related analogue 7 for
comparison, Table 4). With the urea group (in combination
with a biaryl amide) now firmly established as essential for
binding and selectivity, the structural requirements of the
aryl urea motif itself were next probed. It was found that
direct attachment (without a spacer) of the urea to the meta
position (preferred over para) on the aromatic ring was
optimal (see 8 and SI for additional analogues). In addition,
no significant advantages were gained by changing the ethyl
group pendant to the urea terminus (see SI for further
examples) although this strategy has shown to increase
selectivity for v6 in cyclic peptides.[23] Consequently, the
biaryl meta urea template as contained in 6 was kept
unchanged to investigate how increases in potency could be
achieved.
Structure
Potency and
Physchem
properties
Selectivity#
10
11
MW 518
ChromLogD 2.5
BEI* 13.1 (17.1)
v1 pIC50 6.8
v3, 5, 6, 8 pIC50 <5
v1 pKi 8.89
41 pKi 6.71
(x151 sel. for v1)
MW 501
v1 7.8
ChromLogD 1.6
BEI* 15.6 (19.8)
v6 5.3
v8 6.1
v3, 5 pIC50 <5
v1 pKi 9.94
41 pKi 7.52
(x263 sel. for v1)
v1 pIC50 7.9
v3, 5, 6 pIC50 <5
v8 pIC50 6.2
12
MW 502
ChromLogD 1.2
BEI* 15.7 (19.5)
v1 pKi 9.78
41 pKi 7.11
(x468 sel. for v1)
We further hypothesized that reducing the electron
density of the terminal phenyl ring could serve to increase
the urea binding interactions in two ways; (i) by increasing
urea NH polarization to effect stronger hydrogen bonding to
Asp218 and (ii) by increasing favourable π-π interactions
# GSK Radioligand data (pKi) – mean of 2 assay test occasions; cell adhesion data
(pIC50) – mean of at least 2 assay test occasions. *based on v1 cell adhesion and
(radioligand) data.
The meta-F analogue 10 conferred the highest potency
(pIC50 6.8, pKi 8.9) for only a slight increase in MW compared
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900359
ChemMedChem
COMMUNICATION
to hit 6 (F for H), translating to a higher BEI (17.1 vs. 16.4).
v1 over 1, approaching an impressive 500-fold
Importantly, the trend for increased selectivity over 1 window. For optimized leads 11 and 12, the BEI was
significantly improved over hit urea 6, by >x1000 fold,
allaying initial concerns over the low efficiency start points.
The potency of pyridine 11 is very similar to that of
pyridazine 12, explained by the stabilisation energy of the
ligand-Asp218 interaction for 12 falling between the
energies of protonated and deprotonated pyridine forms of
11. In addition, the lowest energy state urea conformations
could also be a minor factor to help explain potency
differences between urea 6 and the less electron rich
systems of 10, 11 and 12. (See SI for dihedral energy plots).
The trans urea conformation was calculated to be more
stable in phenyl urea 6 (+0.28 kcal/mol for cis) thus causing
an energy penalty for delivering bidentate binding to Asp218.
In the case of compound 10, the energy difference between
cis and trans was negligible (+0.03 kcal/mol), but for potent
ureas 11 and 12, the cis conformation was highly favoured
(+0.52 & +1.48 kcal/mol for trans), allowing effective
bidentate interaction with Asp218.
continued, mainly by increasing potency at v1 − compare
fluoro phenyl urea analogue 10 (v1:1 sel. x151) with
the unsubstituted phenyl urea 6 (v1:1 sel. x21).
Further empirical support that the hydrogen bond donating
ability of the urea is crucial for binding to Asp218 was
observed with the electron deficient trifluoromethyl urea 9
(Table 4), which gave an increase in potency of 0.5 log
compared with hit urea 6.
In order to raise the v1 potency and selectivity still
further, with the hypothesis on urea hydrogen bonding
strength appearing to be valid, it was envisaged that electron
deficient aromatic groups such as pyridines could serve to
further enhance binding interactions. This was supported by
DFT calculations (See Table 6 and SI for methodology).
Table 6. Calculated complexation energies of the aryl urea with Asp218 and Tyr178
Ligand-Asp218
(kcal/mol)
-19.31
Ligand-Tyr178
(kcal/mol)
-8.26
6
10
-22.84
-4.77
It is anticipated that the developability of the urea series
in vivo could take a similar path to the structurally related
templates of historic 1 molecules, where high oral
bioavailability required an ester prodrug.[25] Preliminary
work however, showed that several urea parent compounds
had good passive permeability, such as 10 (MDCKII-MDR1
Pexact 77 nm/s). Compound 10 also demonstrated low in
vitro clearance in rat and human liver S9 fractions (<0.6
ml/min/g tissue) which suggests a reasonable level of
metabolic stability − (see SI for assay details). The wider
selectivity profile for compound 10 against numerous
targets was also very good (See SI for full list).
In conclusion, this work describes for the first time, the
identification of non-RGD dual 1-v1 ligands and their
subsequent conversion into highly potent and selective
RGD v1 inhibitors, by the incorporation of a non-basic aryl
urea as the Arg mimetic. This lead urea series has
developability advantages, such as ease of synthesis and
high selectivity over 41, compared to previously
described molecules. Additionally, this non-zwitterionic and
non-peptidic template could serve benefit for the design of
good oral properties – this remains the subject of future
optimization.
11 protonated
(deprotonated)
12
-32.13
(-23.81)
-25.49
-5.89
(-4.79)
-5.31
The trend in the ligand-Asp218 complexation energy was
similar to the measured activity, whereas no trend was
observed with the ligand-Tyr218 complexation energy,
although clearly this interaction contributes to the potency.
Each pyridine isomer was synthesized in the same way as
previous molecules (Figure 3) and immediately, elevated
potencies were obtained, compared with the fluorobenzene
matched pairs − in all but one isomer (pyridyl isomers: v1
pIC50 5.5-7.8 vs. fluorobenzene isomers: v1 pIC50 6.4-6.8,
see Table 5 and SI). Pyridine 11 delivered the highest potency
at v1 (pIC50 7.8, pKi 9.94) – an increment of >1.5 log units
from lead urea 6 (pKi 8.22). The selectivity over
1 continued to widen (x263 fold). Interestingly, pyridyl
isomer 11 was considerably more basic than the other
pyridyl isomers (measured pKa 6.3 vs. 3.4-4.9), with the
protonated form expected to give a substantial reduction in
electron density and deliver stronger interactions with the
receptor by ~8.5 kcal/mol. It is interesting to note that the
functional groups leading to increased v1 binding also
introduced weak activity at v6 and v8, reinforcing the
point that subtle changes in the electron density and
structure of the ligand in the v-subunit binding site may
impact on potency and selectivity across all the vx family
members. Alternative heteroaromatic groups, as in the case
of the non-basic pyridazine heterocycle 12 (measured pKa
4.0) also delivered high potency, presumably the result of
the large dipole moment[24] compared to 6, enabling
stronger interactions with Asp218. Thus far, pyridazine 12
delivered the best combination of potency and selectivity for
Acknowledgements
The authors wish to acknowledge the helpful comments made by the
reviewers.
Conflict of interest
The authors declare no conflict of interest.
Keywords: v1 integrin • RGD integrin inhibitors • drug design
• 1 integrin • medicinal chemistry
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900359
ChemMedChem
COMMUNICATION
[1] a) T. A. Wynn, J. Pathol. 2008, 214, 199–210; b) S. L.
Friedman, D. Sheppard, J. S. Duffield, S. Violette, Sci
Transl. Med. 2013, 5, 167sr1.
[2] C. B. Nanthakumar, R. J. D. Hatley, S. Lemma, J.
Gauldie, R. P. Marshall, S. J. F. Macdonald, Nat. Rev. Drug
Discov. 2015, 14, 693-720.
[3] N. C Henderson, D. Sheppard, Biochim. Biophys. Acta.
2013, 1832, 891–896.
[4] J. Takagi, B. M. Petre, T. Walz, T. A. Springer, Cell
2002, 110, 599–611.
[18] N. I. Reed, Y-Z. Tang, J. McIntosh, Y. Wu, K. S.
Molnar, A. Civitavecchia, D. Sheppard, W. F. DeGrado, H.
Jo. ACS Med. Chem. Lett. 2016, 7, 902–907.
[19] W. K. Hagmann et al. Bioorg. Med. Chem. Lett. 2001,
11, 2709–2713.
[20] I. Sircar et al. Bioorg. Med. Chem. 2002, 10, 2051–
2066.
[21] R. J. Young, D. V. S. Green, C. N. Luscombe, A. P.
Hill, Drug Discov. Today 2011, 16, 822−830.
[22] a) K. Urbahns, M. Härter, M. Albers, D. Schmidt, B.
Stelte-Ludwig, U. Brüggemeier, A. Vaupel, C. Gerdes,
Bioorg. Med. Chem. Lett. 2002, 12, 205–208; b) K.
[5] a) S. L. Goodman and M. Picard, Trends in Pharm. Sci.
2012, 33, 405–412; b) D. Cox, M. Brennan, N. Moran, Nat.
Rev. Drug Discov. 2010, 9, 804–820; c) R. J. D. Hatley, S. J. Urbahns, M. Härter, A. Vaupel, M. Albers, D. Schmidt, U.
F. Macdonald, R. J. Slack, J. Le, S. B. Ludbrook, P. T.
Lukey, Angew. Chem. Int. Ed. 2018, 57, 3298–332.
[6] a) G. S. Horan et al. Am. J. Respir. Crit. Care Med.
2008, 177, 56–65; b) Y. Popov et al. J. of Hepatology 2008,
48, 453–464; c) B. Wang, B. M. Dolinski, N. Kikuchi, D. R.
Leone, M. G. Peters, P. H. Weinreb, S. M. Violette, D.
Montgomery Bissell, Hepatology 2007, 46, 1404–1412.
[7] a) N. I. Reed, H. Jo, C. Chen, K. Tsujino, T. D. Arnold,
W. F. DeGrado, D. Sheppard, Sci. Transl. Med. 2015, 7,
288ra79; b) Y. Chang, W.L. Lau, H. Jo, K. Tsujino, L.
Gewin, N.I. Reed, A. Atakilit, A.C.F. Nunes, W.F.
DeGrado, D. Sheppard, J. Am. Soc. Neph. 2017, 28, 1998–
2005.
Brüggemeier, B. Stelte-Ludwig, C. Gerdes, H. Tsujishita,
Bioorg. Med. Chem. Lett. 2003, 13, 1071–1074; c) S.
Nagashima, S. Akamatsu, E. Kawaminami, S. Kawazoe, T.
Ogami, Y. Matsumoto, M. Okada, K-I. Suzuki, S-I.
Tsukamoto, Chem. Pharm. Bull. 2001, 49, 1420–1432.
[23] T. G. Kapp, Maximilian Fottner, O. V. Maltsev, H.
Kessler, Angew. Chem. Int. Ed. 2016, 55, 1540–1543.
[24] N. A. Meanwell, J. Med. Chem. 2018, 61, 5822–5880.
[25] H. Chanteux, L. Staelens, V. Mancel, B. Gerin, D.
Boucaut, C. Prakash, J-M. Nicolas, Drug Metab. Dispos.
2015, 43, 1381–1391.
[8] A. L. Wilkinson, J. W. Barrett, R. J. Slack, Eur. J.
Pharmacol. 2019, 842, 239–247.
[9] M. E. Duggan, J. H Hutchinson, Expert Opinion on
Therapeutic Patents, 2000, 10, 1367–1383.
[10] L. M. Miller, J. M. Pritchard, S. J. F. Macdonald, C.
Jamieson, A. J. B. Watson, J. Med. Chem. 2017, 60, 3241–
3251
[11] J. B. Baell, G. A. Holloway, J. Med. Chem. 2010, 53,
2719–2740.
[12] G. X. Yang and W.K. Hagmann, Medicinal Research
Reviews, 2003, 23, 369–392.
[13] a) S. T. Barry, S. B. Ludbrook, E. Murrison, C. M. T.
Horgan, Biochem. Biophys. Res. Commun. 2000, 267, 764–
769; b) S. B. Ludbrook, S. T. Barry, C. J. Delves, C. M. T.
Horgan, Biochem. J. 2003, 369, 311–318; c) S. J. F.
Macdonald et al. ACS Med. Chem. Lett. 2014, 5 1207-1212.
[14] a) D. M. Nowlin, F. Gorcsan, M. Moscinski, S-L.
Chiang, T. J. Lobl, P. M. Cardarelli, J. Biol. Chem, 1993,
268, 20352–20359; b) P. M. Cardarelli, R. R. Cobb, D. M.
Nowlin, W. Scholz, F. Gorcsan, M. Moscinski, M.
Yasuhara, S-L. Chiang, T. J. Lobl, J. Biol. Chem, 1994, 269,
18668–18673; c) P. Sánchez-Aparicio, C. Dominguez-
Jiménez, A. Garcia-Pardo, J. Cell Biol. 1994, 126, 271–279.
[15] C. Abad-Zapatero, J. T. Metz, Drug Discov. Today
2005, 10, 464–469.
[16] J. S. Desgrosellier and D. A. Cheresh, Nat. Rev. Cancer
2010, 10, 9–22.
[17] Y-z. Xu et al. Bioorg. Med. Chem. Lett. 2013, 23,
3070–3074.
This article is protected by copyright. All rights reserved.

10.1002/cmdc.201900359
ChemMedChem
COMMUNICATION
COMMUNICATION
Author(s), Corresponding Author(s)*
Page No. – Page No.
Title
Novel v1 inhibitors: The discovery and subsequent optimization of non-RGD
dual v1−41 ligands resulted in the identification of novel, highly potent and
selective aryl urea inhibitors of the RGD integrin, v1.
This article is protected by copyright. All rights reserved.