Selective non zinc binding inhibitors of MMP13

Article abstract of DOI:10.1016/j.bmcl.2011.05.075

Directed screening has identified a novel series of MMP13 inhibitors that possess good levels of activity whilst possessing excellent selectivity over related MMPs. The binding mode of the series has been solved by co-crystallisation and demonstrates an interesting mode of inhibition without interaction with the catalytic zinc atom.

Full text of DOI:10.1016/j.bmcl.2011.05.075

Bioorganic & Medicinal Chemistry Letters 21 (2011) 4215–4219
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Selective non zinc binding inhibitors of MMP13
⇑
Chris De Savi, Andrew D. Morley , Attilla Ting, Ian Nash, Kostas Karabelas, Christine M. Wood,
Michael James, Stephen J. Norris, Galith Karoutchi, Neil Rankine, Gordon Hamlin, Philip A. MacFaul,
David Ryan, Sarah V. Baker, David Hargreaves, Stefan Gerhardt
AstraZeneca, Mereside, Alderley Park, Macclesfield, Cheshire SK10 4TG, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
Directed screening has identified a novel series of MMP13 inhibitors that possess good levels of activity
whilst possessing excellent selectivity over related MMPs. The binding mode of the series has been solved
by co-crystallisation and demonstrates an interesting mode of inhibition without interaction with the
catalytic zinc atom.
Received 19 April 2011
Revised 18 May 2011
Accepted 20 May 2011
Available online 27 May 2011
Ó 2011 Elsevier Ltd. All rights reserved.
Keywords:
MMP13
Selectivity
Non zinc binder
Lead Generation
Matrix metalloproteases (MMPs) are a class of zinc and calcium
dependent endopeptidases that are responsible for the cleavage of
a variety of extracellular matrix proteins.1 More than 20 family
members have been identified and these have been sub-divided
dependent on their preferential substrate specificity.2 For example,
the sub group termed collagenases (MMPs 1, 8, 13, 18) have been
shown to favour cleavage of various types of collagen. The enzyme
family is usually tightly regulated, both at a transcription and local
level,3,4 in the latter cases via tissue inhibitors of metalloproteases
(TIMPs).5,6 However, when this equilibrium breaks down, it can
lead to diseases such as arthritis, chronic obstructive pulmonary
disease, cancer, periodontal disease and multiple sclerosis.4,7–11
MMP13 (collagenase 3) has been shown to be the key enzyme in
the cleavage of type II collagen and believed to play a pivotal role
in the breakdown of cartilage in osteoarthritic joints.12
in the S10 region of MMP13.17,18 This type of approach formed
the basis for our own project strategy.
O
O
O
N
O
N
H
N
H
N
N
N
N
N
2
1
CO2H
N
O
Although many inhibitors of various MMPs have been identified
and progressed into clinical development, all have failed, primarily
due to lack of therapeutic margins.13,14 The rate limiting effect is
musculoskeletal syndrome, which results in joint stiffness and ten-
donitis and the effect has been attributed to a lack of selectivity
over other MMP family members. Most published MMP inhibitors
contain a metal binding motif which interacts with the catalytic
zinc atom, a conserved feature of the target class.1,2,15,16 Identifying
inhibitors which form enzyme specific interactions (e.g., 1 and 2)
has been shown to generate compounds with superior selectivity
profiles over traditional MMP inhibitors through binding deeply
O
N
H
3
MMP13
MMP2
pIC50 = 5
pIC50 = 4.3
MMP9/12/14 pIC50 <4
A directed screen of the AstraZeneca corporate collection (19 K)
against MMP13 was undertaken at 30 M (duplicate testing), gen-
l
erating a 15% hit rate. Actives were tested in a four point dose
response against MMPs 2, 9, 12, 13, 14 to get a view of compound
potency and selectivity against key MMP isoforms.19 A large
⇑
Corresponding author. Tel.: +44 1509 644772.
E-mail address: andy.morley@astrazeneca.com (A.D. Morley).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.05.075

4216
C. D. Savi et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4215–4219
number of non-selective inhibitors and false positives were identi-
fied as part of the downstream evaluation. One of the few
compounds that did inhibit MMP13 with modest levels of selectiv-
ity was compound 3. Binding to the enzyme was confirmed using
NMR chemical shift perturbation mapping with 15N-labelled
The pyridyl or aromatic motifs per se are not essential for
MMP13 activity. The poor potency of the unsubstituted benzyl
amide (5) was unclear. Homologation removed MMP13 activity
and methylation to generate the tertiary amide showed a slight
reduction in potency. Smaller and more polar analogues, such as
10 and 12 were less active but demonstrated improved ligand effi-
ciency (LE)23 and/or lipophilic ligand efficiency (LLE).24 These val-
ues were still well below what we deemed acceptable for hit
identification start points. The data hinted that the compounds
were inhibiting MMP13 through a novel binding mode, as the gen-
erated SAR was inconsistent with reported literature data for allo-
steric modulation or traditional zinc binders.17,18
Evaluation of the SAR of the phenoxyphenyl motif was then
undertaken (Table 2). The terminal phenyl ring was required for
compounds to show activity within the testing concentration
range. Extending the linker between the ether and aromatic was
detrimental to activity, whereas removal of the ether was toler-
ated, albeit with some reduction in MMP13 activity (i.e., 15). Initial
attempts from available starting materials to introduce substitu-
ents in the ortho or meta positions of the terminal ring were not
encouraging, whereas para substitution was tolerated, with 21
showing the most encouraging LLEs and comparable LE to 3.
At this stage it was essential to find some additional interac-
tions that generated significant MMP13 potency increases to
warrant progression of the series into lead identification. Based
on the SAR that had been generated, a focused exploration was
undertaken, around the para substituent of the terminal phenyl
ring. Most of the synthesised compounds were devoid of MMP13
MMP13 (180 l
M) on a 600 MHz Varian Inova spectrometer.20
Compound 3 induced specific amide resonance shifts in slow
exchange on the NMR time scale and saturated at a 1:1 molar
stoichiometry, indicative of binding in the <10 lM Kd range.
N
N
O
22 MMP13 pIC50 7.2 R2 =
23 MMP13 pIC50 6.9 R2 =
O
N
H
N
R2
OH
Compound 3 does not appear to be an obvious zinc binder,
although the pyridine ring could be acting as a metal chelator.
Alternatively, there are similarities to 1, and a reasonable overlay
can be generated.21 Preliminary exploration focused on probing
the structure activity relationship (SAR) of the amide substituent
and a small set of analogues were synthesized. This initial assess-
ment was limited due to the relatively low solubility of the com-
pounds compared to their MMP13 potency. This made detailed
evaluation at this stage risky and results were challenging to
unequivocally interpret. These are summarised in Table 1.
Table 1
Early amide SAR
O
R
O
Mb
LEc
LLEd
1.1
a
Compd
R
MMP13 pIC50
5
c Log P
Sol
10
l
N
H
3
3.9
0.21
N
N
H
4
4.6
3.9
NT
0.19
0.7
N
N
H
5
6
7
<3.4
4.6
5.4
5.4
3.9
20
5
—
—
MeO
N
H
0.18
0.17
ꢀ0.8
1.0
N
4.3
NT
Me
N
N
H
N
8
9
<3.4
4.3
4.0
6.1
4
—
—
N
H
NT
0.18
ꢀ1.8
H
N
10
11
12
4.5
2.9
4.1
3.5
260
0.22
—
1.6
—
HO
H
<3.4
4.6
2800
10
Me2N
N
H
N
0.26
1.1
Me
NT – Not tested.
a
Values mean of n = 2 experiments.
b
c
Thermodynamic solubility data.22
LE (Ligand efficiency)23 Values are calculated as pIC50/Heavy atom count.
d
LLE (Lipophilic ligand efficiency)24 Values are calculated as pIC50-c log P.

C. D. Savi et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4215–4219
4217
Table 2
Phenoxyphenyl SAR
N
O
R1
N
H
Compd
R1
MMP13 pIC50
c log P
Sol
lM
LE
LLE
13
14
15
OMe
OPh
Ph
<3
4.6
4.4
1.5
3.4
3.2
2200
29
11
—
0.2
0.21
—
1.2
1.2
16
17
O
O
Ph
<4
<4
3.3
3.6
9
—
—
—
—
Ph
35
O
O
NO2
NO2
18
19
4.7
3.2
3.1
3
0.18
—
1.6
—
<3.4
NT
OMe
NC
20
21
<3.4
5.3
2.9
2.1
79
19
—
—
O
O
CONHMe
0.2
3.2
activity, but one analogue, 22, provided
improvement in potency.
a
hundred fold
to fix the loop in a specific conformation. It is assumed that
lipophilic interactions between methylene units of the quinucli-
dine ring and proline 255 contributes to MMP13 potency. This
could also play a key role in generating good selectivity over
MMP2 (where the corresponding amino acid is serine). The bridge-
head hydroxyl substituent is shown to form a network of hydrogen
bonds with adjacent water molecules.
An approximate 4 fold difference in MMP13 activity exists
between the two enantiomers. The improved MMP13 activity for
the (S)-isomer appears the major reason for the increased MMP
selectivity profile as potency for both enantiomers against the
In order to expand the data set around this new observation, 23
was synthesized, showing a similar MMP13 profile to 22. Broader
testing of 23 gave selectivity over MMP2 and MMP12 of 22 and
35 fold respectively, with no inhibition against a wider panel of
MMPs and related metalloproteases (see supplementary
information).
At this stage we were successful in co-crystallizing 23 with
MMP13.25 The (S)-enantiomer (24) preferentially bound to the
enzyme in the co-crystallisation process. This preference was con-
firmed when both isomers were prepared26 and tested (Table 3).
The binding mode and key interactions are summarized in
Figure 1. Compound 24 binds with the majority of the molecule
positioned within the S10 pocket. Both the amide carbonyl and
NH form hydrogen bonds to leucine 185 and proline 242 respec-
tively. The diarylether fits tightly into the barrel that forms the
neck of the S10 pocket. The pyridyl motif sits in the groove where
the peptide backbone of the enzyme substrate would bind, proxi-
mal to the zinc atom, but without making any specific interactions
with the metal. This region of MMP13 is quite lipophilic so there
are reasonable Van der Waals interactions between the aryl group
of 24 with the enzyme, particularly leucine 184. The quinuclidine
motif binds to the S10 loop. This region appears to be quite flexible
when comparing structures across the MMP family. The basic
nitrogen forms an interaction with threonine 247, which appears
Table 3
Enantiomeric biological profiles
N
O
R3
N
H
N
O
Compd
R3
MMP13 pIC50
MMP2, 9, 12, 14 fold selectivity
23
24
25
(R/S)-OH
(S)-OH
(R)-OH
6.9
7.1
6.5
25, >75, 35, >75
94, >135, >135, >135
15, >30, >30, >30
Figure 1. Structure of 24 bound to MMP13.

4218
C. D. Savi et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4215–4219
Table 4
Table 5
Linker SAR
Profile of 26
O
N
R2
O
N
H
N
N
OH
N
H
O
N
N
O
Compd R2
MMP13 pIC50 MMP2, 9, 12, 14 fold selectivity
N
Parameter
Lead criteria
26
26
27
28
29
7
>100, >100, >100, >100
20, >25, >25, >25
>4, >4, >4, >4
MMP13 pIC50
MMP2, 9, 12, 14 fold selectivity
Cytotoxicity pIC50
Log D
Mol Wt
LE
pIC50 >7
>50
<4.3
<3.0
<450
7
HO
HO
>100
<4.3
1.8
454
0.21
5.1
N
a
6.4
5.6
6.6
N
N
LLE
Solubility
HO
l
M22
>100
lM
>800 lM
Hu/Rat Prot Binding
% free
24/23
<2 l/min/mg
l/min/106cells
25, >40, >40, >40
Hu Micsb
<30
<15
l
l
l/min/mg
l
HO
Rat Hepsc
l/min/106cells
3 l
N
Chem Stability29
t
>100 h
>1000 h
<5 (5/5)
<4.3 (5/5)
<4.3
½
d
30
31
5.5
6.3
>3, >3, >3, >3
Cyp Inhib pIC50
pIC50 <5
pIC50 <5
pIC50 <5
t1/2 >1 h
MeO
CypTDIe
N
hERG pIC50
Rat PK Cl/t1/2/Vdss
6, >20, >20, >20
f
17/2.4/2.4
pIC50 values are the means of at least three experiments.
a
Inhibition of THP-1 cell viability.
b
Human microsome metabolism intrinsic clearance (l .
L/min/mg).27
other MMP isoforms is similar. Compound 24 has ꢁ100 fold selec-
tivity, or better, against all the family members that were tested.
The presence of the 4-pyridyl motif did mean that 23 inhibited
Cyp3A4, 2C9 and 2C19 with pIC50s >5. Replacing this group with a
4-pyrimidinyl unit maintained the biological profile and gave clean
Cyp inhibition (pIC50 <5) for all isoforms tested. Attention then
turned to exploring the SAR of the linker between the aryl and
quinuclidine motifs in more detail. Key data is summarized in
Table 4.
SAR suggests all the interactions observed in the crystal struc-
ture provide important contributions towards MMP13 activity
and any modifications in this region reduced affinity and subse-
quently selectivity over MMP2, in particular. Several other linkers
were also explored (data not shown) but all failed to generate
acceptable biological profiles (MMP13 pIC50s <5). Compound 26
was chosen as our lead candidate and profiled extensively (Table
5).
c
Rat Sprague–Dawley hepatocyte metabolism intrinsic clearance (l
L/min/106
cells).28
d
Inhibition of cytochrome P450 isoforms: 1A2, 2C9, 2C19, 2D6 and 3A4.
TDI –Time dependant inhibition.
Compounds dosed at 2 mg/kg iv, n = 2 animals (Cl ml/min/kg, t1/2 h, Vdss l/kg).
e
f
metabolites). Therefore, representative members of the series were
evaluated in several stability assays to assess suitability to progress
into lead optimisation. All tested analogues were stable to heating
at pH range 4–10, extrapolating to t1/2 of >1000 h in all cases.29 No
changes were observed when compounds were subjected to our in
house photolytic stability assay.30 Evaluation for reactivity in the
presence of glutathione (both with and without the S-transferase)
showed the compounds to be inert.31 The series also showed no
time dependent inhibition of five human CYPs. The electron with-
drawing effects of the alkyne and oxygen functions limit the quinu-
clidine pKa to 8.5, giving the series attractive physicochemical and
DMPK profiles. Metabolite identification studies showed the N-
oxide of the quinuclidine to be the only metabolite identified and
The presence of the alkyne unit could be perceived as a
potential risk with respect to chemical reactivity (directly or via
R2
CO2Me
Br
CO2Me
Br
CO2H
a
b or c
O
O
O
R2
=
32
OR4
N
OR4
N
d,e or e
O
R2
CO2H
O
R2
R
R2
R
f
N
H
g
N
H
O
O
O
R2
=
31
23-30
OR4
N
OR4
N
N
OR4
Scheme 1. Reagents and conditions: (a) 5–10% HCl/MeOH, 65 °C, 7 h, 90%; (b) quinuclidine alkyne, Pd(Cl)2(PPh3)2, Cu(I)I, PPh3, Et3N, DMF, 70 °C, 83%; or (c) quinuclidine
alkene, Pd(OAc)2, TBACl, (o-tolyl)3P, Et3N, DMF, 130 °C, 1 h, MW, 59%; (d) 10% Pd/C, H2, EtOH, 48 h, 90%; (e) 2.5 N aq NaOH (or LiOHꢂH2O) H2O/MeOH, 50 °C, 2–16 h, 84%; (f)
amine, HATU, DIPEA, DMF, rt, 15–84%; (g) formic acid, 120 °C, 2.5 h, 21–96%

C. D. Savi et al. / Bioorg. Med. Chem. Lett. 21 (2011) 4215–4219
4219
15. Jacobsen, J. A.; Major Jourden, J. L.; Miller, M. T.; Cohen, S. M. Biochim. Biophys.
Acta 2010, 1803, 72.
16. Li, N.-G.; Shi, Z.-H.; Tang, Y.-P.; Wang, Z.-J.; Song, S.-L.; Qian, L.-H.; Qian, D.-W.;
Duan, J.-A. Curr. Med. Chem. 2011, 18, 977.
17. Johnson, A. R.; Pavlovsky, A. G.; Ortwine, D. F.; Prior, F.; Man, C.-F.; Bornemeier,
D. A.; Banotai, C. A.; Mueller, W. T.; McConnell, P.; Yan, C.; Baragi, V.; Lesch, C.;
Roark, W. H.; Wilson, M.; Datta, K.; Guzman, R.; Han, H.-K.; Dyer, R. D. J. Biol.
Chem. 2007, 282, 27781.
18. Engel, C. K.; Pirard, B.; Schimanski, S.; Kirsch, R.; Habermann, J.; Klingler, O.;
Schlotte, V.; Weithmann, K. U.; Wendt, K. U. Chem. Biol. 2005, 12, 181.
19. In vitro fluorescent assay for matrix metalloproteinase family including for
example MMP-13: Recombinant human proMMP-13 may be expressed and
purified as described by Knauper et al. [V. Knauper et al., Biochem. J. 1996, 271,
1544]. The purified enzyme was used to monitor inhibitors of activity as
follows: purified proMMP-13 was activated using 1 mM amino phenyl
mercuric acid (APMA) for 20 h at 35 °C in assay buffer (0.1 M tris–HCl, pH7.5,
containing 0.1 M NaCl, 20 mM CaCl2, 0.02 mM ZnCl and 0.05% (w/v) brij 35
containing the synthetic substrate Mca-Pro-cyclohexyl-Ala-Gly-Nva-His-Ala-
Dap(Dnp)-NH2 (Bachem) in the presence or absence of inhibitors. Activity was
determined by measuring the fluorescence at ex328 nm and em393 nm. An
IC50 value for the inhibitor was calculated by plotting the raw data in a dose
response curve fitting application (e.g., Origin). A similar protocol with the
above substrate was used for expressed and purified pro MMP1, 2, 3, 7, 8, 9, 12,
14. Similar protocols were applied for measuring the activity of ADAM-17 and
ADAM-10 using the substrate dimethoxyfluorescein-Ser-Pro-Leu-Ala-Gln-Ala-
Val-Arg-ser-Ser-Arg-Cys-fluorescein (Bachem).
the profile was consistent in both rodent and human systems. The
series generally scaled well from in vitro (rat hepatocytes) to
rodent in vivo studies, with all tested compounds being within
two fold of experimental prediction. In vivo clearances were mod-
est and reasonable half lives were generated. Biliary and renal
clearances were negligible for tested compounds. This overall pro-
file make the series a useful start point for more detailed lead opti-
misation exploration.
The syntheses of key compounds are shown in Scheme 1. Inter-
mediate, 32, was used in palladium catalysed coupling processes to
generate alkyne and alkene linkers, which can be readily reduced
to the corresponding alkanes. Dehydration of the alcohol motif
(OR4) is achieved using acidic conditions to generate 31. The syn-
thesis of most of the exemplified quinuclidine intermediates has
previously been reported.32–34
We have identified a novel series of selective MMP13 inhibitors
and determined their binding mode through co-crystallisation
studies. Compounds exert their biological effects without interac-
tion with the catalytic zinc atom, which enables good levels of
selectivity over related enzyme family members to be achieved.
The series was subsequently progressed into lead optimisation
and these studies will be reported in the future.
20. Moy, F. J.; Chanda, P. K.; Chen, J. M.; Cosmi, S.; Edris, W.; Levin, J. I.; Powers, R. J.
Mol. Biol. 2000, 302, 671.
21. The model of compound 3 was built using Maestro molecular modelling
program, docked using Glide into MMP13 structure (PDB ID = 1xuc) and
overlay with compound 1 in Maestro. Glide and Maestro were licensed from
Schrödinger, LGG (www.schrodinger.com). The final conformation of
compound 3 was optimised in MMP13 protein using Szybki (MMFF94 force
field) licensed from Openeye (http://www.eyesopen.com/).
22. Buttar, D.; Colclough, N.; Gerhardt, S.; MacFaul, P. A.; Phillips, S. D.; Plowright,
A.; Whittamore, P.; Tam, K.; Maskos, K.; Steinbacher, S.; Steuber, H. Bioorg. Med.
Chem. 2010, 18, 7486.
Acknowledgements
We acknowledge the work of Ben Goult in generating the NMR
information and Ken Page and Gillian Smith for the TDI data.
23. Hopkins, A. L.; Groom, C. R.; Alex, A. Drug Discov. Today 2004, 9, 430.
24. Leeson, P. D.; Springthorpe, B. Nat. Rev. Drug. Discov. 2007, 6, 881.
25. The X-ray crystal ligand–protein structure of compound 24 (PDB ID = 2YIG) has
been deposited into the RCSB.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.05.075.
Protein data bank (http://www.pdb.org/pdb/home/home.do).
26. Coope, J. F.; Main, B. G. Tetrahedron: Asymmetry 1995, 1393, 6.
27. McGinnity, D. F.; Parker, A. J.; Soars, M.; Riley, R. J. Drug Metab. Dispos. 2000, 28,
1327.
References and notes
28. Soars, M. G.; Grime, K.; Sproston, J. L.; Webborn, P. J. H.; Riley, R. J. Drug Metab.
Dispos. 2007, 35, 859.
29. MacFaul, P. A.; Ruston, L.; Wood, J. M. Med. Chem. Commun. 2011, 2, 140.
1. Whittaker, M.; Floyd, C. D.; Brown, P.; Gearing, A. J. H. Chem. Rev. 1999, 99,
2735.
2. Hu, J.; Van den Steen, P. E.; Sang, Q.-X. A.; Opdenakker, G. Nat. Rev., Drug Discov.
2007, 6, 480.
30. The photolytic stability of compounds is assessed by irradiating 50 lM solution
of the compounds in phosphate buffer at pH 7 in the presence of 20% ethanol
co-solvent using a Hanau Suntest (Atlas Instruments) for up to 4 h. Any
changes to the parent compound are assessed by LC–UV–MS. Compounds with
measured half-lives of less than 30 min are regarded as being unstable.
31. MacFaul, P. A.; Morley, A. D.; Crawford, J. J. Bioorg. Med. Chem. Lett. 2009, 19,
1136.
3. Nuti, E.; Tuccinardi, T.; Rossello, A. Curr. Pharm. Des. 2007, 13, 2087.
4. Nagase, H.; Woessner, J. F., Jr. J. Biol. Chem. 1999, 274, 21491.
5. Nagase, H.; Visse, R.; Murphy, G. Cardiovasc. Res. 2006, 69, 562.
6. Gomez, D. E.; Alonso, D. F.; Yoshiji, H.; Thorgeirsson, U. P. Eur. J. Cell Biol. 1997,
74, 111.
7. Martel-Pelletier, J. Osteoarthritis Cartilage 1999, 7, 371.
8. Glodring, M. B.; Glodring, S. R. J. Cell Physiol. 2007, 213, 626.
9. Churg, A.; Wang, R. D.; Tai, H.; Wang, X.; Xie, C.; Dai, J.; Shapiro, S. D.; Wright, J.
L. Am. J. Respir. Crit. Care Med. 2003, 167, 1083.
10. Hautamaki, R. D.; Kobayashi, D. K.; Senior, R. M.; Shapiro, S. D. Science 1997,
277, 2002.
11. Coussens, L. M.; Fingleton, B.; Matrisian, L. M. Science 2002, 295, 2387.
12. Mitchell, P. G.; Magna, H. A.; Reeves, L. M.; Lopresti-Morrow, L. L.; Yocum, S. A.;
Rosner, P. J.; Geoghegan, K. F.; Hambor, J. E. J. Clin. Invest. 1996, 97, 761.
13. Bigg, H. F.; Rowan, A. D. Curr. Opin. Pharmacol. 2001, 1, 314.
14. Brown, P. D. Expert Opin. Inv. Drug 2000, 9, 2167.
32. Brown, G. R.; Clarke, D. S.; Foubister, A. J.; Freeman, S.; Harrison, P. J.; Johnson,
M. C.; Mallion, K. B.; McCormick, J.; McTaggart, F.; Reid, A. C.; Smith, G. J.;
Taylor, M. J. J. Med. Chem. 1996, 39, 2971.
33. Brown, G. R.; Hollinshead, D. M.; Stokes, E. S. E.; Clarke, D. S.; Eakin, M. A.;
Foubister, A. J.; Glossop, S. C.; Griffiths, D.; Johnson, M. C.; McTaggart, F.;
Mirrlees, D. J.; Smith, G. J.; Wood, R. J. Med. Chem. 1999, 42, 130.
34. Brown, G. R.; Foubister, A. J.; Freeman, S.; McTaggart, F.; Mirrlees, R. A. C.;
Smith, G. J.; Taylor, M. J.; Thomason, D. A.; Whittamore, P. R. O. Bioorg. Med.
Chem. Lett. 1997, 7, 597.