Hit-to-Lead studies: The discovery of potent, orally bioavailable thiazolopyrimidine CXCR2 receptor antagonists

Article abstract of DOI:10.1016/j.bmcl.2005.10.091

A Hit-to-Lead optimisation programme was carried out on a high throughput screening hit, the thiazolopyrimidine 1, resulting in the discovery of the potent, orally bioavailable CXCR2 antagonist 29.

Full text of DOI:10.1016/j.bmcl.2005.10.091

Bioorganic & Medicinal Chemistry Letters 16 (2006) 960–963
Hit-to-Lead studies: The discovery of potent, orally bioavailable
thiazolopyrimidine CXCR2 receptor antagonists
Andrew Baxter,* Anne Cooper, Elizabeth Kinchin, Kerry Moakes, John Unitt and
Alan Wallace
AstraZeneca R&D Charnwood, Bakewell Road, Loughborough LE11 5RH, UK
Received 1 June 2005; revised 26 October 2005; accepted 27 October 2005
Available online 15 November 2005
Abstract—A Hit-to-Lead optimisation programme was carried out on a high throughput screening hit, the thiazolopyrimidine 1,
resulting in the discovery of the potent, orally bioavailable CXCR2 antagonist 29.
ꢀ 2005 Elsevier Ltd. All rights reserved.
The use of high throughput screening (HTS) and Hit-to-
Lead (HtL) is now widespread in the pharmaceutical
industry.1 The lead identification process has been more
accurately defined as two distinct steps—Active-to-Hit
(AtH) and Hit-to-Lead.
tion that once the 100 nM level has been achieved then
further increases in potency to the nanomolar level are
normally possible. The physicochemical parameters are
a reflection of the lipophilicity and molecular weight
increases normally seen in LO programmes and are
based on an extension of the arguments put forward
by Teague et al.,2 i.e., lower than the molecular weight
(<500) and lipophilicity (logP < 5) drug-like targets out-
lined by Lipinski et al.3 In vitro and in vivo DMPK tar-
gets have been equated to clearances of approximately
half liver blood flow, detectable oral bioavailability
and non-limiting plasma protein binding. In this way,
lead compounds should give some oral exposure to aid
an in vivo hypothesis test of the target mechanism and
be a solid starting point for a LO project leading to an
orally active CD. Lead compound should fulfil the
majority of the lead criteria—an explanation of any
missed targets should be provided. It is also important
at the lead stage that the series has a large scope for fur-
ther optimisation including areas of unexplored SAR
that will enable the LO project to increase potency
and improve other properties going forward to CDs.
HTS and HtL should aim to provide multiple series at
the lead stage and this is seen as an important factor
in de-risking future LO work. In this way, consistent
standards are maintained for entry into LO and this
should provide compound series that have optimisable
potency and DMPK.
AtH and HtL are based on our definitions of Actives,
Hits and Leads.
• Actives are defined as being the raw output from a
HTS and will be in the form of an IC50 using a
HTS assay on a solution in DMSO of a sample from
our compound collection.
• Hits are defined in terms of a profile that includes
confirmation of chemical structure and biological
activity as well as physical chemistry data and drug
metabolism and pharmacokinetic (DMPK) measure-
ments. DMPK data are used at this early stage of the
drug discovery process to avoid failure at a late stage
due to poor pharmacokinetics (metabolic instability
and/or poor oral absorption).
• Leads are described in terms of a target profile of
chemical and biological data. This profile was chosen
so that a Lead Optimisation (LO) programme would
have a reasonable chance of producing a Candidate
Drug (CD) in a target time of two years.
The Lead target profile used is shown in Figure 1. The
potency target was arrived at by the empirical observa-
In our previous paper on the HtL carried out on triazol-
ethiols,4 the biological background and rationale in
disease were presented. In this paper, we present a
HtL study that was carried out on a chemically distinct
Keywords: CXCR2 antagonist; Hit-to-Lead.
*
Corresponding author. Tel.: +44 1509 644772; fax: +44 1509
645513; e-mail: andrew.jg.baxter@astrazeneca.com
0960-894X/$ - see front matter ꢀ 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2005.10.091

A. Baxter et al. / Bioorg. Med. Chem. Lett. 16 (2006) 960–963
961
Figure 1. Hit-to-Lead generic Lead target profile.
Table 1. Hit profile of thiazolopyrimidine 1
recombinant CXCR2 (hrCXCR2) expressed in HEK
293 membranes using a scintillation proximity assay
(SPA). Subsequently at the AtH stage, compounds with
binding inhibition seen in the HTS SPA were initially
validated in a more conventional filter wash hrCXCR2
OH
S
N
H2N
N
[
125I]IL8 binding assay. Confirmation of functional
N
S
antagonism was shown by blockade of GROa stimu-
lated intracellular calcium mobilisation in isolated hu-
man neutrophils using a fluorescence imaging plate
reader (FLIPR).5
Generic lead criteriaa
Compound 1
Binding IC50 < 0.1 lM
Ca flux IC50 < 0.1 lM
Rat hepatocyte Cl < 14
Human microsome Cl < 23
Molecular weight < 450
Solubility > 10 lg/mL
clogP < 3.0
10 lM
2.0 lM
49
18
The thiazolopyrimidine 1 was one of the hits from the
CXCR2 HTS. This novel compound was part of our
compound collection and, while having weak activity
in the binding assay (IC50 10 lM), had comparable
potency in a functional FLIPR assay (IC50 2.0 lM).
The profile of this CXCR2 hit is shown in Table 1
compared with the generic lead target profile. Our plan
was to examine in turn the three thiazolopyrimidine
substituents looking for SAR leading to increasing
potency. The poor in vitro rat hepatocyte clearance
270
27
2.0
2.9
logD < 3.0
a Units as Figure 1 where not stated.
hit compound to provide a second lead series of CXCR2
antagonists. A HTS was undertaken to identify com-
pounds that blocked the binding of [125I]IL8 to human
OH
OH
OH
N
S
i
N
ii
NC
N
H2N
N
SH
OH
N
H2N
N
SR
H2N
OH
N
SR
S
N
N
AcHN
S
iii
N
SR
OH
H
vii
7
N
N
SR
iv
OH
5
viii
ix
S
S
N
N
N
H2N
H2N
OH
N
v
N
SR
N
SH
S
N
N
1-4, 19-20
Br
N
SR
6
x
vi
X
Cl
OH
S
N
S
xi
S
N
N
N
N
H2N
H2N
MeHN
N
N
N
SR
SR
N
SR
10-18, 23-29
X = amine or thiol
8
9, 14, 21-22
Scheme 1. Reagents and conditions: i—RBr, NaOH, EtOH; ii—KSCN, pyridine, DMF, 65 ꢁC then cool to 5 ꢁC, add Br2, 2 h; iii—DMF, H2O, 10 h,
120 ꢁC; iv—t-BuONO, THF, 65 ꢁC; v—TMSBr, t-BuONO, CH3CN, rt, 16 h; vi—MeNH2, MeOH, rt; vii—AcCl, pyridine, CH2Cl2; viii—AlBr3,
toluene, 60 ꢁC, 6 h; ix—RBr or RI, t-BuOK, DMSO, 3 days, rt; x—POCl3, PhNEt2, reflux, 2 h; xi—amine or thiol, THF or DMSO, 65 ꢁC.

962
A. Baxter et al. / Bioorg. Med. Chem. Lett. 16 (2006) 960–963
was associated with possible conjugation of the hydrox-
yl group and the presence of the lipophilic S-pentyl was
also seen as a site of potential metabolic instability.
of amines and thiols to give compounds 10–13 and
15–18. In order to vary the 5-alkylthio more widely we
prepared the thiol by de-benzylation of 4 with alumini-
um tribromide.7 This thiol was then re-alkylated to give
19 and 20, which were chlorinated as above to give 21
and 22. Reaction of the chloro compounds 14, 21 and
22 with the appropriate hydroxyamines gave the
compounds 23–29.
The synthetic routes to the thiazolopyrimidines pre-
pared in this paper are shown in Scheme 1. Alkylation
of 6-aminothiouracil followed by thiocyanation and
cyclisation6 gave the alkylthiothiazolopyrimidines 1–4.
Diazotisation with tert-butyl nitrite in THF or in the
presence of TMSBr gave the 2-protio- and 2-bromo-
thiazolopyrimidines 5 and 6. Acetylisation of 1 gave 7
and reaction of the bromo compound 6 with methyl-
amine gave the methylamino analogue 8. The hydroxy
on thiazolopyrimidines 1 and 4 was converted to the
chloro (compounds 9 and 14) using phosphorus oxy-
chloride. This chloro was readily displaced by a variety
The initial focus of HtL was to examine the effects of
simple changes to the substituents on the thiazolopyrim-
idine core. Variation of the S-alkyl 5 substituent was
undertaken via synthesis of each analogue starting from
aminothiouracil (Scheme 1). CXCR2 antagonist activity
(Table 2) was lost in the change from pentyl (1) to meth-
yl (2) and ethyl (3). An encouraging increase in potency
was seen though with the benzylthio analogue 4. Some
simple transformations were undertaken on the 2-amino
group. Only replacement by hydrogen (5) showed any
activity, bromo (6), acetylamino (7), and methylamino
(8) were all inactive. The 2-amino group was kept
constant in all subsequent analogues prepared in HtL.
The hydroxyl in the 7 position is amenable to variation
via the readily prepared chloro compounds 9 and 14.
Displacement with a wide variety of amines was under-
taken in a combinatorial experiment. Compounds that
showed activity as well as two inactive analogues were
re-synthesised as solid characterised compounds. Simple
amino substituents such as diethylamino (compound 10)
and cyclopentylamino (compound 11) were inactive,
amino (compound 18) had reduced potency but
hydroxyethylamino (compound 12) retained activity.
This activity was sensitive to chain length as the
hydroxypropyl analogue (13) was inactive. Sulfur linked
analogues (15–17) were also inactive. This preliminary
examination of SAR indicated that areas worthy of
urther exploitation were the 5-benzylthio and the
7-hydroxyethylamino substituents.
Table 2. CXCR2 antagonist binding potencies
X
7
S
N
2
Y
R
N
5
N
S
X
Y
R
CXCR2 IC50 (lM)a
1
2
3
4
5
6
7
8
9
OH
OH
OH
OH
OH
OH
OH
OH
Cl
NH2
NH2
NH2
NH2
H
Pentyl
Me
Et
10
NA
NA
CH2Ph 1.3
Pentyl
Pentyl
10
Br
AcNH Pentyl
MeNH Pentyl
NA
NA
NA
NA
NA
NA
7
NH2
NH2
Pentyl
Pentyl
Pentyl
Pentyl
Pentyl
10 NEt2
11 NH-Cyclopentyl NH2
12 NH(CH2)2OH
13 NH(CH2)3OH
14 Cl
NH2
NH2
NH2
NH2
NH2
NH2
NH2
NA
Variation of the 5-benzyl substitution pattern was
undertaken combinatorially using the reaction of the
7-hydroxy-5-thiol with a variety of benzyl halides.
Two substitution patterns were chosen as they gave
increased potency, the 3-phenoxy and 2,3-difluoro. Con-
firmation (Table 3) was obtained by the resynthesis of
CH2Ph NA
CH2Ph NA
CH2Ph NA
CH2Ph NA
CH2Ph 10
15 SCH2CO2Me
16 SCH2CO2H
17 S(CH2)2OH
18 NH2
a NA = <50% inhibition at 10 lM.
Table 3. CXCR2 antagonist binding potencies
X
7
S
N
N
5
2
H2N
R
N
S
Compound, CXCR2 IC50 (lM)a
R = PhCH2
R = (3-PhOPh)CH2
R = (2,3-DiFPh)CH2
X = OH
X = Cl
4, 1.3
19, 0.48
21, NA
25, 1.4
20, 0.27
22, NA
14, NA
23, 3.9
24, 0.12
X = (S)-NHCH(Et)CH2OH
X = (R)-NHCH(Et)CH2OH
X = NHC(Me)2CH2OH
26, 0.35
27, 0.050
28, 0.0095
29, 0.014
a NA = <50% inhibition at 10 lM.

A. Baxter et al. / Bioorg. Med. Chem. Lett. 16 (2006) 960–963
963
Table 4. Thiazolopyrimidine DMPK data
Compound Rat hepatocyte Cl (lL/min/106 cells) Human microsome Cl (lL/min/mg)
Rat pharmacokinetics iv and po
Cl (mL/min/kg) Vss (L/kg) T1/2 (h) F (%)
1
4
49
68
3
18
<3
2
20
24
29
9.5
40
25
0.9
2.2
1.9
3.5
1.0
1.2
14
5
6
31
15
Table 5. Lead profile of thiazolopyrimidine 29
In vitro metabolic stability and in vivo pharmacokinet-
ics were determined for key compounds throughout this
study. The results are shown in Table 4. The initial
pentylthio compound 1 was metabolised in vitro by rat
hepatocytes indicating a conjugation mechanism. The
benzylthio analogue 4 has a similar profile but the diflu-
orobenzyl compound 20 is much more stable to both rat
hepatocytes and human microsomes. This indicated that
oxidation followed by conjugation was probably the
metabolic route and that could be blocked by fluorina-
tion of the phenyl ring. This profile is maintained in vivo
in the rat, compound 20 having low clearance and ade-
quate intravenous half-life. A slight increase in metabol-
ic instability is seen with the hydroxyamino derivatives
24 and 29 but these compounds otherwise have an
encouraging profile. In particular, compound 29 (Table 5)
fulfils most of the lead DMPK criteria including oral
bioavailability even though the solubility is low. A slight
increase in human microsome clearance was observed
but other close analogues (20 and 24) were more stable.
The thiazolopyrimidines exemplified by compound 29
formed the basis of a new Lead Optimisation project.
OH
HN
S
N
F
H2N
N
F
N
S
29
Generic lead criteriaa
Compound 29
Binding IC50 < 0.1 lM
Ca flux IC50 < 0.1 lM
Rat hepatocyte Cl < 14
Human microsome Cl < 23
Rat iv Cl < 35
0.014 lM
0.04 lM
4
31
25
Rat iv Vss > 0.5 L/kg
Rat iv T1/2 > 0.5 h
Rat po bioavailability > 10%
1.9
1.2
15%
98.4%
397
Plasma protein binding < 99.5%
Molecular weight < 450
Solubility > 10 lg/mL
clogP < 3.0
0.5
3.1
logD < 3.0
3.4
a Units as Figure 1 where not stated.
References and notes
1. Michne, W. F. Pharm. News 1996, 3, 19.
2. Teague, S. J.; Davis, A. D.; Leeson, P. D.; Opera, T.
Angew. Chem. 1999, 38, 3743.
3. Lipinski, C. A.; Lombardo, F.; Dominy, B. W.; Feeney, P.
J. Adv. Drug Deliv. Rev. 1998, 23, 3.
4. Baxter, A.; Bennion, C.; Bent, J.; Boden, K.; Brough, S.;
Cooper, A.; Kinchin, E.; Kindon, N.; McInally, T.;
Mortimore, M.; Roberts, B.; Unitt, J. Bioorg. Med. Chem.
Lett. 2003, 13, 2625.
5. For full details of the biological assays used see: Austen, R.;
Baxter, A.; Bonnert, R.; Hunt, F.; Kinchin, E.; Willis, P.
Int. Patent Appl. WO 00/09511.
solid, characterised samples of compounds 19 (IC50
0.48 lM) and 20 (IC50 0.27 lM). The chloro compounds
(14, 23 and 24) were then reacted with a set of
2-hydroxyamines in a combinatorial experiment. A large
number of alkyl substituted hydroxyamines are readily
available in chiral form as they are derived from amino
acids. The most potent of these were chosen for resyn-
thesis together with the optical isomer. Comparison of
compounds 19 and 25 with 20 and 26 shows the chiral
preference for (R) over (S) in this substituent. Com-
pounds 28 and 29 were the most potent prepared in this
study.
6. Baker, J. A.; Chatfield, P. V. J. Chem. Soc. C 1970, 2478.
7. Lanzilotti, A. E.; Ziegler, J. B.; Shabica, A. C. J. Am. Chem.
Soc. 1968, 76, 3666.