1618
R. P. Austin et al. / Bioorg. Med. Chem. Lett. 25 (2015) 1616–1620
access to sulfamides (R4 = substituted amino) as well as sulfon-
amides (R4 = alkyl, (hetero)aryl).
pounds,9 and optimal potency was achieved with (R)-alaninol as
the amine substituent, together with a 2,3-dihalobenzyl on the
sulfur.
Initial results are shown in Table 1. Monocyclic analogues 6, 9
and 16, retaining key (R)-alaninol and S-benzyl units from our pre-
vious studies,9 all showed a marked improvement in solubility
compared to bicyclic compounds 1 and 2, as well as promising
activity against CXCR2. The benzoic acid 6 and its acyl sulfonamide
analogue 9 have lower activity than the sulfonamide 16, indicating
that the acidic centre may not be optimally positioned for CXCR2
receptor binding. Additionally, the high plasma protein binding
for 6 relative to logD was predicted to cause difficulties in achiev-
ing sufficiently high free blood concentrations. The pyrimidine sul-
fonamide 16 therefore became a new starting point for lead
optimisation.
We then investigated the effect of varying the sulfonamide
(Table 3). Replacement of methyl with higher alkyl analogues
(compounds 31 and 32) offered little advantage. Potency was not
significantly improved and in the case of the most lipophilic benzyl
analogue 32, plasma protein binding was very high. The tri-
fluoromethyl analogue 33 had pIC50 8.0 which is probably due to
its increased acidity.7b This is also true of the benzene sulfonamide
analogues 34–36, where the observed potencies against the CXCR2
receptor parallel the relative acidities of the compounds. Unfortu-
nately, the more acidic analogues suffer from high plasma protein
binding. An interesting observation is that primary sulfamide 37 is
inactive at CXCR2. We postulate that this is due to intramolecular
hydrogen bonding between the NH2 and the pyrimidine N to form
a six-membered ring,18 and suggest that this is not an active con-
formation. Further evidence is provided by the positive activity
data for a similar sulfamide 38 in which intramolecular H-bonding
is not possible.
Optimisation of 16 began by investigation of the S-benzyl and
4-amino substituents (Table 2). SAR in the sulfonamide series
was found to be parallel to that seen for the earlier bicyclic com-
Table 1
Monocycles versus bicycles: CXCR2 binding potency,13 solubility,14 pKa,15 human
plasma protein binding (hPPB)16 and logD17 for 1, 2, 6, 9, 16
Encouraged by these results, we investigated the benzene sul-
fonamide and the sulfamide series further, choosing to use 2,3-
difluorobenzyl and 2-fluoro-3-chlorobenzyl substituents on sulfur,
which were optimal for potency (Table 4). Changes to the sulfon-
amide were well tolerated in terms of potency and solubility,
and offered an opportunity to modulate the overall physicochem-
ical properties of this series. Introduction of heteroatoms into the
benzene sulfonamides provided more polar analogues with lower
plasma protein binding, illustrated by imidazole analogues 39
and 40. Unfortunately, these compounds showed a significant
drop-off in potency when tested in a CXCR2 cell-based assay.19
An explanation of this could be that the compounds do not pene-
trate the cell membrane effectively during the time frame of the
in vitro assay, and that binding to the receptor is at an intracellular
site as suggested by previously reported pharmacology.20 More
significant cell potency drop-off is seen with piperazine 42. The
addition of a basic centre, creating a zwitterionic compound, is tol-
erated in terms of receptor binding but reduced cell permeability
results in a drop in cell potency. A range of other sulfamides
(43–48) have excellent overall profiles with required potency in
both binding and cellular assays, solubility and hPPB. From a
metabolic stability viewpoint, azetidine (43 and 44) and morpho-
line (47 and 48) were more stable than pyrrolidine (45) and piper-
idine (46) analogues. Metabolite identification studies in human
hepatocytes showed that these sulfamides have three major sites
of oxidative metabolism: debenzylation of sulfur, N-dealkylation
of the sulfamide, and oxidation of the (R)-alaninol primary alcohol.
The increased stability of the azetidine and morpholine analogues
is probably a composite of both lower lipophilicity and reduced
potential for N-dealkylation.
Key compounds were tested in in vivo pharmacokinetic studies
in rat23 and compared to earlier bicyclic compounds (Table 5 and
Fig. 2). Whilst maintaining comparable binding and cell potency,
the new pyrimidine sulfamide compounds showed large improve-
ments in solubility over 1 and 2 and improved overall pharmacoki-
netic profiles. Furthermore, compounds 43 and 47 have lower
intrinsic clearances in human hepatocytes, offering the potential
for in vivo profiling of this series in animal models of inflammatory
disease and the possibility of profiling in man.
In conclusion, we have disclosed a novel monocyclic series of
pyrimidine sulfamide CXCR2 receptor antagonists. These are
potent and soluble compounds with favourable pharmacokinetic
and physicochemical properties. Key findings from this program-
me of research showed that breaking the bicyclic core of our pre-
viously disclosed CXCR2 antagonists was an effective way of
significantly improving solubility. These compounds represent
Compound
CXCR2 pIC50
Solubility (
l
M)
pKa
hPPB, % free
logD
1
2
6
9
8.4
9.0
6.4
6.1
7.1
3.7
0.3
150
>120
>100
—
2.6
0.3
1.0
7.0
1.9
3.3
3.2
À0.7
À0.7
1.9
6.7
5.5
4.9
6.9
16
Table 2
Investigation of methane sulfonamide SAR: CXCR2 potency13 and solubility14 for 16,
24–30
R3
H
N
N
O
O
N
R5
S
N
S
H
Compound R3
R5
CXCR2
pIC50
Solubility
M)
(l
OH
OH
16
24
25
26
7.1
5.1
5.5
6.4
>100
>130
59
OH
—
OH
OH
F
F
F
27
28
29
30
7.5
7.9
6.8
6.2
30
21
OH
OH
OH
Cl
Cl
Cl
F
100
94
O
F
Austin, Rupert P.
